v»EPA
EPA-240-R-24-001 | December 2024
Guidelines for Preparing Economic Analyses
Third Edition
r
Office of Policy
National Center for Environmental Economics
www.epa.gov/economics
-------
Chapter 1 - Introduction 1-1
1.1 Background 1-1
1.2 The Scope of the Guidelines 1-2
1.3 Economic Framework for Analysis 1-2
1.4 Principles for Conducting Economic Analysis 1-6
Chapter 1 References 1-8
Chapter 2 - Statutory and Executive Directives for Conducting Economic Analyses 2-1
2.1 Executive Orders 2-3
2.2 Statutes 2-10
Chapter 2 References 2-13
Chapter 3 - Need for Regulatory Action and Evaluation of Policy Options 3-1
3.1 The Statement of Need 3-1
3.2 General Guidance on Policy Options to Evaluate 3-6
Chapter 3 References 3-12
Chapter 4 - Regulatory and Non-Regulatory Approaches to Environmental Policy
4.1 Traditional Command-and-Control or Prescriptive Regulation 4-1
4.2 Market-Based Approaches 4-5
4.3 Hybrid and Other Approaches 4-15
4.4 Voluntary Programs 4-22
4.5 Cross-Cutting Issues When Comparing Regulatory and Non-Regulatory Approaches 4-24
Chapter 4 References 4-30
Chapter 5 - Setting the Foundation; Scope, Baseline and Other Analytic Design Considerations.5-1
5.1 Scope of Analysis 5-1
5.2 Baseline 5-6
5.3 Multiple Rules 5-12
5.4. Time Horizon of Analysis 5-16
5.5 Representing Economic Behavior 5-19
5.6 Uncertainty 5-29
Chapter 5 References 5-35
-------
Chapter 6 - Discounting Future Benefits and Costs 6-1
6.1 Mechanics and Methods for Discounting 6-2
6.2 Background and Rationales for Social Discounting 6-10
6.3 Intergenerational Social Discounting 6-20
6.4 The Role of Private Discounting in Economic Analysis 6-26
6.5 Recommendations and Guidance 6-28
Chapter 6 References 6-31
Chapter 7 - Analyzing Benefits 7-1
7.1 The Benefits Analysis Process 7-2
7.2 Economic Value and Types of Benefits 7-12
7.3 Economic Valuation Methods for Benefits Analysis 7-25
7.4 Benefit Transfer 7-55
7.5 Accommodating Benefits that Cannot Be Quantified and/or Monetized 7-61
Chapter 7 References 7-64
Chapter 8 - Analyzing Costs 8-1
8.1 The Economics of Social Cost 8-2
8.2 Estimating Social Cost 8-11
8.3 Models Used in Estimating the Costs of Environmental Regulation 8-25
8.4 Modeling Decisions and Challenges 8-36
Chapter 8 References 8-43
Chapter 9 - Economic Impacts 9-1
9.1 Background 9-2
9.2 Statutes and Policies 9-2
9.3 Connections between Economic Impacts and Frameworks of Distributional Effects 9-3
9.4 Analytic Components of an Economic Impact Analysis 9-10
9.5 Impact Categories 9-16
Chapter 9 References 9-41
Chapter 10 - Environmental Justice and Life Stage Considerations,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 10-1
10.1 Executive Orders, Directives and Policies 10-1
10.2 Environmental Justice 10-2
10.3 Environmental Health for Children and Older Adults 10-21
Chapter 10 References 10-25
ii Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 11 - Presentation of Analysis and Results
11.1 Presenting Results of Economic Analyses 11-1
11.2 Communicating Sources of Uncertainty 11-13
11.3 Use of Economic Analyses 11-15
Chapter 11 References 11-17
Appendi oiionile Theory A-l
A.1 Market Economy A-l
A.2 Reasons for Market or Institutional Failure A-5
A.3 Benefit-Cost Analysis A-8
A.4 Measuring Economic Impacts A-9
A.5 Optimal Level of Regulation A-17
A.6 Conclusion A- 21
Appendix A References A-22
Appendix 1 - Mortality Risk Valuation Estimates 1-1
B.l Central Estimate ofVSL B-2
B.2 Other VSL Information B-2
B.3 Benefit Transfer Considerations B-4
B.4 Adjustments Associated with Risk Characteristics B-4
B.5 Effects on WTP Associated with Demographic Characteristics B-5
B.6 Conclusion B-7
Appendix B References B-9
Glossary i
iii Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 1 - Introduction
The Guidelines for Preparing Economic Analyses are part of the U.S.
Environmental Protection Agency's (the EPA's] commitment to improve the
preparation and use of sound science in economic analysis to inform decision
making. Written primarily for the economic analyst, the main purpose of this
document is to define and describe best practices for economic analysis
grounded in the economics literature. It also describes Executive Orders [EOs]
and other documents that impose analytic requirements and provides
detailed information on selected important topics for economic analyses.
1.1 Background
Thorough and careful economic analysis is an important component for informing and developing
sound environmental policies. High-quality economic analyses can greatly enhance the
effectiveness of environmental policy decisions by providing policy makers and the public with
data-driven information needed to systematically assess the consequences of various actions or
options.1 An economic analysis of a rulemaking is a positive exercise, as opposed to a normative
one, that provides information on the potential economic efficiency of policy alternatives and
assesses the magnitude and distribution of an array of impacts through careful investigation.
Economic analysis also serves as a mechanism for organizing information carefully, identifying the
kinds of impacts associated with stated policy alternatives, and projecting who will be affected.
Ultimately, economic analysis based on sound science should lead to better-informed regulatory
and policy decisions.
The Guidelines for Preparing Economic Analyses, hereafter Guidelines¦, focus on the conduct of
economic analysis to inform policy decisions and to meet requirements described by related
statutes, Executive Orders (EOs), and associated implementing guidance of those EOs.2 Based on the
state of science and economics at the time of its writing, this document is intended to ensure high-
quality analyses and consistency in how these economic analyses are prepared, performed and
reported. In so doing, the Guidelines elevate the quality of information shaping environmental
policy decisions and EPA-issued guidance. The Guidelines also describe an interactive development
process between analysts and decision makers; reviews and summarizes environmental economics
theory and the practice of benefit-cost analysis; and emphasizes issues in practical applications.
1 It is important to note that economic analysis is but one component in the decision-making process. Depending
on the statutory context, all or certain components of the economic analysis may not be used by or required for
the legal rationale for the regulation. Other factors that may influence decision makers include statutory
requirements, health risks, distributional considerations, enforceability, technical feasibility, policy priorities and
ethics.
2 Chapter 2 describes many of these statues, EOs and the analytic and/or procedural requirements they impose,
as well as associated guidance materials.
1-1 Guidelines for Preparing Economic Analyses | 3rd edition
-------
i. w TIt* >"f tl ue >>
The Guidelines apply to economic analyses conducted for environmental policies using both
regulatory and non-regulatory management strategies (e.g., support for voluntary programs) as
well as Agency-issued guidance. Separate EPA guidance documents exist for related analyses, such
as risk assessments, which can be inputs to economic analyses. No attempt is made here to
summarize such guidance materials. Instead, their existence and content are noted in the
appropriate sections.
The Guidelines assume the reader has some background in microeconomics as applied to
environmental and natural resource policies. To fully understand and apply the approaches and
recommendations presented in the Guidelines; readers should be familiar with basic applied
microeconomic analysis, the concepts and measurement of consumer and producer surplus, and
the economic foundations of benefit-cost analysis. Appendix A provides a brief review of
economic foundations, and the Glossary defines selected key terms.
The Guidelines are designed to assist staff with the preparation of economic analyses but are not a
rigid blueprint nor a detailed set of step-by-step directions for all economic analyses. The most
productive and illuminating technical approaches for an analysis will depend on case-specific
factors and will require professional judgment. The Guidelines are a summary of analytical
methodologies, empirical techniques, best practices, and data sources that can assist in identifying
and implementing those approaches.
Finally, it is important to note that while the Guidelines apply to all types of economic analysis, the
focus is on benefit-cost analysis and economic impact analysis -- two mainstays of the EPA's
economic analyses. Typically, these economic analyses are not independent from other analyses.
Assessing the effects of environmental policy is an inherently complex process in which results
from various disciplines are integrated and inform one another. Taken together, they are used to
predict environmental and behavioral outcomes and their economic consequences.
1," >11, f.
Conceptually, the ideal economic framework for assessing the effects of policy actions is one of
general equilibrium that defines the allocation of resources and interrelationships for an entire
economy with all its diverse components (e.g., households, firms, government). Potential
regulatory alternatives are then modeled as economic changes that move the economy from a
state of equilibrium absent the regulation (the baseline), to a new state of equilibrium with the
regulation in effect The differences between the old and new states are measured as changes in
prices, quantities of goods, services and factors produced and consumed, including
environmental quality, as well as wealth, income, and other economic metrics. These
measurements may then be used to characterize the net welfare change for each affected group to
inform questions of efficiency and distribution, based on individuals' expected changes in their
own welfare.
Questions about efficiency focus on aggregate changes in welfare. Economists generally define
benefits as positive changes in welfare and costs as the opportunities foregone, or reductions in
welfare. To assess efficiency under this scenario, we add these changes in welfare measured in
monetary terms across all affected individuals. In the ideal, general equilibrium framework, we can
estimate and sum all benefits and costs; so, a policy is a movement toward efficiency if the sum is
1-2 Guidelines for Preparing Economic Analyses | 3rd edition
-------
positive and a movement away from efficiency if the sum is negative. The policy that maximizes
this sum, i.e., net benefits, is considered economically efficient.3
Questions about the distribution of benefits and costs examine how specific groups of households
and industries are affected by the policy. The ideal framework would answer questions framed in
terms of welfare changes for groups of individuals (e.g., is the policy welfare-improving for a
specific group?) or in terms of specific economic factors (e.g., how much will prices change for some
goods?). These assessments of distributional outcomes are often important, apart from analysis of
benefits and costs (i.e., economic efficiency).
In practice, of course, capturing this idealized framework empirically can be difficult, if not
impossible, due to data availability and in most cases, it is not possible to monetize all benefits and
costs. No single modeling tool allows us to answer all policy-relevant questions about efficiency and
distributional effects.4 As a practical matter, most economic analyses assemble a set of models to
address these issues separately, but, even then, not all effects can be monetized. If limitations are
appropriately described, however, it is still informative to present comparisons of benefits and
costs that can be monetized and qualitatively characterized, as well as evaluations of effects on
specific groups.
As detailed more fully in Chapter 2, economic analysis of benefits, costs and distributional impacts
are required by EO 12866 for economically significant rules. Although EO 12291 in 1981 was the
first to require an economic assessment of significant regulatory actions in a regulatory impact
analysis (RIA), these analyses were not as extensive as the economic analyses required now by EO
12866. A complete economic analysis today, though it may still at times be labeled as an RIA,
consists of a benefit-cost analysis and any related cost-effectiveness analyses and assessments of
economic and distributional impacts. The Office of Management and Budget (OMB) has a useful
checklist (shown in adapted form in Text Box 1.1) for all components of an economic analysis
conducted under EO 12866 (OMB 2010).5
i - i sin Dnriii-i Lini ier* -irli h -n-i'lir O - i vnalysis
(BCA)
Benefit-cost analyses assess economic efficiency using the Potential Pareto criterion: is it
theoretically possible for those who gain from the policy to fully compensate those who lose, and
remain better off? When the answer to this question is "yes," then net benefits (benefits minus
costs) are positive and the policy is a movement toward economic efficiency.6
While conceptually identical, benefits and costs are often evaluated separately due to practical
considerations. The benefits of reduced pollution are often attributable to changes in outcomes not
3 Appendix A provides a conceptual overview of the economic theory of welfare changes and benefit-cost
analysis.
4 As discussed in Chapter 8, computable general equilibrium (CGE) models capture most, or all, modeled market
benefits and costs, but may not include non-market benefits. In practice, CGE models may be unable to analyze
relatively small sectors of the economy. See Chapter 8, Section 4.6.
5 The questions in Text Box 1.1 have been reproduced with minor modification from the OMB checklist without
the extensive footnotes. The footnotes and other details about the checklist can be found at
h tips:/'/www, whitehouse. aov/wp-
content/uploads/leaacy drupal files/omb/inforea/inforea/reapol/RIA Checklistpdf.
6 Appendix A describes the underlying economic theory in greater detail.
1-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
exchanged in markets, such as improvements in public health. In contrast, the costs are generally
measured through changes in outcomes that are exchanged in markets, such as pollution control
equipment. As a result, different techniques are often used to estimate benefits and costs.7
Social benefits analyses evaluate the total expected welfare gains individuals experience resulting
from the regulation or policy action. From the perspective of an action that reduces pollution or
environmental contaminants, many of these benefits come from improvements in environmental
quality. Once the changes in pollution levels or other environmental effects resulting from a policy
are estimated, these changes are translated into health outcomes or other relevant outcomes using
information provided by risk assessment and other disciplines. Benefits analyses then apply a
variety of economic methodologies to estimate the value of these anticipated health improvements
and other types of environmental benefits, but it is important to note that even those benefits that
cannot be quantified or put into dollar terms should be described in a benefits analysis. Chapter 7
provides details on methods for estimating social benefits. Within a benefits assessment, pollution
exposure may increase for some, e.g., emissions of a pollutant other than the one being regulated
may increase, or when the policy is deregulatory. Such costs may be presented as negative benefits
and may be described as disbenefits or foregone benefits provided that the analysis is internally
consistent.
Social cost analyses evaluate the total expected welfare losses experienced by individuals resulting
from the regulation or policy action. In most instances, these costs are measured by higher prices
for goods and services for consumers and lower earnings for producers and factors of production.
Sometimes one modeling effort can be used to estimate both social costs and inputs for benefits
analyses, such as predicted changes in pollution from regulated sources. Chapter 8 provides
detailed information on methods for estimating social costs. As with benefits, costs that cannot be
quantified or put into dollar terms should be described. Also, some costs may decrease due to the
regulation. For example, profits may increase for certain related entities or when the action is
deregulatory. These outcomes may be presented as negative costs and may be described as avoided
costs, again, provided that the analysis is internally consistent Ultimately, from the perspective of
economic theory, the treatment of disbenefits and avoided costs in the analysis is primarily a
communications issue and should not affect efficiency analysis and whether net benefits are
positive or negative.
7 These Guidelines are organized from the perspective of an action that is designed to achieve health and
environmental protection benefits, albeit at some cost. Chapter 7 (Estimating Benefits) therefore focuses
primarily on how to evaluate improvements in health and environmental quality, while Chapter 8 (Social Costs)
focuses on evaluating the costs associated with actions to achieve those benefits. However, the methods described
in these chapters are equally applicable to evaluating decrements in health or environmental quality, and for
cost savings if that is appropriate for the policy being evaluated (e.g., for deregulatory actions).
1-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 1.1 - Agency Checklist for Regulatory Impact Analysis
Does the RIA include a reasonably detailed description of the need for regulatory action?
Does the RIA include an explanation of how the regulatory action will meet that need?
Does the RIA use an appropriate baseline (i.e., best assessment of how the world would look in
the absence of the proposed action)?
Is the information in the RIA based on the best reasonably obtainable scientific, technical, and
economic information and is it presented in an accurate, clear, complete, and unbiased
manner?
Are the data, sources, and methods used in the RIA provided to the public on the internet so
that a qualified person can reproduce the analysis?
To the extent feasible, does the RIA quantify and monetize the anticipated benefits from the
regulatory action?
To the extent feasible, does the RIA quantify and monetize the anticipated costs?
Does the RIA explain and support a reasoned determination that the benefits of the intended
regulation justify its costs (recognizing that some benefits and costs are difficult to quantify)?
Does the RIA assess the potentially effective and reasonably feasible alternatives? Does the
RIA assess different regulatory provisions separately if included in the rule?
Does the RIA assess at least one alternative that achieves additional benefits and at least one
alternative that costs less?
Does the RIA consider setting different requirements for large and small firms?
Does the selected/finalized option have the highest net benefits (including potential economic,
environmental, public health and safety, and other advantages; distributive impacts; and
equity), unless a stature requires a different approach?
Does the RIA include an explanation of why the planned regulatory action is preferable to the
identified potential alternatives?
Does the RIA use appropriate discount rates for benefits and costs that are expected to occur
in the future?
Does the RIA include, if and where relevant, an appropriate uncertainty analysis?
Does the RIA include, if and where relevant, a separate description of distributive impacts and
equity?
Does the RIA provide a description/accounting of transfer payments?
Does the RIA analyze relevant effects on disadvantaged or vulnerable populations (e.g.,
persons with disabilities and low-income groups)?
Does the analysis include a clear, plain language executive summary, including an accounting
statement that summarizes the benefit and cost estimates for the regulatory action under
consideration, including qualitative and non-monetized benefits and costs?
Does the analysis include a clear and transparent table presenting (to the extent feasible)
anticipated benefits and costs (quantitative and qualitative)?
Adapted from OMB's Agency Checklist: Regulatory Impact Analysis (2010).
1-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
i , C v essin I zmomic and Distribution i iii i cts
The assumptions and modeling framework developed for the BCA often do not include or allow for
detailed examination of impacts on specific groups. Understanding the nature and magnitude of
policy impacts and who will gain or lose from a regulation can be important to policy evaluation,
and this requires analyses to supplement BCA.
The EPA addresses economic and distributional impacts of environmental policy through two sets
of analyses:
Economic Impact Analyses (EIAs) provide insight into how compliance costs, transfers and
other policy outcomes are distributed across groups. EIAs describe and often quantify
outcomes such as changes in employment, plant closures or local government tax revenues
that provide insight into the economic consequences of regulation. Economic impacts may
fall on groups such as industry sectors, small businesses, state or local governments,
consumers or workers that may benefit or be harmed by a policy. Chapter 9 provides
information on analyzing economic impacts.
Other analyses evaluate the distribution of changes in environmental risks or health
outcomes due to regulation from environmental justice (i.e., on minority, low-income or
Indigenous populations) and life stage (i.e., on children, the elderly) perspectives.
Consideration of costs may also be relevant in such analyses. Chapter 10 provides
information on how to analyze impacts from these perspectives.
i -1 Pj inripff"f [-^f[¦ -vi[i :miI /ns
Many aspects of an economic analysis will vary depending on the purpose, area of focus, available
data, and needed level of detail for the analysis. That said, the following are core principles that
apply to all economic analyses:
Economic analyses should be based on sound economics and science. Economic
analyses should be grounded in well-established economic methods, theory, and principles.
The effects considered in BCA, for example, should follow from economic principles and are
independent of what is considered in legal or policy analyses, or what may be defined by
science policy in other disciplines. Economic analysis should also be flexible enough to
incorporate new information and advances in theory and the practice of economics.
Economic analyses often rely upon or draw from the tools and results of other scientific
analyses. These analyses should also be grounded in the principles, theories, and methods
appropriate to their discipline.
Economic analyses should be objective and avoid bias. The goal of the economic
analysis is to provide objective information about the consequences of policy decisions.
Professional judgments and assumptions are generally required for economic analyses, but
these judgments and assumptions should not be based on the preferences of the analyst or
policy maker. Economic analyses should seek to capture the expected behavioral responses
of households, firms, and governments to incentives and options created by the actual
requirements of the regulation or other context being analyzed as accurately as possible.
Analyses should be unbiased and should not be framed or performed in a manner to obtain
predetermined results or to defend a particular policy decision. In addition, judgments or
assumptions should not be made to favor one conclusion over another. For instance,
sensitivity analysis can be used to explore a range of possible outcomes but should examine
both higher and lower values rather than only one or the other.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Economic analyses should be transparent and replicable. Economic analysis requires
choices about data sources, methods, models, and assumptions. The reasons for these
choices should be presented explicitly and clearly, along with appropriate justification.
Economic analysis should also explicitly acknowledge and characterize important
uncertainties in the analysis, state the judgments and decisions associated with these
uncertainties, and should identify the implications of these choices. Specific references
should be made to all data sources and models, and publicly available data and models
should be used to the maximum extent possible. The analysis should provide enough
information for readers to clearly see how final empirical estimates and conclusions were
reached.
Key Best Practic * mi-." m Hi - sidelines
Key best practices that apply to all or most economic analyses are also covered in these Guidelines.
These are listed below along with the chapter in which they are covered:
Economic analyses produced by the EPA should be responsive to directives from applicable
statutes and executive orders (Chapter 2).
Analyses should describe the economic basis for the policy action and evaluate multiple
options to arrive at the most desirable decision (Chapter 3).
Economics and economic analysis can also inform the consequences of different regulatory
designs under consideration, identifying those that are likely to be most cost-effective
(Chapter 4).
The economic impact and consequences of policy must be evaluated relative to some
alternative setting, generally one without the policy action. This alternative setting is called
the analytic baseline. Specifying a baseline can sometimes be challenging, but it is essential
for sound and informative economic analysis. The scope of the analysis should also be
clearly defined, and uncertainties in the analysis should be evaluated and characterized
(Chapter 5).
The economic effects of policies typically occur over several years. As such, consistent
application of discounting is needed to make these effects comparable (Chapter 6).
Analysis of benefits and costs should be grounded in sound, well-established economic
principles and approaches, should capture all relevant outcomes to the extent possible, and
should incorporate advances in the field where warranted (Chapter 7 and Chapter 8).
Analysis of the distribution of impacts associated with policy decisions should adhere to the
same high standards of an economic analysis, should start with the same baselines as the
economic analysis, and should provide a balanced accounting of who gains and who loses as
a result the policy action (Chapter 9 and 10).
Finally, an economic analysis must be clearly and effectively communicated for it to be
valuable for decision-making (Chapter 11).
1-7
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chr,pt«r 1 F pcj- rf,^,
OMB. 2023. Circular A-4, Regulatory Analysis, November 9, 2023. Available at:
https://www.whitehouse.gov/wp-content/uploads/2023/ll/CircularA-4.pdf (accessed June 13, 2024).
OMB. 2010. Agency Checklist: Regulatory Impact Analysis, October 28, 2010. Available at:
https://www.whitehouse.gov/wp-
content/uploads/legacy drupal files/omb/inforeg/inforeg/regpol/RIA Checklistpdf (accessed June 13,
2024).
1-8
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 2 - Statutory and Executive
Directives for Conducting Economic
Analyses
Federal agencies are subject to statutes and executive orders (EOs] that direct
them to conduct specific types of economic analyses. Many are potentially
relevant for all U.S Environmental Protection Agency [EPA] programs; others
target individual programs. The scopes of the directives calling for economic
analyses vary substantially. In some cases, a statute or EO may be limited in its
applicability to those regulatory actions that exceed a specified threshold in
significance or impact. To determine whether a regulatory action meets such a
threshold and is covered by the statutory or EO provisions, the agency may
need to conduct a preliminary economic analysis. Covered regulatory actions
may need:
Economic analysis (e.g., analysis of benefits and costs pursuant to EO
12866, "Regulatory Planning and Review"];
Procedural steps (e.g., consultation with affected state and local
governments pursuant to EO 13132, "Federalism"]; or
A combination of both economic analysis and procedural steps.
This chapter identifies directives for conducting economic analyses that may
apply to all EPA programs (see Table 2.1] and thresholds that trigger an
economic analysis or additional procedural steps for a regulatory action.1 It
also summarizes general provisions calling for economic analyses in selected
statutes and EOs and provides direction for analysts seeking guidance on
compliance with them. References to applicable Office of Management and
Budget (OMB] and EPA guidelines for each EO or statute are provided. For
further information about the type and scope of analysis directed, the
program's Office of General Counsel (OGC] attorney is a good resource.2 This
chapter does not address provisions of the statutes and EOs that do not call
for economic analysis.
1 Although not discussed here, analysts should carefully consider the relevant program-specific statutory
requirements when designing and conducting economic analyses, recognizing that these requirements may
mandate specific economic analyses.
2 For OGC's reference guide on cross-cutting statutory and EO reviews that may apply to rules, see U.S. EPA
(2003b, 2005).
2-1 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Table 2.1 - Overview of Executive Orders and Statues
Executive Order/Statute
Economic
Threshold*
Guidance/
Information
Available
EO 12866, Regulatory Planning and Review (1993) as amended
by Executive Order 14094, "Modernizing Regulatory Review"
(2023)
Specific
EPA, OMB
EO 12898, Federal Actions to Address Environmental Justice in
Minority Populations and Low-Income Populations (1994)
General
EPA
EO 13045, Protection of Children from Environmental Health
Risks and Safety Risks (1997)
Specific
EPA
EO 13132, Federalism (1999)
Specific
EPA
EO 13175, Consultation and Coordination with Indian Tribal
Governments (2000)
General
EPA, OMB
EO 13211, Actions Concerning Regulations that Significantly
Affect Energy Supply, Distribution, or Use (2001)
Specific
OMB
EO 13563, Improving Regulation and Regulatory Review (2011)
Specific
OMB
EO 13707, Using Behavioral Science Insights to Better Serve the
American People (2015)
General
White House
Memo
EO 14096, Revitalizing Our Nation's Commitment to
Environmental Justice for All (2023)
General
EPA
Regulatory Flexibility Act (RFA), as Amended by the Small
Business Regulatory Enforcement Fairness Act of 1996
(SBREFA)
Specific
EPA
Unfunded Mandates Reform Act of 1995 (UMRA)
Specific
EPA, OMB
Paperwork Reduction Act of 1995 (PRA)
Specific
EPA, OMB
The Foundations for Evidence-Based Policymaking Act of 2018
None
OMB
* Economic Threshold: "Specific" ifEO or statute provides specific numeric threshold or detailed
criteria; "General" ifEO or statute provides only general description or statement.
2-2 Guidelines for Preparing Economic Analyses | 3rd edition
-------
, I h ¦ Mih vh-i-i s
C.i.i li wi e Order i C , |-'-:; atoi. [Manning and Review" as
amenc -' II' - 'un -- >V-i 14094, "Modernizi ilatory
Review"
Threshold: Significant regulatory actions as defined by the EO. A "significant regulatory action"
is defined by Section 3 (f)(1)-(4) as any regulatory action that is likely to result in a rule that may:
1. Have an annual effect on the economy of $200 million or more (adjusted every 3 years by
the Administrator of OIRA for changes in gross domestic product); or adversely affect in a
material way the economy, a sector of the economy, productivity, competition, jobs, the
environment, public health or safety, or State, local, territorial, or tribal governments or
communities;4
2. Create a serious inconsistency or otherwise interfere with an action taken or planned by
another agency;
3. Materially alter the budgetary impact of entitlements, grants, user fees or loan programs or
the rights and obligations of recipients thereof; or
4. Raise legal or policy issues for which centralized review would meaningfully further the
President's priorities or the principles set forth in this Executive order, as specifically
authorized in a timely manner by the Administrator of OIRA in each case.
EO 12866 does not distinguish between regulatory and deregulatory actions. Meeting one or more
of the threshold criteria triggers the classification of a regulatory action as "significant." OMB
categorizes a regulatory action that meets the first criterion as significant under Section 3(f)(1) of
Executive Order 12866 (as amended) (formerly referred to as "economically" significant).5 The
determination of significance under Section 3(f)(1) is multi-faceted. Rules that have an annual
effect that meets the $200 million threshold (as adjusted every 3 years) are deemed significant
under Section 3(f)(1). OMB clarified that they interpret the EO 12866 threshold as being based on
the annual costs, benefits, or transfers of the regulatory action in any one year.6
The word "or" is important: $200 million (updated every 3 years) in annual benefits, or costs, or
transfers is sufficient to meet the threshold.7 Note that the threshold determination is not based on
net effects, so if any category meets the threshold, the rule would be significant under 3(f)(1). For
example, suppose Congress passes a new law that requires the EPA to collect user fees from an
industry that manufactures chemicals. The user fees will be used to defray EPA's costs associated
with an existing obligation to conduct risk evaluations of new chemicals. Previously, funds to pay
the EPA's costs to conduct these evaluations were provided by Congress through its annual
congressional appropriation. This new rule requires the EPA to recoup these costs from industry.
Assume that the fees to be collected from industry total $220 million per year. In this case, no new
3 EO 13563, "Improving Regulation and Regulatory Review," issued in January 2011, supplements and reaffirms
the provisions ofEO 12866. It emphasizes the importance of reducing regulatory costs and burdens and
maintaining flexibility and freedom of choice. See Section 2.1.7 in this chapter for more information on EO
13563.
4 EO 14094 increased the 12866 threshold from $100 to $200 million and added the inflation adjustment every
three years.
5 See OMB 2023.
6 OMB 2023a.
7 OMB 2023a.
2-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
burden is being placed on society. The $220 million is simply a transfer of payments from
businesses to government; however, because the transfer is more than $200 million annually, this
action is 3(f)(1) significant. By contrast, a rule with $120 million in benefits and $120 million in
costs would not be sufficient to meet the dollar threshold under Section 3(f)(1).
In addition, rules that "adversely affect in a material way the economy, a sector of the economy,
productivity, competition, jobs, the environment, public health or safety, or State, local, territorial,
or tribal governments or communities" are also deemed 3(f)(1) significant. These criteria are
independent of the $200 million threshold to trigger the "significant under 3(f)(1)" designation.
It is important to note that meeting the $200 million threshold can include consideration of
unquantified effects as well as quantified effects. There may be impacts that are unquantified due to
lack of data or valuation methods, but if the judgement of the EPA, or ultimately OMB, is that the
combined quantified and unquantified annual effects are likely to exceed $200 million, the
regulation would be considered 3(f)(1) significant. OMB clarifies that the threshold determination
should also consider effects that may seem "indirect" or "ancillary."8
In practice, while the threshold for 3 (f)(1) significance is important, the level of analysis can vary.
OMB clarifies, "Different regulations may call for different emphases in the analysis, depending on
the nature and complexity of the regulatory issues and the sensitivity of the benefit and cost
estimates to the key modeling choices."9
Per amendments made to EO 12866 by EO 14094, OMB will automatically update the threshold for
3(f)(1) significance every three years, indexed to GDP growth.10
Analyses contingent on threshold: Regulatory actions designated "significant" are subject to EO
12866 review by OMB. The process of making this determination is discussed in "EPA's Action
Development Process: Guidance for EPA Staff on Developing Quality Actions."11 For all significant
regulatory actions, the agency shall provide to OMB a statement of the need for the regulatory
action and an assessment of potential benefits and costs (Section 6(a)(3)(B)). The analysis of
benefits and costs increases in complexity and detail for 3(f)(1) significant rules (i.e., those that fall
under the definition in the first bullet above). For these rules, the EO directs that, in addition to
assessing potential costs and benefits, agencies must include the underlying analysis informing that
assessment, quantify benefits and costs to the extent feasible, assess the benefits and costs of
potentially effective and reasonably feasible alternative approaches, and provide the underlying
analysis of that alternatives assessment (Section 6(a)(3)(C)). OMB's Circular A-4 (discussed below)
states that analysts should generally analyze at least three options for each key attribute or
provision: the proposed or finalized option; at least one option that achieves additional benefits;
and at least one option that costs less.12
Guidance: OMB's Circular A-4 (2023) provides guidance to federal agencies on the development of
regulatory analysis for 3(f)(1) significant rules as directed by EO 12866 as well as for other
regulatory analysis either required or undertaken at the agency's discretion. Circular A-4 is
8 OMB 2023a.
9 OMB 2023, p. 4.
10 EO 12866 did not provide for an inflation adjustment, resulting in the $100 million threshold becoming more
stringent as inflation increased over the years.
11 U.S. EPA 2024.
12 OMB 2023, p. 21.
2-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
intended to assist analysts in conducting high-quality and evidence-based regulatory analysis and
to standardize the way benefits and costs of federal regulatory actions are measured and reported.
Parts of Circular A-4 guidance are standardized for rules that are 3 (f)(1) significant For example,
agencies are asked to provide a prominent standardized accounting statement, with one or more
tables summarizing costs and benefits (including monetized; quantified, but not monetized; and
unquantified), at a standardized consumption discount rate (updated every 3 years) for the main
analysis along with reporting of the undiscounted annual stream of benefits and costs.13 In other
respects, OMB notes that "you cannot conduct a good regulatory analysis according to a formula.
Conducting high-quality analysis requires competent professional judgment" 14 OMB published
additional supporting information in a separate document entitled OMB Circular No. A-4:
Explanation and Response to Public Input.15
The Guidelines provide more in-depth Agency guidance, building on the OMB's guidance with a
focus on approaches and methods that are relevant to environmental regulations. Chapters 3
through 8 of this document provide more detailed guidance for fulfilling the EO 12866 benefit-cost
analysis provisions, consistent with directions in OMB's Circular A-4. Chapters 9 and 10 provide
guidance on addressing distributional effects of environmental regulation, with a focus on
economic impact analysis examining compliance costs effects (e.g., profitability, employment,
prices) in Chapter 9 and on environmental justice and life stage considerations in Chapter 10.16
12898, "Federal Actions to Address
Environmental Justi in [ line i ir. Populations and Low-Income
[ v filiations [I -cutive Order 14096, ,rP- -mli jh;; I ion's
1*^ Erriintel tic 1" vi 'Ml"
Threshold: No specific threshold; EO 12898 directs each agency, to the greatest extent
practicable and permitted bylaw, to "make achieving environmental justice part of Its mission."
EO 14096 supplements EO 12898 and calls on the federal governmentto "build upon and
strengthen its commitment to deliver environmental justice."
Analyses contingent on threshold: EO 12898 directs agencies, to the greatest extent practicable
and permitted by law, to "identify[] and address[],... disproportionately high and adverse human
health or environmental effects of its programs, policies and activities on minority populations
and low-income populations."17 Among other directives and consistent with EO 12898, EO 14096
calls on agencies to, as appropriate and consistent with applicable law, "identify, analyze, and
address":
13 See Chapter 11 of this document, Presentation of Analysis and Results, for agency guidance on presenting
economic analysis results.
14 OMB 2023, p. 4.
15 OMB 2023a.
16 In its Statement of Regulatory Philosophy, EO 12866 states that agencies should consider the distributional
and equity effects of a rule (Section 1(a)).
17 See EO 12898 Sec. 1-101. EO 14096 uses the phrase "disproportionate and adverse" instead of
"disproportionately high and adverse," as used in EO 12898. According to the White House Fact Sheet on EO
14096, these phrases have the same meaning. Removing the word "high" is intended in to eliminate potential
2-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
1. Disproportionate and adverse human health and environmental effects..., including those
related to climate change and cumulative impacts of environmental and other burdens on
communities with environmental justice concerns;
2. Historical inequities, systemic barriers, or actions related to any Federal regulation, policy,
or practice that impair the ability of communities with environmental justice concerns to
achieve or maintain a healthy and sustainable environment; and
3. Barriers related to Federal activities that impair the ability of communities with
environmental justice concerns to receive equitable access to human health or
environmental benefits, including benefits related to natural disaster recovery and climate
mitigation, adaptation, and resilience.18
Guidance: The EPA's "Technical Guidance for Assessing Environmental Justice in Regulatory
Analysis" is designed to outline analytic expectations and discuss technical approaches and
methods that can be used by EPA analysts to evaluate the environmental justice (EJ) effects of
regulatory actions.19 This technical guidance is also useful for understanding what role analysis
can play in ensuring that EJ concerns are appropriately considered and addressed in the
development of regulatory actions, to the extent practicable and permitted by law. Chapter 10 of
this document addresses environmental justice analysis, including guidance on considering the
distribution of exposure, health outcomes, benefits and/or costs when evaluating impacts on
these specific populations.
C i , I'-Lin 'j - -i iWm" "Protection of Chil -11 i"i ~itt
Environmental Health Risks a »ks"
Threshold: Economically significant regulatory actions as described by EO 12866 (now referred
to as 3 (f)(1) significant, per above) that involve environmental health risk or safety risk that an
agency has reason to believe may disproportionately affect children.
Analyses contingent on threshold: An evaluation of the health or safety effects of the planned
regulation on children (section 5(a)) or an explanation of why not conducted. The agency shall also
provide an explanation of why the planned regulation is preferable to other potentially effective
and reasonably feasible alternatives the agency is considering (Section 5(b)).
Guidance: The EPA has prepared guidance to assist EPA staff on the implementation of EO
13045.20 The EPA's Children's Health Valuation Handbook discusses special issues related to
estimation of the value of health risk reductions to children.21 The Office of Children's Health
Protection also provides
misunderstanding that agencies should only be considering large disproportionate effects. See FACT SHEET:
President Biden Signs Executive Order to Revitalize Our Nation's Commitment to Environmental Justice for All,
The White House (Apr. 21, 2023), https://www.whitehouse.gov/briefing-room/statements-
releases/2023/04/21/fact-sheet-president-biden-signsexecutive-order-to-revitalize-our-nations-commitment-
to-environmental-justice-for-all/. EO 14096 also includes a definition of "environmental justice" which expands
on the demographic categories laid out in EO 12898, such as by including Tribal affiliation and individuals with
disabilities. See EO 14096 Sec. 2(b).
18 See EO 14096 Sec. 3(a)(i), (iii), and (iv).
19 U.S. EPA 2016.
20 U.S. EPA 2024.
21 U.S. EPA 2003a.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
online information with links to resource materials on guidance and tools.22 Guidance in Chapter
10 of this document addresses analyses of impacts on children.
C i -i - tive Order r r "..J. "Federalism"
Threshold: Rules that have "federalism implications" that either impose substantial compliance
costs on state and local governments or preempt state or local law. According to EPA policy, rules
are considered to impose substantial compliance costs if:
The action is likely to result in the expenditure by state and local governments, in the aggregate, of
$25 million or more in any one year; or
The action is likely to result in expenditures by small governments that equal or exceed 1% of their
annual revenues.23
Exception: An action that imposes substantial compliance costs (meets the $25 million threshold
or the 1% test) does not have a federalism implication if: (1) the action is expressly required by
statute (without any discretion by the EPA); or (2) there are federal funds available to cover the
compliance costs.
Analyses contingent on threshold: For actions with federalism implications, agencies shall
conduct pre-proposal consultation with elected state/local officials or their representative national
organizations. Rules must include a Federalism Summary Impact Statement in the preamble, and a
signed Federalism Certification from the Agency's designated official should be provided to OMB for
rules subject to OMB review under EO 12866 along with any written communications that the EPA
received from state or local officials.
Guidance: Specific guidance on EO 13132 can be found in the internal EPA document "Guidance on
Executive Order 13132: Federalism".24
C i * I'-Lin >s v -i i :.n^ "Consvh vion ai ' ? *rn i v on whir
)l Governments"
Threshold: Regulations that have substantial direct effects on one or more American Indian tribe,
on the relationship between the federal government and tribes, or on the distribution of power and
responsibilities between the federal government and tribes and that: (1) impose substantial direct
compliance costs on Indian tribal governments that are not required by statute, or (2) preempt
tribal law.
Analyses contingent on threshold: To the extent practicable and permitted by law, EO 13175
directs the Agency to either provide the funds necessary to pay the Tribal governments' direct
compliance costs, if applicable, or prior to the formal promulgation of the regulation, to (1) consult
with Tribal officials early in the process of developing the proposed regulation; (2) make any
written communications submitted to the Agency by Tribal officials available to the Director of
OMB; and (3) include in the preamble of the regulation a Tribal Summary Impact Statement. The
22 See https://www.epa.gov/children/guidance-tools-and-glossary-key-terms (accessedJuly 31, 2024).
23 U.S. EPA 2008.
24 U.S. EPA 2008.
2-7 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Statement should include a description of the extent of the Agency's prior consultation with Tribal
governments; a summary of the nature of the Tribe's concerns and the Agency's position
supporting the need to issue the regulation; and a statement of the extent to which the concerns of
Tribal governments have been met
Guidance: OMB issued Guidance for Implementing E013175in2010to provide direction for
compliance and documentation,25 and the White House issued Presidential Memoranda in 2009,
2021 and 2022 to support implementation of EO 13175.26 The 2021 Presidential Memo (Tribal
Consultation and Strengthening Nation-to-Nation Relationships) reaffirms the policy in the 2009
Presidential Memo (Tribal Consultation) and directs agencies to submit detailed plans of action to
implementthe policies and directives. The 2022 Presidential Memo (Uniform Standards for Tribal
Consultation) establishes uniform minimum standards to be implemented across all agencies
regarding how Tribal consultations are to be conducted. The EPA updated its Policy on Consultation
and Coordination with Indian Tribes in 2023 to establish national guidelines and institutional
controls for consultation across the EPA. This policy states, "The U.S. Environmental Protection
Agency's policy is to consult on a government-to-government basis with federally recognized Tribal
governments when EPA actions or decisions may affect Tribes." [emphasis added].27 Chapter 10 of
this document addresses environmental justice analyses focusing on people of color, low-income
populations, and/or Indigenous populations.
C i ^ I'-Lin <- - -1 'f r;.,2'f 1'. "Action ' ncernii ilationsthat
" io 11i"i - -ii111 \ Ni" ir I ner " i.i| |:f\. Distributirn. Use"
Threshold: Rules that are significant regulatory actions under EO 12866 and that are likely to have
significant adverse effects on the supply, distribution, or use of energy.
Analyses contingent on threshold: Submission of a detailed Statement of Energy Effects to OMB.
The Statement of Energy Effects must address any expected adverse effects on energy supply,
distribution or use, the reasonable alternatives to the action, and the expected effects of such
alternatives on energy supply, distribution, and use.
Guidance: OMB Issued Memoranda in 2001 (M-01-27 Guidance for Implementing EO 13211) and
2021 (M-21-12 on Furthering Compliance with Executive Order 13211).28 M-21-12 affirms and
amends M-01-27 to reflect changes in market conditions since 2001 with additional examples of
qualifying "adverse effects" of regulatory actions.
25 OMB 2010.
26 The White House 2021 and 2022.
27 U.S. EPA 2023, p. 1.
28 OMB 2001 and OMB 2021.
2-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
C i " I' «¦ "Lin uv -i 13563, "iiii-j i oving Regulation and Regulatory
Review"
Threshold: Significant regulatory actions under EO 12866 as amended by EO 14094 (reaffirms EO
12866 and includes additional provisions).29
Analyses contingent on threshold: As mentioned, EO 13563 supplements and reaffirms the
provisions of EO 12866 (as amended). EO 13563 states, "Our regulatory system must protect public
health, welfare, safety, and our environment while promoting economic growth, innovation,
competitiveness, and job creation." It emphasizes the importance of reducing regulatory costs and
burdens and maintaining flexibility and freedom of choice. The EO highlights the importance of
scientific integrity, and retrospective analyses of existing rules.
Among other directives, agencies must use best available techniques to quantify costs and benefits,
give the public meaningful opportunity to comment online, include relevant scientific and technical
findings in the rulemaking docket, consider the combined effects of their regulations on particular
sectors and industries and promote coordination across agencies. With regard to existing
regulations, EO 13563 instructs agencies to periodically review their significant regulations with
the goal of making their regulatory programs more effective or less burdensome. Per OMB
guidance, agencies are particularly encouraged to identify actions for review that will significantly
reduce existing regulatory burdens and promote economic growth and job creation. Chapter 5
includes a discussion of retrospective review and analysis; see Text Box 5.1 on Retrospective
Analysis.
Guidance: OMB issued implementation guidance in three memos: M-ll-10 February 2, 2011; M-
11-19 April 25; 2011; M-ll-25 June 14, 2011.30
C i * id1 - tive Order i, \> " "Us I- lavioral Science Insights to
I- ¦- it-1 ' IIv- Niiericn i K - ~| He"
Threshold: No specific threshold; the EO encourages agencies to "identify policies, programs, and
operations where applying behavioral science insights may yield substantial improvements in
public welfare, program outcomes, and program cost effectiveness..."
Analyses contingent on threshold: Agencies are encouraged to use behavioral science insights
when designing policies and specifically when determining access to programs, presenting
Information to the public, structuring choices within programs and designing incentives.
Guidance: The White House Social and Behavioral Sciences Team issued implementation guidance
in a memo on September 15, 2016.31 Chapter 4 of this document includes a discussion of
behavioral economics.
29 OMB 2011a.
30 See EO 13563 and OMB 2011a, 2011b, 2011c.
31 Executive Office of the President, Office of Science and Technology Policy. 2016.
2-9 Guidelines for Preparing Economic Analyses | 3rd edition
-------
iites
R-;;i.uatOi. F[>5xihi[iV\ -"t >,RF - k .¦?? ? i i r -mded by T "mall
Business Reguls n I in cemem K"n n- s i,-".PREF- »i U.S.\.
Threshold: Regulations that may have a "significant economic impact on a substantial number of
small entities," (SISNOSE), including small businesses, governments and non-profit organizations.
The RFA does not define the terms significant or substantial.
Analyses contingent on threshold: For rules that may have a SISNOSE, agencies are required to
prepare an initial regulatory flexibility analysis (IRFA) and a final regulatory flexibility analysis
(FRFA) examining potential adverse economic impacts on small entities and complying with a
number of procedural and analytical requirements to solicit and consider flexible regulatory
options that minimize adverse economic impacts on small entities and address significant issues
raised in public comments. The IRFA and FRF A, or summaries thereof, are published with the
proposed and final rules, respectively,
Guidance: The EPA has issued specific guidance for complying with RFA/SBREFA requirements
in the
"EPA Final Guidance for EPA Rulewriters: Regulatory Flexibility Act as amended by the Small
Business Regulatory Enforcement Fairness Act".32 The guidance identifies approaches for
determining whether a specific rule may have a SISNOSE but provides flexibility to use alternative
methods or reach different conclusions where appropriate in the context of a specific rule. See also
Chapter 9 of this document on economic impact analysis.
ZZ^Unfunrin, hViiTr^ r «i i| IP V « : i i :-T I* L
Threshold one (Sections 202 and 205 of UMRA): Regulatory actions that include federal
mandates "that may result in the expenditure by State, local, and tribal governments, in the
aggregate, or by the private sector, of $100 million or more (adjusted annually for inflation) in any
one year."33 An action contains a federal mandate if it imposes an enforceable duty on state, local
or tribal governments or the private sector.
Analyses contingent on threshold one: Section 202 of UMRA requires preparation of a written
statement that includes the legal authority for the action; a BCA; a distributional analysis; estimates
of macroeconomic impacts; a description of an agency's pre-proposal consultation with elected
representatives of the affected state, local or tribal governments; and a summary of concerns raised
and how they were addressed. Section 205 of UMRA requires an agency to consider a reasonable
number of regulatory alternatives and select the least costly, most cost-effective, or least
burdensome alternative that achieves the objectives of the rule, or to publish with the final rule an
explanation from the agency head of why such alternative was not chosen.
32 U.S. EPA 2006a
33 Note that the threshold in this case is adjusted annually for inflation (since enactment). Generally, the EPA
uses the U.S. Bureau of Economic Analysis gross domestic product implicit price deflator to adjust the $100
million UMRA threshold for inflation each year (e.g., the UMRA threshold was $186 million in 2024$). Note that
EO14094 increased the 12866 threshold from $100 to $200 million and added an inflation adjustment every
three years.
2-10 Guidelines for Preparing Economic Analyses | 3rd edition
-------
OMB's Circular A-4 (2023) notes that the analytical concepts under EO 12866 are similar to the
analytical concepts under UMRA, and "an analysis produced pursuant to Executive Order 12866 will
usually satisfy the analytic requirements for a written statement under the Unfunded Mandates
Reform Act."
Threshold two (Section 203 of UMRA): Regulatory requirements that might "significantly" or
"uniquely" affect small governments. Small governments include governments of cities, counties,
towns, townships, villages, school districts or special districts with a population of less than 50,000.
Requirements contingent on threshold two: Agencies must solicit involvement from, and
conduct outreach to, potentially affected elected officers of small governments (or their designated
employees) during development and implementation.
Guidance: The EPA has issued "Interim Guidance on the Unfunded Mandates Reform Act of 1995"
1995), and OMB issued a memo on "Guidance for Implementing Title II of S.l" that provides general
guidance on complying with requirements contingent on each of the two thresholds under
UMRA.34
2,2,3 The Paperwork Reduction
Threshold: Any action that requires or requests record-keeping, reporting or disclosure or
includes other information collection activities imposed upon or posed to 10 or more persons,35
other than federal agency employees.
Requirements contingent on threshold: The agency must submit an information collection
request (ICR) to OMB for review and approval and meet other procedural requirements including
public notice and opportunity for comment The ICR should: (1) describe the information to be
collected, (2) give the reason the information is needed and (3) estimate the time and cost for the
public to respond to the request.
Guidance: Both guidance and templates for completing an ICR and associated Federal Register
(FR) notices can be found on the EPA's intranet site, "ICR Center."36
2.2.4 The Foundatic i" vi li dence-Bas ' K ilicymakinf -vi- ^
1
Threshold: No specific threshold.
Requirements contingent on threshold: The Foundations for Evidence-Based Policymaking Act
of 2018 ("Evidence Act"), mandates federal evidence-building activities, where evidence is broadly
defined and includes foundational fact finding, performance measurement, policy analysis and
program evaluation.37 The act does not specify what evidence-building activities agencies should
conduct but instead calls on agencies to significantly rethink how they currently plan and organize
34 U.S. EPA 1995 and OMB 1995
35 Exceptions include "listening sessions with interested parties; asking non-standardized questions on a
particular process, theme, or issue...; directly observing the experiences of program applicants and participants."
[OMB 2022).
36 See https://work.epa.aov/icr (accessed August 1,2024, internal EPA websitej.
37 OMB 2019.
2-11 Guidelines for Preparing Economic Analyses | 3rd edition
-------
evidence building, data management and data access functions to ensure they have the evidence
they need for informed decision making. Prospective and retrospective economic analyses of
agency programs and regulations are evidence-building activities under the Evidence Act and data
used or produced in economic analyses may be subject to Title II of the Evidence Act (the Open
Government Data Act), including the requirement of being open by default
Guidance: In July 2019, OMB issued a memorandum on Phase 1 Implementation of the
Foundations for Evidence-Based Policymaking Act of 2018: Learning Agendas, Personnel and
Planning Guidance. OMB notes that in their annual evaluation plans, "agencies should also discuss
any evaluation activities that relate to its proposed regulatory actions in the Unified Agenda of
Federal Regulatory and Deregulatery Actions, recognizing that these activities often need to occur
well before the development of economically significant regulatory actions".38
38 OMB 2019, p. 34. In 2021, OMB Issued a guidance memorandum on "Evidence-Based Policymaking: Learning
Agendas and Annual Evaluation Plans" (see OMB 2021a) that reaffirms and expands on previous OMB guidance
on Learning Agendas and Annual Evaluation Plans, including OMB M-19-23.
2-12 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 2 References
2 U.S.C. 48: Unfunded Mandates Reform Act (P.L. 104-4), March 22,1995. Available at:
https://www.congress.gov/104/plaws/publ4/PLAW-104publ4.pdf (accessed October 23, 2024).
5 U.S.C. 101: Foundations for Evidence-Based Policymaking Act (P.L. 115-435), January 22, 2019. Available at:
https: //www.c0ngress.g0v/l 15/statute/STATIJTE-l 32/STATUTK-132-Pg5529.pdf (accessed October 23,
2024).
5 U.S.C. 601-612: The Regulatory Flexibility Act (P.L. 96-354), as amended by the Small Business Enforcement
Fairness Act (P.L. 104-121), March 29,1996. Available at: https: //www.govinfo.gov/content/pkg/PLAW-
104publl21/pdf/PLAW-104publl21.pdf (see page 11, Title II) (accessed October 23, 2024).
44 U.S.C. 3501: Paperwork Reduction Act (P.L. 104-13), May 22,1995. Available at:
https://www.govinfo.gOv/content/pkg/PLAW-104publl3/pdf/PLAW-104publl3.pdf (accessed October 23,
2024).
Executive Office of the President, Office of Science and Technology Policy. 2016. Implementation Guidance for
Executive Order 13707: Using Behavioral Science Insights to Better Serve the American People, September
15, 2016. Available at:
https://sbst.gOv/download/Executive%200rder%2013707%20Implementation%20Guidance.pdf (accessed
October 23, 2024).
Executive Order 12866: Regulatory Planning and Review, Section 1(a), October 4,1993. Available at:
https://www.archives.gov/files/federal-register/ executive-orders/pdf/12866.pdf (accessed October 23,
2024).
Executive Order 12898: Federal Actions to Address Environmental Justice in Minority Populations and Low-
Income Populations, February 11,1994. Available at: https: //www.archives.gov/files/federal-
register/executive-orders/pdf/12898.pdf (accessed October 23, 2024).
Executive Order 13045: Protection of Children from Environmental Health Risks and safety Risks, April 23,
1997. Available at: https: //www.govinfo.gov/content/pkg/FR-1997-04-23/pdf/97-10695.pdf (accessed
October 23, 2024).
Executive Order 13132: Federalism, August 10,1999. Available at:
https://www.govinfo.gOv/content/pkg/FR-1999-08-10/pdf/99-20729.pdf (accessed October 23, 2024).
Executive Order 13175: Consultation and Coordination with Indian Tribal Governments, November 6,2000.
Available at: https: //www.federalregister.gOv/documents/2000/ll /09 /00-29003 /consultation-and-
coordination-with-indian-tribal-governments (accessed October 23, 2024).
Executive Order 13211: Actions Concerning Regulations that Significant Affect Energy Supply, Distribution,
Use, May 18, 2001. Available at: https://www.govinfo.gov/content/pkg/WCPD-2001-05-21 /pdf/WCPD-
2001-05-21-Pg769.pdf (accessed October 23,2024).
Executive Order 13563: Improving Regulation and Regulatory Review, January 18,2011. Available at:
https://obamawhitehouse.archives.gOv/the-press-office/2011/01/18/executive-order-13563-improving-
regulation-and-regulatorv-review (accessed October 23, 2024).
Executive Order 13707: Using Behavioral Science Insights to Better Serve the American People, September
15,2015. Available at: https://www.federalregister.gov/documents/2015/09/18/2015-2363Q/using-
behavioral-science-insights-to-better-serve-the-american-people (accessed October 23,2024).
Executive Order 14094: Modernizing Regulatory Review, April 6, 2023. Available at:
https://www.federalregister.gOv/documents/2023/04/ll/2023-07760/modernizing-regulatory-review
(accessed October 23, 2024).
Executive Order 14096: Revitalizing Our Nation's Commitment to Environmental Justice for All, April 21,
2023. Available at: https: //www.federalregister.gov/documents/2023 /04/26/2023-08955 /revitalizing-our-
nations-commitment-to-environmental-iustice-for-all (accessed October 23, 2024).
2-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
OMB. 1995. Memorandum M-95-09, Guidance for Implementing Title II of S. 1, March 31,1995. Available at:
https://www.whitehouse.gov/wp-content/uploads/legacy drupal files/omb/memoranda/1995-1998/m95-
09.pdf (accessed October 23, 2024).
OMB. 2001. Memoranda 01-27, Guidance for Implementing E.0.13211, July 13, 2001. Available at:
https://www.whitehouse.gOv/wp-content/uploads/2017/ll/2001-M-01-27-Guidance-for-Implementing-
E.0.-13211.pdf (accessed October 23, 2024).
OMB. 2010. Guidance for Implementing E.0.13175, "Consultation and Coordination with Indian Tribal
Governments," July 30, 2010. Available at: https://work.epa.gov/sites/default/files/2021-07/trihalguidance-
omb-csunstein-07-30-10.pdf (accessed October 23, 2024).
OMB. 2011a. Executive Order 13563, Improving Regulation and Regulatory Review, February 2,2011.
Available at: https://obamawhitehouse.archives.gov/sites /default/files /omb/memoranda/2 Oil /ml 1-
10.pdf (accessed October 23, 2024).
OMB. 2011b. Retrospective Analysis of Existing Significant Regulations, April 25,2011. Available at
https://ohamawhitehouse.archives.gov/sites /default/files /omb /memoranda /2 011 /ml 1 -19.pdf (accessed
October 23, 2024).
OMB. 2011c. Final Plans for Retrospective Analysis of Existing Rules, June 14, 2011. Available at
https://obamawhitehouse.archives.gov/sites/default/files/omb/memoranda/2011/mll-25.pdf (accessed
October 23, 2024).
OMB. 2019. Phase 1 Implementation of the Foundations for Evidence-Based Policymaking Act of 2018:
Learning Agendas, Personnel, and Planning Guidance, July 10, 2019. Available at:
https://www.whitehouse.gov/wp-content/uploads/2019/07/M-19-23.pdf (accessed October 23, 2024).
OMB. 2021. Memorandum M-21-12, Furthering Compliance with Executive Order 13211, January 13, 2021.
Available at: https: //www.whitehouse.gov/wp-content/uploads/2021/01/M-21-12.pdf (accessed October
23,2024).
OMB. 2021a. Memorandum M-21-27, Evidence-Based Policymaking: Learning Agendas and Annual
Evaluation Plans, June 20, 2021. Available at: https: //www.whitehouse.gov/wp-
content/uploads/2021 /06/M-21-27.pdf (accessed October 23, 2024).
OMB. 2022. Memorandum M-22-10, Improving Access to Public Benefits Programs Through the Paperwork
Reduction Act, April 13, 2022. Available at: https: //www.whitehouse.gov/wp-content/uploads/2022 /04/M-
22-10.pdf (accessed November 18, 2024).
OMB. 2023. Circular A-4, Regulatory Analysis, November 9, 2023. Available at:
https://www.whitehouse.gov/wp-content/uploads/2023/ll/CircularA-4.pdf (accessed October 23, 2024).
OMB. 2023a. OMB Circular No. A-4: Explanation and Response to Public Input, November 9, 2023. Available
at: https: //www.whitehouse.gov/wp-content/uploads/2023 /ll /CircularA-4Explanation.pdf (accessed
October 22, 2024).
The White House. 2021. Memorandum on Tribal Consultation and Strengthening Nation-to-Nation
Relationships, January 26, 2021. Available at: https: //www.whitehouse.gov/briefing-room/presidential-
actions/2021/01/26/memorandum-on-tribal-consultation-and-strengthening-nation-to-nation-
relationships/ (accessed October 23, 2024).
The White House. 2022. Memorandum on Uniform Standards for Tribal Consultation, November 30, 2022.
Available at: https: //www.whitehouse.gov/hriefing-room/presidential-actions/2022 /ll /30/memorandum-
on-uniform-standards-for-tribal-consultation/ (accessed October 23, 2024).
The White House. 2023. Fact Sheet: President Biden Signs Executive Order to Revitalize Our Nation's
Commitment to Environmental Justice for All, April 21,2023. Available at:
https://www.whitehouse.gOv/hriefing-room/statements-releases/2023/04/21/fact-sheet-president-hiden-
signs-executive-order-to-revitalize-our-nations-commitment-to-environmental-iustice-for-all/ (accessed
October 23, 2024).
2-14 Guidelines for Preparing Economic Analyses | 3rd edition
-------
U.S. EPA. 1995. Interim Guidance on the Unfunded Mandates Reform Act of 1995, memorandum from the
Office of General Counsel, March 23,1995. Available at: https://work.epa.gov/sites/default/files/2021-
07/umraguidance-03-2 5-95.pdf (accessed October 23,2024, internal EPA document).
U.S. EPA. 2003a. Children's Health Valuation Handbook, October 2003. Available at:
https: //www.epa.gOv/sites/production/files/2017-09/documents/ee-0474-01.pdf (accessed October 23,
2024).
U.S. EPA. 2003b. Office of General Counsel (OGC) Checklist of Statutory and Executive Order Reviews, July
2003. Available at: http://intranet.epa.gov/ogc/memoranda/checklist703.pdf (accessed October 23, 2024,
internal EPA document).
U.S. EPA. 2005. OGC Desktop Reference Guide: Partial Summary of Cross-Cutting Statutory and Executive
Order Reviews that May Apply to Agency Rulemakings, June 2005. Available at:
http://intranet.epa.gov/ogc/memoranda/desktoprefguide.pdf (accessed October 23, 2024, internal EPA
document).
U.S. EPA. 2006a. EPA's Action Development Process: EPA Final Guidance for EPA Rulewriters: Regulatory
Flexibility Act as amended by the Small Business Regulatory Enforcement Fairness Act, November 2006.
Available at: https: //www.epa.gov/sites/production/files/2015-06/documents/guidance-regflexact.pdf
(accessed October 23, 2024).
U.S. EPA. 2008. EPA's Action Development Process: Guidance on Executive Order 13132: Federalism,
November 2008. Available at: https: //work.epa.gov/sites/default/files/2021-07/federalismguidell-00-
08.pdf (accessed October 3, 2024, internal EPA document).
U.S. EPA. 2016. Technical Guidance for Assessing Environmental Justice in Regulatory Actions, June 2016.
Available at: https://www.epa.gov/sites/production/files/2016-06/documents/eitg 5 6 16 v5.1.pdf
(accessed October 23, 2024).
U.S. EPA. 2024. EPA's Action Development Process: Guidance for EPA Staff on Developing Quality Actions,
October 2024. Available at: https://work.epa.gov/sites/default/files/2024-
10/ADP%20Guidance%202024%20Update%2OFinal.pdf (accessed November 5,2024, internal EPA
document).
U.S. EPA. 2023. EPA Policy on Consultation with Indian Tribes, December 7, 2023. Available at:
https://www.epa.gov/system/files/documents/2Q23-12/epa-policy-on-consultation-with-indian-tribes-
2023.pdf (accessed October 23,2024).
U.S. EPA. 2024. Guide to Considering Children's Health When Developing EPA Actions: Implementing EPA's
Policy on Children's Health and Executive Order 13045, February 2024. Available at:
https://work.epa.gov/sites/default/files/2024-
02/EPA%20ADP%20CEH%20Guide%20FINAL%20Februarv%202024.pdf (accessed October 22, 2024, internal EPA
document).
The White House. 2021. Memorandum on Tribal Consultation and Strengthening Nation-to-Nation
Relationships, January 26, 2021. Available at: https: //www.whitehouse.gov/briefing-room/presidential-
actions/2021/01/26/memorandum-on-tribal-consultation-and-strengthening-nation-to-nation-
relationships/ (accessed October 23, 2024).
The White House. 2022. Memorandum on Uniform Standards for Tribal Consultation, November 30, 2022.
Available at: https: //www.whitehouse.gov/briefing-room/presidential-actions/2022 /ll /30/memorandum-
on-uniform-standards-for-tribal-consultation/ (accessed October 23, 2024).
The White House. 2023. Fact Sheet: President Biden Signs Executive Order to Revitalize Our Nation's
Commitment to Environmental Justice for All, April 21,2023. Available at:
https: / / www.whitehouse.gov/briefing-room/statements-releases /2023/04/21/ fact-sheet-president-biden-
signs-executive-order-to-revitalize-our-nations-commitment-to-environmental-iustice-for-all/ (accessed
October 23, 2024).
2-15 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 3 - Need for Regulatory
Action and Evaluation of Policy
The essential components of an economic analysis are [1] a clear statement of
the need for regulatory action describing the problem to be addressed by the
policy and [2] a detailed evaluation of policy options. The statement of need
should include a description of the market, institutional, or behavioral
distortions being addressed, an explanation of why the market and other
institutions have failed to correct these problems, and a justification for
federal action to address them.
The economic analysis should consider and evaluate multiple policy options
that address the environmental problem. This is true for analyses of proposed
and final rules, even when the Agency has settled on a specific option. When
identifying policy options, the analysis should describe any statutory or
judicial requirements that must be considered. The options should include
those permissible under the relevant statutory authority and may include
those that are unavailable but with other advantages. The options may differ
in their levels of stringency, compliance dates, and requirements based on
entity size and location, or they may represent entirely different regulatory
approaches. Detailing possible options is a necessary step in establishing why
the selected option is the appropriate choice.
A i The I Uwf
Consistent with Executive Order (EO) 12866 and Office of Management and Budget (OMB) Circular
A-4 (2023), each economic analysis should include a statement of need that provides: (1) a clear
description of the problem being addressed and the significance of that problem, (2) the failures of
private markets or public institutions that warrant agency action, and (3) an assessment of whether
Federal regulation is the best way to correct the problem.1 This statement sets the stage for the
1 EO 12866 states, "Federal agencies should promulgate only such regulations as are required by law, are
necessary to interpret the law, or are made necessary by compelling need, such as material failures of private
markets to protect or improve the health and safety of the public, the environment, or the well-being of the
American people..." (emphasis added). The Office of Management and Budget's guidance for how to comply with
EO 12866, Circular A-4 (OMB 2023), provides recommendations to federal agencies on the development of
economic analyses supporting regulatory actions. OMB (2023, p. 14) states that "including a summary in
regulatory analyses of the needs being addressed may provide useful background and help ensure that the
description of the needs informs the scope of the analyses (and vice versa) to the extent relevant, appropriate,
and consistent with the best available evidence and best practices for objective analysis."
3-1 Guidelines for Preparing Economic Analyses | 3rd edition
-------
subsequent benefit-cost analysis (BCA) and allows one to judge whether the policy adequately
addresses the problem.
o. i". i" Ff-'hl-ifTT L'-.:'Tiption
The statement of need should begin with a brief review of the problem or public need to be
addressed by the policy. While not always the case, the compelling public need for U.S
Environmental Protection Agency (EPA) regulations is generally to address an environmental
problem. In this case, the following considerations are often relevant:
The primary environmental contaminants causing the problem and their concentrations.
The media through which exposures or damages take place.
Private and public sector sources responsible for creating the problem.
Human exposures involved and the health effects due to those exposures.
Non-human resources affected and the resulting outcome.
The expected change in the environmental problem over time, absent additional regulation.
Available and potential abatement and mitigation techniques and technologies.
The amount or proportion of the environmental problem likely to be corrected by federal
action.
Any existing state, local and other federal activities that partially or fully address the
problem.
, i C I-1 - isons for R - f itory Action
After defining the problem, the statement of need should examine the reasons why the market and
other public and private sector institutions have failed to correct it That is, it should define the
reason or social purpose for the regulatory action. This identification is an important component of
policy development because the underlying failure itself often suggests the most appropriate
remedy for the problem (see Chapter 4). A regulation can be promulgated for a number of social
purposes. For pollution problems, the social purpose is commonly to correct a "market failure."
Other potential reasons for regulatory action include addressing behavioral biases; improving the
efficiency and effectiveness of government operations; promoting distributional fairness and
advancing equity; and protecting civil rights and civil liberties.
I . I I | - l \ -
A market failure occurs when the allocation of goods and services by the free market is not
economically efficient The most common causes of market failure are externalities, overutilization
of common property resources, under-provision of public goods, market power, and inadequate or
asymmetric information.2 While there are other social purposes for government regulation,
2 For further discussion of market failure, types of market failures and externalities see Scitovsky (1954), Bator
(1958), Buchanan and Stubblebine (1962), Mishan (1969), Baumol and Oates (1988), Comes and Sandler
(1996), Hanley etal. (2019), Perman etal. (2003), and Tietenberg and Lewis (2014). OMB (2023) also describes
different categories of market failure as well as other reasons for regulation. Section A-2 of these Guidelines
provides further discussion of externalities.
3-2 Guidelines for Preparing Economic Analyses | 3rd edition
-------
correcting a market failure, particularly addressing an externality, is most likely the driver behind
environmental policy.
As defined by Keohane and Olmstead (2016), "An externality results when the actions of one
individual (or firm) have a direct, unintentional, and uncompensated effect on the well-being of
other individuals or the profits of other firms."3 Technically, externalities occur when the outputs
and inputs chosen by one individual enter the utility or production function of another without
passing through markets or contracts. Put another way, externalities occur when the market does
not account for the effect of one party's activities on another party's well-being without
compensation.
Consider, for example, a factory that produces smoke as a by-product of manufacturing that, in turn,
affects individuals living downwind. The factory does not weigh the costs of its actions on the
downwind community when making production decisions. Although the factory imposes an
externality on the downwind community, the mere existence of an externality is not enough to
justify a regulation. Under certain conditions, namely, the ability to bargain, availability of complete
information, and presence of low transaction costs, externalities can be internalized by the free
market (Coase 1960). Text Box 3.1 describes this Coasian solution in more detail.
It is important to differentiate externalities from other external effects when an individual or firm is
affected by the behavior of others. For example, a negative outcome caused by another individual is
not an externality if the affected individual rationally and willingly accepts the risk of that outcome
through a private transaction between them. This may occur when a worker accepts a job with a
greater risk of injury in exchange for a higher wage. However, this assumes complete and perfect
markets with full information and that the transaction stipulations reflect and incorporate the
expected risk such that no externality is associated with increased risk of injury. Similarly, external
effects that function through the price system (e.g., higher prices faced by certain consumers
because of rising demand) or zero-sum transfers from one person to another (e.g., through taxes or
redistribution of consumer and producer surplus) are not externalities by definition and do not
constitute a market failure. For example, if person A outbids person B in an auction, person B may
be made worse off than if they had won the auction but were unwilling to pay the higher bid. This is
a result of the price system working to ensure scarce resources go to those willing to pay the most
for them, avoiding an inefficient allocation of resources.4
3 Keohane and Olmstead (2016) go on to say, "Note three keywords in the definition: direct, unintentional, and
uncompensated. For example, because your health and happiness depend in part on how clean the air is,
automobile drivers have a direct effect on your well-being. Unintentional is included in the definition to rule out
acts of spite or malice. (It is the effect rather than the action that is unintentional. I may decide deliberately to
use a gasoline-powered lawnmower, without the intent of my action being to pollute the air or disturb the
neighbors.) Finally, uncompensated implies that the responsible actor does not compensate the damaged parties
(or is not fined) for his actions. This rules out market transactions or bargaining between individuals" [emphasis
in original].
4 External effects operating through the price system are referred to as pecuniary externalities.
3-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 3.1 - Coasian Solution
Government intervention for the control of environmental externalities may not be necessary if
parties can work out an agreement between themselves. Coase (1960) outlined conditions
under which transaction costs are low enough that a private agreement between affected
parties might result in the attainment of a welfare-maximizing level of pollution without
government intervention. First, property rights must be fully and clearly defined and
transferable. In situations where the resource in question is not "owned" by anyone, there is no
ability to negotiate, and the offending party can "free ride," or continue to pollute, without facing
the costs of its behavior, and a Coasian solution is not possible.
When property rights can be defined and have been allocated, a welfare-maximizing solution
can be reached regardless of which party is assigned the property rights, although the
distribution of the gains from bargaining will differ. Take for example a farm whose pesticide
application to its crops pollutes the well water of nearby homeowners. If property rights of the
watershed are assigned to the homeowners, and information is available to them about
potential damages from the pollution, then the farm may negotiate with the homeowners about
its continued use of the pesticide. Potential compensation from the farm to the homeowners
agreed upon through such negotiations need not be in the form of cash but could involve
investments to reduce the water contamination or land swaps, (e.g., Deryugina etal. 2021), If
property rights of the watershed are given to the farm, then the homeowners could negotiate to
pay the farm to stop applying the pesticide.
The effectiveness of such agreements is contingent on meeting additional conditions: bargaining
must be possible, damages must be known, and transaction costs must be low. These conditions
are more likely to be met when there are only a small number of individuals involved. If either
party is unwilling to negotiate or faces high transaction costs, then no private agreement will be
reached. Asymmetric information or bargaining power can also hinder a socially optimal
solution. Going back to the example, consider a case where there are many farms in the
watershed using the pesticide on their crops, and it may be difficult to identify the relative
contribution of each farm's effluent on damages experienced by the homeowner. Clearly,
homeowners would have more difficulty in negotiating an agreement with many farms than
they would in negotiating with a single farm. However, technological advances in data sharing
and networking can increase the likelihood of finding a Coasian solution. Advances in internet
search and the availability of monitoring devices that can lower transactions costs and reduce
information asymmetries, and social networks can make it easier for groups to communicate
and arrive at bargained solutions. Deryugina etal. (2021) discusses several Coasian solutions to
actual environmental problems.
When left unaddressed, externalities prevent the market from achieving economic efficiency and
reduce economic welfare. This can occur in the presence of high transaction costs that make it
difficult for private parties to internalize the cost of damages through bargaining, legal action, or
other means such that both parties are no worse off. High transaction costs may result when
activities that pose environmental risks are difficult to link to the resulting damages because they
occur over long periods or occur in a different location than where the pollution originates.5 If these
5 The concept of an externality is closely tied to the concept of a public good, which is a good that either can be
used simultaneously by many (i.e., nonrival) or that is difficult to prevent others from using (i.e., nonexcludable).
The environment is a classic example of a public good.
3-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
high transaction costs are overcome and the parties can internalize the cost of the damage, then
scarce resources will again be efficiently allocated by the market. If the cost of damages cannot be
internalized, then government intervention may be necessary to fully address the externality.6
But even the presence of an unaddressed externality is not enough to justify a regulation; what is
required is a compelling need for government intervention at any level of government (federal,
state or local). That is, there must be some form of evidence that government intervention can
improve economic welfare.7 Government regulation may not be warranted if the benefits of
regulation do not justify the costs. Circumstances where this may occur include when a regulation
designed to reduce a negative externality (e.g., direct emission controls) exacerbates pre-existing
distortions. In this case, government intervention could make things worse.
There should also be some evidence that the externality will persist If the market will correct itself
through innovation and technological change or the externality will cease to exist through private
transactions, then government intervention may not be necessary. A BCA can determine whether
government intervention to remove the externality can improve economic efficiency even if the
externality only exists for a short time absent additional regulation (i.e., it is resolved in the
baseline after the short time). Furthermore, even if an externality warrants government
intervention, it may not warrant direct, prescriptive regulation. Some externalities maybe
addressed more efficiently through other means such as providing information, requiring firms to
carry insurance, defining legal liability, or assigning property rights. The nature of the externality
may determine the best approach for government action (see Chapter 4).
,1 her Social i i _ ttory Action
While correcting a market failure, particularly an externality, is the most common justification for
environmental regulation, there are other underlying institutional or behavioral distortions that
may justify regulatory action or government intervention. These include addressing behavioral
biases; improving the efficiency and effectiveness of government operations, promoting
distributional fairness and advancing equity; and protecting civil rights and civil liberties.
Additionally, regulation may be justified for multiple interconnected reasons, such as addressing a
market failure and promoting distributional fairness.
If the social purpose of a regulation is not to address a market failure (e.g., to improve Agency
processes or solely to define a statutory term), then the statement of need still should include a
description of the problem being addressed and an explanation of why government action is
necessary to address this problem. If the purpose of a regulation is to protect sensitive
subpopulations or address other distributional impacts rather than, or in addition to, addressing a
market failure, that should be stated in the statement of need.
One possible social purpose is addressing behavioral biases. The behavioral economics literature
has documented situations in which individuals appear to act in ways that are inconsistent with
6 As shown in Section A-5, there is an optimal level at which an externality should be addressed by a regulation.
At this optimal level, further reduction in the externality is inefficient. Therefore, in the simple case where there is
only one externality and it is controlled by an existing regulation, the existing regulation is not sufficiently
stringent if the additional benefits from reducing the externality further will exceed the additional cost, and
therefore additional regulation would be net-beneficial. Similarly, an existing regulation may be too stringent
such that additional regulation would lead to negative net benefits.
7 husk (2013) provides a useful nine-point checklist for externalities that require prescriptive regulation.
3-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
rational choice, sometime referred to as "behavioral failures" or "behavioral anomalies" (Shogren
and Taylor 2008). In such situations, it is possible that government intervention could lead to a
more efficient allocation of resources than the free market outcome. However, because the mission
of EPA is to protect human health and the environment, behavioral failure absent an environmental
externality is not a typical justification for regulation at EPA. If insights from behavioral economics
are used as a justification for regulation, analysts should provide robust empirical evidence
supporting the existence of behavioral anomalies in the affected market and rules out other
explanations consistent with rational behavior, such as hidden costs. Chapter 4 includes more
discussion of behavioral economics and its implication for policy design.
ral Action
The final component of the statement of need for the regulatory action is an evaluation and
explanation of why a federal remedy is preferable to actions by private and other public-sector
entities, such as the judicial system or state and local governments.8 Federal involvement is often
required for environmental problems that cross jurisdictional boundaries (e.g., when pollution in
one state affects the population of another). In some cases, federal involvement is mandated by
statute or directed by an EO as described in Chapter 2. A federal regulation could be justified by
comparing its expected performance to realistic alternatives that rely on other institutional
arrangements. This component of the statement of need for regulatory action, justifying federal
regulation, should verify that the policy action is necessary, within the jurisdiction of the relevant
statutory authorities, and yields results that will be preferable to no action. Finally, the statement of
need should identify those aspects of the regulation necessitated by statutory requirements and
those that are discretionary.
< 1 K mi l> 11 ¦ nl 11 \ |, tv> B
3.2,1 Need to Assess llultif rtions
Each analysis should evaluate multiple policy options. Following the statement of need, the
economic analysis should identify and describe in detail all policy options or potential regulatory
alternatives that were considered. This includes clearly explaining which options were selected for
emphasis and further analysis and why other important options were not Since the BCA informs
the public, stakeholders and Congress and other decision makers of the effects of the policy
assessing a robust set of policy options is important
The identification of policy options should describe any statutory or judicial requirements that
must be considered when designing the regulation, how those requirements may influence the
8 As discussed in Chapter 2, EO 13132, "Federalism," describes principles of federalism and identifies
requirements for federal preemption of state or local law. Also, there is a robust economics literature on the pros
and cons of regulating environmental quality at different jurisdictional levels that may be informative when
determining whether federal regulation is appropriate as a substitute or complement to state or local regulation
(e.g., Oates 2002). See also Circular A-4 (OMB 2023) on "Showing Whether Federal Regulation Is the Best Way to
Solve the Problem."
Guidelines for Preparing Economic Analyses | 3rd edition
-------
options considered and how the proposed or finalized option satisfies them.9 For example, the
description should identify any economic considerations (e.g., costs incurred by regulated entities)
and discretionary provisions in the statute that may be used to shape the form and stringency of the
regulation. The analysis may also identify options that are more efficient or cost-effective even if
the regulatory approaches may be prohibited by statutory or judicial requirements (see also OMB
2023). For example, the Supreme Court has held that the Clean Air Act requires that National
Ambient Air Quality Standards be set based on health or welfare considerations only; the Act bars
EPA from considering the costs of implementing them when setting the standards.
At a minimum, the economic analysis should fully assess and present three options for
consideration: the proposed or finalized option; a more stringent option; and a less stringent
one.1011 The incremental benefits and costs for each option, as well as other important criteria (e.g.,
distributional consequences), should be compared across the options. Measuring the incremental
benefits and costs of successively more stringent regulatory options provides a clear indication of
the most economically efficient option, provided important benefits and costs can be quantified and
monetized. If options cannot be characterized by regulatory stringency (e.g., they differ by the
provisions included), the economic analysis should still analyze at least three options, including one
that achieves greater benefits and one that costs less than the proposed or finalized option (see also
OMB 2023).
Assessing at least three options applies in any circumstance. It is not adequate to evaluate only the
selected option, even for a final rule that establishes the option to be promulgated. Similarly, in
cases where the design of the regulation is dictated by statute, presenting multiple options is still
necessary when the regulation is proposed or finalized -- even though the Agency may have no
discretion in its design. Assessing multiple options helps inform the public about the anticipated
benefits and costs of the Agency's final action compared to options not pursued, it is imperative that
the analysis assesses multiple options.
The analysis should also consider whether there are alternatives to federal regulation that may
address the market failure or other regulatory objective (e.g., distributional concern) more
efficiently. Alternatives may include using existing product liability rules to encourage firms to
internalize the costs of the environmental damages, introducing market-oriented approaches such
as fees, penalties, subsidies, marketable permits, and offsets, or the potential for state or local
regulation. Even if options are not available due to statutory restrictions, the economic analysis
should discuss the limitations of the statutory requirements and, if possible, estimate the
9 Often, consideration of different regulatory options is required or encouraged by statute (e.g., different
stringencies of emissions standards). Any qualitative or quantitative analysis that supports these considerations
should be summarized in the BCA, even if estimates of the benefits and costs of those options were not produced.
10 An exception may occur if the proposed or finalized option is at or near the limit of technical feasibility, in
which case the analysis might not need to examine a more stringent option. However, it is possible that even if
abatement of an environmental contaminant using on-site controls is technically infeasible, the value of the good
or activity whose production creates the contaminant may be less than the harm the contaminant causes. In such
circumstances, a more stringent option that shifts production away from the good or activity should be
evaluated.
11 While developing a regulation, the decision maker may choose the more stringent or less stringent option
after weighing the results of the analysis. Doing so demonstrates the usefulness of the analysis. In this
circumstance, the analysis should include an additional option to satisjy this guidance if time allows. If there is
insufficient time to evaluate an additional option, the other two options should still be presented, and the
analysis should explain why the central option was not selected.
3-7 Guidelines for Preparing Economic Analyses | 3rd edition
-------
opportunity cost of not being allowed to pursue these options. There is no prohibition against
analyzing these options.12
When a rule includes several distinct regulatory provisions, the benefits and costs of each provision
should be analyzed both separately and jointly (i.e., as a package of provisions).13 Doing so may
yield insights such as identifying unnecessary or otherwise undesirable regulatory requirements.
For example, evaluating provisions independently may identify those provisions for which their
costs exceed their benefits, even when the benefits of a regulation in its entirety exceed its costs.
Jointly analyzing multiple provisions becomes more complicated when the existence of one
provision affects the benefits or costs arising from another. Even so, it is still possible to evaluate a
specific provision by estimating the net benefits of a regulatory option with and without that
provision.
Ultimately, the number of options to evaluate and their design is a matter of judgment, but the
analysis should strive for a balance between thoroughness and analytic capacity. Realistically,
analyzing all possible combinations of provisions is impractical if their number is large and
interactions between provisions are common. Generally, some options can be eliminated through a
preliminary and less rigorous analysis, leaving a more manageable number to be evaluated in the
formal BCA. For a proposed rule, it may be useful to provide an economic analysis that illuminates
important tradeoffs associated with key specific aspects of the rule on which the Agency is soliciting
comment
r:.Z,2 Policy Desirn Options
The analysis should carefully describe the policy design being evaluated and, when the costs or
benefits vary substantially with alternative policy designs, assess alternative design options.14The
policy design includes the core regulatory approach as well as key features of its implementation
and structure. Prescriptive regulation (e.g., technology, design, or performance standards) is
common in Federal environmental regulations. Performance standards, which specify the allowable
limit but not the way regulated entities must achieve that limit, are generally less costly than
standards that dictate technologies or techniques. Economic analyses may include assessments of
policy designs that currently are not statutorily allowed to highlight potential tradeoffs between the
required approach and other more desirable approaches (for example, more flexible market-based
approaches such as emissions taxes and allowance trading systems that may be prohibited).
12 OMB Circular A-4 (2023) states, "Your analysis of the effects of the regulation should not presuppose that
there is a need for the regulation, and your analysis of the potential need for the regulation should not
presuppose the effectiveness of your regulation." (p.14) and "If legal or other constraints prevent the selection of
a regulatory action that best satisfies the philosophy and principles of Executive Orders 12866, you may consider
identifying these constraints and estimating their opportunity cost (and effects more generally). Such
information may, for example, be useful to Congress under the Regulatory Right-to-Know Act or in considering
statutory reforms." (pp.22-23)
13 When the benefits or costs of a regulation or one of its provisions are highly uncertain, an option may include
a voluntary program or pilot project or additional data collection prior to regulation. See Chapter 4 for further
discussion of these options.
14 Chapter 4 provides a detailed description of different regulatory approaches, including a detailed discussion
of considerations for selecting among different regulatory approaches.
3-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Aspects of the market failure may help identify which types of regulatory approaches to consider. A
key principle in the design of environmental regulations is that the regulatory structure and
incentives should align with the environmental objective. For example, if the effect of emissions on
human health depends on the proximity to the emission, then generally the optimal regulation
should more stringently control emissions from emitters that are closer to population centers.
Another example is that regulations should impose requirements on emissions rather than the
inputs associated with the emissions provided emissions monitoring costs are not too high relative
to the costs of monitoring input use.
Evaluating regulatory features other than stringency and regulatory approach may also help
identify better policy designs. Options that vary these regulatory features, both alone and in
combination, should be considered (see also OMB 2023). These features include the entities that
are subject to the regulation.15 By varying policy design features in the options considered, the
analysis may identify approaches that increase net benefits or reduce the impact on certain groups.
These features include but are not limited to:
Compliance dates: Providing more time before a regulation takes effect may reduce costs
by allowing the regulated entities additional planning time, which can be weighed against a
possible reduction in benefits.
Enforcement methods: Alternatives include regular on-site inspections, random
monitoring, periodic reporting and noncompliance penalties, which may have different
costs and efficacy.
Requirements for different-sized firms or facilities: In some cases, small firms or
facilities may face proportionately higher compliance costs, especially if there are large,
fixed compliance costs.16 When a market-based approach cannot be used, varying the
regulatory stringency or pollution requirement by firm size may increase economic
efficiency.
Requirements for different geographic regions: Differentiating requirements by region
may be desirable if there is significant regional variation in pollution reduction benefits or
the costs of compliance.
Requirement for facilities of different vintages: New facilities may face lower costs of
compliance than older facilities because of the relative ease with which abatement methods
can be integrated into their production processes. Also, pollution control investments may
be in use longer at new facilities, and therefore may yield greater benefits over time.
15 The coverage of a regulation may include different market sectors or different entities within a sector.
Generally, the statutes that EPA implements identify the groups of similar emitting sources that would be subject
to a particular regulation, although there is often some flexibility in defining the types of entities included in
each group, the requirements for different subgroups and some regulatory choices may influence subsequent
requirements for multiple sectors.
16 Chapter 2 describes analysis for examining potential adverse economic impacts on small entities and
procedures to solicit and consider flexible regulatory options that minimize adverse economic impacts on small
entities under the Regulatory Flexibility Act ofl 980 (RFA), as amended by The Small Business Regulatory
Enforcement Fairness Act of1996 (SBREFA) (5 U.S.C. 601-612). These are required for rules with a "significant
economic impact on a substantial number of small entities." Chapter 9 outlines the analytic tasks associated with
complying with the RFA.
3-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
However, stricter requirements for new facilities than old ones may lead to inefficient
investment patterns (e.g., firms delaying investments to avoid stricter regulation).17
It is important to account for and present both the total benefits and costs of each option and the
incremental benefits and costs among the options. As discussed in depth in Chapter 5, it is
important to account for all of the benefits and costs for all policy options because any options
where benefits exceed costs is an improvement in economic efficiency according to the potential
Pareto principle.1819 By this standard, selecting any option with positive net benefits would
improve societal welfare. However, the most economically efficient option is the one that produces
the largest increase in net benefits. While the option with the highest net benefits is obvious from
the presentation of total benefits and costs, presenting the incremental benefits and costs of each
option compared to the next less-stringent alternative helps to indicate if there is an even more
economically efficient option other than those being considered. In general, economic efficiency is
maximized (i.e., net benefits are highest) when incremental benefits are equal to incremental
costs.20
Determining which option is the most economically efficient may be more challenging when there
are consequences that society would be willing to pay for (or avoid) but that cannot be quantified
or monetized. As discussed In Chapter 5 and elsewhere in these Guidelines, effects should be
quantified, even if they cannot be monetized, and discussed qualitatively if not Differences between
consequences that are not quantified or monetized should also be compared among policy options.
In particular, different policy options may have different distributional impacts, even without
significantly changing the benefits and costs of the regulation, and this difference may not be
obvious when only evaluating the total costs and benefits. It may be important to consider which
regulatory alternatives may generate important differences in distributional effects.
Furthermore, carefully detailing the sources of the benefits and costs of a rule, rather than looking
only at its total net benefits, may help identify other policy options. For example, a regulation that is
designed to reduce releases of one contaminant may result in an increase or decrease in releases of
other contaminants. Again, the benefits and costs from all the changes in contaminant levels should
be accounted for in a BCA. However, when an action produces benefits from reductions in
contaminants other than those related to the statutory objective of the regulation, and the benefits
associated with these reductions in other contaminants are a large share of total benefits, or when
net-benefits would be negative without them, then the analysis should identify other policy options
that include directly regulating those contaminants.21
17 Chapter 8 provides additional discussion of the advantages and disadvantages of vintage-differentiated
regulations. Chapter 4 describes regulatory designs that can address some of the disadvantages.
18 The potential Pareto principle, or the compensation principle, states that economic welfare is improved by an
action if the benefits of the action outweigh the costs (provided both benefits and costs can be measured
accurately) because the gainers (those who benefit) could, theoretically, compensate the losers (those who bear
the costs) and still be better off. Section A.3 of these Guidelines provides a further description of the potential
Pareto principle.
19 Executive Order 12866 and OMB (2023) also consistently affirm that all benefits and costs should be assessed
in BCA of regulatory actions.
20 The proposed or finalized option should also be reasonably robust to alternative potential baseline conditions.
See Section 5.6 on uncertainty.
21 The statutory objective of the regulation is the specific objective of the statutory provision under which the
regulation is promulgated.
3-10 Guidelines for Preparing Economic Analyses | 3rd edition
-------
In addition, an analysis of a policy option in which the other contaminant(s) are regulated directly,
either separately or simultaneously with the regulation being analyzed, may be warranted.22 If
there are interactions in the control of contaminants, the most economically-efficient approach to
their control requires simultaneously determining the appropriate policy design for each (e.g.,
Tietenberg, 1973). If there are important interactions in the control of multiple contaminants,
options that jointly consider the appropriate design for each should be identified and may be
analyzed, even if such considerations are not currently permissible. Correspondingly, there may be
costs from increases in other environmental contaminants that are not associated with the
statutory objective of the regulation. 23 If the effects of these increases due to the regulation are
large, analysis of options to mitigate them may be warranted.
22 Chapter 5 provides further discussion and guidance on how to treat in an economic analysis those benefits
from environmental contaminants other than those related to the statutory objective.
23 Such costs attributable to increases in other pollutants (and other environmental contaminants) should be
accounted for even if future regulation might reduce them.
3-11 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 3 References
Bator, F.M., 1958. The anatomy of market failure. The Quarterly Journal of Economics, 72(3): 351-379.
Baumol, W.J. and W.E. Oates. 1988. The Theory of Environmental Policy. Cambridge: Cambridge University
Press.
Buchanan, J. M. and W.C. Stubblebine. 1962. Externality. Economica 29(116): 371-384.
Coase, R. 1960. The Problem of Social Cost. The Journal of Law and Economics, 3:1-44.
Cornes, R. and T. Sandler. 1996. The Theory of Externalities, Public Goods, and Club Goods. Cambridge:
Cambridge University Press.
Deryugina, T., F. Moore, and R.S. Tol. 2021. Environmental Applications of the Coase Theorem. Environmental
Science & Policy, 120: 81-88.
Executive Order 12866: Regulatory Planning and Review, October 4,1993. Available at:
https: //www.archives.gov/files/federal-register/executive-orders/pdf/12866.pdf (accessed December 22,
2020).
Executive Order 13132: Federalism, August 10,1999. Available at:
https://www.govinfo.gOv/content/pkg/FR-1999-08-10/pdf/99-20729.pdf (accessed December 22,2020).
Hanley, N., J. Shogren, and B. White. 2019. Introduction to Environmental Economics. Oxford: Oxford University
Press.
Keohane, N.O. and S.M. Olmstead, 2016. Markets and the Environment. Island Press.Lusk, J.L. 2013. Lunch
with Pigou: Externalities and the "Hidden" Cost of Food. Agricultural and Resource Economics Review, 42(3):
419-435.
Mishan, E.J. 1969. The Relationship Between Joint Products, Collective Goods, and External Effects. Journal of
Political Economy, 77(3): 329-348.
Oates, W. E. 2002. A Reconsideration of Environmental Federalism. Recent Advances in Environmental
Economics, ed. J. List and A. De Zeeuw. Cheltenham, UK: Edward Elgar Publishing. 1-32.
0MB. 2023. Circular A-4, Regulatory Analysis, November 9, 2023. Available at:
https://www.whitehouse.gov/wp-content/uploads/2023/ll ZCircularA-4.pdf (accessed June 13, 2024).
Perman, R., M. Common, J. McGilvray, J., and Y. Ma. 2003. Natural Resource and Environmental Economics. 3rd
Ed. Essex: Pearson Education Limited.
Scitovsky, T. 1954. Two Concepts of External Economies. Journal of Political Economy, 62(2): 143-151.
Shogren, J. and L. Taylor. 2008. On Behavioral-Environmental Economics. Review of Environmental Economics
and Policy, 2(1): 26-44.
Tietenberg, T. 1973. Specific Taxes and the Control of Pollution: A General Equilibrium Analysis. The Quarterly
Journal of Economics, 86: 503-522.
Tietenberg, T. and L. Lewis. 2018. Environmental and Natural Resource Economics. 11th Ed. New York:
Routledge.
3-12 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 4 - Regulatory and Non-
Regulatory Approaches to
Environmental Policy
This chapter describes several regulatory and non-regulatory approaches
used in environmental policymaking. It also highlights a few key advantages
and disadvantages of each approach, provides an overview of cross-cutting
policy design issues, and offers references for those interested in a more in-
depth discussion. This chapter covers four general approaches to
environmental policymaking: [1] command-and-control regulation; [2]
market-based approaches; [3] hybrid and other approaches; and [4]
voluntary programs.1 While command-and-control regulation has been a
commonly used approach to environmental regulation in the United States,
market-based and hybrid approaches can sometimes offer increased
flexibility and lower costs. Voluntary programs may encourage environmental
improvements or allow new approaches to be tested in areas not traditionally
regulated by the U.S. Environmental Protection Agency [EPA].
The policy approaches discussed here are conceptually distinct, but they can
sometimes be designed in ways to achieve similar benefits and costs. The
approaches can also be combined into hybrid policy instruments, and multiple
instruments can be used in tandem to address environmental problems
caused by multiple market failures.2 As such, the approaches discussed in this
chapter represent an overlapping continuum of policy design tools.
1.1 Tjrv*"SitK>n,\j '"[ Ff'-'Uv'f jptr-i
ion
A prescriptive regulation is a policy that stipulates how much pollution an individual source or
plant is allowed to emit and/or what types of control equipment or approaches it must use to
reduce pollution. Prescriptive regulations are also known as "direct regulatory instruments" or
"command-and-control" regulations (Goulder and Parry 2008; Ellerman 2006). Despite the
introduction of potentially more cost-effective approaches for regulating emissions, this type of
1 Baumol and Oates (1988), particularly Chapters 10-14; Kolstad (2010); Field and Field (2021); Tietenberg and
Lewis (2014); and Phaneufand Requate (2016) are useful references on the economic foundations of many of the
approaches presented here.
2 This chapter uses the terms "approaches" and "instruments" interchangeably when discussing various policy or
regulatory tools.
4-1 Guidelines for Preparing Economic Analyses | 3rd edition
-------
regulation is still often used and is sometimes required by law. It is almost always available as a
"backstop" if other approaches do not achieve desired pollution limits.
A common approach to prescriptive regulation is to issue a license or permit to an individual facility
or firm that specifies the allowable level of pollution and the conditions under which it can be
released into the environment. For instance, a permit issued to a hazardous waste treatment facility
typically stipulates what waste management activities can be conducted at the site. It may also
include requirements for safety and training, insurance, monitoring and reporting. The EPA may
also set minimum standards when the licenses or permits are issued by another authority, such as
states or tribes.
It is also common for a prescriptive regulation to be defined in terms of a source-level emission
rate, which means that it does not directly control the aggregate emission level. In such cases,
aggregate emissions will depend on the number of polluters and the output of each polluter. As
either production or market size increase, so will aggregate emissions. Even when the standard is
defined in terms of an emission level per polluting source, aggregate emissions will still be a
function of the total number of polluters.
When abatement approaches and costs are similar across regulated sources, a source-level
standard may be reasonably cost-effective. However, when abatement costs vary substantially
across polluters, reallocating abatement activities so that some polluters abate more than others
could lead to substantial cost savings. For example, if reallocation were possible through a less
prescriptive market-oriented approach, a polluter facing relatively high abatement costs could
continue to emit at its current level but would have to pay an emissions tax or purchase allowances,
while a polluter with relatively low abatement costs could reduce its emissions, allowing it to avoid
the tax or sell its allowances (see Section 4.3 for more discussion of these approaches).3 A
prescriptive regulation usually does not allow for reallocation of abatement activities to take place
each entity is expected to achieve a specified emission rate or use certain abatement
technologies.
Prescriptive regulations can involve restrictingor in the most stringent case, prohibitingthe
production, use, or disposal of specific products or substances. For instance, the EPA has banned
most uses of chlorofluorocarbons (CFCs) and certain pesticides. This approach to regulation is
potentially useful in cases where the level of pollution that maximizes social welfare is at or near
zero. Prescriptive regulations include technology or design standards and performance-based
standards, discussed below.
>gy or Design Standards
A technology or design standard mandates the use of specific control technologies or production
processes an individual facility must use. This type of standard constrains firm behavior by
mandating how a source must reduce pollution, regardless of whether such an action is cost-
effective. Technology standards may be particularly useful in cases where the cost of emissions
monitoring is high but determining whether a specific technology or production process has been
3 Tietenberg and Lewis (2014) discussed empirical studies on the cost-effectiveness of prescriptive air pollution
regulations. Of the 10 studies included, eight found that prescriptive regulations cost substantially more than the
most cost-effective strategy. Harrington et al. (2004) compared the costs and outcomes ofcommand-and-control
and market-based approaches in the United States and Europe. Newell and Stavins (2003) generated rules of
thumb to help determine when market-based incentives may result in cost savings over prescriptive regulations.
4-2 Guidelines for Preparing Economic Analyses | 3rd edition
-------
put in place (and is operating properly) to meet a standard is relatively easy. However, since these
types of standards specify the abatement technology required to reduce emissions, sources do not
have an incentive to invest in more cost-effective types of abatement or to explore new and
innovative abatement strategies that are not permitted by regulation. Also, because permitting
authority is often delegated to the states, approval of a technology in one state does not ensure
its use is allowed in another.
Key Advantages4
Technology or design standards can yield environmental improvements with a high level of
certainty.
Technology or design standards can approximate an economically efficient outcome if the
regulated industry has relatively homogeneous abatement costs across firms.
If it is costly or infeasible to directly monitor emissions or environmental damages,
technology standards may provide an easier approach to monitor compliance with
regulatory requirements.
Key Disadvantages
Technology or design standards are less likely to be economically efficient when there are a
large number of diverse firms with varying abatement options because they do not allow for
flexibility in the approach to pollution reduction or in the distribution of pollution reduction
across sources.
These standards reduce incentives for innovation of new technologies and approaches to
achieve the environmental improvements at lower cost.
These standards could motivate rent-seeking by firms producing pollution control
technologies.
mce-Based Standards
A performance-based standard requires that polluters meet a source-level emission standard but
allows a polluter to choose among available methods to comply with the standard. At times, the
available methods are constrained by additional criteria specified in a regulation. Performance-
based standards that are technology-based do not specify a particular technology, but rather
consider what is possible for available and affordable technology to achieve when establishing a
limit on emissions.5
A performance-based standard can be defined in terms of an emission level or an emission rate (i.e.,
emissions per unit of output or input). A standard that specifies an emission level allows a source to
choose to reduce output or to reduce emissions per unit of output by changing the technology or
input mix. Therefore, an emission rate can be more restrictive depending on how it is defined. If the
emission rate is defined per unit of output, then it only allows a source to meet the standard by
reducing emissions per unit of output and not through reducing output. In some applications, it
4 The discussion of key advantages and disadvantages of each approach is intended to highlight a few notable
features but is not intended to be exhaustive.
5 As an example, Reasonably Available Control Technology (RACT) specifies that the technology used to meet the
standard should achieve "the lowest emission limit that a particular source or source category is capable of
meeting by application of control technology that is reasonably available considering technological and
economic feasibility."
4-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
may even create incentives to increase output (Holland et al. 2009). If the rate is defined as an
average amount of emissions over a certain time period, then the source may reduce output or
emissions per unit of output to meet the standard.
Regulators can account for some variability in costs by allowing prescriptive regulations to vary
according to size of the polluting entity, production process, geographic location, or other firm or
product attributes that are not the direct target of the regulation. If the attribute is correlated with
compliance costs, this type of attribute-based standard can improve economic efficiency relative to
a uniform standard by helping to equate marginal compliance costs across firms in the absence of
trading. However, the efficiency gains from equalizing marginal costs can be partially offset if firms
comply with the standard by modifying the attribute instead of reducing pollution. For example, in
2011 the U.S. introduced a "footprint based" vehicle fuel economy standard, which subjected larger
cars to a less stringent standard. The footprint-based standard was intended to discourage
automakers from downsizing vehicles. However, researchers have found that such standards can
incentivize automakers to design larger cars instead, eroding some of the gains in fuel savings
expected from the standard (Whitefoot and Skerlos 2012; Ito and Sallee 2018).6
While performance-based standards encourage firms to meet the standard at lower cost than
technology standards, they generally do not provide incentives to reduce pollution beyond
what is required. There is still limited incentive for regulated firms to develop new, less
expensive and potentially superior technologies compared to market-based policies (Swift
2000; Johnstone etal. 2010).
Key Advantages
Performance standards, like technology and design standards, can yield environmental
improvements with a high level of certainty.
Performance standards can allow more flexibility to achieve environmental benefits at
lower cost compared to technology or design standards.
Performance standards create greater incentives for technological innovation than
technology or design standards.
Key Disadvantages
Performance standards are unlikely to be as economically efficient as market-based
policies if abatement costs vary substantially across sources.
Performance standards do not incentivize sources with low abatement costs to make
environmental improvements beyond what the standard requires.
If technological innovation yields lower-cost abatement opportunities in the future, the
standard may need to be tightened over time to more closely approximate an
economically efficient outcome.
6 Attribute-based standards have been used in both prescriptive, non-tradable performance standards and
market-based, tradable performance standards, of which the corporate average fuel economy standards for
vehicles are an example. Tradable performance standards are a type of hybrid instrument discussed in Section
4.3.1.2. Making the standard tradable negates the potential economic efficiency gains associated with attribute-
based standards (Ito and Sallee 2018).
4-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
i. w I' Varket Fv,:>vf : ppj res
Market-based regulatory approaches create an incentive for the private sector to incorporate
pollution abatement into production or consumption decisions and prompt innovation to explore
cheaper methods of abatement. Market-based approaches can differ from more traditional
regulatory approaches in terms of economic efficiency, cost-effectiveness and the distribution of
benefits and costs. Because market-based approaches do not mandate that each polluter meet a
given emission standard, they typically allow firms more flexibility than prescriptive regulations
and capitalize on the heterogeneity of abatement costs across polluters to reduce aggregate
pollution efficiently. Environmental economists generally favor market-based policies because they
tend to be less costly, they place a lower information burden on the regulator, and they provide
incentives for technological advances.
Market-based policies create incentives for regulated firms to find the cheapest way to reduce
pollution. This may involve a reduction in output (and in the extreme, exiting the industry), a
change in inputs, the installation of pollution control equipment or a process change that prevents
the creation of pollution. Polluters decide individually how much to control their emissions based
on the costs of control and the financial incentives created by the policy.
Four market-based approaches are discussed in this section:
Allowance trading systems;
Emissions taxes;
Environmental subsidies; and
Tax-subsidy combinations.7
While operationally different, these market-based approaches put similar incentives in place. This
is particularly true of emissions taxes and cap-and-trade systems, which can be designed to achieve
the same goal at equivalent cost.
4,2.1 Allowanc hi hp' stems
v*Jr J
Several forms of emissions trading exist, including cap-and-trade and project-based trading
systems. The common element across these programs is that sources can trade credits, offsets or
allowances so that those with opportunities to reduce emissions at lower costs have an incentive to
do so. Emission-rate trading systems, a hybrid approach between tradable allowances and
command-and-control, are discussed in Section 4.3.1.3.
ii »it i fi
-------
For a uniformly mixed pollutant where marginal damages are identical for all sources and in all
locations, if the cap is set at the economically efficient level, then the equilibrium price of allowances
adjusts so that it equals the marginal damages from a unit of pollution. This equivalency implies that
any externality associated with emissions is completely internalized by the firm. For polluters with
marginal abatement costs greater than the allowance price, the cheapest option is to purchase
allowances and continue to emit For polluters with marginal abatement costs less than the
allowance price, the cheapest option is to reduce emissions and forgo purchasing allowances (or to
sell any allowances that they own at the market price). As long as the price of allowances differs
from individual firms' marginal abatement costs, firms will continue to buy or sell them. Trading will
occur until marginal abatement costs equalize across all firms.8 Assuming no other market failures,
an allowance price that is lower than the marginal damages from pollution implies that the cap is set
at an inefficiently high level.
When the government sells allowances at auction, the revenue represents a transfer from the
purchasers to the government Allowance auctions can be designed in a variety of ways. Typically,
allowances are purchased through a bidding process that reveals buyers' willingness to pay, with
allowances going to the highest bidder.
The government could also decide to allocate allowances to polluters for free according to a
specified rule either at the outset of the program or on an annual or other ongoing basis. This
represents a transfer from the government to polluting firms, some of which may find that the value
of allowances exceeds the firm's aggregate abatement costs (i.e., rents). Economic rents are any
payment to the owner of capital or a resource above what it would cost to induce them to engage in
a certain behavior.9 The way in which allowances are allocated can also affect firm entry, exit, and
production decisions. For example, allocating allowances based on historical emissions can create
perverse incentives for old, dirty plants to continue to operate to qualify for allowances. Another
alternative is to allocate allowances based on current production, which encourages firms to
increase output to capture a greater number of future allocations (Fischer and Fox 2007; Lange and
Maniloff 2021).
Additional considerations in designing an effective cap-and-trade system include the number of
market participants, transaction costs, banking and hotspots. The United States' experience
suggests that a market characterized by low transaction costs and being "thick" with many buyers
and sellers is critical if pollution is to be reduced at the lowest cost This is because small numbers
of potential traders in a market can inhibit competitive behavior, and fewer trading opportunities
result in lower cost savings. Likewise, the number of trades that occur could be significantly
hindered by burdensome requirements that increase the transaction costs associated with each
trade. Text Box 4.1 provides an example where thin markets resulted in few trades.
8 Schmalensee and Stavins (2017) provide an overview of emission trading programs and lessons learned
regarding implementation, system design and performance.
9 Tietenberg (2006) defined scarcity rent as, "producer's surplus which persists in long-run competitive
equilibrium." In the context of a cap-and-trade market, these rents occur because firms are given allowances
that can be bought and sold in the market. For a discussion of scarcity rents created by environmental
regulations through pollution restrictions and captured by firms in the form of higher profits, see Fullerton and
Metcalf(2001). Buchanan and Tullock(1975) discussed the potential for scarcity rents under a cap-and-trade
system where permits are distributed for free.
4-6 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 4,1 Water Quality Trading of Nonpoint Sources
In 2003, the EPA issued a "Water Quality Trading Policy" (U.S. EPA 2003} that encouraged
states and tribes to develop and implement voluntary water-quality trading to control
nutrients and sediments in areas where it is possible to achieve these reductions at lower
costs. A 2019 memo announced additional flexibilities available to states and tribes to further
facilitate the uptake of water quality trading particularly between point and nonpoint sources.
The memo cited the increased availability of effective nonpoint emission reducing
technologies and practices and enhanced monitoring capabilities as reasons to modernize the
2003 policy (U.S. EPA 2019).
Under the Clean Water Act, states are required to establish Total Maximum Daily Loads
(TMDLs) of pollutants for impaired water bodies. The TMDL is not a regulation and does not
establish an enforceable cap on discharges to the watershed, but it does provide an approach
for allocating pollutant discharges among point and nonpoint sources. Point sources are
regulated under the Clean Water Act by the EPA and, as such, are required to hold National
Pollutant Discharge Elimination System (NPDES) permits that limit discharges. Where a TMDL
exists, the point source NPDES discharge limit is informed by the TMDL allocation. Nonpoint
sources are not regulated under the Clean Water Act However, many water bodies are still
threatened by pollution from these sources. Nutrients and sediment from urban and
agricultural runoff have led to water quality problems that limit recreational uses of rivers,
lakes and streams; create hypoxia in the Gulf of Mexico and other coastal waters; and decrease
fish populations in the Chesapeake Bay and other areas.
To account for uncertainties and differences associated with nonpoint source pollution,
trading ratios are often applied. These ratios account for the differential effects resulting from
a variety of factors, which may include:
Location of the sources in the watershed relative to the downstream area of concern;
Distance between the allowance buyer and seller;
Uncertainty about nonpoint source reductions;
Equivalency of different forms of the same pollutant discharged by the trading
partners; and
Additional water quality improvements above and beyond those required by
regulation.
Trading can allow continued growth in production while providing nonpoint sources with an
incentive to reduce pollution through participation in the market. If it is cheaper for a
nonpoint source to reduce pollution than to forgo revenues earned from the sale of any
unused credits to point sources, economic theory predicts that the nonpoint source will
choose to emit less pollution.
As of 2014, the EPA had identified 19 nutrient trading programs in 11 states, with the majority
of trades occurring in just three states Connecticut, Pennsylvania and Virginia (GAO 2017).
Trading has been limited in many of these programs for several reasons. First, as previously
mentioned there is no enforceable cap on discharges that applies to both point and nonpoint
sources within a watershed. Reductions by nonpoint sources are voluntary absent state-level
mandates. Point-source dischargers often explore trading as a way to expand production
while meeting the requirements of their individual permits, but there is no general signal in
the market to do so, and it can be challenging to encourage nonpoint source involvement.
Second, these are often thin markets (i.e., markets with few trades). The lack of participants
can make it difficult or expensive for an entity to identify and complete a trade. Third, while
4-7 Guidelines for Preparing Economic Analyses | 3rd edition
-------
best management practices (BMPs) are typically used to define a pollution reduction credit
from a nonpoint source, uncertain or changing climatic conditions, river flow and stream
conditions make it difficult to measure the effect of a BMP on downstream water quality. This
uncertainty makes it difficult to define appropriate trading ratios between point and nonpoint
sources (Morgan and Wolverton 2008; U.S. EPA 2008). Such uncertainty also makes
measuring and enforcing a pollution reduction from a nonpoint source difficult.
Banking introduces increased flexibility into a trading system by allowing polluters to save unused
allowances for future use. A firm may reduce emissions below the allowance level earlier and bank
remaining allowances to cover excess emissions or sell to another polluter at a later time. In this
way, polluters that face greater uncertainty regarding future emissions or that expect increased
regulatory stringency can bank allowances to offset potentially higher future marginal abatement
costs.
Cap-and-trade systems for non-uniformly mixed pollutants have the potential to create temporal or
spatial spikes or "hotspots" areas with particularly high pollution concentrations. Market-based
policy should therefore be carefully designed and consider localized effects. While one potential
solution to this problem is to adjust trading ratios (i.e., the rate at which allowances from one
source can be traded with another) to equalize impacts, determining the appropriate adjustments
to these ratios can be costly and difficult. Another possible solution is zone-based trading. Two
reviews of the literature on SO2 and NOx trading programs (see Text Box 4.2) found little evidence
of spatial or temporal spikes in pollution (Burtraw et al. 2005; Harrington et al. 2004). In fact, they
have led to smoothing of emissions across space in some cases. However, it is still important to
consider the potential for localized effects when designing cap-and-trade policies. See Section 4.5.1
for a discussion of distributional concerns.
4.2.1.2 Project-Ba 3 Systems
Offsets and bubbles (sometimes known as "project-based" trading systems) allow restricted forms
of emissions trading across or within sources to allow sources flexibility in complying with
emission limits or facility-level permits.10 A bubble allows a facility to consider all sources of
emissions of a specific pollutant within the facility to achieve an overall target level of emissions or
environmental improvement. To meet air quality standards for particulate matter, EPA employed a
compliance bubble alternative that allowed pulp and paper mills to set site-specific emission limits
as long as total emissions from all sources within the site were less than or equal to the standard.
This flexibility resulted in lower the compliance costs (Morgan et al. 2014).
10 Bennear and Coglianese (2012) evaluated how these types of flexibilities have worked in the United States.
4-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 4.2 - Acid Rain Trading Program for Sulfur Dioxide (SO2)
In 1995, Title IV of the 1990 Clean Air Act Amendments established a cap-and-trade system for
S02 emissions to address acid rain. The 263 highest S02-emitting units at 110 electric utility
plants were selected to participate in Phase I of the trading program to limit S02 emissions to
8.7 million tons in 1995. Most of the plants that participated in Phase I were coal-fired units
located east of the Mississippi River. Allowances were allocated to units on a historical basis for
plants to use, sell to other units, or "bank" for use in later years. Phase I plants were required to
install continuous emission monitoring systems, allowing for easy monitoring and enforcement
of emission restrictions in accordance with the allowances. The second phase of the program,
initiated in 2000, imposed a national S02 emissions cap of 10 million tons and brought almost
all S02-emitting units into the system.
Evaluations of the Phase I suggest that the S02 trading system significantly reduced emissions.
Compliance costs were estimated to be 15 to 90% lower than an equally stringent command-
and-control alternative. The success of the program continued into Phase It Chan et al. (2018)
estimated Phase II annual cost savings at $700 million compared to a simulated uniform
performance standard.
S02 Caps and Emissions, 1988-2010
O
CO
w
C
o
c
o
18
16
14 "
12 -
10 -
8 -
6 -
4 -
Emission limits (cap)
O Actual emissions
<$
¦00000^8.0000
¦13
1988
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Copyright American Economic Association; reproduced with permission of the Journal of Economic Perspectives
In the figure above, Schmalensee and Stavins (2013) reported that emissions declined by 36%
between 1990-2004, even as coal-fired electricity generation increased. One reason for such
large emission reductions was the ability to bank allowances for future use. In addition,
incentives to innovate continued to reduce abatement costs over time (Bellas and Lange 2011;
Frey 2013). Railroad deregulation and investment by utilities in mining and infrastructure also
played a role by making low-sulfur coal cheaper. That said, researchers observed less inter-firm
trading than expected, meaning that marginal abatement costs were not equalized across plants
(Swift 2001; Swinton 2004). Estimates of the S02 allowance program's annual benefits range
from $59-116 billion with estimated annual costs of $0.5 to $2 billion (in 2000$) (Schmalensee
and Stavins 2013).
Congress did not grant the EPA the authority to adjust the cap in response to new information
on either the costs or benefits of reducing emissions. For this reason, the EPA pursued additional
4-9 Guidelines for Preparing Economic Analyses | 3rd edition
-------
reductions in S02 emissions via more traditional regulatory approaches, which restricted the
ability of sources to trade and reduced allowance prices to zero by 2012 (Schmalensee and
Stavins 2013).
For more information, see Chestnut and Mills (2005), U.S. EPA (2007), Schmalensee and Stavins
(2013), Chan et al. (2018), and Evans and Woodward (2013).
An offset allows a new polluter to negotiate with an existing source to secure a reduction in the
latter's emissions. Offsets, which entail cross-firm emissions trading, have at times been hindered
by high administrative and transaction costs. However, regulators can improve the economic
efficiency of offsets by allowing third parties, who are not themselves polluters, to participate in the
market. For instance, evidence suggests that some water quality offset programs that operated
through an intermediary or clearinghouse eliminated the need for direct negotiations between
buyer and seller and lowered these costs substantially (Woodward and Kaiser 2002; Morgan and
Wolverton 2008). Offsets have also been included in cap-and-trade programs for greenhouse gas
emissions such as the Clean Development Mechanism of the Kyoto Protocol. Such systems allow
entities covered under the cap to purchase offsets for emission reductions or carbon sequestration
from firms in industries or locations not covered under the program, increasing the flexibility and
reducing the costs of meeting the aggregate greenhouse gas emissions target.
Key Advantages
Like other market-based policies, tradable allowance systems can be more economically
efficient than prescriptive regulations, particularly when there are many heterogenous
market participants.
Like other market-based policies, tradable allowances create incentives for innovation as
firms compete for new ways to reduce emissions at the lowest cost
Tradable allowance systems provide more certainty about the total level of emissions than
emissions taxes or subsidies; as such, they may be preferable to emissions taxes when
marginal damages increase with the level of emissions.
Key Disadvantages
If the pollution cap is set at an inefficiently high (low) level, then allowance prices will be
lower (higher) than the marginal damages from pollution, and an inefficiently low (high)
level of abatement will occur.
Tradable allowance systems raise complicated issues regarding the distribution of
allowances, including auction design and rent-seeking by regulated firms.
Tradable allowance systems can result in emissions spikes or hotspots. Regulators can set
trading ratios or use additional instruments to avoid hotspots or to address heterogeneous
pollution damages, but such additional requirements raise analytical and administrative
challenges.
4.2.2 Emissions Taxes
Emissions taxes are a charge per unit of pollution that is imposed by the government. Under an
emissions tax, the polluter will abate emissions up to the point at which the additional cost of
abating one more unit of pollution is equal to the tax. For any remaining emissions, the pollute r
4-10 Guidelines for Preparing Economic Analyses | 3rd edition
-------
prefers to pay the tax rather than to abate further. The tax will result in an economically efficient
outcome if it is set equal to the external damage caused by the last unit of pollution emitted.11
User or product charges are a variation on emissions taxes. These charges may be imposed on
users of publicly operated facilities or on intermediate or final products whose use or disposal
harms the environment. User or product charges may be effective approximations of an
emissions tax when the product is closely related to environmental damage. User and product
charges will not result in an efficient level of pollution if they are set at a level sufficient to
recover only the private costs of operating a public system, rather than incorporating the marginal
social damages of pollution.
Emissions taxes, like tradable allowance systems that distribute the allowances using an auction,
raise revenue for the government. The welfare and distributional effects of an emissions tax
depend on how the revenues are used and how the tax interacts with other distortions in the
economy. If distributed to households or firms, the revenues can be used to compensate
individuals made worse off by the policy or to address other distributional priorities of the
policymaker, though it can be difficult to accurately target individuals for compensation (Cronin,
Fullerton and Sexton 2019). If the revenues are instead used to reduce other distortionary taxes,
such as labor taxes, then this "revenue recycling" could yield economic gains due to a resulting
increase in employment or investment (e.g., Goulder 2000). However, emissions taxes or
allowances can also exacerbate pre-existing tax distortions, causing an increase in deadweight
loss. Analysts should consider the opportunity costs associated with collecting and spending public
funds. Section 8.3.1 of these Guidelines discusses general equilibrium approaches to examine these
types of economy-wide effects.
Emissions taxes should lead to outcomes similar to those from allowance trading systems when
both are designed to achieve the same level of emissions. Rather than specifying the total quantity
of emissions, taxes specify the effective "price" of emitting pollutants. However, these two types of
policy instruments differ in their usefulness when there is uncertainty about the costs or benefits of
abatement Section 4.5.5 discusses instrument choice under uncertainty.
Key Advantages12
Like tradable allowances, emissions taxes are an economically efficient approach to
incentivize pollution reduction, allowing flexibility to reduce emissions multiple ways
and/or to pay the tax for remaining emissions.
Like tradable allowances that are distributed via auction, emissions taxes raise revenue that
can be used to compensate individuals made worse off by the policy or to offset other
distortionary taxes, increasing economic efficiency throughout the economy.
Emissions taxes are advantageous in situations where there is uncertainty about abatement
costs, but damages do not change much with additional pollution.
Key Disadvantages
Emissions taxes do not set source-level or aggregate limits on emissions and do not
eliminate the potential for emissions spikes or hotspots.
Emissions taxes may be difficult to implement efficiently when pollution damages vary over
space and time.
11 These taxes are caiied "Pigovian" after the economist, Arthur Pigou, who first formalized them (Pigou 1932).
12 See Fullerton, Leicester and Smith (2010) for more discussion of the advantages and disadvantages of
emissions taxes.
4-11 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Emissions taxes are less well-suited to situations in which contaminant releases are difficult
to measure and are not directly related to a marketed input or output.
4,2,3 Environmental Subsidies
A subsidy is a payment or financial assistance made to encourage a certain behavior. Subsidies paid
by the government to firms or consumers for technology-neutral reductions in pollution create
similar abatement incentives as emissions taxes. Economic theory predicts that firms will reduce
pollution up to the point where the additional private costs are equal to the subsidy.
Unlike an emissions tax, an environmental subsidy lowers a firm's total and average costs of
production, encouraging production by both existing and new firms. The result may be a decrease
in emissions from individual polluters but a smaller net decrease (or even an increase) in overall
pollution.13 However, it is possible to minimize the entry and exit of firms resulting from subsidies
by redefining the subsidy as a partial repayment of verified abatement costs, instead of defining it
as a per-unit payment for emissions reductions relative to a baseline. Defining the subsidy in this
way also minimizes strategic behavior because no baseline must be specified.14 An environmental
subsidy also differs from an emissions tax because it requires government expenditure (versus
generating government revenue).
Government funding for research and development of technologies to reduce pollution and
improve environmental quality is another form of subsidy. The private market does not always
have an incentive to invest in the socially optimal level of innovation and diffusion of new
technologies because these activities can create positive information spillovers that benefit other
firms. In addition, network externalities, which occur when the net benefits of adopting a new
technology increase as the number of users increases, can limit the spread of otherwise promising
innovations (Jaffe, Newell and Stavins 2005).15 Subsidies for technology development and
demonstration can be used to address these types of market failure, complementing other
environmental policy approaches. Research on new technologies and approaches to improve
environmental quality may also yield data that could be useful in future analyses of regulatory or
non-regulatory approaches to environmental policy.
Cost-sharing constitutes another type of subsidy, with examples that include reduced interest rates,
accelerated depreciation, direct capital grants, loan assistance or guarantees for investments, and
government "buy-backs." Under a buy-back program, the government offers a payment for the
return of an older, high-polluting product or a rebate on a new, cleaner substitute if the older model
is turned in. For example, the EPA has funded changeout programs to encourage the replacement of
old wood stoves with EPA-certified gas, electric or wood appliances that reduce indoor air pollution
13 See Sterner and Coria (2012) and Goulder and Parry (2008) for a discussion and examples of environmental
subsidies.
14 Strategic behavior is a problem common to any instrument or regulation that measures emissions relative to
a baseline. In cases where a firm or consumer may potentially receive funds from the government, they may
attempt to make the current state look worse than reality to receive credit for large improvements. If firms or
consumers are responsible for paying for emissions above a given level, they may try to lobby for that level to be
set at a fairly high level so that they pay less in fines or taxes.
15 Electric vehicle adoption provides one example of network externalities. The cost and convenience of electric
vehicle use depends on the availability of a network of electric charging stations. Spreading the cost of this
infrastructure across many users lowers the costs for each individual user.
4-12 Guidelines for Preparing Economic Analyses | 3rd edition
-------
(U.S. EPA 2014). In 2009, the U.S. ran a program called "Cash for Clunkers" that offered rebates for
trading in old, fuel-inefficient, but still drivable vehicles for new, fuel-efficient vehicles to stimulate
auto sales during a recession.
The effectiveness of subsidies depends on the degree to which they motivate behavior that would
not have already occurred without the subsidy (an effect called "additionality"). In the Cash for
Clunkers program, researchers estimated that most of the funds were received by consumers who
would have purchased a vehicle in 2009 regardless, though the program did induce sales of more
fuel-efficient vehicles than would have been purchased without the subsidy (Li, Spiller and Lin
2013). Similar to allowance trading systems, auctions can be incorporated into subsidy programs to
incentivize participants to reveal their opportunity costs and avoid payments in excess of this
amount In these programs, sometimes referred to as conservation or reverse auctions, subsidies
are awarded to the lowest bidder (de Vries and Hanley 2016). The effectiveness of subsidies
targeted to households also depends on transaction costs and other non-financial barriers such as
lack of trust and inattention. For example, programs subsidizing home energy efficient retrofits and
lead water pipe replacements have had low adoption rates, even when coupled with information
and outreach, possibly due to the hassles of completing extensive paperwork or having
construction work in the home (Fowlie etal. 2015; Klemick etal. 2024).
A subsidy for specific technologies a policy approach that is sometimes termed "picking
winners" is typically not as economically efficient as a subsidy (or tax) per unit of emission
reduction or other environmental outcomes. This is because, similar to prescriptive regulation,
such programs do not encourage flexibility in the way firms or individuals reduce their adverse
environmental impacts, and the government may not have good information on what technology
will ultimately be the most efficient abatement option.
Key Advantages
Technology-neutral environmental subsidies are an economically efficient way to
encourage pollution reduction because they create incentives to reduce emissions up to the
point at which marginal abatement costs equal the subsidy.
Subsidies for research and development of new pollution abatement technologies and
approaches can help mitigate market failures that inhibit technological innovation.
Subsidies provide flexibility to polluters about whether and how much to abate and impose
no mandatory requirements on the public.
Key Disadvantages
Subsidies have limited effectiveness if most market participants would have undertaken the
environmentally beneficial action without the subsidy in this case, the subsidy acts as a
transfer and results in no net social benefit.
Subsidies for specific technologies are typically less efficient than technology-neutral
subsidies because they allow less flexibility for achieving environmental improvements.
Like emissions taxes, subsidies provide less certainty that source-specific or aggregate
emissions will remain below a particular level; emission spikes or hotspots could occur.
-i C-i r >' jbs nbinatioi i
Emissions taxes and subsidies can be combined to achieve the same level of abatement as when
each instrument is used alone. One example of this type of instrument is a deposit-refund system.
Under a deposit-refund system, firms or consumers pay an upfront deposit that serves as a tax on
the production or use of certain goods. A refund is then provided if firms or consumers
4-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
demonstrate that they used a cleaner form of production or engaged in proper disposal, acting as a
subsidy.16
A tax and subsidy combination functions best when there is a direct relationship between use of a
product and emissions. For instance, a tax on the production or use of hydrochlorofluorocarbons
(HCFCs) combined with a refund for HCFCs recycled or collected in a closed system is a good proxy
for an HCFCs emissions tax.
The main advantage of a combined tax and subsidy is that both parts apply to a market transaction.
Because the taxed and subsidized items are easily observable in the market, this type of economic
instrument is appealing when it is difficult to measure emissions or to control illegal dumping. In
addition, polluters have an incentive to reveal accurate information on abatement activity to qualify
for the subsidy. Because firms have access to better information than the government does, they
can measure and report their actions with greater precision and at a potentially lower cost.
A disadvantage of the combined tax-subsidy system is potentially high implementation and
administrative costs. In addition, while it is possible to adjust an emissions tax to account for
variation in marginal damages, a tax on output cannot be matched temporally or spatially to
emissions during production. Likewise, if inputs contribute differentially to environmental
damages, then it is necessary to tax them at different rates to achieve economic efficiency. When
firms are heterogeneous and select a different set of inputs or abatement options based on firm-
specific cost considerations, then the subsidy needs to be adjusted for these differences. Given these
complications, other market-based approaches may have lower implementation costs when
emissions are easily monitored.
Conceptually similar to the tax-subsidy combination is the requirement that firms post-
performance bonds that are forfeited in the event of damages, or that firms contribute upfront
funds to a pool. Such funds may be used for pollution abatement or to compensate individuals
harmed by pollution if environmental damages occur. If the company demonstrates it has fulfilled
certain obligations, the contribution is usually refunded.17
Key Advantages18
Like tradable allowances and emissions taxes, tax-subsidy combinations can be an
economically efficient approach to achieve environmental improvements.
A tax-subsidy combination can be useful when it is less costly to observe market outputs
and inputs than it is to observe emissions or environmental damages.
Performance bonds, a conceptually similar approach, create a pool of funds that can be used
to abate pollution or to compensate individuals affected by environmental damages.
Key Disadvantages
Tax-subsidy combinations can involve high implementation and administrative costs.
16 When a deposit-refund encourages firms to use a less-polluting input, a deposit on output induces the firm to
reduce its use of all inputs, both clean and dirty (i.e., the output effect). The refund provides the firm an incentive
to switch to a specific input such as a cleaner fuel (i.e., the input substitution effect).
17 For more information on the use of financial assurance or performance bonds, see Davis (2015), Dana and
Wiseman (2013), and Boyd (2002).
18 The main advantages and disadvantages of deposit-refund systems are discussed in Walls (2013) and
Fullerton and Wolverton (2001,2005). Fullerton and West (2010), Walls (2013), and Sterner and Coria (2012)
provide more discussion and examples of tax-subsidy combinations.
4-14 Guidelines for Preparing Economic Analyses | 3rd edition
-------
It is difficult to adjust tax-subsidy combinations to account for heterogeneity in
environmental damages.
Like other market-based policies, the lack of a limit on individual sources means that
hotspots with a high concentration of pollution can occur.
! H In1' 11 l> \ \l i 3f : ppf "Vviivi:;,
In addition to the instruments discussed above, several other approaches have been used alone or
in combination. This section discusses the following approaches:
Combining prescriptive and market-based approaches;
Liability rules and insurance requirements;
Information disclosure; and
Behavioral economics and "nudge" approaches.
-f.i \'-MiThininr R - ;riptive and Market-Bas ' ,l| | i ' ches
Some policies combine aspects of prescriptive and market-based policies. As such, they may not
represent the most economically efficient approach. The cost of the policy is likely to be greater
than what would be achieved using a pure market-based approach. Nevertheless, such approaches
are appealing to policymakers because they combine the certainty associated with a standard or
technology with some flexibility, allowing firms to comply at a lower cost. Combining standards and
pricing and tradable performance standards are two hybrid approaches.
i ' i i tmbmmij «\v' Rising
Emissions taxes restrict costs by allowing polluters to pay a tax on the amount they emit rather
than undertake excessively expensive abatement. Taxes, however, do not set a limit on the quantity
of emissions and leave open the possibility that pollution may be excessively high. Some
researchers suggest a policy that limits both costs and pollution by combining quantity and pricing
instruments, such as a "safety-valve" or price collar (Jacoby and Ellerman 2004; Fell and
Morgenstern 2010; Fell et al. 2012). In the case of a prescriptive standard and tax combination, an
emission standard is imposed on all polluters, but polluters can pay a unit tax for emissions above
the standard. Safety-valve systems can also be entirely market-based, by combining a cap-and-trade
policy with an emissions tax (i.e., an allowance price ceiling and/or floor) if allowance prices go
above or below a certain level (Burtraw, Palmer and Kahn 2010).
This policy combination has several attractive features. First, it allows for more certainty in the
expected environmental and health effects of the policy than would occur with a pricing approach
alone. 19 Second, overall abatement costs are lower than under a prescriptive standard because
polluters with low abatement costs reduce pollution while polluters with high abatement costs pay
taxes.
19 Section 4.5.5 elaborates on instrument choice under uncertainty.
4-15 Guidelines for Preparing Economic Analyses | 3rd edition
-------
adable Performance Standards
A tradable performance standard establishes a standard or emission rate, as described in section
4.1.2, but introduces credit trading and banking as additional flexibilities. Sources that perform
better than the standard can earn credits and sell them to sources that perform worse or can save
them for future use. A credit allows a source to emit one unit of a pollutant in excess of what would
normally be permitted (e.g., reducing emissions below a baseline or cap).
In rate-based trading systems, sources able to reduce their emission rate at low cost have an
incentive to do so since they can sell the resulting credits to sources facing higher abatement costs.
Rate-based trading programs have been used in the United States to phase out lead in gasoline
(Newell and Rogers 2006; Schmalensee and Stavins 2017) and to control emissions of light-duty
and heavy-duty vehicles (Bento etal. 2020; Leard and McConnell 2017). Similarly, state Renewable
Portfolio Standards require the use of renewable energy sources such as wind or solar for
electricity generation, but they incorporate tradable credits so that firms can meet the overall
standard at least cost. These approaches encourage cleaner transportation or electricity, but they
do not allow reducing output or consumption as a way to comply with the standard (e.g., reducing
vehicle miles traveled or electricity consumption). Emissions may increase under these programs if
sources increase their production or if new sources enter the market. The regulating authority may
need to periodically impose new standards to maintain the desired emission target, which may lead
to uncertainty in the long term for regulated sources.
Key Advantages
Combining prescriptive and market-based approaches can achieve a particular emission
rate or technology adoption target at least cost
Combining approaches can increase certainty about achieving an emission rate or
technology adoption target.
"Safety-valve" systems that combine cap-and-trade with an emissions tax (a price floor
and/or price ceiling) achieve the economic efficiency of market-based policies while
mitigating uncertainty about abatement costs and emission reductions.
Key Disadvantages
A combined prescriptive and market-based policy is typically not the most economically
efficient approach because it limits flexibility in the way that environmental improvements
are achieved.
A tradable emission rate is not the most efficient approach to improving environmental
quality because it does not create incentives to reduce output or consumption.
Like prescriptive regulations, if the hybrid approach does not set an overall limit on
emissions across the regulated sector, then it is possible for total emissions to increase even
if source-level emissions decline.
-f.o.: Information Disci * m -
Market failure due to imperfect information occurs when firms or consumers are unable to make
optimal decisions due to lack of information on emission levels, health and ecological risks, or
approaches to mitigate these risks. Asymmetric information exists when one party in the
transaction has more information than others, which can also yield suboptimal outcomes.
Regulations requiring disclosure of environmental information can minimize inefficiencies
4-16 Guidelines for Preparing Economic Analyses | 3rd edition
-------
associated with imperfect or asymmetric information.20 Information disclosure can also be an
important component of non-regulatory EPA programs. By collecting and making information
publicly available, firms, government agencies and consumers can become better informed about
the environmental and human health consequences of their production and consumption decisions.
In some cases, the availability of this information may encourage more environmentally benign
activities and discourage environmentally detrimental ones. For example, warning labels on
hazardous substances that describe risks or safe-handling procedures may encourage consumers to
take greater precautions or switch to less-damaging substitutes. A community with information on
a nearby firm's pollution activity may exert pressure on the firm to reduce emissions, even if
regulations to limit pollution are weak or nonexistent21
Requirements for information disclosure need not be tied to an emission standard. However, such
requirements might allow members of the public to easily understand the level of emissions in the
context of existing standards. As with market-based instruments, polluters have the flexibility to
respond to community pressure by reducing emissions in the cheapest way possible.
The use of information disclosure or labeling rules has other advantages. When expensive
emissions monitoring is required to collect information, reporting requirements that switch the
burden of proof for monitoring and reporting from the government to the firm might result in
lower costs, because firms are often in a better position to monitor their own emissions. However,
random inspections may be needed to ensure that monitoring equipment functions properly and
that firms report results accurately. Information disclosed to regulators or the public through such
programs could be useful for analysis of other potential regulatory approaches in the future.
Information disclosure alone does not typically result in an economically efficient level of pollution
when externalities are present. Several conditions are necessary for it to be effective and welfare-
improving. The information must be complete and accurate. Consumers must be able to access the
information and understand it22 In addition to complete information, Coase (1960) identified low
transaction costs and the possibility of bargaining as two conditions necessary for a private
agreement between affected parties to lead to an efficient level of pollution (see Text Box 3.1). A
community's ability to bargain with or exert pressure on an emitting plant may be related to
socioeconomic status. Lower income, less-educated populations may face more barriers to political
participation and be less likely to have their concerns addressed than richer, well-educated
populations (Hamilton, 1993; Arora and Cason, 1999; Earnhart, 2004). The effect that public
pressure has on behavior may also vary by firm and depend on factors such as the firm's market
power and societal reputation. Finally, even if information is complete and consumers can access it
readily which may be strong assumptions individuals do not always act to further their own
best interests, as discussed in Section 4.3.4.
The most studied environmental disclosure program is the Toxics Release Inventory (TRI), but
researchers offer a mixed view on the extent to which it has changed firm behavior. Some studies
have found that high-polluting firms experienced stock price declines on the day the TRI was
publicly released, and that those with the largest drop in stock prices reduced reported emissions
20 See OMB (2010b) for guidance issued to regulatory agencies on the use of information disclosure and
simplification in the regulatory process.
21 For more information on how information disclosure may help resolve market failures, see Pargal and
Wheeler (1996), Tietenberg (1998), Tietenberg and Wheeler (2001), and Brouhle and Khanna (2007).
22 As noted in Sunstein (2011), "accurate disclosure of information can be ineffective if the information is too
abstract, vague, detailed, complex, poorly framed, or overwhelming to be useful."
4-17 Guidelines for Preparing Economic Analyses | 3rd edition
-------
the most in subsequent years (Hamilton 1995; Konar and Cohen 1997).23 Others found no evidence
of a negative stock price effect (Bui 2005). Bae et al. (2010) found that making raw TRI data
available did not significantly change environmental risk, even when emissions declined. However,
when data were processed and presented to the public in more digestible terms, they found a
significant decline in environmental risk. The economics literature has also found evidence that
consumers respond to product labels in specific cases.24
Key Advantages
Information disclosure requirements can help address market failures due to imperfect or
asymmetric information.
Information programs can complement other approaches, including emission standards,
market-based approaches and nudges.
Reporting requirements that make new information available to the public and to
government agencies can yield new data useful for developing improved regulatory
analyses in the future.
Key Disadvantages
On their own, information disclosure requirements are not well-suited to addressing
market failures due to externalities; conditions necessary for bargaining to result in an
economically efficient outcome are not typical of most markets.
It may be particularly difficult for disadvantaged communities to influence polluters to
reduce their environmental damages in response to information disclosure.
Information programs have not been studied extensively, and empirical evidence on their
effectiveness is mixed.
-f.o.o UaHHn Rfjl'-.: "Ti >uran< merits
Liability rules impose a legal responsibility for polluters to pay for environmental damages after
they occur. These instruments serve two main purposes: (1) to create an economic incentive for
firms to incorporate the cost of environmental damages into their decision-making processes; and
(2) to compensate harmed individuals when damages occur. These rules are used to guide
compensation decisions when the court rules in favor of the victim. To the extent that polluters are
aware that they will be held liable before the release occurs, they have an incentive to minimize
damages to others.
While a liability rule can be constructed to mimic an economically efficient market solution in
certain cases, there are reasons to expect that this efficiency may not be achieved. First, payments
need not reflect the social damages. The amount that polluters are required to pay after damages
have occurred is dependent on the legal system and may be limited by an inability to prove the full
extent of damages or by the ability of the firm to pay. Second, liability rules can generate large
23 Khanna, Quimio and Bojilova (1998), Bui and Mayer (2003), Banzhafand Walsh (2008), and Mastromonaco
(2015) also have investigated how the TRI has affected firm behavior, stock market valuation and housing
markets.
24 For example, Teisl, et.al (2002) and Bj0rner, et. al (2004) studied the effects of labels for dolphin-safe tuna
and paper products, respectively, on consumer purchases. Brounen and Kok (2011) examined the extent to which
energy performance labels are capitalized into housing prices.
4-18 Guidelines for Preparing Economic Analyses | 3rd edition
-------
transaction costs, both in terms of assessing the environmental damage caused and the resources
used to take legal action.25
Liability rules are most useful in cases where damages requiring compensation are infrequent (e.g.,
accidental releases) and where monitoring compliance with other regulatory approaches is
difficult Finally, the scope of liability may affect overall economic efficiency. Under the
Comprehensive Environmental Response, Compensation and Liability Act of 1980 (CERCLA), for
example, new owners of contaminated land are defined to be potentially responsible parties that
can be held liable for past pollution, creating disincentives for the redevelopment of contaminated
land (Jenkins, Kopits and Simpson 2009).26 Depending on the effectiveness of liability rules to
provide incentives to firms to minimize environmental damages, they can be either an alternative
or a complement to other regulatory approaches.
Strict liability and negligence are two types of liability rules. Under strict liability, polluters are held
responsible for all health and environmental damages caused by their pollution, regardless of
actions taken to prevent the damages. Under negligence, polluters are liable only if they do not
exhibit "due standard of care." Regulations that impose strict liability on polluters may reduce the
transactions costs of legal actions brought by affected parties. This approach may induce polluters
to alter their behavior to reduce the probability of a pollution release that causes damages (Akey
and Appel 2021).
Requiring polluters to carry insurance is another approach that can be used to reward risk-
reducing and penalize risk-increasing behavior through the setting and adjustment of insurance
premiums. This instrument also generates a pool of money that can be used for remediation when
contamination occurs. Dana and Wiseman (2013) discussed this approach in the context of oil and
gas well development. Insurance has also been discussed to pool risk against extreme weather
events in the context of climate change (Linnerooth-Bayer and Hochrainer-Stigler 2015).
Key Advantages
Liability rules and insurance requirements can incentivize polluters to adopt behaviors that
reduce the risk of environmental damages.
Liability rules and insurance requirements are most useful for situations in which
environmental damages are infrequent and monitoring compliance with other types of
regulatory requirements is costly.
Liability rules and insurance requirements both involve ways to compensate those harmed
by contaminant releases.
Key Disadvantages
Payments by polluters to harmed individuals under liability rules are determined by the
legal system and need not be equal to social damages; therefore, on their own, they may not
create an incentive for polluters to undertake an economically efficient level of mitigation.
Insurance requirements will only yield an efficient level of environmental protection if
premiums are set to encourage firms to undertake abatement up to the level at which
marginal costs equals marginal social damages.
25 Segerson (1995) and Alberini and Austin (2001) discussed different types of liability rules and their efficiency
properties.
26 The Small Business Liability Relief and Brownfields Revitalization Act eased some ofCERCLA's liability
provisions to encourage the redevelopment of potentially contaminated industrial sites, known as brownfields.
4-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Determining payments through the legal system entails high transaction costs, including
resources used in the legal process and to measure environmental damages.
lavioral Economics and "Nudge" Approaches
The neoclassical economics paradigm that has helped inform the design of market-based and other
policy instruments makes several simplifications about human behavior for instance, that people
are rational, well-informed, self-interested and disciplined. While these may be reasonable
assumptions in many contexts, they do not always hold in the real world. Behavioral economics is a
subfield at the intersection of economics and psychology that examines departures from the
neoclassical economics model. Such behavioral anomalies include cognitive limitations, altruism,
inequality aversion, procrastination, status quo bias and loss aversion, among others.27
Behavior that is altruistic, short-sighted or inattentive may have important implications for the way
environmental policies are designed and enforced.28 Inattentive or impatient behavior may help
explain some consumers' reluctance to invest in energy-saving appliances or fuel-efficient cars that
cost more upfront but save money in the long run. Altruism and social norms may lead people to
purchase eco-labeled products even absent regulation or price signals.
Insights from behavioral economics can be relevant to the design of many types of policy
instruments. In addition, they present the opportunity to design policies that "nudge" people to
make choices that improve their well-being. Nudges have been proposed as an approach to
encourage socially beneficial actions by making small changes to the context in which people make
decisions. Thaler and Sunstein (2021) define a nudge as "any aspect of the choice architecture that
alters people's behavior in a predictable way without forbidding any options or significantly
changing their economic incentives," elaborating that, "the intervention must be easy and cheap to
avoid. Nudges are not taxes, fines, subsidies, bans, or mandates."
While market-based policies are typically designed to correct externalities, nudges may be
especially relevant in situations where the market under-provides environmental quality due to
lack of information, cognitive limitations, procrastination, or other behavioral anomalies. In
contrast to the use of information disclosure alone, a nudge or nudges emphasize the visual design,
timing, delivery method and other aspects of the way information is presented to make it more
salient and useful. Other strategies that have been used as nudges include default rules that require
individuals or firms to opt out of a program instead of opting in, moral persuasion or pro-social
messages that appeal to a sense of altruism or fairness, ordering choices to put the most beneficial
option first, and the use of social norms that tap into individuals' desire to match or outperform
their peers.29 Examples of nudges outside the realm of environmental policy include automatic
27 Loss aversion occurs when individuals facing risky choices place greater weight on losses compared to gains
of an equivalent value. Empirical research suggests that many people tend to give losses double the weight of
gains (Kahneman and Tversky 1979, Tversky and Kahneman 1992). Loss aversion can contribute to status quo
bias, which describes a preference for avoiding any change from the current situation.
28Shogren and Taylor (2008), Shogren, Parkhurst, and Banerjee (2010), and Croson and Treich (2014) provide
in-depth discussions of the intersection between behavioral economics and environmental economics.
29 Executive Order 13707, "Using Behavioral Insights to Better Serve the American People" (The White House,
Sept. 15,2015), encouraged federal agencies to consider behavioral science strategies with particular attention
to access to programs, presentation of information to the public, the structure of choices within programs and
the design of financial and non-financial incentives.
4-20 Guidelines for Preparing Economic Analyses | 3rd edition
-------
enrollment of employees into retirement savings plans (Madrian and Shea 2001) and rearranging
cafeterias to make healthy foods more convenient or eye-catching (Hanks etal. 2012).
There are many potential applications of nudges to environmental policy. For example, research
has shown that providing residential consumers with real-time information about electricity
consumption and prices can reduce electricity use, which can lead to decreased pollution from
fossil-powered electricity generation. Signals conveyed visually, such as a "glowing orb" that
changes color to reflect changes in prices or demand, have been shown to be particularly effective.30
Residential consumers who received reports comparing their own consumption of water or
electricity to that of their neighbors also reduced their resource consumption (Allcott 2011b,
Ferraro and Price 2013). Text Box 4.3 describes a few EPA examples.
Nudges that are effective in one situation are not always transferable to different contexts. For
example, the residential electricity consumption reports mentioned above led to larger reductions
in electricity use for high-user households and for environmentalists, while they have been less
effective for other households (Allcott 2015). In addition, research on electricity consumption has
yielded mixed results on the effectiveness of combining various nudges and financial incentives
(Pellerano etal. 2017; Brandon etal. 2019). These examples highlight the importance of using
rigorous empirical approaches such as randomized controlled trials to test the effectiveness of new
nudges before adopting them on a wide scale (List and Metcalfe 2014; Allcott and Mullainathan
2010; Hahn and Metcalfe 2016).
Beyond nudges, behavioral economics insights can be applied in the design of other policy
instruments. The implementation of plastic bag taxes provides one example. Standard economic
models predict that individual consumers will respond similarly to market incentives regardless of
whether they are presented as a tax on damaging activities or a subsidy for beneficial activities.
However, research has found that consumers faced with a fee for disposable bags cut their bag use
by more than 40%, but no change occurred in response to a subsidy for reusable bag use (Homonoff
2018). This result is consistent with loss aversion and suggests that consumer responsiveness to
market-based policies can depend on how the incentives are framed.
Key Advantages
Nudges can address environmental problems that occur or are exacerbated due to
inattention, impatience or other behaviors inconsistent with rational choice theory.
Nudges can complement other approaches, particularly information disclosure
requirements.
Nudges are low-cost and impose no mandatory requirements on the public.
Key Disadvantages
On their own, nudges are not well-suited to addressing market failure due to externalities.
Nudges are not well-suited to addressing sectors in which rational, profit-maximizing
behavior is well documented.
Empirical evidence on nudges' effectiveness in improving environmental outcomes is
limited.
30 Allcott (2011a) and Jessoe and Rapson (2014) focused on real-time electricity pricing, while Houde etal.
(2013) examined the effect of real-time electricity consumption information.
4-21 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 4.3 - Nudging Through Labels
Product labels represent an intriguing opportunity to examine whether the way information is
presented can nudge consumers toward environmentally friendly purchases. For example, some
research has found that the EPA's ENERGY STAR logo encourages investments in energy
efficient appliances more effectively than information on energy use and expenditures alone
(Newell and Siikamaki 2014).
The EPA also collaborated with the U.S. Department of Transportation in 2011 on the redesign
of labels to convey the fuel efficiency and environmental attributes of light duty vehicles. They
considered elements like color, layout, graphics and alternative rating scales. One issue they
confronted was which metric to use to represent fuel economy. Research by Larrick and Soil
(2008) pointed out that miles per gallon (mpg) can mislead consumers about fuel expenses and
tailpipe emissions because mpg is not linearly related to fuel consumption. Consumers are
especially likely to undervalue small changes in mpg for less fuel-efficient vehicles because most
are not aware that shifting from 10 to 12 mpg, for example, saves more fuel than increasing from
33 to 50 mpg for the same number of miles driven. Larrick and Soil proposed "gallons per 100
miles" as an alternative measure of fuel economy that is linear in fuel consumption. The agencies
used focus group testing to compare the different metrics and found that many participants
preferred mpg due to its familiarity (U.S. EPA 2010a, 2010b). For the final label, the agencies
kept the mpg metric, as required by law, but also included the gallons per 100 miles information
in smaller print. In addition, the label prominently featured fuel cost savings compared to the
average new vehicle, a highly relevant metric for consumers that allows for easy comparisons
across vehicles.
4.4 Voluntary Programs
The EPA has sometimes used voluntary programs as an alternative to regulations to reduce
emissions and other environmental hazards. Many EPA voluntary programs encourage polluting
entities to go beyond what is mandated by regulation. Other voluntary programs address
environmental quality in areas that may be regulated in the future but are currently not regulated.31
Voluntary programs can offer the EPA the opportunity to pilot new approaches or to work with
new industries before implementing a regulation with mandatory requirements.
The EPA typically designs voluntary programs through consultation with affected industries or
consumers. In many cases, voluntary programs facilitate problem solving between the EPA and
industry because information on practices that reduce pollutants and waste are shared through the
consultative process. Voluntary programs also frequently encourage peer education and
information sharing among participants. Data on abatement costs that are generated or disclosed
through voluntary programs could help to inform future programs, analysis or regulatory action in
the sector.
31 While this chapter only discusses EPA-led voluntary program, other government agencies, industry, non-
profits and international organizations have also organized voluntary programs to address environmental
issues.
4-22 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Voluntary programs can have either broad environmental objectives targeting a variety of firms
from different industries or focus on specific environmental problems relevant to a single industrial
sector.32 They often use one or more of the following four approaches:
Encourage firms or facilities to set specific environmental goals. Goals can be EPA-
specified, program-wide targets designed to provide a consistent objective across firms. In
other cases, goals are qualitative and process-oriented so that a firm may set a unique
target.
Promote firm environmental awareness and encourage process change within firms.
Programs designed to promote environmental awareness and process change often involve
implementing a system to evaluate firms' operations and to provide information on new
technologies. These programs may also promote or recognize use of third-party industry
standards for products and materials.
Publicly recognize firm participation. Voluntary programs that publicly recognize firm
participation are designed to provide green consumers and investors with new
information that may alter their consumption and investment patterns in favor of
cleaner firms. Firms may also use their environmental achievements to differentiate
their products from competitors' products.33
Use labeling to identify environmentally responsible products. Product labeling can be
applied to either intermediate inputs or final goods. Labels on intermediate goods
encourage firms to purchase environmentally responsible inputs. Labels on final goods
allow consumers to identify goods produced using a relatively clean production process.
Section 4.3.4 and Text Box 4.3 discuss how labeling can be made more effective by using
behavioral economics concepts.
The economics literature has not systematically evaluated the effectiveness of these four
approaches. Like mandatory information disclosure programs, economic theory suggests that
approaches involving sharing information among firms or labelling consumer products may be
most useful in situations where imperfect or asymmetric information leads to adverse
environmental outcomes.
Most empirical studies of EPA voluntary programs have focused on a few large, multi-sector
programs such as 33/50, Green Lights and ENERGY STAR. They have found mixed evidence
regarding the extent to which these programs have reduced emissions. Many studies failed to
account for what would have occurred absent the program, potentially overstating reductions. The
potential for beneficial information or technology spillovers from program participants to other
firms in the target industry can make it difficult to measure a program's impact (Lyon and Maxwell
2007).34 Many smaller regulatory programs remain unstudied.
32 See Brouhle et al. (2005), Lyon and Maxwell (2007), Borck and Coglianese (2009), and Prakash and Potoski
(2012) for discussions of how voluntary programs have been used in U.S. environmental policy.
33 See Konar and Cohen (2001), Videras and Alberini (2000), Brouhle, Griffiths and Wolverton (2005),
Morgenstern and Pizer (2007), and Borck and Coglianese (2011 )for more information on the main arguments
for why firms participate in voluntary programs.
34 One thread of literature points to the role a regulatory threat plays in improving voluntary program
effectiveness. When the threat of regulation is weak, abatement levels are lower. However, when the threat of
regulation is strong, Segerson and Wu (2006) showed that levels achieved are closer to those that would be
achieved under a standards-based approach. See also Morgenstern and Pizer (2007); Brouhle, Griffiths and
4-23 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Key Advantages
Voluntary programs allow agencies the opportunity to pilot new approaches to working
with industries or on environmental problems not yet subject to regulation, which could be
particularly useful if there is substantial uncertainty about the benefits or costs of
regulation.
Voluntary programs involving data gathering and reporting by program participants could
yield new data useful for future analyses or regulatory actions.
Voluntary programs with a labeling or information disclosure component could be well-
suited to address market failures due to asymmetric or imperfect information.
Key Disadvantages
If voluntary programs only attract participants that are already industry leaders in
environmental protection, they may not yield significant improvements in environmental
outcomes relative to a baseline without the voluntary program.
Economic theory suggests that firms or individuals are unlikely to participate if the private
costs of participating exceed the private benefits, even if the social net benefits of
participating are positive.
Empirical studies on the effectiveness of voluntary programs in improving environmental
outcomes are limited, and the available evidence is mixed.
4.5 Cross-Cutting Issues When Comparing Regulatory and
I ^«si f ¦ ,;mI i i ^ H 11 ¦ ^
Using a simplified theoretical framework, Fullerton (2001) demonstrated that a variety of
regulatory and non-regulatory approaches can be designed to achieve the same level of economic
efficiency.35 In practice, there are likely to be important tradeoffs across approaches. Economic
analysis can play an important role in identifying these tradeoffs.
Analysts can provide insight into the approaches that maximize net benefits and how they vary in
efficiency over time. One regulatory feature that reduces economic efficiency is "grandfathering"
a practice in which older polluters are exempted from new regulations or are subjected to a less
stringent standard than newer polluters. Grandfathering creates a bias against constructing new
facilities and investing in new pollution control technology or production processes. As a result,
grandfathered older facilities with higher emission rates tend to remain active longer than they
would if the same emission standard applied to all polluters (Helfand, 1991; Stavins, 2006).
In general, varying regulatory requirements by firm age, size, location or other attributes can
increase efficiency by helping to equalize heterogeneous abatement costs or benefits, but it can also
reduce the efficiency of a policy if the design creates perverse incentives to shift production away
from more regulated firms or products toward those that are less regulated. Further distortions are
introduced when a small change in behavior results in a discrete change in regulatory
requirements. Such discontinuities or "notches" incentivize firms to strategically avoid going over
Wolverton (2009); Lange (2009); Vidovic and Khanna (2011); Kim and Lyon (2011); Brouhle, Graham and
Harrington (2013); and Ferrara and Lange (2014).
35 Fullerton (2001) assumed no administrative costs, perfect information, no enforcement issues, perfect labor
mobility and competitive firms.
4-24 Guidelines for Preparing Economic Analyses | 3rd edition
-------
the specified threshold to minimize the cost of compliance (Ito and Sallee 2018). For example, some
policies exempt firms below a certain size, which can discourage firms from expanding or
consolidating to avoid the regulation (Blinder and Rosen 1985; Sallee and Slemrod 2012). Chapter
3 provides more discussion of these policy design features.
There are several other cross-cutting issues that may be useful to analyze when evaluating
potential tradeoffs across approaches. These include distributional impacts and environmental
justice; administrative, monitoring and enforcement costs; interactions with other distortions;
degree of flexibility; information requirements and uncertainty; and the nature of the
environmental problem.36 Analysts can evaluate these factors using methods discussed in Chapters
7 through 10. Stringency is another important consideration for regulatory design that is discussed
in Chapter 3.
-i * i Distribution i inrj" cts an i-ii ii iii isntal Justice
The distribution of costs and benefits across firms, workers, governments, households and
individuals over time and space is often of interest to decision makers.37 For example, market-
based instruments that directly affect the price of the goods produced by polluting firms will likely
have different distributional and environmental justice consequences than prescriptive regulations
(Berck and Helfand 2005; Rozenberg et al. 2020; Zhao and Mattauch 2022). A commonly expressed
concern is that market-based policies to reduce greenhouse gas emissions may lead to increases in
other pollutants in already overburdened neighborhoods.38
The distribution of economic rents may also differ across approaches. If allowances are auctioned
or sold to polluters, the distributional consequences of a cap-and-trade policy will be similar to
those of emissions taxes. If allowances are instead distributed for free, distributional consequences
will depend on the allocation approach (e.g., historical output or inputs; updating), who receives the
allowances and the ability of the recipients to pass the costs onto their customers. Likewise, for
approaches that raise revenue such as emissions taxes or a cap-and-trade policy that auctions
allowances the way revenue is used will affect the distributional outcomes (Burtraw et al.
2010).39
36 Many of these criteria are also highlighted in Fullerton (2001). Another criterion discussed by Fullerton
(2001) is political and ethical considerations. The approach ultimately chosen will also depend on statutory and
other legal limitations. This chapter does not expand on these considerations because analysts have a limited
role to play in evaluating them.
37 See Chapter 9 for approaches to quantify the economic impacts of approaches under consideration. See
Chapter 1 Ofor discussion of impacts on minority, low-income or Indigenous populations and on children and
older adults.
38 In the context of California, research has not reached consensus on the degree to which its carbon cap-and-
trade policy has exacerbated existing criteria air emissions in communities of color or low-income communities
(e.g., Fowlie et al, 2012; Grainger and Ruangmas, 2017; Mansur and Sheriff, 2021; Hernandez-Cortez and Meng,
2023; Cushing etal, 2018).
39 To explicitly weight economic efficiency alongside distributional or environmental justice considerations,
analysts would need to employ a social welfare function that aggregates welfare across individuals into a single
value to allow an explicit ranking of different policy options (see Adler2008, 2012). However, a social welfare
function is based on a normative judgement, and while it makes the criteria explicit regarding how society
4-25 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Differing treatment applied to sources based on age, size, location or other attributes can also affect
the distribution of revenues, expenses and rents within the economy under both prescriptive and
market-based approaches.
-i * C v i t irative, Monitr d Enforcement v
Analysts can help shed light on differences in the cost of administering, monitoring and enforcing
the approaches under consideration. For instance, what are the costs and foreseeable challenges for
ensuring compliance? Is pollution observable, or will it need to be estimated based on inputs and
technology used? Are technologies available to decrease the costs of monitoring and reporting?
When pollutant emissions or concentrations can be easily measured, it is more feasible to directly
regulate the level of the pollutant. For example, continuous emissions monitoring equipment at
power plants allowed for direct measurement of pollution and facilitated the use of a cap-and trade
system to regulate SO2 emissions (see Text Box 4.2). If a source has fewer allowances than the
monitored emission levels at the end of a compliance period, it is in noncompliance and the source
must provide allowances to cover its environmental obligation and pay a penalty.40
If monitoring and enforcement costs are high, a regulation may fail to deliver environmental
benefits due to widespread noncompliance (e.g., illegal dumping).41 In these cases, it may be easier
to regulate a related input or output to leverage approaches that incentivize sources to reveal
information about their production or abatement processes (e.g., a tax and subsidy combination).
Mandating the use of specific abatement technologies can sometimes reduce the cost of monitoring
compliance, as noted in Section 4.1.1. In addition, it may be easier to monitor and enforce
regulations on a smaller number of "upstream" sources (e.g., oil refineries) rather than a larger set
of "downstream" sources (e.g., gasoline consumers) (Mansur 2012).
tactions with Other Distortions
Analysts should consider the potential distortionary effects of any policy option considered. Even if
a policy is relatively efficient on its own, it may interact with pre-existing environmental, trade, tax
or agricultural policies in ways that exacerbate distortions in the economy and result in additional
social costs. One such distortion occurs when imperfect competition due to market power results in
lower output than would occur in a competitive market, which results in a loss in economic welfare.
Policy instruments that cause firms to restrict output (e.g., an emissions tax) may create additional
inefficiencies in sectors where firms have some degree of market power (Baumol and Oates 1988;
Fowlie et al. 2016). A combination of market-based instruments may work more effectively than a
single instrument in this instance.
prefers to distribute resources across individuals, there is no consensus regarding those preferences. Thus,
distributional information is typically analyzed and presented separate from efficiency considerations.
40 The U.S. Acid Rain Trading Program has high levels of compliance and requires fewer than 50 EPA staff to
administer since penalties are automatically levied for each ton of excess emissions (Napolitano et ah, 2007).
41 However, Sigman (2012) presented a theoretical model showing that compliance need not decrease when
regulations are broadened beyond industries with low-cost monitoring to include those where monitoring costs
are higher, but abatement costs are lower.
4-26 Guidelines for Preparing Economic Analyses | 3rd edition
-------
If costs differ between existing and new firms, the use of certain instruments may cause a change in
market structure that favors existing firms by creating barriers to entry and allowing existing firms
a certain amount of control over price. Cap-and-trade systems that set aside a certain number of
allowances for new firms may guard against such barriers.
Instruments that involve the government collecting revenue, such as auctioned allowances or taxes,
may create opportunities to reduce distortions.42 At the same time, society also incurs a welfare
loss from raising revenues through taxes due to the difference between the value of an additional
dollar raised by the government and the value of that dollar to a private individual (termed the
marginal cost of public funds). See Chapter 8 and Appendix A for more discussion of analyzing
welfare effects under sectoral or economy-wide distortions.
-i -i L* -11 - - -f Kl - Hihhn -ii n n?iTri v tment
Even if a regulation is set at an economically efficient level at the outset, changing conditions over
time can result in inefficient levels of pollution control. To what extent does the approach allow for
automatic adjustments in requirements or stringency over time in response to new information or
technological improvements? Is the approach flexible enough to accommodate transition costs?
Does the approach encourage innovation in abatement techniques that decrease the cost of
compliance with environmental regulations over time?43
For instance, market-based approaches often differ from prescriptive approaches to regulation by
encouraging firms to find the cheapest way to reduce emissions. The incentive to innovate means
that the marginal abatement cost curve may shift downward over time as cheaper compliance
options become available. If innovation causes the cost of pollution control to fall, the marginal cost
of decreasing pollution levels could drop below the marginal benefit A cap-and-trade approach
incorporating a price floor below which allowances are removed from the market is one approach
to dynamically adjusting a regulation. Similarly, a price ceiling above which additional allowances
are introduced to the market can be used to ensure that marginal costs do not rise too far above
marginal benefits (Fell et al. 2012). Features such as banking and borrowing also afford regulated
plants some flexibility in the timing of reductions.
-f." Information Requirements n In- r
What information is required to implement the approach? How well does the approach perform
under imperfect or asymmetric information, or when there is uncertainty about costs and/or
benefits? Can the approach be designed in a way that will reveal new information about costs and
benefits that can reduce uncertainty if additional analysis or regulatory action is considered in the
future?
When abatement costs and benefits are certain, price-based instruments (e.g., emissions taxes) and
quantity-based instruments (e.g., cap-and-trade) are theoretically equivalent and can be designed
to achieve the same outcomes. However, this result may not hold when there is uncertainty about
42 For more information on the how revenues raised via market-based instruments affect social welfare, see
Bovenberg and Goulder (1996), Goulder (2013), Jorgenson etal. (2013), and McKibbin etal. (2015).
43 For a theoretical analysis of incentives for technological change, see Jung et al. (1996) and Montero (2002).
Empirical analyses can be found in Jaffe and Stavins (1995), Kerr and Newell (2003), Requate (2005), and
Newell (2010).
4-27 Guidelines for Preparing Economic Analyses | 3rd edition
-------
the benefits and costs of pollution control, or when marginal benefits and costs change substantially
with the stringency of the pollution control target (Weitzman 1974).44 If uncertainty associated
with the abatement costs exists but damages do not change much with additional pollution, then
policymakers can limit costs by using a price instrument without having much impact on the
benefits of the policy. If, on the other hand, there is more uncertainty associated with the benefits of
controlling pollution, and policymakers wish to guard against high environmental damages, a
quantity instrument is preferable. In some circumstances, this may come to resemble a more
prescriptive approach that specifies zero allowable source-level emissions to avoid potentially
costly or damaging mistakes. Hybrid approaches that combine features of price and quantity
instruments or add other flexibilities to address uncertainty can also vary with respect to dynamic
efficiency (Pizer 2002; Weitzman 2020; see Section 4.3.1.3).
Other types of regulatory and non-regulatory approaches may reveal information about emissions,
abatement approaches, or abatement costs that can facilitate both retrospective analysis to
understand how well an approach is working and prospective analysis of potential future
regulatory actions.45 Monitoring and reporting requirements can be used to compel regulated
entities to release data on emissions or abatement approaches. Allowance trading systems can
reveal information to regulators and the public about abatement costs because the equilibrium
allowance price indicates the marginal cost of compliance with the regulation. Subsidy programs
may also require participants to reveal information about abatement activities.
In some instances where there is a high level of uncertainty about costs and/or benefits, a voluntary
program or pilot project may be a compelling alternative to regulation. Such an approach
encourages environmental improvements but allows the government and regulated community to
test out different abatement technologies or process changes and gather information on what
works and what does not Research and development efforts may also contribute to better
understanding the costs and benefits of regulating. As technology improves or more data become
available, analysts will be better able to analyze a variety of approaches. Value of information
analysis could be used to examine whether more resources should be invested to reduce
uncertainty before developing a regulation and which aspects of uncertainty to prioritize (Finkel
and Evans 1987).46
-i * ^ fli - Natu - vi" ill -i Environmental Problem
Another important issue is the type of environmental problem being addressed. Are the sources
heterogeneous? Does the pollutant vary across time and space? Do emissions derive from a point
source or a nonpoint source? Do the pollutants persist in the environment or dissipate rapidly?47
Point sources, which emit at identifiable and specific locations, are typically easier to control than
44 Pezzey and Jotzo (2012) built on Weitzman (1974) by examining how revenue recycling affects the welfare
implications of a price- versus quantity-based market instrument under uncertainty.
45 Chapter 5 (Text Box 5.1) provides more discussion of retrospective analysis.
46 For more discussion and examples of value of information analysis in environmental policy, see Cullen and
Frey (1999), Dekay etal. (2002), Keisleretal. (2014), Marchese etal. (2018), Thompson and Evans (1997), and
Yokota and Thompson (2004).
47 For a more discussion of how the nature of the environmental problem affects instrument choice, see Kahn
(2005); Goulderand Parry (2008); Parry and Williams (1999); Sterner and Coria (2012); Tietenberg and Lewis
(2014); and Xabadia, Goetz and Zilberman (2008).
4-28 Guidelines for Preparing Economic Analyses | 3rd edition
-------
diffuse, numerous nonpoint sources and are often responsive to a variety of approaches.
Monitoring and control of nonpoint source emissions are more challenging (see Text Box 4.1). In
instances where both point and nonpoint sources contribute to a pollution problem, a case can be
made for a tax-subsidy combination (with taxes directed toward point sources and subsidies to
nonpoint sources) or an allowance trading system with offsets.
Flow pollutants that dissipate quickly are responsive to a wide variety of market and hybrid
instruments. In contrast, stock pollutants that persist in the environment may require strict limits
to prevent bioaccumulation or detrimental health effects at small doses, making direct regulation
appealing. Approaches that set a limit on the overall quantity of pollution may be also preferred if
there are discontinuities or threshold values above which sudden or large changes in
environmental damages could occur (Pindyck 2007). For pollutants that do not mix uniformly, it is
important to account for differences in baseline pollution levels and in emissions across more- and
less-polluted areas. Damages can also vary by time of day or season. For example, health impacts
associated with vehicle emissions may be larger during rush hour because roads are congested, and
cars spend time idling or in stop-and-go traffic. Differential pricing of resources used by these
mobile sources (such as higher tolls on roads or greater subsidies to public transportation during
rush hour) is a potentially useful tool.
4-29 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 4 References
Akey, P. and I. Appel. 2021. The Limits of Limited Liability: Evidence from Industrial Pollution. The Journal of
Finance 76(1): 5-55.
Alberini, A. and D. Austin. 2001. Liability Policy and Toxic Pollution Releases. The Law and Economics of the
Environment, ed. A. Heyes. Cheltenham, U.K.: Edward Elgar Publishing.
Allcott, H. 2011a. Rethinking Real-Time Electricity Pricing. Resource and Energy Economics, 33: 820-842.
Allcott, H. 2011b. Social Norms and Energy Conservation. Journal of Public Economics, 95(9-10): 1082-1095.
Allcott, H. 2015. Site Selection Bias in Program Evaluation. Quarterly Journal of Economics, 130(3): 1117-
1165.
Allcott, H. and S. Mullainathan. 2010. Behavior and Energy Policy. Science, 327(5970): 1204-1205.
Arora, S. and T. Cason. 1999. Do Community Characteristics Influence Environmental Outcomes? Evidence
from the Toxics Release Inventory. Southern Economic Journal, 65(4): 691-716.
Bae, H., P. Wilcoxen, and D. Popp. 2010. Information Disclosure Policy: Do State Data Processing Efforts Help
More than the Information Disclosure Itself?" Journal of Policy Analysis and Management, 29(1): 163-182.
Banzhaf, S. and R. Walsh. 2008. Do People Vote With Their Feet? An Empirical Test of Teibout. American
Economic Review, 98(3): 843-863.
Baumol, W.J. and W.E. Oates. 1988. The Theory of Environmental Policy. 2nd Ed. New York: Cambridge
University Press.
Bellas, A. and I. Lange. 2011. Evidence of Innovation and Diffusion under Tradable Permit Programs.
International Review of Environmental and Resource Economics, 5(1): 1-22.
Bennear, L. and C. Coglianese. 2012. The Oxford Handbook of U.S. Environmental Policy, ed. S. Kamieniecki and
M. Kraft. Oxford: Oxford University Press.
Bento, A.M., M.R. Jacobsen, C.R. Knittel, and A.A. van Benthem. 2020. Estimating the Costs and Benefits of Fuel-
Economy Stands. Environmental and Energy Policy and the Economy, 1:129-157.
Berckand Helfand. 2005. The Case of Markets Versus Standards for Pollution Policy. Natural Resources
Journal, 45(2): 345-368.
Bj0rner, T., L. Hansen., and C. Russell. 2004. Environmental Labeling and Consumers' Choicean Empirical
Analysis of the Effect of the Nordic Swan .Journal of Environmental Economics and Management, 47(3): 411-
434.
Blinder, A. and H. Rosen. 1985. Notches. American Economic Review 75(4): 736-747.
Borck, J. and C. Coglianese. 2009. Voluntary Environmental Programs: Assessing Their Effectiveness. Annual
Review of Environment and Resources, 34: 305-324.
Borck, J. and C. Coglianese. 2011. Beyond Compliance: Explaining Business Participation in Voluntary
Environmental Programs, Explaining Compliance: Business Responses to Regulation, ed. C. Parker and V.
Nielsen. Northampton: Edward Elgar Publishing.
Brounen, D. and N. Kok. 2011. On the Economics of Energy Labels in the Housing Market .Journal of
Environmental Economics and Management, 62(2): 166-179.
Bovenberg, A.L. and L.H. Goulder. 1996. Optimal Taxation in the Presence of Other Taxes: General-
Equilibrium Analysis. American Economic Review, 86(4): 985-1000.
Boyd, J. 2002. Financial Responsibility for Environmental Obligations: Are Bonding and Assurance Rules
Fulfilling Their Promise? Research in Law and Economics, 20: 417-486.
4-30 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Brandon, A., J.A. List, R.D. Metcalfe. M.K. Price, and F. Rundhammer. 2019. Testing for Crowd Out in Social
Nudges: Evidence from a Natural Field Experiment in the Market for Electricity. Proceedings of the National
Academy of Sciences, 116(12): 5293-5298.
Brouhle, K., B. Graham, and D. Harrington. 2013. Innovation Under the Climate Wise Program Resources and
Energy Economics, 35: 91-112.
Brouhle, K., C. Griffiths, and A. Wolverton. 2005. The Use of Voluntary Approaches for Environmental
Policymaking in the U.S. The Handbook of Environmental Voluntary Agreements, ed. E. Croci. The Netherlands:
Kluwer Academic Publishers. 107-134.
Brouhle, K., C. Griffiths, and A. Wolverton. 2009. Evaluating the Role of EPA Policy Levers: An Examination of a
Voluntary Program and Regulatory Threat in the Metal Finishing Industry. Journal of Environmental
Economics and Management, 57(2): 166-181.
Brouhle, K. and M. Khanna. 2007. Information and the Provision of Quality-Differentiated Goods. Economic
Inquiry, 45(2): 377-395.
Bui, L. and Mayer, C. 2003. Regulation and Capitalization of Environmental Amenities: Evidence from the
Toxic Release Inventory in Massachusetts. Review of Economics and Statistics, 85(3): 693-708.
Bui, L. 2005. Public Disclosure of Private Information as a Tool for Regulating Emissions: Firm-level
Responses by Petroleum Refineries to the Toxics Release Inventory. Working paper 05-13, US Census Bureau.
Burtraw, D., D.A. Evans, A. Krupnick, K. Palmer, and R. Toth. 2005. Economics of Pollution Trading for S02 and
NOx. Annual Review of Environment and Resources, 30: 253-289.
Burtraw, D., K. Palmer, and D. Kahn. 2010. A Symmetric Safety Valve. Energy Policy, 38(9): 4921-4932.
Burtraw, D., M. Walls, and J.A. Blonz. 2010. Distributional Impacts of Carbon Policies in the Electricity Sector.
US Energy Tax Policy, ed. G. Metcalf. Cambridge: Cambridge University Press.
Chan, H., A. Chupp, M. Cropper, and N. Muller. 2018. The Impact of Trading on the Costs and Benefits of the
Acid Rain Program. Journal of Environmental Economics and Management, 88:180-209.
Chestnut, L.G. and D.M. Mills. 2005. A Fresh Look at the Benefits and Costs of the U.S. Acid Rain Program.
Journal of Environmental Management, 77(3): 252-256.
Coase, R. 1960. The Problem of Social Cost. The Journal of Law and Economics, 3:1-44.
Cronin, J.A., D. Fullerton, and S. Sexton. 2019. Vertical and Horizontal Redistributions from a Carbon Tax and
Rebate. Journal of the Association of Environmental and Resource Economists, 6(S1): S169-S208.
Croson, R. and N. Treich. 2014. Behavioral Environmental Economics: Promises and Challenges.
Environmental and Resource Economics, 58(3): 335-351.
Cullen, A. and H.C. Frey. 1999. Probabilistic Techniques in Exposure Assessment: A Handbook for Dealing with
Variability and Uncertainty in Models and Inputs. New York: Plenum Press.
Dana, D. and H. Wiseman. 2013. A Market Approach to Regulating the Energy Revolution: Assurance Bonds,
Insurance, and the Certain and Uncertain Risks of Hydraulic Fracturing Iowa Law Review, 99:1523-1594.
Davis, L. 2015. Policy Monitor - Bonding Requirements for U.S. Natural Gas Producers. Review of
Environmental Economics and Policy, 9(1): 128-144.
de Vries, F. and N. Hanley. 2016. Incentive-Based Policy Design for Pollution Control and Biodiversity
Conservation: An Overview. Environmental and Resource Economics, 63: 687-702.
Dekay, M.L., M.J. Small, P.S. Fischbeck, R.S. Farrow, A. Cullen, J.B. Kadane, L.B. Lave, M.G. Morgan, and K.
Takemura. 2002. Risk-Based Decision Analysis in Support of Precautionary Policies. Journal of Risk Research,
5(4).
Earnhart, D. 2004. The Effects of Community Characteristics on Polluter Compliance Levels. Land Economics,
80(3): 408-432.
4-31 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Ellerman, D. 2006. Are Cap-and-Trade Programs More Environmentally Effective than Conventional
Regulation? Moving the Markets in Environmental Regulation: Lessons from Twenty Years of Experience, ed. J.
Freeman and C. Kolstad. Oxford: Oxford University Press.
Evans, D.A. and R. Woodward. 2013. What Can We Learn From the End of the Grand Policy Experiment? The
Collapse of the National S02 Trading Program and Implications for Tradable Permits as a Policy Instrument.
Annual Review of Resource Economics, 5: 325-348.
Executive Order 13707: Using Behavioral Science Insights to Better Serve the American People, September
15, 2015. Available at: https: //ohamawhitehouse.archives.gov/the-press-office/2015/09/15/executive-
order-using-behavioral-science-insights-better-serve-american (accessed December 22, 2020).
Fell, H. and R. Morgenstern. 2010. Alternative Approaches to Cost Containment in a Cap-and-Trade System.
Environmental and Resource Economics 47: 275-297.
Fell, H., D. Burtraw, R. Morgenstern, and K. Palmer. 2012. Soft and Hard Price Collars in a Cap-and-Trade
System: A Comparative Analysis. Journal of Environmental Economic and Management, 64(2): 183-198.
Ferrara, A. and I. Lange. 2014. Voluntary Programs to Encourage Diffusion: The Case of the Combined Heat-
and Power Partnership. Energy Journal, 35(1): 161-173.
Ferraro, P. and M. Price. 2013. Using Nonpecuniary Strategies to Influence Behavior: Evidence from a Large-
Scale Field Experiment. The Review of Economics and Statistics, 95(1): 64-73.
Field, B. and M. Field. 2021. Environmental Economics. 8th Edition. New York: McGraw-Hill.
Fischer, C., and A. Fox. 2007. Output-Based Allocation of Emissions Permits for Mitigating Tax and Trade
Interactions Land Economics 83(4): 575-599.
Fowlie, M., S.P. Holland, and E.T. Mansur. 2012. What Do Emissions Markets Deliver and to Whom? Evidence
from Southern California's NOx Trading Program. American Economic Review 102(2): 965-93.
Fowlie, M., M. Greenstone, and C. Wolfram. 2015. Are the Non-Monetary Costs of Energy Efficient Investments
Large? Understanding Low Take-up of a Free Energy Efficiency Program. American Economic Review: Papers
& Proceedings 105(5): 201-204.
Fowlie, M., Reguant, M. and Ryan, S.P. 2016. Market-Based Emissions Regulation and Industry Dynamics.
Journal of Political Economy, 124(l):249-302.
Frey, E. 2013. Technology Diffusion and Environmental Regulation: The Adoption of Scrubbers by Coal-Fired
Plants. Energy Journal, 34(1): 177-205.
Fullerton, D. 2001. A Framework to Compare Environmental Policies. Southern Economic Journal, 68(2): 224-
248.
Fullerton, D., A. Leicester, and S. Smith. 2010. Environmental taxes. Dimensions of Tax Design, ed. S. Adam, and
others. Oxford: Oxford University Press.
Fullerton, D. and G.E. Metcalf. 2001. Environmental Controls, Scarcity Rents and Pre-Existing Distortions.
Journal of Public Economics, 80(2): 249-267.
Fullerton, D. and S. West. 2010. Tax and Subsidy Combinations for the Control of Car Pollution. The BE Journal
of Economic Analysis and Policy, 10(1): 1-33.
Fullerton, D. and A. Wolverton. 2001. The Case for a Two-Part Instrument: Presumptive Tax and
Environmental Subsidy. The Economics of Household Garbage and Recycling Behavior, ed. Fullerton, D. and T.
Kinnaman. Cheltenham, UK: Edward Elgar Publishing.
Fullerton, D. and A. Wolverton. 2005. The Two-Part Instrument in a Second-Best World. Journal of Public
Economics, 89:1961-1975.
GAO (US Government Accountability Office). 2017. Water Pollution: Some States Have Trading Programs to
Help Address Nutrient Pollution, but Use Has Been Limited. GAO-18-84, October 2017.
4-32 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Goulder, L.H. 2000. Environmental Policy Making in a Second-Best Setting. Economics of the Environment:
Selected Readings. 4th Ed, ed. R. Stavins. New York: W.W. Norton.
Goulder, L.H. 2013. Climate Change Policy's Interactions with the Tax System. Energy Economics, 40(S1]: S3-
Sll.
Goulder, L.H. and I.W.H. Parry. 2008. Instrument Choice in Environmental Policy Design. Review of
Environmental Economics and Policy, 2:152-174.
Grainger, C., & Ruangmas, T. (2018). Who Wins from Emissions Trading? Evidence from California.
Environmental and Resource Economics 71: 703-727.
Hahn, R. and R. Metcalfe. 2016. The Impact of Behavioral Science Experiments on Energy Policy. Economics of
Energy and Environmental Policy, 5(2).
Hamilton, J. 1993. Politics and Social Costs: Estimating the Impact of Collective Action on Hazardous Waste
Facilities. RAND Journal of Economics, 24(1): 101-125.
Hamilton, J. 1995. Pollution as News: Media and Stock Market Reactions to the Toxics Release Inventory Data.
Journal of Environmental Economics and Management, 28: 98-113.
Hanks, A., D. Just, L. Smith, and B. Wansink. 2012. Healthy Convenience: Nudging Students Toward Healthier
Choices in the Lunchroom. Journal of Public Health, 34(3): 370-376.
Harrington, W., R. Morgenstern, and T. Sterner. 2004. Choosing Environmental Policy: Comparing Instruments
and Outcomes in the United States and Europe. Washington, DC: Resources for the Future.
Helfand, G. 1991. Standards versus Standards: The Effects of Different Pollution Restriction. American
Economic Review, 81(4): 622-634.
Hernandez-Cortes, D & Meng, K. C. (2023). Do Environmental Markets Cause Environmental Injustice?
Evidence from California's Carbon Market. Journal of Public Economics, 217,104786.
Holland, S., J. Hughes, and C. Knittel. 2009. Greenhouse Gas Reductions Under Low Carbon Fuel Standards?
American Economic Journal: Economic Policy 1(1): 106-146.
Homonoff, T. 2018. Can Small Incentives Have Large Effects? The Impact of Taxes versus Bonuses on
Disposable Bag Use. American Economic Journal: Economic Policy, 10(4): 177-210.
Houde, S. A. Todd, A. Sudarshan, J. Flora and C. Armel. 2013. Real-Time Feedback and Electricity
Consumption: A Field Experiment Assessing the Potential for Savings and Persistence. The Energy Journal,
34(1): 87-102.
Ito, K., and J. Sallee. 2018. The Economics of Attribute-Based Regulation: Theory and Evidence from Fuel-
Economy Standards. Review of Economics and Statistics 100(2): 319-336.
Jacoby, H. and A.D. Ellerman. 2004. The Safety Value and Climate Policy. Energy Policy, 32(4): 481-491.
Jaffe, A.B. and R.N. Stavins. 1995. Dynamic Incentives of Environmental Regulations: The Effects of Alternative
Policy Instruments on Technology Diffusion. Journal of Environmental Economics and Management, 29: S43-
S63.
Jaffe, A.B. R. Newell, R.N. Stavins. 2005. A Tale of Two Market Failures: Technology and Environmental Policy.
Ecological Economics, 54:164-174.
Jenkins, R., E. Kopits, and R.D. Simpson. 2009. Policy MonitorThe Evolution of Solid and Hazardous Waste
Regulation in the United States. Review of Environmental Economics and Policy, 3(1): 104-120.
Jessoe, K. and D. Rapson. 2014. Knowledge is (less) Power: Experimental Evidence from Residential Energy
Use. American Economic Review, 104(4): 1417-1438.
Johnstone, N., Hascic, I., and M. Kalamove. 2010. Environmental Policy Characteristics and Technological
Innovation. Economia Politico, 27(2): 275-299.
4-33 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Jorgenson, D., R. Goettle, M. Ho, and P. Wilcoxen. 2013. Double Dividend: Environmental Taxes and Fiscal
Reform in the United States. Cambridge, MA: MIT Press.
Jung, C., K. Krutilla, and R. Boyd. 1996. Incentives for Advanced Pollution Abatement Technology at the
Industry Level: An Evaluation of Policy Alternatives. Journal of Environmental Economics and Management,
30: 95-111.
Kahn, J. 2005. The Economic Approach to Environmental and Natural Resources. 3rd Ed. London:
Thomson/South-Western.
Kahneman, D. and A. Tversky. 1979. Prospect Theory: An Analysis of Decision Under Risk. Econometrica,
47(2): 263-291.
Keisler, J.M., Collier, Z.A., Chu, E., Sinatra, N., and Linkov. I. 2014. Value of Information Analysis: The State of
Application. Environmental Systems and Decisions 34(1): 3-23.
Keohane, N. and S. Olmstead. 2016. Markets and the Environment. Washington, DC: Island Press.
Kerr, S. and R. Newell. 2003. Policy-Induced Technology Adoption: Evidence from the U.S. Lead Phasedown.
Journal of Industrial Economics, 51(2): 317-343.
Khanna, M., W. Quimio, and D. Bojilova. 1998. Toxics Release Information: A Policy Tool for Environmental
Protection. Journal of Environmental Economics and Management, 36: 243-266.
Kim, E. and T. Lyon. 2011. Strategic Environmental Disclosure: Evidence form the DOE's Voluntary
Greenhouse Gas Registry. Journal of Environmental Economics and Management, 61: 311-326.
Klemick, H., A. Wolverton, B. Parthum, K. Epstein, S. Kutzing, and S. Armstrong. 2024. Factors Influencing
Participation In a Program to Replace Lead Pipes for Drinking Water. Environmental and Resource Economics
87: 791-832.
Kolstad, C. 2010. Environmental Economics. 2nd Edition. New York: Oxford University Press.
Konar, S. and M. Cohen. 1997. Information as Regulation: The Effect of Community Right-to-Know Laws on
Toxic Emissions .Journal of Environmental Economics and Management, 32:109-124.
Konar, S. and M. Cohen. 2001. Does the Market Value Environmental Performance? Review of Economics and
Statistics, 83(2): 281-89.
Lange, I. 2009. Evaluating Voluntary Measures with Treatment Spillovers; The Case of Coal Combustion
Products Partnership. B.E. Journal of Economic Analysis and Policy, 9(1): Article 36.
Lange, I., and P. Maniloff. 2021. Updating Allowance Allocations in Cap-and-Trade: Evidence from the NOx
Budget Program. Journal of Environmental Economic and Management 105:102380.
Larrick. R. and J. Soil. 2008. The MPG illusion. Science, 320(5883): 1953-1594.
Leard, B. and V. McConnell. 2017. New Markets for Credit Trading Under U.S. Automobile Greenhouse Gas and
Fuel Economy Standards. Review of Environmental Economics and Policy, 11(2): 207-226.
Li, S., E. Spiller, and J. Linn. 2013. Evaluating 'Cash-for-Clunkers': Program effects on auto sales and the
environment. Journal of Environmental Economics and Management, 65(2): 175-193.
Linnerooth-Bayer, J. and S. Hochrainer-Stigler. 2015. Financial Instruments for Disaster Risk Management
and Climate Change Adaptation. Climatic Change, 133: 85-100.
List, J. and R. Metcalfe. 2014. Field Experiments in the Developed World: An Introduction. Oxford Review of
Economic Policy, 30(4): 585-596.
Lyon, T. and J. Maxwell. 2007. Environmental Public Voluntary Programs Reconsidered. The Policy Studies
Journal, 35(4): 723-750.
Madrian, B. and D. Shea. 2001 The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior.
Quarterly Journal of Economics, 116(4): 1149-1187.
4-34 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Mansur, E. 2012. Upstream versus Downstream Implementation of Climate Policy. The Design and
Implementation of U.S. Climate Policy, ed. D. Fullerton and C. Wolfram. Chicago: University of Chicago Press:
179-193.
Mansur, E. and G. Sheriff. 2021. On the Measurement of Environmental Inequality: Ranking Emissions
Distributions Generated by Different Policy Instruments. Journal of the Association of Environmental and
Resource Economists 8(4): 721-758.
Marchese, D.C., M.E. Bates, J.M. Keisler, M.L. Alcaraz, I. Linkov, and E.A. Olivetti. 2018. Value of Information
Analysis for Life Cycle Assessment: Uncertain Emissions in the Green Manufacturing of Electronic Tablets.
Journal of Cleaner Production, 197, parti: 1540-1545.
Mastromonaco, R. 2015. Do Environmental Right-to-Know Laws Affect Markets? Capitalization of Information
in the Toxic Release Inventory. Journal of Environmental Economics and Management, 71: 54-70.
McKibbin, W., A. Morris, P. Wilcoxen, and Y. Cai 2015. Carbon Taxes and U.S. Fiscal Reform. National Tax
Journal, 68(1): 139-156.
Montero, J. 2002. Permits, Standards, and Technology Innovation. Journal of Environmental Economics and
Management, 44: 23-44.
Morgan C, Pasurka C, Shadbegian R. 2014. Ex Ante and Ex Post Cost Estimates of the Cluster Rule and MACTII
Rule. Journal of Benefit-Cost Analysis, 5(2):195-224.
Morgan, C. and A. Wolverton. 2008. Water Quality Trading in the United States: Trading Programs and One-
Time Offset Agreements. Water Policy, 10 (1): 73-93.
Morgenstern, R.D. and W.A. Pizer. 2007. Reality Check: The Nature and Performance of Voluntary
Environmental Programs in the United States, Europe, and Japan. Washington, DC: RFF Press.
Napolitano, S., et al. 2007. The U.S. Acid Rain Program: Key Insights from the Design, Operation, and
Assessment of a Cap-and-Trade Program. Electricity Journal, 20(7): 47-58.
Newell, R. 2010. The Role of Markets and Policies in Delivering Innovation for Climate Change Mitigation.
Oxford Review of Economic Policy, 26(2): 253-269.
Newell, R. and K. Rogers (2006). The U.S. Experience with the Phasedown of Lead in Gasoline. Moving to
Markets in Environmental Regulation: Lessons from Twenty Years of Experience, ed. J. Freeman and C. Kolstad.
Newell, R. and J. Siikamaki. 2014. Nudging Energy Efficiency Behavior: The Role of Information Labels. Journal
of the Association of Environmental and Resource Economists, 1(4): 555-598.
Newell, R. and R. Stavins. 2003. Cost Heterogeneity and the Potential Savings from Market-Based Policies.
Journal of Regulatory Economics, 23(1): 43-59.
OMB. 2010b. Disclosure and Simplification as Regulatory Tools. Available at:
https://ohamawhitehouse.archives.gov/sites/default/files/omh/assets/inforeg/disclosure principles.pdf
(accessed March 1, 2019).
Olmstead, S. 2012. The Role of Market Incentives in Environmental Policy. The Oxford Handbook of U.S.
Environmental Policy, ed. M. E. Kraft, and S. Kamieniecki. New York: Oxford University Press.
Pargal, S. and D. Wheeler. 1996. Informal Regulation of Industrial Pollution in Developing Countries: Evidence
in Indonesia .Journal of Political Economy, 104(6): 1314-1327.
Parry, I.W.H. and R.C. Williams. 1999. A Second-Best Evaluation of Eight Policy Instruments to Reduce Carbon
Emissions. Resource and Energy Economics, 21: 347-373.
Pellerano, J., M. Price, S. Puller, and G. Sanchez. 2017. Do Extrinsic Incentives Undermine Social Norms?
Evidence from a Field Experiment in Energy Conservation. Environmental and Resource Economics, 67(3):
413-428.
Pezzey, J.C.V. and F. Jotzo. 2012. Tax-versus-trading and Efficient Revenue Recycling as Issues for Greenhouse
Gas Abatement. Journal of Environmental Economics and Management, 64(2): 230-236.
4-35 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Phaneuf, D. and T. Requate. 2016. A Course in Environmental Economics: Theory, Policy, and Practice. New
York: Cambridge University Press.
Pigou, A. 1932. The Economics of Welfare. 4th Ed. London: MacMillan and Company.
Pindyck, R. 2 007. Uncertainty in Environmental Economics. Review of Environmental Economics and Policy,
1(1): 45-65.
Pizer, W. 2002. Combining Price and Quantity Controls to Mitigate Global Climate Change. Journal of Public
Economics, 85(3): 409-434
Prakash, A. and Potoski, M. 2012. Voluntary Environmental Programs: A Comparative Perspective. Journal of
Policy Analysis and Management, 31(1): 123-138.
Requate, T. 2005. Dynamic Incentives by Environmental Policy Instruments - a Survey. Ecological Economics,
54(2-3): 175-195.
Rozenberg, J., A. Vogt-Schilb, and S. Hallegatte. 2020. Instrument Choice and Stranded Assets in the Transition
to Clean Capital. Journal of Environmental Economics and Management 10:102183.
Sallee, J., and J. Slemrod. 2012. Car Notches: Strategic Automaker Responses to Fuel Economy Policy. Journal
of Public Economics 96(11-12): 981-999.
Schmalensee, R. and R. Stavins. 2013. The S02 Allowance Trading System: The Ironic History of a Grand
Policy Experiment .Journal of Economic Perspectives, 27(1): 103-122.
Schmalensee, R. and R. Stavins. 2017. Lessons Learned from Three Decades of Cap-and-trade Programs.
Review of Environmental Economics and Policy, 11(1): 59-79.
Segerson, K. 1995. Liability and Penalty Structures in Policy Design. The Handbook of Environmental
Economics, ed. D.W. Bromley. Cambridge, MA: Blackwell Publishers. 272-294.
Segerson, K. and J. Wu. 2006. Nonpoint Pollution Control: Inducing First-Best Outcomes Through the Use of
Threats. Journal of Environmental Economics and Management, 51: 165-184.
Sheriff, G. (2024). California's GHG Cap-and-Trade Program and the Equity of Air Toxic Releases. Journal of
the Association of Environmental and Resource Economists, 11(1), 137-170.
Shogren, J. and L. Taylor. 2008. On Behavioral-Environmental Economics. Review of Environmental Economics
and Policy, 2(1): 26-44.
Shogren, J., G. Parkhurst, and P. Banerjee. 2010. Two Cheers and a Qualm for Behavioral Environmental
Economics. Environmental and Resource Economics, 46(2): 235-247.
Sigman, H. 2012. Monitoring and Enforcement of Climate Policy. The Design and Implementation of U.S.
Climate Policy, ed. D. Fullerton and C. Wolfram. University of Chicago Press: 213-225.
Stavins, R. 2006. Vintage-Differentiated Environmental Regulation. Stanford Environmental Law Journal, 25:
29-66.
Sterner, T. and J. Coria. 2012. Policy Instruments for Environmental and Natural Resource Management, 2nd
Edition. New York, New York: RFF Press.
Sunstein, C. R. 2011. Empirically Informed Regulation. University of Chicago Law Review 78; 1349.
Swift, B. 2000. How Environmental Laws Can Discourage Pollution Prevention: Case Studies of Barriers to
Innovation. PPI Policy Report. Washington, DC: Progressive Policy Institute.
Swift, B. 2001. How Environmental Laws Work: An Analysis of the Utility Sector's Response to Regulation of
Nitrogen Oxides and Sulfur Dioxide Under the Clean Air Act. Tulane Environmental Law Journal, 14: 309-425.
Swinton, J. 2004. Phase I Completed: An Empirical Assessment of the 1990 CAAA. Environmental and Resource
Economics, 27(3): 227-246.
Thaler, R. and C. Sunstein. 2008. Nudge: Improving Decisions about Health, Wealth, and Happiness. Yale
University Press.
4-36 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Thompson, K.M. and J.S. Evans. 1997. The Value of Improved National Exposure Information for
Perchloroethylene(Perc): A Case Study for Dry Cleaners. Risk Analysis 17(2): 253-271.
Teisl, M., B. Roe, and R. Hicks. 2002. Can Eco-Labels Tune a Market? Evidence from Dolphin-Safe Labeling.
Journal of Environmental Economics and Management, 43(3): 339-359.
Tietenberg, T. 1998. Disclosure Strategies for Pollution Control. Environmental and Resource Economics, 11(3-
4): 587-602.
Tietenberg, T. and L. Lewis. 2014. Environmental and Natural Resource Economics. 10th Edition. New York:
Routledge.
Tietenberg, T. and D. Wheeler. 2001. Empowering the Community: Information Strategies for Pollution
Control. Frontiers of Environmental Economics, ed. H. Folmer, H. Gabel, S. Gerking, and A. Rose. Cheltenham,
UK: Edward Elgar Publishing. 85-120.
Tversky, A. and D. Kahneman, 1992. Advances in Prospect Theory: Cumulative Representation of Uncertainty.
Journal of Risk and Uncertainty, 5: 297-323.
U.S. EPA. 2003. Water Quality Trading. January 13. Available at:
https: //archive.epa.gov/ncer /events /calendar /archive /web /pdf/finalpolicv2 003.pdf (accessed March 1,
2019).
U.S. EPA. 2007. Acid Rain and Related Programs: 2006 Progress Report. EPA-430-R-07-011. Office of Air and
Radiation, Clean Markets Division.
U.S. EPA. 2008. EPA Water Quality Trading Evaluation: Final Report. October 2008. Available at:
https://www.epa.gov/sites/production/files/2015-09/documents/epa-water-quality-trading-evaluation.pdf
(accessed March 1, 2019).
U.S. EPA. 2010a. Environmental Protection Agency Fuel Economy Label: Phase 1 Focus Groups. EPA Office of
Transportation and Air Quality and National Highway Traffic Safety Administration. EPA-420-R-10-903.
August 2010.
U.S. EPA. 2010b. Environmental Protection Agency Fuel Economy Label: Phase 2 Focus Groups. EPA Office of
Transportation and Air Quality and National Highway Traffic Safety Administration. EPA-420-R-10-904.
August 2010.
U.S. EPA. 2014. How to Implement a Wood-Burning Appliance Changeout Program. Available at:
https://www.epa.gov/sites/production/files/2015-
08/documents/howtoimplementawoodstovechangeout.pdf (accessed February 28. 2019).
U.S. EPA. 2019. Updating the Environmental Protection Agency's (EPA) Water Quality Trading Policy to
Promote Market-Based Mechanisms for Improving Water Quality. Memorandum from Assistant
Administrator for Water David P. Ross, February 6, 2019. Available at:
https://www.epa.gOv/sites/production/files/2019-02/documents/trading-policy-memo-2019.pdf (accessed
April 2, 2020).
Videras, J. and A. Alberini. 2000. The Appeal of Voluntary Environmental Programs: Which Firms Participate
and Why? Contemporary Economic Policy, 18(4): 449-61.
Vidovic, M. and N. Khanna. 2011. Is Voluntary Pollution Abatement in the Absence of a Carrot or Stick
Effective? Evidence from Facility Participation in the EPA's 33/50 Program. Environmental and Resource
Economics, 52 (3): 369-393.
Walls, M. 2013. Deposit-Refund Systems in Practice and Theory. Encyclopedia of Energy, Natural Resources
and Environmental Economics, ed. Jason Shogren. Amsterdam: Elsevier.
Weitzman, M. 1974. Prices Versus Quantities. Review of Economic Studies, 41(4): 477-491.
Weitzman, M. 2020. Prices or Quantities Can Dominate Banking and Borrowing. The Scandinavian Journal of
Economics. 122(2): 437-463.
4-37 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Whitefoot., K. and S. Skerlos. 2012. Design Incentives to Increase Vehicle Size Created from the U.S. Footprint-
Based Fuel Economy Standards. Energy Policy 41: 402-411.
Woodward, R. & Kaiser, R. 2002. Market Structure for US Water Quality Trading. Review of Agricultural
Economics, 24(2): 366-383.
Yokota, F. and K.M. Thompson. 2004. Value of Information Analysis in Environmental Health Risk
Management Decisions: Past, Present, and Future. Risk Analysis, 24(3): 635-650.
Xabadia, A., R.U. Goetz, and D. Zilberman. 2008. The Gains from Differentiated Policies to Control Stock
Pollution When Producers Are Heterogeneous. American Journal of Agricultural Economics, 90(4): 1059-1073.
Zhao, J. and L. Mattauch. 2022. When Standards Have Better Distributional Consequences Than Carbon Taxes.
Journal of Environmental Economics and Management 116:102747.
4-38 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 5 - Setting the Foundation:
Scope, Baseline and Other Analytic
Design Considerations
This chapter provides an overview of a broad set of issues related to the
design of an economic analysis. These include [1] the appropriate scope of a
benefit-cost analysis (BCA], [2] how to specify the baseline, [3] how to
account for behavioral and technological change, [4] what to assume about
regulatory compliance and [5] how to address analytic uncertainty, among
others. Identifying key issues or questions surrounding these decisions early
in the regulatory development process is important because they can have a
profound impact on analytic outcomes. Subsequent chapters on benefits
(Chapter 7], costs (Chapter 8], economic impacts (Chapter 9] and
environmental justice and other distributional analyses (Chapter 10] delve
into these considerations in more depth. The discussion of analytic design
considerations focuses on their application to prospective analyses, though
they are equally applicable to retrospective analysis of existing regulations
(see Text Box 5.1 for more discussion].
5.1 Scope of Analysis
Several early analytic decisions determine the scope of a BCA of a regulation. Specifically, analysts
must consider whose costs and benefits to count in a regulatory analysis and the types of markets
and non-market effects that should be evaluated, including those that cannot be quantified.
A comprehensive approach to benefit-cost analysis is required to assess whether it is conceivable
for those who experience a net gain from a regulatory action to potentially compensate those who
experience a net loss.1 These benefits and costs may occur in private markets as well as through
changes in externalities. Analysts should carefully consider how various benefits and costs may
materialize as a result of the regulatory action by looking beyond effects on regulated entities and
changes in the regulated contaminant(s). Without a comprehensive accounting of benefits and
costs, the analysis may provide misleading conclusions regarding the sign and magnitude of net
benefits and the relative rankings of the analyzed regulatory options (Farrow 2013).2 As discussed
1 These gains and losses are measured by an individual's willingness to pay or willingness to accept. See Section
A-3for a discussion of the Kaldor-Hicks potential compensation test that underlies the economic practice of BCA.
2 EO12866 and OMB's Circular A-4 (2023) require and affirm that all benefits and costs resulting from a policy
change should be considered in a BCA. For example, Circular A-4 states, "Your analysis should look beyond the
obvious benefits and costs of your regulation and consider any important additional benefits or costs, when
5-1 Guidelines for Preparing Economic Analyses | 3rd edition
-------
in later chapters, the BCA also should clearly identify each source of benefits and costs and present
it in a disaggregated and informative way (see Chapter 11).
While in principle the analyst should account for all benefits and costs, in practice, not all changes in
economic welfare can be quantified and monetized due to limitations in tools, data and resources.
In these cases, analysts are advised to prioritize quantifying those effects that are likely to have the
greatest influence on net benefits and the relative ranking of the options under consideration. Since
the results of a BCA are therefore likely incomplete, they should be presented and interpreted with
care. The BCA should identify effects that could not be quantified or monetized (along with an
explanation of why they were not included), describe evidence on the potential magnitude of the
benefits and costs from these effects, and explicitly document and discuss any other analytic
limitations and omissions. Furthermore, equal effort should be made to account for both benefits
and costs so the analysis provides an assessment of net benefits that is balanced and as accurate as
possible. While this section provides guidance on the scope of a BCA, Chapters 9 and 10 provide
guidance on the scope of economic impact and distributional analysis.3
'* i i r
One of the first scoping questions an analyst must answer is, who has economic standing? Or put
another way, whose gains and losses should be accounted for in the analysis? The most inclusive
answer is all persons who may be affected by the policy regardless of where (or when) they live.
Regulatory analysis often focuses on the costs that accrue to regulated sources, regardless of the
nationality of the owners of affected physical assets, and the benefits to individuals that reside
within the country's national boundaries. This approach reflects the fact that these are the two
groups primarily affected by most regulations.4
feasible. An additional benefit may be a favorable regulation that is unrelated to the main purpose of the
regulation..., while an additional cost may be an adverse impact...that occurs due to a regulation and is not
already accounted for in the obvious costs of the regulation. These sorts of effects sometimes are referred to by
other names: for example, indirect or ancillary benefits and costs, co-benefits, or countervailing risks."
3 While Section 5.1 focuses on the scope of a BCA, the same set of issues applies broadly to economic impact and
distributional analysis. An exception is that it may be worthwhile to estimate certain welfare effects for a
distributional analysis even when those effects do not fundamentally change the net benefits of regulatory
options under consideration.
4 Regulations often only apply to activities within a national border by residents and firms who have consented
to adhere to the same set of rules and values for collective decision-making. In addition, most domestic policies
are expected to have relatively negligible effects on other countries (Gayer and Viscusi, 2016; Kopp et al. 1997;
Whittington etal. 1986).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 5.1 - Retrospective Analysis
The principles for prospective analysis also apply to estimating benefits, costs, or economic
impacts from existing regulations. A retrospective analysis can provide an opportunity to
understand whether a regulation has achieved its objectives for example, whether the
regulation improved societal welfare as expected. Retrospective analysis may identify
compliance pathways, behavioral responses, or consequences that may not have been fully
anticipated at rule promulgation. Retrospective analyses may also suggest ways to improve
prospective analysis for instance, if certain consequences of regulation are routinely
underestimated ex-ante, methods to anticipate these effects may be developed. Ultimately,
retrospective analysis may result in improvements in regulatory design.
While the importance of retrospective analysis in policy evaluation and regulatory reform is
well-recognized, ex-post studies of EPA regulations are relatively rare (U.S. EPA 2014; Aldy
2014; Morgenstern 2018; Fraas etal 2023). Absent systematic data collection, retrospective
analyses of the benefits, costs, or economic impacts of regulations have been conducted
opportunistically (Fraas etal 2023; Aldy etal 2022; Cropper etal 2018). In addition,
retrospective assessments have struggled with issues such as "how to evaluate a highly
heterogeneous industry with a limited set of information, how to form a reasonable
counterfactual, and how to disentangle the costs [or benefits] of compliance from other factors"
(U.S. EPA, 2014). Another challenge has been identifying metrics that can be measured ex post
that are relevant to the regulatory outcomes of interest (Morgenstern 2018).
Because of the many challenges inherent in conducting robust retrospective analysis, studies of
EPA regulations are often relatively narrow in scope in that they only evaluate a subset of the
questions of interest For example, a study may examine how emissions have changed post-
regulation but due to data limitations, may not evaluate the extent to which changes in risk or
health outcomes have occurred. Likewise, researchers may identify the mix of compliance
strategies that were used or offer insights into specific aspects of unit costs but not have enough
information to assess their costs in aggregate (Fraas et al 2023).
Given sufficient data, analysts can use a variety of techniques to conduct rigorous retrospective
review. One approach is to use statistical techniques to control for other exogenous factors that
affected firm or consumer behavior over time. If a set of similar facilities remained unregulated
over the time period, then it may also be possible to compare the regulated firms' behavior to a
reasonable counterfactual. If data for several years before and after the regulation became
effective is available, it may also be possible to analyze how benefits or costs changed over time.
This would potentially allow one to evaluate whether a regulation induced technological change
or affected employment, for example. Though used less in published retrospective analysis,
another approach is to use computational models to address statistical and data challenges.
Even when the model chosen is scientifically defensible, fit for purpose, appropriately
parameterized and reasonably transparent, separating out the effects of the regulation from
other changes that would have occurred anyway (i.e., in the counterfactual) is still a challenge.
The EPA is exploring additional steps to better institutionalize the practice of conducting
retrospective review and analysis. For example, this could be through the development of a
systematic approach to identifying the types of rules most amenable to retrospective analysis,
best practices for retrospective analysis, and how to identify analytic requirements for such
analysis. Data needs could be identified and avenues for ex-post data collection integrated into
the regulation (while also accounting for the cost and time needed for firms to collect such
information). In this way, the EPA could learn from past experience and improve both policy
designs in future regulatory actions and analytic approaches in future prospective analyses.
5-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
In certain contexts, however, it may be important to include effects beyond national boundaries.
This is particularly relevant to consider when evaluating a regulation's impact on a global public
good.5 It is also important to be cognizant of analytic challenges when attempting to disaggregate
benefits and costs accruing to domestic and foreign citizens and residents. For example, to limit
economic standing to citizens and residents of the United States, one may need to determine how to
treat multinational firms with plants in the United States but shareholders elsewhere and how to
estimate the extent to which impacts on foreign companies and citizens have feedback effects on
U.S. citizens and residents.6
The basis for the decision about the scope of the analysis should be transparent and clear and
should focus on capturing the significant effects of a regulation. Analysts should ensure that the
application is supported by the available data and that standing is consistently applied when
estimating costs and benefits; in other words, if a group has standing for estimating costs, it should
also have standing for benefit estimation.
i". C Market I ll- if
Another scoping question is: which markets will be affected by the regulation? The ways in which a
regulation may affect different markets helps inform the analytic approach to take (see Chapters 7
and 8 for more discussion). Ideally, the analyst should comprehensively capture all costs and
benefits of a regulation. In practice, this may not always be feasible due to limitations in available
data, methodologies, or resources. When prioritizing which costs and benefits to include, consider
the effect of the regulation on related markets.
A "distorted" market is one in which factors such as pre-existing taxes, externalities, regulations, or
imperfectly competitive markets move consumers or firms away from what would occur under
perfect competition.7 In the absence of market distortions, focusing on the impacts within the
market may be sufficient. While a policy may have effects on other markets, market-clearing
conditions ensure that they are effectively canceled out from an aggregate welfare perspective
(Farrow and Rose 2018; Justetal. 2004).
Every market is distorted to some degree. In particular, effects in related markets are important to
consider when there are both pre-existing distortions in these markets and there are significant
5 For example, when emissions of a pollutant contribute to damages around the world regardless of where they
are emitted, it is important to consider how U.S. mitigation activities may affect international reciprocity and
cooperation in addressing the same pollutant, as any international mitigation actions will provide a benefit to
U.S. citizens and residents. There may also be cases where international or domestic legal obligations require or
support calculation of regulatory effects accruing beyond national boundaries. For more discussion of when the
effects of U.S. policy on non-residents might be relevant in regulatory analysis, see OMB (2023).
6 For example, impacts that occur outside U.S. borders can impact the welfare of individuals and the profits of
firms that reside in the U.S. because of their connection to the global economy. This can occur through effects on
supply chains, international markets, trade, tourism, and other activities. Other challenges might include how to
account for leakage due to regulatory requirements that are not harmonized across countries or how to treat
impacts on U.S. citizens or assets residing outside U.S. borders. See National Academies (2017) for a detailed
discussion of challenges in the context of quantifying the effects of changes in greenhouse gas emissions.
7 Perfectly competitive markets are characterized by the following conditions: all economic agents have
complete information; there are no barriers to entry or exit; firms have constant returns to scale; and there are
no taxes, subsidies or policies that create a wedge between the price suppliers receive for a good and the price
consumers pay for it. The term "externality" is discussed in Chapter 3.
5-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
cross-price effects between the regulated sector and these other economic sectors (Harberger
1964; Boardman etal. 2018; U.S. EPA 2017). Related markets may include those for major inputs to
the regulated sector, products that use the regulated sector's output as an input, and products that
are substitutes or complements to the regulated sector's output. A key question for the analyst to
consider given market distortions is: when is it reasonable to assume away these effects (e.g., Hahn
and Hird 1990)? Evidence suggests that effects outside of the regulated sector, and therefore
changes in welfare, may be substantial even with a relatively small sector-specific regulation
(Marten et al. 2019; Goulder and Williams 2003). The presence of a distortion alone, however, may
not warrant a broader analytic approach, particularly if the value of information from accounting
for its effect on costs and benefits is relatively small. Analysts should take special care to justify
their choice of which markets to explicitly analyze as part of the regulatory analysis and identify
key assumptions and limitations underlying this choice.89
?rnalities
BCA should aim to comprehensively evaluate all benefits and costs resulting from the regulation,
which includes welfare effects from all changes in externalities due to changes in environmental
contaminants as well as any other externalities.10 If some of these effects cannot be quantified or
monetized, they should be evaluated qualitatively (including a discussion of their potential
magnitude).
Welfare effects from changes in externalities could be favorable or adverse. Analogous to how a
regulation's interactions with existing market distortions (e.g., pre-existing taxes, asymmetric
information) could lead to additional social costs, a regulation could ameliorate or exacerbate other
pre-existing externalities. Changes in other environmental contaminants may arise from the
compliance methods of regulated sources. For example, the use of an abatement technology by
regulated sources to reduce emissions of a pollutant into one medium (e.g., air) may change the
emissions of another pollutant into the same medium (e.g., from the same smokestack) or cause
changes in emissions of pollutants into another medium (e.g., water).
Changes in other environmental contaminants may also occur as a result of market interactions
induced by the regulation. For example, more stringent vehicle emissions standards can lead to
changes in upstream oil refinery emissions. Section 5.5.6 discusses the importance of ensuring that
projected changes in contaminants are consistent with expected market behavior, considers
8 Analysts should also keep in mind that even in cases where effects in other sectors contribute little to the
overall social cost or benefits of the policy, they may have important distributional consequences that warrant a
broader analytic treatment than one that focuses solely on the directly regulated market. See Chapters 9 and 10
for more discussion.
9 Choosing the model that is most appropriate for capturing the key impacts of a policy is sometimes referred to
as "horses for courses. "Just as the best horse for a race depends on the features of the course, the best economic
model(s) to evaluate the benefits and costs of a regulation depend on the features of the regulation and the
affected markets. Text Box 5.3 discusses model selection criteria more generally.
10 These effects are among the distortions discussed in Section 5.1.2 as the presence of an externality represents
a deviation from perfect and complete markets. Such a deviation may be ameliorated or exacerbated by
behavioral changes induced by a regulation. The costs and benefits from unaddressed externalities differ from
the costs and benefits of the production of marketed goods in that the welfare effects due to changes in an
externality are not reflected in the market prices of those sectors and activities that cause the externality. See
Chapter 3 for further discussion of externalities.
5-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
interactions with other regulations and provides several common examples of how changes in
other contaminants arise in analyses of U.S. Environmental Protection (EPA) regulations. This
guidance also applies to expected changes in externalities other than those associated with
environmental contaminants. For example, changes in vehicle emissions standards may reduce the
marginal cost of driving due to greater fuel efficiency and lead to an increase in vehicle miles
traveled that affects road safety, congestion, and other transport-related externalities. These
welfare effects should also be accounted for in the BCA and, if they cannot be accounted for because
of limited resources, data, and other limitations, they should be described qualitatively.
When presenting the results of the BCA, identifying benefits and costs that are specifically
contemplated by the statutory provision under which the regulation is being promulgated when
it is possible to do so provides transparency. For example, in a BCA of a regulation promulgated
under a Clean Air Act provision whose objective is reducing hazardous air pollutants (HAPs), it is
helpful to clearly distinguish the air pollution benefits resulting from reductions in HAP emissions
from other welfare effects resulting from the expected compliance strategies of regulated entities.11
Yet, when calculating net benefits all welfare effects should be included, as it is the total willingness
to pay for all changes induced by a regulation that determines whether the regulation increases
economic efficiency.
5.2 Baseline
Establishing the baseline of an economic analysis is a critical step for accurately evaluating benefits
and costs. Because a BCA considers the impact of a policy or regulation in relation to the baseline,
its specification can have a profound influence on the results of the analysis. The level of detail
presented in the baseline is also an important determinant of the type of analysis that can be
conducted when evaluating regulatory options.
* C i I; > v Ini1- L" - I'iiiivion
The baseline is defined as the best assessment of the way the world would evolve absent the
proposed regulation. It is the primary point of comparison for assessing the effects of the regulatory
options under consideration. Specifically, the BCA models two states of the world: the expected
state without the regulation (the baseline scenario) and the expected state with the proposed
regulation in effect (the policy scenario(s)). The effects of each policy scenario are measured by
examining the differences in net benefits between the scenarios and the baseline.
The baseline describes the expected future of the environmental problem and level of
environmental contaminants along with the affected markets and population in the absence of the
proposed regulation. While the policy scenario is described in a similar fashion to the baseline, it
reflects different environmental and/or market outcomes.
Figure 5.1 illustrates the difference between the baseline and a policy scenario, although there may
be multiple policy scenarios under consideration. An economic analysis begins with a description of
the state of the world in the current period as a foundation before any analytic scenarios are
constructed. The current state of the world includes a description of the environmental problem as
11 This means that if the air pollution reduction also reduces harmful deposition of the pollutant into the water,
the benefits from reducing water pollution should be distinguished from the benefits arising from the reduction
of the pollutant in the air.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
well as other variables such as the level of environmental contaminants; the number and
characteristics of the affected markets, firms, consumers and state and local governments; the
consumption and production of affected goods within and beyond the regulated market;
characteristics of the exposed or otherwise affected population; and existing federal, state and local
regulations that may affect the environmental problem. Based on the description of the current
state of the world, the next step is to develop a projection of the future state of the world without
the regulation, which is referred to as the baseline. This step is done by characterizing how
economic and environmental conditions are expected to change over time. Changes may occur in
demographics, the pace and direction of technology, energy and other prices, sector-specific
economic activity, consumer behavior and other related policies and programs that are already in
place. The baseline should reflect likely outcomes, or "business as usual" not an outlier scenario.
The policy scenario is evaluated in a similar fashion, but the economic and environmental
conditions reflect the future state of the world with the regulation in place. The two scenarios are
then compared.
It is important to note that the comparison of the world with the policy, to the world without the
policy is distinct and quite different from a comparison of the state of the world before the
action to the state of the world after the action. In other words, the baseline is a future scenario
without the regulatory program under consideration; it is not a scenario assuming no change from
current conditions. The economy and other factors may change over the time horizon of analysis
even in the absence of regulation, so a proper baseline should incorporate assumptions about the
changes in the economy that may affect relevant benefits and costs.
In most cases, future economic and environmental conditions in the baseline are expected to have
changed solely in response to factors unrelated to the regulation under consideration. On occasion
this may not be the case. For example, a regulation under consideration may extend the compliance
period of an existing regulation. In this case, the baseline specification might incorporate the
expiration of the existing program. However, changes between the baseline and policy scenario
should be solely attributable to the introduction of the regulation. The economic and environmental
characteristics specified in the baseline should be used in the policy scenario unless the policy
scenario is anticipated to change those characteristics. This is what makes the baseline the relevant
point of comparison for the policy. In general, the construction of the baseline needs to be balanced
to equally identify factors that may meaningfully affect both benefits and costs. For example, the
analyst should not assume a great deal of technological innovation in one sector (e.g., the pollution
abatement sector) and ignore potential technology improvements in other sectors.
The final step in an analysis, as illustrated in Figure 5.1, is to use the information from the baseline
and policy scenarios as a basis for estimating the benefits, costs, economic impacts, and
distributional impacts of the regulatory option(s) under consideration. The damages from exposure
to environmental contaminant levels in the baseline and policy scenarios can be valued using
appropriate economic techniques (see Chapter 7: Analyzing Benefits). The value of the change in
damages in the policy scenario are the benefits of the policy. The new compliance activities and
other effects identified in the policy scenario can be used to quantify the costs of the policy (see
Chapter 8: Analyzing Costs). The figure provides examples of economic and distributional impacts
that may occur (for additional examples and explanation, see Chapters 9 and 10).
5-7 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Figure 5.1 - Structure of a Benefit-Cost Analysis
Current Period
- Current environmental problem/level of environmental contaminants
- Number and characteristics of affected markets, firms, consumers and state/local governments
- Consumption and production of affected goods
- Characteristics of exposed population
- Federal/state/local regulations that have beaming on the environmental problem
- Demographic changes
- Technological changes
- Future economic activity
- Consumer behavior changes
- Other policies and programs
Future Period
Baseline: Without Regulation
-Expected extent of future environmental problem/level of
environmental contaminants
- Number and characteristics of affected markets, firms,
consumers and state/local governments
- Consumption and production of affected goods
- Characteristics of exposed population
- Anticipated federal/state/local regulations that have bearing
on the environmental problem
Future Period
Policy Scenario: With Regulation
- Expected new compliance activities
- Expected new environmental condi tions/level of
environmental contaminants
- Possible new market configurations
- Number and characteristics of affected markets, firms,
consumers, and state/local governments
- Consumption and productions of affected goods
- Characteristics of exposed population
- Anticipated federal/state/local regulations that have bearing
on the environmental problem
- Damages from environmental contaminants on exposed
population
- Valuation of damages
- Damages from environmental contaminants on exposed
population
- Cost of new compliance activities
- Valuation of damages
- Benefits = [Baseline valuation of damage] - Valuation of damages with policy!
- Costs = [Policy cost of controlling environmental contaminants]
- Net benefit = Benflts - Costs
- Economic impacts = [Baseline market conditions] - [Mark_L conditions with policy]
- Distributional impacts = [Baseline exposures] - [Exposures with policy]
5-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
: ¦-Vi^inr R nviplv i I- seline Specification
In specifying the baseline, analysts should employ the following guiding principles:
1. Clearly specify the environmental problem that the regulation addresses and the regulatory
approach being considered in the statement of need.
2. Identify all required variables for the analysis.
3. Clearly specify the current and future state of relevant economic and regulatory variables.
4. Focus on the components of the analysis that have the greatest influence on the results.
5. Clearly identify all assumptions made in specifying the baseline conditions.
6. Detail all aspects of the baseline specification that are uncertain.
7. Use the baseline assumptions consistently throughout the analysis of a regulation.
Though these principles exhibit a common-sense approach to baseline specification, the analyst is
advised to provide statements on each of these points. Failure to do so may result in a confusing
presentation and misinterpretation of the economic results.
Clearly specify the environmental problem that the regulation addresses and the regulatory
approach being considered in the statement of need. As discussed in Chapter 3, the analysis
should begin with a statement of need for regulatory action and an evaluation of policy options. The
statement of need provides a description of the problem being addressed and the significance of
that problem, the failures of private markets or public institutions that warrant agency action, and
an assessment of whether Federal regulation is the best way to correct the problem. This statement
should also include a description of the current regulatory environment and the regulated entities
and other affected parties. The evaluation of policy options should describe all policy options or
potential regulatory or non-regulatory approaches that were considered and how they were
chosen.12 The statement of need and description of the policy options will help clarify the
appropriate baseline to be used.
In general, the baseline will assume no change in behavior to comply with the new regulation or
existing regulations; but in some cases, a different baseline may be considered. For example, if an
industry is certain to be regulated by some other method (e.g., by court order or state action) but
that regulation has not yet been implemented, then the baseline should include it. Also, it is
common practice to assume full compliance with existing regulatory requirements in the baseline
even if there is noncompliance, although a separate analysis assuming less-than-full compliance
may determine the implication of this assumption (see Sections 5.5.4 and 5.6.1 for more discussion
of this issue).
Identify all required variables for the analysis. To ensure that the baseline scenario can be
compared to the policy scenario, there should be a clear understanding of the path from regulation
to economic behavior to environmental changes to impacts on humans or ecosystems. The models,
parameters and variables required for the baseline analysis should be chosen so that they can
inform all subsequent analyses.
Differences between the baseline and policy scenario may include changes in use or production of
toxic substances, production processes and costs, pollutant emissions and ambient concentrations,
and incidence rates for adverse health and environmental outcomes associated with exposure to
pollutants. This does not mean that the analyst must identify all the variables that could possibly
change, but the analyst should recognize all relevant variables needed to compare the baseline
12 See Chapter 4 for a description of various regulatory and non-regulatory approaches.
5-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
scenario to the policy scenario. At a minimum, the analyst should identify the variables that are
expected to have the largest impact on costs and benefits within and across policy options.
Specify the current and future state of relevant economic and regulatory variables. Future
baseline trajectories of certain types of economic variables such as energy prices, the level and
growth of economic activity and population growth may be important for modeling the effects of a
regulation. Even small changes in the rate of economic growth may, over time, result in
considerable differences in emissions and control costs. Assuming no change in the baseline
economic activity may lead to incorrect results.13 Likewise, assumptions about the future growth
and age distribution of the population affected by a regulation are important for predicting the
number of individuals exposed or even the magnitude of aggregate damages. Other variables, such
as broad trends in consumer spending patterns and technological growth, are also important for
modeling the effects of a regulation but are more difficult to estimate. In these cases, the analyst
should specify the baseline levels for these variables and changes over time and explicitly discuss
all assumptions. If other policies or programs influence baseline conditions, they should also be in
the baseline. For example, changes in farm subsidy programs may influence future pesticide use.
Accounting for the way existing regulations affect compliance behavior and economic and
environmental outcomes of a new regulation assures that the BCA properly accounts for the
cumulative effects of all relevant regulations. In an ideal analysis, all potential influences on
baseline conditions, and on the costs and benefits of policy options, would be examined and
estimated. However, it is up to the analyst to determine if these influences warrant consideration in
the regulatory analysis (e.g., because they may change the rank ordering of the analyzed options). If
certain influences are known but not considered significant enough to be included in the
quantitative analysis, they should be discussed qualitatively. However, in certain circumstances it
may be worthwhile to quantify them to confirm or demonstrate that they are small.
Concentrate on the components that have the greatest influence on the results. The analyst
should concentrate analytic efforts on components (e.g., assumptions, data, models) of the baseline
that are most important to the analysis, taking into consideration factors such as the time given to
complete the analysis, the person-hours available, the cost of conducting the analysis, and the
availability of models and data. If several components of the baseline are uncertain, the analyst
should concentrate on components that have the greatest influence on the costs and/or benefits
and can be refined through additional analysis or data collection. Analysts should pay special
attention to the components that will be used to calculate costs and benefits and those that are
important in the evaluation and selection of a policy option.
Identify all assumptions made in specifying the baseline conditions. The analyst should
explain key assumptions in detail, including those related to changes in consumer and producer
behavior, and how these trends may be affected by the regulatory options. Analysts should look for
trends in economic activity or pollution control technologies that occur for reasons unrelated to
environmental regulation. For example, as a consumer's income increases over time, demand for
some commodities may grow at rates faster than the rate of change in income, while demand for
other goods may decrease. Where these trends are expected to have significant influence on the
evaluation of regulatory alternatives, the analyst should explain and identify the assumptions used
13 For example, if the regulated industry is in significant decline, or is moving overseas, this information should
be accounted for in the baseline. In such cases, incremental costs to the regulated community (and
corresponding benefits from the regulation) are likely to be less than if the targeted industry were stable or
growing.
5-10 Guidelines for Preparing Economic Analyses | 3rd edition
-------
in the analysis, with the goal of laying out the assumptions so that other analysts (with access to the
appropriate models) would be able to replicate the baseline specification.
Detail all aspects of the baseline that are uncertain. Because the analyst does not have perfect
foresight, baseline conditions cannot be characterized with certainty. To the extent possible,
estimates of current values should be based on actual data and estimates of future values should be
based on clearly specified models and assumptions. Where reliable projections of future economic
activity and demographics are available, this information should be used and referenced. In general,
uncertainties underlying the baseline conditions should be treated in the same way as other types
of uncertainties in the analysis.
It is also important to discuss information that was not included in the analysis due to scientific
uncertainty. For example, a regulated pollutant may have a suspected health or ecological effect but
no available human dose-response function. In this case, the effects generally are not quantified in
the analysis, but why the effects were excluded should be discussed especially if the expected
magnitude is such that it could significantly affect the net benefit calculation. Analysts should also
explain how scientific uncertainty affects model choice and parameter values. Important aspects of
the analysis which are not included in the baseline due to scientific uncertainty should be included
in an uncertainty section(s) of the analysis (see Section 5.6 below). Significant uncertainty in
important variables may require the construction of alternative baselines (discussed below). While
sensitivity analysis is usually a better choice, multiple baselines may provide insights when
evaluating different policy options.
Use the baseline assumptions consistently for all analyses of a regulation. The economic and
environmental characteristics used in the baseline should be consistent with those used for the
policy scenario (s). For example, the calculation of both costs and benefits should draw upon
estimates derived using the same underlying assumptions about future economic and
environmental conditions. If the benefits and costs are derived using multiple economic and
environmental models, then the baseline conditions applied in those models should be compared to
ensure that they are consistent. Likewise, when comparing and ranking alternative regulatory
options, comparison to the same baseline should be used for all options under consideration.14
In some cases, it may be useful to single out a sector for more detailed analysis, or a follow-on
analysis might be needed to assess impacts on a specific set of households based on their
socioeconomic characteristics, region, or sector. In this case, it may not be possible to specify a
baseline that is fully consistent with the primary analysis, but the analyst should endeavor to make
them as similar as possible. The analyst also should explicitly describe the differences between the
two baselines and any uncertainty associated with them.
Use consistent dollar years across the baseline and policy scenarios. The baseline and policy
scenarios should be presented consistently and should use a recent common dollar year throughout
the analysis. The dollar year is the year to which the purchasing power of a dollar is indexed. This is
important because inflation decreases the purchasing power of money. So, if costs and benefits are
reported in 2022 dollars, for example, this means that the value of those costs and benefits are
denominated to be comparable to market prices in 2022. All nominal values, which are those not
adjusted for inflation, should be converted to real values by adjusting them to the same dollar year
14 In the less common case in which more than one baseline scenario is modeled, the analyst must avoid the
mistake of combining analytic results obtained from different baseline scenarios. To limit confusion on this point,
if multiple baseline scenarios are included in an analysis, the presentation of economic information should
clearly describe and refer to the specific baseline scenario being used.
5-11 Guidelines for Preparing Economic Analyses | 3rd edition
-------
using an appropriate index of inflation, and the index(es) used should be explicitly stated.15
Similarly, if the costs in an analysis are reported in a particular dollar year (e.g., 2020 dollars) but
the benefits are reported in a different dollar year (e.g., 2022 dollars), one of the estimates should
be adjusted for inflation so that they are reported in the same dollar year. The choice of dollar year
should always be made clear. In addition, the reporting year for annual costs and benefits, distinct
from the dollar year, should be made clear in both the text and tables. For example, if an economic
analysis is using a 2022 dollar year, but the costs and benefits for the rule are reported for the year
2024, both the text and tables should be clear that the values are for 2024, in 2022 dollars.
selines
In most cases, a single, well-defined baseline is generally all that is needed as a point of comparison.
However, there are a few situations where it may be informative to compare the policy options to
more than one baseline. Multiple baseline scenarios are needed when it is difficult to identify a
single, reasonable description of the world in the absence of the proposed regulation. For instance,
if the current level of compliance with existing regulations is not known and may substantially
influence the net benefits, then it may be necessary to compare the policy scenario to both a full
compliance baseline (the standard assumption) as well as a partial compliance baseline. Also, if the
impact of other rules currently under consideration fundamentally affects the analysis of the rule
being analyzed, then multiple scenarios with and without these rules in the baseline may be
necessary. For example, for the 2019 rule to repeal the 2015 rule defining "Waters of the United
States," the degree to which states would continue to regulate their waters at the 2015 standard
was uncertain. Since the states' decisions dramatically affected the avoided costs and forgone
benefits of the repeal, multiple baselines were used to illustrate the range of potential impacts (U.S.
EPA 2019).
The decision to include multiple baselines should not be taken lightly since it may result in a
complex set of modeling choices and analytic findings. Multiple baselines increase the possibility of
erroneous comparisons of costs and benefits if the modeling choices and results are not
communicated clearly. The number of baselines should be limited but still cover the key dimensions
of the analysis and any phenomena in the baseline that are uncertain. Each baseline-to-policy
comparison should be internally consistent in its definition and use of baseline assumptions.
5.3 Multiple Rules
Although regulations that have been finalized clearly belong in the baseline of a proposed rule, the
baseline specification may be complicated by regulations other than the one being promulgated
nearing completion. It is important to consider how these other regulations affect market
conditions and the degree to which they might influence the costs or benefits associated with the
policy of interest. This is true not only for multiple rules promulgated by the EPA, but for rules
passed by other federal, state, and local agencies. In addition to agencies that regulate
environmental behavior, other agencies that regulate consumer and industrial behavior, such as the
15 Commonly used indices include the Bureau of Labor Statistics' Consumer and Produce Price Indices (CPI and
PPI), the Bureau of Economic Analysis' Gross Domestic Product (GDP) deflator, and engineer cost indices. The
most appropriate index will depend on the application.
5-12 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Occupational Safety and Health Administration (OSHA), Department of Transportation (DOT) and
Department of Energy (DOE), develop rules that may affect some of the same entities as EPA
regulations.
-------
rule are incurred in the process of complying with the first rule, then these costs should be included
in the baseline and should not be considered as costs of the second rule. Only the incremental
benefits and costs should be included in the second rule. For example, in 2005, the baseline of the
Clean Air Mercury Rule (CAMR) included mercury emission reductions from the previously issued
Clean Air Interstate Rule (CAIR) (see Text Box 5.2).
The assumption commonly made when rules cannot be evaluated together is to consider the actual
or statutory timing of the rules and use this to establish the sequence in which they are analyzed.
However, this may not always be possible. For example, a rule may be phased in over time,
complicating the analysis of a new rule going into effect during that same period. For this case, the
baseline for the new rule should include the timing of each stage of the phased rule and its resulting
environmental, health and economic changes.
In the absence of an orderly sequence of events that allows the attribution of changes in behavior to
a unique regulation, there may be no clear way to allocate the costs and benefits of a package of
policies being developed at the same time to each individual regulation. By implication, there is no
theoretically correct order for conducting a sequential analysis of multiple policies that are
promulgated simultaneously. In this case, analysts should make a reasonable assumption and
explain it, detailing which rules are included in the baseline (see Text Box 5.2). If the impact of
other rules on the costs and benefits of the rule under consideration is small, then this may be all
that is necessary; it may not be worth additional time and resources to reconcile the baseline of
rules being developed at the same time. On the other hand, when the impact on the costs and
benefits is large or if there are few overlapping rules, then a sensitivity analysis can be included to
test the implications of including or omitting other regulations.
In this sensitivity analysis, it may be possible to use the overlapping nature of the regulations to
allow for some regulatory flexibility in compliance dates and regulatory requirements.
Furthermore, if the benefits and costs of each rule in the sequence are expected to differ
significantly based on the order in which they are evaluated, a sensitivity analysis that changes the
order of their evaluation may provide insights into how to design each to maximize the net benefits
of the rules collectively.
5.3.3 Accounting for Ben "i >ts that Ac i ^ r - I
Rules
When the EPA targets the same contaminants or industries through a sequence of regulations, the
benefits and costs of these actions are additive. To ensure consistency in regulatory accounting,
regulatory analyses should fulfill an "adding-up condition" when comparing a single large
regulation to multiple smaller regulations that imply the same requirements for the same set of
entities. The adding-up condition means that the sum of the estimated incremental benefits (and
costs) from a set of small regulations analyzed separately should be the same as the incremental
benefits (and costs) from the same actions evaluated jointly in a single regulation. Benefits and
costs from previous rules should be included in the baseline so that they are not double counted in
a new regulation.
The adding-up condition was originally proposed in the context of contingent valuation studies
(Diamond andHausman 1994; Kling and Phaneuf2018) and has been applied to valuation of water
quality improvements (Newbold et al. 2018). If analysts do not impose an adding-up condition and
fail to account for improved environmental quality in the baseline when valuing incremental
improvements from successive regulations, then inconsistent results could arise if people value
marginal improvements more when the environmental good is scarce.
5-14 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 5.2 - Accounting for Other Regulations in the Baseline
Because the benefit and cost estimates of one regulation may be affected by those of others, it is
important to consider if they should be incorporated into the baseline. As a rule, analysts should
be transparent and use objective reasoning when deciding to account for other regulations in a
baseline. Transparency requires that all assumptions are clearly stated. Objective reasoning
requires that speculation be avoided. If there is uncertainty about an anticipated rule, then two
baselines one with the anticipated rule and one without might be considered. If only one
baseline is considered due to time or resource constraints, then it should be constructed using
only final rules and, in some cases, imminent rules that are expected with a high degree of
certainty in the absence of EPA action. General guidelines to follow are given below.
All final rules, including those that have not fully taken effect, should be included: The analysis
should assume firms will comply with already promulgated rules. For example, on March 15, 2005,
the EPA promulgated the Clean Air Mercury Rule (CAMR) to reduce mercury emissions from coal-
fired power plants (U.S. EPA 2005b). Five days earlier, on March 10, 2005, the EPA finalized the
Clean Air Interstate Rule (CAIR) (U.S. EPA 2005a). While the primary purpose of CAIR was to
reduce sulfur dioxide (SO2) and nitrogen oxides (NOx), the control technologies necessary to
achieve these reductions also lowered mercury emissions. Because the final CAIR rule had been
issued, the analysis for CAMR assumed that the mercury reduction from CAIR was in the baseline.
This meant that the estimated incremental reduction in mercury from CAMR was much smaller
than if CAIR had not been included in the baseline.
Including imminent final rules may be appropriate if the impacts are known with a high
degree of certainty: If another (final) rule is imminent and will take effect prior to the effective
date of the new rule under consideration, then the imminent rule should be included in the
baseline, but only if its requirements and impacts are known with a high degree of certainty. The
analyst should not speculate that another rule will be implemented. In addition, the analyst should
be clear as to what assumptions have been made to include the imminent rule in the baseline.
Proposed rules should not be in the primary baseline: While a proposed rule signals the intent
to issue a final rule and the Agency maintains a schedule to do so, there is no guarantee that the
final rule will be issued or that it will follow the planned schedule. Even if the Agency does issue a
final rule, it may differ significantly from the proposed rule, which means that the assumptions
embedded in a baseline using a proposed rule will not accurately reflect the likely future effects of
the final rule. An alternative baseline for a proposed rule may have another proposed rule in it,
however, if the two rules are expected to be finalized in the same sequence and the existence of
the first rule may influence the benefits and costs of the second substantially.
Future regulatory actions of other jurisdictions should be considered carefully: Actions by
state and local governments and even international organizations can affect the costs and benefits
of federal rules, particularly if they are regulating the same sector or pollutant In this case, the
analyst must use professional judgment to determine what would happen in the baseline (i.e., in
the absence of EPA action) and how the regulatory response of other jurisdictions may change in
the policy scenario.
State regulations that have been finalized should be included in the baseline. The more difficult
case occurs when a state has a legal obligation to implement a regulation but either has not done
so or is in the process of doing so. For example, the EPA occasionally issues rules establishing
numeric water quality standards for some states when the states themselves have not done so.
One might argue that the state regulation should be in the baseline since they had the legal
obligation to issue the criteria, but this is not the case. The EPA's justification for action is that it
assumes the state will not act. In this example, only if the state would issue the water quality
5-15 Guidelines for Preparing Economic Analyses | 3rd edition
-------
standard in the absence of EPA action can a reasonable case be made for including the state
action in the baseline.
Compliance with a finalized international agreement cannot simply be assumed in the baseline,
especially if some EPA action (such as codifying the international standard) is required for it to
become effective. The costs and benefits associated with any behavioral response by firms to
the EPA action should be part of the policy scenario. In the case where firms will meet the
international standard on their own, even without EPA action, then the compliance with the
standard can be included in the baseline but establishing that this behavioral response will
occur requires justification.
In some cases, environmental regulations yield relatively small changes in health or the
environment that may not be noticeable to the public until multiple regulations have achieved a
large aggregate improvement. Just as it is important to account for small average costs imposed by
regulations which can be economically significant when aggregated over a sufficiently large
population it is conceptually correct to account for small improvements in public health and the
environment For instance, the EPA's Science Advisory Board (U.S. EPA 1998) noted that, "small
effects distributed across a large population exert large total health effects," and recommended that
the Agency quantify changes in IQ resulting from regulations that reduce lead exposure, including
changes of less than a single IQ point on average per child.
Some benefits only occur after a threshold has been reached. However, a specific benefits threshold
may not be met with a single rule. In such cases, it is reasonable to account for the benefits of
making progress toward a goal, even if the threshold is not met in the rule under consideration.
Otherwise, if the benefits are associated only with the rule that passes the threshold, it may be
impossible to justify the previous rules that made incremental progress.
For example, the EPA has calculated the benefits associated with improving river miles for various
designated uses (e.g., swimming, fishing and boating) in several rules. In each case, some river
miles were improved for the designated use, while other miles were improved, but not enough to
change their designated use. Analyses of earlier rules claimed benefits only if a river mile changed
its designation, implicitly giving a value of zero to partially improved river miles. More recent
analyses have included estimates of the partial benefits from incremental improvements toward
the threshold. Either approach can be used, but accounting for the benefits of partial gains provides
useful information to decision makers and the public and allows the Agency to justify incremental
progress to a threshold. Once partial gains have been valued in one rule, then subsequent rules
cannot claim full credit for crossing the threshold. Doing so would double count those benefits.
In the special case when new data or methods make estimates of benefits or costs for earlier rules
obsolete, the analyst should develop a baseline based on the new information and discuss all
changes made since the previous regulatory analysis.
5.4. Time Horizon of Analysis
The time horizon of analysis is the period over which the baseline and policy scenarios are
compared. The time horizon is defined by the starting and ending points. 16 A guiding principle is
that the time horizon should be chosen to capture all the benefits and costs for the policy
16 The time horizon for analysis may also be called the "time period of analysis" or "time frame of analysis."
5-16 Guidelines for Preparing Economic Analyses | 3rd edition
-------
alternatives analyzed, subject to available resources.17 This principle is consistent with the
requirement that a BCA sufficiently reflects the welfare outcomes of those affected by the policy. If
the time horizon is too short, the estimate of the net benefits will be of incorrect magnitude and
perhaps of the wrong sign because benefits and costs often occur over different periods of time. The
analysis should clearly describe the time horizon used for the analysis and it should be clearly
identified whenever present or annualized values are reported (see Chapter 6).
The appropriate time horizon will depend on the economic and legal conditions unique to the
regulatory context under consideration. In many cases, the time span of the physical effects that
drive the benefit estimates, duration of market effects from compliance activities, the duration of
impacts on other externalities, and the economic lifetime of any pollution control investments will
be key factors in its determination. Legal conditions that affect the time horizon of analysis include
the timing of compliance dates. While selecting the appropriate time horizon is challenging, the
analysis should identify the time horizon chosen and explain why it is expected to capture all
benefits and costs. It should also identify the extent to which the sign of net benefits or the ranking
of policy options by their magnitude of net benefits may be sensitive to the choice of the analytic
time horizon.18
The starting point for the analysis should be based on when conditions between the baseline and
policy scenarios diverge, and thus benefits and costs of the regulation begin to be realized. Two
possible choices for the starting point are when an enforceable regulatory requirement becomes
effective or when the final rule is promulgated. These dates are convenient starting points because
they are clearly defined under administrative procedures and represent specific deadlines.
However, the starting point of the analysis should precede the date when regulatory requirements
become effective if firms or households are expected to make anticipatory investments or other
behavioral changes after the rule is finalized and leading up to the effective date.19 Likewise, for a
regulation with requirements that become effective over time, benefits and costs should be
accounted for during the period prior to when the legal requirements are fully implemented. A time
horizon of the analysis that begins when a regulation is fully implemented is insufficient for
accounting for all benefits and costs in the case where behavior changes prior to compliance dates,
and thus the starting point of the analysis should be earlier.
The duration of when costs and benefits occur should generally be used to determine the ending
point for the analysis. In theory, the longer the time horizon, the more likely the analysis will
capture enough of the benefits and costs of the regulation to reliably estimate net-benefits and
compare alternatives. However, other factors, such as the relative uncertainty in projecting
17 Chapter 6 provides a formal method of identifying the ending point of the time horizon of analysis. A
symmetric method may be used to identify the starting point. In addition, Chapters 7 and 8 also provide detailed
guidance on selecting the time horizon of the analysis for benefits and costs, respectively.
18 To compare the benefits and costs of a proposed policy, the analyst should estimate the present discounted
values of the total costs and benefits attributable to the policy over the time horizon of analysis. Chapter 6
provides guidance on how to discount benefits and costs.
19 In most circumstances, a starting point that precedes final rule promulgation is unnecessary, but an earlier
starting point may be desirable if significant behavioral changes were made in anticipation of the final rule. Two
possible starting points that precede promulgation of the final rule and are clearly defined legal milestones are
when authorizing legislation was signed into law and when the EPA formally proposed the rule. However, when
using a starting point that precedes regulatory requirements, it is important for the analysis to identify which
behaviors occurred specifically because of the anticipated federal rule versus those that happened for other
reasons. This will likely be difficult to do.
5-17 Guidelines for Preparing Economic Analyses | 3rd edition
-------
conditions in the distant future, may also need to be considered. Forecasts of economic,
demographic, and technological trends are required over the entire time horizon of the analysis.
Because long term forecasts are less reliable than near term forecasts, the analyst should balance
the advantages of capturing important effects against the disadvantages of decreased reliability of
forecasts further out in time, although those sources of uncertainty may meaningfully affect
benefits or costs and should be accounted for if so. The period in which a regulation is fully
implemented should not be used as the ending point if benefits and costs will occur thereafter.
Furthermore, regulated entities will consider expected future conditions when choosing their
compliance strategies, and a longer time horizon will capture the information they will use when
choosing their compliance approaches.
Analysts should ensure consistent accounting of benefits and costs considering differences in when
they accrue over time. To ensure consistent accounting, all the costs from activities that lead to
quantified benefits should be accounted for in the analysis and vice versa. Ensuring consistency
implies that the ending point may differ for assessing costs than for assessing benefits when the
accrual of costs and benefits does not coincide.20 For example, the human health benefits of a policy
to reduce leachate from landfills may not occur for many years after the cost of compliance is
incurred either because decreases in groundwater contamination take time or because even after
contamination is reduced some health improvements do not manifest immediately. In other
contexts, while control costs are incurred upfront, changes in pollution may lead to health and
ecological benefits that continue to accrue over time.
Generally, the analysis should account for costs until at least the end of the economic lifetime of any
pollution control methods adopted for regulatory compliance.21 Costs will then be consistent with
the total abatement, and in turn benefits, achieved by these pollution control methods.22 Similarly,
the length of the cost analysis should capture any turnover in markets for regulated goods (e.g.,
vehicles) and the length of time those goods are in use. This guidance may be challenging to
implement in an analytic framework that captures the possibility of additional regulated sources
appearing in the future, but the possibility of entry and exit of sources should still be included.
Again, the analysts should weigh the value of additional benefit and cost information gleaned from a
longer time horizon of analysis against uncertainty about future economic conditions.
Some statutory provisions have schedules for when regulations need to be reviewed, and an ending
point corresponding to this review date may be a tempting choice. However, care should be taken
when using regulatory or statutory deadlines to determine the ending point of the time horizon of
analysis. For example, these provisions may not envision the regulation being loosened but only
tightened, and therefore the requirements under consideration are expected to persist over time, at
20 However, as explained in Chapter 6 annualized benefits and costs should be calculated using the same
assumed time period over which the annualized values apply.
21 The economic lifetime is the length of time a piece of equipment is expected to be operational before it is worn
out and needs to be replaced. This guidance is particularly important when compliance costs are amortized over
an economic lifetime or financing period. When compliance costs are amortized the benefits during one segment
of the amortization period may be notably different than over another segment of the amortization period. The
analysis will be misleading if the choice of segment affects the relative benefit to cost estimate (as well as the
total benefits and costs of the regulation).
22 As discussed in the previous paragraph, if the benefits from these controls do not arise until later (i.e., are
latent), the end date for the benefits analysis should be later than the end date for the cost analysis.
5-18 Guidelines for Preparing Economic Analyses | 3rd edition
-------
least at the promulgated level of stringency, potentially yielding additional benefits and costs.23
Similarly, the benefits and costs of a regulation should be evaluated beyond when a particular
statutory requirement is satisfied if the regulation will continue to affect behavior. A time horizon
that reflects the span over which the baseline diverges from the policy case and accounts for all the
benefits and costs is appropriate even if the period extends beyond the scheduled review.
In certain circumstances where benefits and costs are not expected to notably change over time, it
may be analytically convenient to estimate benefits and costs over a shorter time period (e.g., one
year) if they are representative of the benefits and costs over a longer time horizon of analysis (e.g.,
a decade). In other cases, it may be analytically challenging to estimate benefits and costs for each
period over the entire time horizon; thus, benefits and/or costs are estimated for only a few periods
that are each representative of longer periods.24 In these cases, the analysis should still identify the
entire time horizon over which the representative periods of analysis are applicable and discuss
any limitations or uncertainty introduced by this approach. The representative periods of analysis
should be chosen such that they adequately identify the relative net benefits of the various options
under consideration. Focusing on one or a subset of periods without careful consideration of
whether those periods are representative of all benefits and costs over longer time periods may
lead to potentially misleading findings of the magnitude, and possibly even the sign, of net benefits.
For example, treating the annual benefits and costs in the year a rule becomes fully implemented as
representative of the benefits and costs in all years may lead to a misleading net-benefits estimate if
the annual benefits or costs incurred prior to the full implementation year are quite different25 26
¦7.P.-ip[^ntifi Iivv' k > j
To measure the benefits and costs of a regulation, it is important to characterize the behavior of
firms and households in both the baseline and the policy scenarios. In particular, assumptions
about how firms and households (1) engage in technological change, (2) comply with regulations,
23 Furthermore, if there is a credible reason to assume that the regulation will be loosened in the future then this
possibility should be acknowledged in the analysis and the compliance choices of regulated sources should reflect
this possibility (e.g., regulated sources would be more likely to adopt easily reversible compliance strategies if
they thought the regulation may be loosened in the future). Another reason to evaluate the benefits and costs of
the rule beyond the statutory review date is if the rule currently under consideration is expected to be accounted
for in the baseline of any analysis with a time frame beyond the statutory review date, including the rulemaking
following the statutory review.
24 The representative periods may be chosen to characterize periods of different length. For example, if benefits
and costs increase quickly in the near term and are then generally constant afterward, representative periods
used to characterize the near term are applicable to short period (e.g., a couple of years), while representative
periods used to characterize the long term are applicable to longer periods (e.g., a decade).
25 This outcome is possible even if the benefits and costs in the full implementation year are representative of
later years. If they are not representative of benefits and costs incurred in later years then, again, the net benefits
estimate may be misleading.
26 Comparing an annualized value to an annual value also may be potentially misleading. The annualization
period chosen is arbitrary so long as it is long enough to accounts for all benefits and costs, and a longer
annualization period would lead to lower annualized benefits or costs. For example, comparing an annual
benefit to annualized costs over a long time period may give the impression that net benefits are positive when
they may not be. Also, if annualized values are reported, they should be reported for both benefits and costs. See
Section 6.1.6 for further discussion.
5-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
(3) participate in voluntary actions, and (4) affect levels of other contaminants in the baseline and
policy scenarios can also influence costs and benefits.
* * i I; iavior of Househol -ii y Finn.:.
Predicting firm, household, and other organizational responses to regulation requires a model of
economic behavior. Analysts should assume behavior consistent with utility or profit maximization
unless there is evidence supporting other behavioral assumptions (see Section 4.4 and Section 5.5.2
for more discussion of behavioral anomalies).
When modeling the response to regulation, it is important to capture how regulated firms may
choose to comply with new requirements. For instance, firms could change production practices,
output, location, or even exit the industry. Likewise, it is important to capture household responses,
such as changes in the products they buy, where they live, or the types and frequency of averting
behaviors (e.g., purchasing bottled water or staying inside on bad air quality days). These responses
also may result in changes in market prices and externalities, which could further alter economic
behavior. Behavioral response to the regulation may also precede compliance dates, which can
make it difficult to disentangle how much of the behavior is attributable to the regulation.
Future economic conditions are inherently uncertain, and households, firms, and other
organizations will account for these uncertainties when responding to regulations (as well as in the
baseline). Their decision-making under uncertainty may differ from what would occur if future
conditions were known with certainty. For example, when facing uncertain future economic
conditions, regulated entities may avoid making irreversible investments, which provides them
greater flexibility to adjust their future compliance strategies. This may occur even if under most
likely future economic conditions an irreversible investment is the least-cost compliance strategy.
Without accounting for such uncertainty, an analysis may predict greater investment in an
irreversible compliance method than would be expected to occur.27
Capturing behavior when uncertainties are present is analytically challenging. For example,
information is needed on the risk preferences of households and firms. Economic modeling tools
are considerably more complex (or must sacrifice other details) to model decision-making that
accounts for uncertain future conditions. When uncertain future conditions are likely to have a
significant effect on the behavior of households and firms, the analysis should describe these
sources of uncertainty and how they may affect estimates of benefits and costs.
Depending on the types of behavioral responses that are anticipated, the analyst will need to
identify and select the most appropriate economic and environmental model(s) for the regulatory
analysis. Uncertainty in model results tends to be higher when a model is either exceedingly simple
(e.g., because it misses key interactions or feedbacks) or increasingly complex (e.g., due to data
27 This behavior is an example of option value (Dixit and Pindyck 1994). An option value is the value of delaying
an action to learn if it is the best choice. Regulations may impose benefits and costs by eliminating options in the
future that may have value to society or private firms and households even if those options would not be
exercised under likely future conditions.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
requirements). Analysts should seek balance: "the optimal choice generally is a model that is no
more complicated than necessary to inform the regulatory decision" (U.S. EPA 2009).28
Models used to inform EPA decision-making should be reliable, transparent, defensible, and useful
(U.S. EPA 2009). For instance, any modeled changes in behavior should be supported by empirical
estimates of demand, supply, cross-price, and income elasticities.29 When the literature presents a
range or identifies factors that could significantly affect these estimates, analysts should also
examine the sensitivity of benefit and cost estimates to different elasticity assumptions. See Text
Box 5.3 for a discussion of other considerations when selecting models for estimation of costs
and/or benefits, including the extent to which a model adequately represents key markets of
interest and the representativeness of other significant assumptions. See Section 5.6 for guidance
on how to conduct uncertainty analyses for BCA and other economic analyses.
: P1 in -I i" *i ;t Savings
If firms and households behave in ways consistent with profit and utility maximization, they will
adopt available cost-effective technologies or practices absent regulation. Even if they are not in
widespread use when a new regulation is developed, cost-effective technologies may be adopted
under baseline conditions in the future as information about their effectiveness spreads. When
households and firms voluntarily undertake these changes without the regulation, the regulatory
action cannot be credited with any private cost savings resulting from their adoption. In cases
where a regulation is estimated to result in net private cost savings, it is important to provide
evidence of why these cost-saving measures would not already be undertaken in the baseline.
When evidence to explain this phenomenon is not available, analysts should consider whether the
finding of private cost savings is defensible and whether all costs are being accounted for. For
instance, a regulation may impose "hidden" costs that are not easily quantified in a standard
engineering cost model but still represent welfare losses for firms or households that offset cost
savings from adopting a technology. Lower operating expenditures from a new technology required
by a regulation might be offset by increases in other costs if the new technology breaks down more
frequently, requires special training to operate, or has other undesirable features. If data are
available on such costs, analysts should include them in the analysis.
In some cases, evidence may suggest that firms or households do not adopt cost-saving measures
because of market failures (e.g., asymmetric information). If the regulation addresses these market
failures, it could lead to net private cost savings. In these instances, analysts should provide a clear
description and evidence of the market failure and how the new action addresses it
28 "Models are constructed to provide the simplest analysis possible that allows us to understand the issue at
hand [...] The real world is typically much more complex than the models we postulate. That doesn't invalidate
the model, but rather by stripping away extraneous details, the model is a lens for focusing our attention on
specific aspects of the real world that we wish to understand" (McAfee and Lewis 2009).
29 Demand elasticities show how the quantity of a product purchased changes as its price changes, all else equal.
Cross-price elasticities show how a change in the price of one good can result in a change in the price of another
good (either a substitute or a complement), thereby altering the quantity purchased. Income elasticity allows a
modeler to forecast how much more of a good or service consumers will buy when their income increases. See
Appendix A for more information about elasticities.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 5.3 - Model Choice
When selecting models for use in regulatory analysis, analysts should evaluate the following:
Is the model based on sound science? Prior to use in regulatory analysis, the model should be
subject to credible and objective peer review to ensure that it is consistent with scientific and
economic theory and based on the best available data and empirical evidence. Many of the
questions that follow can also be put to peer reviewers to evaluate the particulars of a specific
model and/or appropriateness of the model within a specific policy context
Is the model "fit for purpose"? Analysts must identify the best model(s) for the analysis and
thoroughly explain why it is applicable given the features and expected effects of the rule. A
model may be based on sound science but still inappropriate for evaluating the features of
interest
Is the model supported by the best available data? The suitability or representativeness of
underlying data to evaluate the effects of a specific policy is also an important consideration. For
instance, data quality and resolution may limit the ability to use some models in a regulatory
context For this reason, it is important to identify what data are available or can be collected to
adequately parameterize the model (U.S. EPA 2009). Analysts should use assumptions and
calibration/estimation of key parameters that are peer reviewed.
Does the model reasonably approximate the systems or market(s) of interest? A model
should capture the most salient details of the policy and the systems or markets affected. A
model selected to evaluate a regulation should be no more complicated than is necessary to
inform decision-making. If model capabilities add complexity without substantially improving
performance, the more transparent option is to eliminate them (NRC 2007].
Is the model transparent? In addition to model tractability, it is important that documentation
of all aspects of the model be publicly available, including details about model structure, key
assumptions, sources and values of key parameters, and limitations. When possible, models and
their underlying data should be publicly available. When a model is not publicly available, for
instance due to the confidential nature of underlying data, it is important to explain the reasons
for relying on these sources of information.
Can key model assumptions or parameter values be evaluated? Analysts should use
sensitivity analysis to explore the robustness of results to key input values, specifications, or
assumptions, particularly when the literature is inconclusive regarding the most defensible
approach or estimates. Sensitivity analysis may be application specific: parameters that may
matter little in one context may be key drivers of results in other contexts (U.S. EPA 2009).
Conducting uncertainty analysis is also important, as it "investigates the effects of lack of
knowledge and other potential sources of error in the model" (U.S. EPA 2009). Sensitivity and
uncertainty analysis inform users about the confidence that can be placed in model results. In
some cases, analysts also may need to rely on multiple models. Section 5.5 provides detailed
guidance regarding when sensitivity and uncertainty analyses are appropriate.
What are the key limitations of the model? Every model has its strengths and weaknesses. It
is important that decision makers and stakeholders understand a model's limitations. What does
the model capture? What is not captured or only captured with large bounds of uncertainty?
These should be communicated within the analysis in a way that is easy for a non-technical
audience to understand.
5-22 Guidelines for Preparing Economic Analyses | 3rd edition
-------
The economics literature has also documented specific instances in which households or firms act
in ways that appear to run counter to their self-interest (Shogren and Taylor 2008; Shogren et al.
2010; Croson andTreich 2014). However, research also indicates that market experience can
eliminate behavior that is inconsistent with profit-maximization in certain settings (List 2003; List
2011). If estimated net private cost savings could be due to widespread suboptimal behavior,
analysts should provide empirical evidence specific to the affected market. In addition, care should
be taken to ensure that assumptions that underlie modeled household behavior are consistent with
actual behavior.30 In the absence of such evidence, analysts should assume rational profit- or utility-
maximizing behavior by firms and households in the primary analysis, which would eliminate the
possibility of estimating net private cost savings as a result of regulation.31 Sensitivity analysis can
be used to consider other behavioral assumptions if warranted.
It is also important for analysts to make consistent assumptions about firm and consumer behavior
under the baseline and policy scenarios unless there is reason to believe the regulation will change
underlying behavioral patterns. For example, the economics literature has found mixed evidence on
whether car buyers fully account for future gasoline expenses when choosing fuel economy.32 A fuel
economy standard could reduce the impact of undervaluation of fuel economy on consumer
decisions, but if such behavior occurs in the baseline, it is likely to persist regardless of regulatory
requirements. Chapter 4 Section 4.4 offers more discussion about possible insights from behavioral
economics for policy design.
)gical Change
It is important to capture future changes in production techniques or pollution control that may
influence the baseline and policy scenarios and consequentially both costs and benefits.
Technological change can be thought of as having at least two components: genuinely new
technological innovation, such as the development and adoption of a new alternative pollution
control method; and learning effects, in which experience leads to cost savings through
improvements in operations, capability or similar factors. Analysts should recognize that the longer
the time horizon, the greater the uncertainty regarding the potential for and characteristics of
technological change (or learning) within a sector. Thus, it is important to balance the need to
account for the effect of innovation on the costs and benefits of regulation against the defensibility
of those analytic assumptions.
Technological change in other sectors of the economy may also be important to account for in the
analysis. For example, while the cost of phasing out ozone-depleting substances has declined over
time due to technological improvements in substitutes, innovation in mitigating factors, such as
improvements in skin cancer treatments and efficacy of sunscreen lotions, have also occurred.
Further, the analysis should include the costs associated with research and development (R&D),
including the potential to crowd out other investments that would have occurred absent the
30 See Ketcham etal. (2016) for an example where the finding that consumers do not act in their own self-
interest was actually driven by the inflexibility of the functional form assumed.
31 An exception would be when the regulation involves a transfer, such as a subsidy or rebate to purchase a
product, that leads to a net-cost savings for the firm or household. However, absent the value of the transfer, the
net-cost savings would still be negative under profit- or utility-maximizing behavior.
32 Recent studies continue to find a wide range in how consumers value future gasoline expenses in their vehicle
purchase decisions (Allcott and Wozny 2014; Busse et al. 2013; Sallee et al. 2016; Gillingham et al. 2021; Leard et
al. 2023).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
regulation, to correctly value cost-reducing technological innovation, but only if the costs are
induced by the regulation. Distinguishing R&D induced by the regulation from changes in other
investment decisions is often difficult. While innovation is expected to occur in the baseline and
policy scenarios, rates of technological change may differ across scenarios due to innovations that
reduce the cost of compliance. In cases where small changes in technology could dramatically affect
the costs and benefits, or where technological change is reasonably anticipated, the analyst should
consider exploring these effects in a sensitivity analysis. This might include probabilities associated
with specific technological changes or adoption rates of a new technology, or it may be an analysis
of the rate required to alter the policy decision. Such an analysis should show the policy significance
of emerging technologies that have already been accepted, or are, at a minimum, in development or
reasonably anticipated.
In some cases, there also may be empirical evidence of reductions in costs as firms accumulate
experience in production or abatement over time. Historic and projected estimates of learning are
often represented by "learning rates". A learning rate is typically defined as the percentage
reduction in costs for each doubling of production (or production capacity). It is not advisable to
assume a constant, generic learning rate or rate of technological progress, even if the rate is small,
simply because the continuous compounding of this rate over time can lead to implausible rates of
technological innovation and cost reduction. Furthermore, while learning may reduce compliance
costs over time, it is not widely believed that cost become negative as discussed in Text Box 5.4.
Before incorporating learning effects, the analyst should carefully examine the existing evidence for
relevance to the specific context. Estimated learning effects can vary due to many factors, including
already accumulated experience with a technology, industry, and the length of the time period
considered. Also, because estimates of learning rates are based on doubling of cumulative
production, including learning effects will have a greater influence on analyses with longer time
horizons. See Chapter 8 for further discussion.
npliance
One aspect of analytic design that can be complex is what to assume about the extent of compliance
with current and future environmental regulations. Assumptions about compliance in both the
baseline and policy cases can significantly affect the results of the analysis and should be clearly
described. Assumptions about compliance rates for a new regulation for a sector should generally
be based on past compliance behavior for related regulations for the sector. When an industry has
not been regulated before, data will not typically be available to gauge the likelihood of compliance
with a new rule, but compliance should be expected to be consistent with similar regulations of
similar entities. In most cases, a baseline and policy scenario that assumes full compliance should
be analyzed along with evidence-backed scenarios including alternative assumptions about
compliance.
5-24 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 5.4 - Technological Change, Induced Innovation, and the Porter
Hypothesis
There are many proposed mechanisms by which environmental regulation could cause
technological change. One mechanism is by induced innovation: the induced innovation
hypothesis states that as the relative prices of factors of production change, the relative rate of
innovation for the more expensive factor will also increase. This idea is well accepted; for
example, Newell et al. (1999} found that a considerable amount of the increase in energy
efficiency over the preceding few decades was caused by the increase in the relative price of
energy over that time.
A similar idea has also been described (somewhat less formally) as the "Porter Hypothesis"
(Porter and van der Linde 1995; Heyes and Liston-Heyes 1999). Jaffe and Palmer (1997)
delineate three versions of the hypothesis: weak, narrow and strong.
The weak version of the hypothesis assumes that an environmental regulation will stimulate
innovation, but it does not predict the magnitude of these innovations or the resulting cost
savings. There is mixed evidence in support of the weak version of the Porter hypothesis
(Ambec et al. 2013; Martinez-Zarazosa, et al. 2019). This version of the hypothesis is very
similar to the induced innovation hypothesis.
The narrow version of the hypothesis predicts that flexible regulation (e.g., incentive-based) will
induce more innovation than inflexible regulation and vice versa. There is empirical evidence
thatthis is the case (Kerr and Newell 2003; Popp 2003; De Santis and Jona Lasinio 2016).
Analysts may be able to estimate the rate of change of innovation under the weak or narrow
version of the hypothesis, or under induced innovation. Note, however, that these types of
innovation may crowd out other forms of innovation. By raising the cost of pollution, the
regulation makes it profitable to find cheaper compliance strategies, but finding these strategies
also has its own opportunity cost (e.g., firms use their engineers, scientists, and other experts to
develop more cost-effective compliance strategies instead of developing some other
technology).
The strong version of the Porter Hypothesis predicts cost savings from environmental
regulation under the assumption that firms do not maximize cost savings without pressure to
do so. While anecdotal evidence of this phenomenon may exist, the available economic
literature has found no statistical evidence supporting it as a general claim (Jaffe et al. 1995;
Palmer etal. 1995; Jaffe and Palmer 1997; Brannlund and Lundgren 2009; Ambec etal. 2013;
Dechezlepetre and Sato 2017). For the strong version to be true, it requires special assumptions
and an environmental regulation combined with other market imperfections that are difficult to
generalize. Thus, analysts should not assume cost savings from a regulation based on the strong
version of the Porter Hypothesis.
When there are significant compliance issues with an existing regulation, an assumption of under-
compliance in the baseline for a new regulation should be included when supported by evidence
from monitors, inspections, or enforcement actions.33 Analysts may establish a "current practice"
33 For example, in the Lead Renovation, Repair, and Painting Program Rule (U.S. EPA 2008), the EPA assumed a
75% percent compliance rate for estimating costs and benefits based on compliance in the construction industry
with previous occupational health and safety regulations.
5-25 Guidelines for Preparing Economic Analyses | 3rd edition
-------
baseline incorporating data on actual compliance rates rather than assume full compliance. Current
practice baselines are particularly useful for regulations intended to address compliance problems
with existing policies. Assuming a full-compliance baseline that disregards under-compliant
behavior could obscure the value of these types of regulations.34 If the policy being evaluated is not
designed to address the underlying reason for non-compliance, then under-compliance data may be
applicable to the policy case as well as the baseline.
If under-compliance is assumed either in the baseline or in the policy case, then identifying the
reason for non-compliance is important and could affect the sign of the regulation's net benefits and
the distribution of benefits and costs. For example, non-compliance could occur selectively where
compliance costs are high. If compliance is not systematically correlated with costs, then the
compliance assumption is less likely to change the sign of the regulation's net benefits.
When analyzing new requirements for an industry subject to existing regulations, it is important to
carefully specify the assumptions about baseline compliance to avoid double counting benefits and
costs. This could arise if the same set of actions occurs across multiple regulations. Assuming full
compliance with existing regulations in the baseline makes it easier for analysts to focus on the
incremental effects of the new regulatory action without double counting. If there is evidence of
under-compliance in the baseline, analysts should consider whether the regulation is structured to
reduce the compliance problem35 or whether the problem is likely to persist in the policy case. If it
will persist and this behavior is not captured, the net benefits of a regulation will not be estimated
correctly. For example, if analysts repeatedly factor under-compliance into the baselines for a
sequence of emissions tightening rules but assume that entities will fully comply under the policy
case, inconsistent results will arise. Summing the benefits and costs from the sequence of rules will
overstate the benefits and costs because each rule takes credit for a portion of the same actions.
Conversely, there may be cases in which firms over-comply with regulations. Over-compliance in
the policy scenario should be assumed in limited circumstances. As with under-compliance, it is
important to identify the reason for over-compliance and assure it is consistent with expected
behavior. The analysis should not typically assume that a regulation will motivate abatement
greater than what is legally required. However, over-compliance may occur if firms wish to reduce
the risk of non-compliance (e.g., facilities may overcontrol due to local pressure) or because least-
cost compliance methods achieve greater reductions than required (e.g., shifting to a different
process that does not pollute rather than installing abatement equipment) among other reasons. In
such cases, the benefits, and costs of over-compliance in the policy case should be accounted for. If
more additional regulations are considered later, current practices can be used to define baseline
conditions for the new regulation unless these practices are expected to change.
To summarize, analysts should include a baseline and policy scenario that assumes full compliance,
but under-compliance in the baseline or policy scenario should also be analyzed when there is
supporting evidence. Over-compliance can be assumed in limited circumstances. Whenever
scenarios other than full compliance are included in regulatory analysis, the analyst should discuss
the sensitivity of the results to the compliance rate assumption.
34 For instance, banning lead from gasoline was precipitated, in part, by the noncompliance of consumers. When
consumers put leaded gasoline in vehicles that required non-leaded fuel, this resulted in increased vehicle
emissions (U.S. EPA 1985).
35 See Section 3.2 for a brief discussion of relevant enforcement methods to consider, Chapter 4 for some
examples, and Section 8.2for a discussion of compliance assumptions in a cost analysis.
5-26 Guidelines for Preparing Economic Analyses | 3rd edition
-------
untary Actions
Occasionally, polluting industries adopt voluntary measures to reduce emissions. Firms or sectors
can undertake such actions independently, or they might participate in formal, government-
sponsored programs. Such voluntary measures are adopted for a variety of reasons, including to
improve public relations, to avoid regulatory controls, to reduce other legal risks or to access
resources associated with joining a formal program. When this is the case, it is important to account
for these actions in the baseline for new regulations and to be explicit about the assumptions of
firms' future actions. If participation in voluntary programs was motivated by the threat of the
regulation, then a new regulation could affect future participation in these programs.
Typically, voluntary emission reductions that are expected to occur without a new regulation may
be included in the baseline consistent with the guidance on over-compliance above. This is not
always possible, however, as voluntary actions are often difficult to measure (Brouhle et al. 2005).
It can be difficult to determine whether pollution reduction measures that precede compliance
dates represent anticipatory effects that are attributable to a regulation or if they are voluntary
measures that would have occurred without the regulation. Sensitivity analysis could shed light on
the importance of assumptions about voluntary emission reductions under the baseline if this is a
significant source of uncertainty.
'.'i,'ing- in 'sYh r Environmental .' ntaminants
It is common for EPA regulations to cause decreases and increases in environmental contaminants
that are not the subject of the regulation. These changes may occur for a variety of reasons that the
analyst should consider. This section provides guidance specifically on identifying and accounting
for changes in these other contaminants by drawing out the implications of properly accounting for
the baseline and behavior discussed above. Projections of changes in the levels of environmental
contaminants should be consistent with expected economic behavior. These changes should be
based on expected outcomes of least cost compliance, existing economic relationships, and
continued compliance with existing regulations. The analysis should take a balanced approach to
identifying increases and decreases in other contaminants that may be affected by the regulation
relative to efforts to account for other welfare changes that may result36 Any benefits or costs from
these changes in other contaminants should be accounted for in a BCA.
As mentioned in Section 5.1.3, changes in environmental contaminants other than those subject to
the regulation may result from the compliance approaches used by regulated entities.37 For
example, the use of an abatement technology to reduce one air pollutant may simultaneously
36 The benefits from changes in environmental contaminants other than those related to the statutory objective
of the regulation have sometimes been called "co-benefits" and these contaminants sometimes called "co-
pollutants". However, these terms are imprecise and have been applied inconsistently in past practice, and as
such should be avoided (unless these terms are used explicitly in statutes). Similarly, benefits from changes in
environmental contaminants other than those related to the statutory objective of the regulation are sometimes
referred to as "ancillary benefits,". This term should be used cautiously in an analysis because it may be
interpreted as having economic, legal or policy meaning that is unintended.
37 Section 5.1.3 also emphasizes that changes in externalities other than those due to changes in environmental
contaminants should be accounted for in a BCA. These other externality changes are not as common across
regulations as changes in other contaminants but may be particularly important in certain regulatory contexts
such as changes in transportation externalities from emissions standards on vehicles (e.g., congestion or safety)
or changes in ambient conditions such as temperature and noise.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
reduce or increase other air pollutants from the same source, and/or could change the emissions of
the same or another pollutant into a different medium (e.g., water). It is also possible for changes in
other environmental contaminants to occur as a result of market interactions. For example, a
regulation may cause consumers or firms to substitute away from one commodity towards another,
whose increased production may be associated with additional emissions of an environmental
contaminant as well as the costs of abating it Other examples include when a regulation induces
beneficial reuse of a waste product and thereby reduces production and the associated emissions
and costs of the product that the waste replaces; a controlled pollutant might be a precursor to
multiple secondary pollutants; or when the use of a hazardous product is banned, and its
replacement also poses a hazard.
Care should be taken when estimating changes in other contaminants to ensure they are consistent
with expected market behavior and technological change. For example, consider an abatement
technology that may potentially reduce emissions of multiple pollutants. The analyst should
consider whether the technology will achieve similar reductions in all of these pollutants in a new
application as it had in previous applications, or if the regulated entities will tailor it to control the
regulated pollutant(s) in the new application so as to reduce the technology's cost
As with estimating changes in contaminants subject to a regulation, analysts should also consider
the implications of existing pollution control regulations on other contaminant levels and costs. For
example, consider the case where a regulation on one pollutant leads to installations of a
technology that reduces a second pollutant, and that second pollutant is subject to an allowance
trading program with a cap that is economically binding (i.e., there is a positive allowance price). In
this case, the regulation may not ultimately lead to reductions in the second pollutant Instead,
reductions in the second pollutant at regulated entities that install the new technology may be
offset by reductions in abatement activities by entities subject only to the cap.38 To the extent that
any new regulation affects the cost of complying with an existing regulation, as would occur in this
example, these changes in cost should be accounted for in the analysis.
If a regulation is expected to increase environmental contaminants not subject to the regulation,
they should be accounted for in a BCA even if an anticipated future regulation is expected to
mitigate them. This guidance follows directly from establishing the baseline and accounting for all
benefits and costs. It is important to account for these changes for completeness, such that the sum
of the benefits and costs of rules evaluated in sequence should sum to the costs and benefits of the
rules if evaluated collectively as discussed in Section 5.3.3.
Finally, as discussed in Chapter 3, if the regulation is expected to induce large benefits from changes
in contaminant(s) beyond those arising from its primary statutory objective, an analysis of a policy
option where those contaminant(s) are regulated, either separately or simultaneously with the
contaminants that are the primary statutory objective of the regulation, may be useful to determine
whether there are more economically efficient ways of obtaining these benefits.
38 There may still be benefits (or negative benefits) from changes in the timing and location of emissions of these
environmental contaminants even if the cap continues to bind. Chapter 4 describes how allowance trading
programs work.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
5.6 Uncertainty
Uncertainty is inherent in BCAs, particularly when estimating and valuing environmental benefits
for which there are no existing markets.39 The primary issue is often not how to reduce uncertainty,
but how to account for it and present useful conclusions to inform policy decisions. While
households and firms can be expected to incorporate uncertainty in decisions and responses will
reflect their risk preferences (see Section 5.5.1), BCA itself should not adopt any particular risk
stance because the Potential Pareto criterion requires BCA to reflect the values of those affected. An
additional imposition of risk preferences in the BCA itself is, therefore, inconsistent with the
underlying basis of BCA.
BCAs should present information on the expected or most plausible outcomes and associated
uncertainty (Dudley et al. 2017). It is important to recognize that point estimates alone cannot
provide policy makers with information about whether these estimates are robust to alternate
assumptions, nor can they convey the full range of potential outcomes. Treatment of uncertainty is
an essential component of analysis that enhances the communication process between analysts and
policy makers.
The guiding principles for assessing and describing uncertainty in analysis are transparency and
clarity of presentation. Although the extent to which uncertainty is treated and presented will vary
according to the specific needs of the analysis, some general minimum requirements apply to most
BCAs. In assessing and presenting uncertainty, analysis should:
Present outcomes or conclusions based on expected or most plausible values;
Provide descriptions of all known key assumptions, biases, and omissions;
Perform sensitivity analysis on key assumptions;
Include sensitivity analyses that examine both higher and lower values rather than only
one or the other;
Justify the assumptions used in the analyses; and
Make full use of available probability distributions of key parameters.
Sensitivity analysis on key assumptions may be all that is needed for an uncertainty analysis, or it
may be only the initial assessment. Statistical confidence intervals and probability distributions, if
available, are used to describe the statistical uncertainty associated with specific variables and to
provide a more complete characterization of uncertainty. The outcome of the initial assessment
may be sufficient to understand the influence of key parameters on outcomes and to inform the
policy decisions. If, however, the implications of uncertainty are not adequately captured in the
initial assessment then a more sophisticated analysis should be undertaken when the data allow.
The need for additional analysis should be clearly stated, along with a description of the methods
used for assessing uncertainty.
Probabilistic methods such as Monte Carlo analysis can be particularly useful because they
explicitly characterize analytical uncertainty and variability (e.g. Brandimarte, 2014). Where
probability distributions of relevant input assumptions are available and can be feasibly and
39 Stemming from definitions given in Knight (1921). economists have often distinguished risk and uncertainty
according to how well one can characterize the probabilities associated with potential outcomes. Risk applies to
situations or circumstances in which a probability distribution is known or assumed, while uncertainty applies to
cases where knowledge of probabilities is absent. However, these definitions are not always adhered to in
economics. Also, note that the economic definitions for these terms may differ from those used in other
disciplines.
5-23 Guidelines for Preparing Economic Analyses | 3rd edition
-------
credibly combined, BCAs should characterize how the probability distributions of the relevant input
assumptions would, on net, affect the resulting distribution of benefit and cost estimates. In this
case, the analysis would consider sources of uncertainty jointly rather than singly.
However, probabilistic methods can be challenging to implement when data needed to characterize
distributions are limited.40 In the absence of data to specify distributions for specific parameter
values, it is more transparent and defensible to use simpler sensitivity analysis. Note that for some
rules OMB Circular A-4 requires a formal quantitative uncertainty analysis that provides some
estimate of the probability distribution of regulatory effects.41
The analysis should make clear that the statistical uncertainty captured by the Monte Carlo or other
probabilistic analysis generally does not account for model uncertainty, the degree to which
mathematical models represent real-world systems. For example, when quantifying changes in a
specific health effect from a reduction in an environmental contaminant, the statistical uncertainty
analysis assumes that a particular dose-response model is the "true" model; that is, as if we are
certain there is a causal relationship and that the dose-response function used in the analysis is the
truth. There are some approaches to incorporating model uncertainty in probabilistic analyses,
such as model averaging.42 More often, model uncertainty (including uncertainty over whether an
environmental contaminant causes a specific type of health impact) will need to be captured and
described independent of the statistical uncertainty analysis. When possible, alternative model
specifications that are supported by or consistent with underlying biological, engineering, or
economic evidence or theory should be used to illustrate the consequences of assuming a different
model.
It is important to recognize that there may be cases where there are competing assumptions,
estimates or models considered as equally plausible that cannot be combined or weighted
probabilistically. In these cases, it can be appropriate for the results driven by these factors to be
presented co-equally in the BCA. However, the number of outcomes will generally grow
multiplicatively with the number of inputs treated equivalently. For example, if there are three
alternative inputs that are evaluated and treated separately and equivalently, and each of these can
take two values, then there would be eight co-equal net benefits estimates to present, making it
difficult to interpret BCA results. Therefore, the presentation of co-equal results in BCA should be
done sparingly, with sensitivity analysis as the preferred treatment where possible. Presenting co-
equal results should be reserved for particularly important analytic inputs and should always be
fully described and justified whenever it is done.
* ^ i I-11* "*11tt111^nsitivity Analys i
Sensitivity analysis is a systematic method for describing how net benefit estimates or other
outputs of the analysis change with assumptions about input parameters. Some basic principles for
sensitivity analysis include:
Identify key parameters. For most applied analyses, a full sensitivity analysis that includes
every variable is not feasible. Instead, the sensitivity analysis will often need to be limited to
40 Jaffe and Stavins (2007) provide a useful overview of probabilistic analysis of uncertainty in regulatory
analysis, including challenges and limitations.
41 See Circular A-4 (OMB 2023) for additional details on this requirement.
42 Moral-Benito (2015) provides an overview of model averaging in economics.
5-30 Guidelines for Preparing Economic Analyses | 3rd edition
-------
those input parameters considered to be key or particularly important, which may be
economic parameters (e.g., valuation estimates) or inputs from other disciplines that feed
into the benefits analysis (e.g., dose-response, exposure). A determination about which
parameters are key should be informed by the range of possible values for input parameters
and each one's functional relationship to the output of analysis. The analyst should specify a
plausible range of values for each key variable and describe the rationale for the range of
values tested.
Vary these key parameters. The most common approach is a partial sensitivity analysis
that estimates the change in net benefits or other economic outcomes while varying a single
parameter, leaving other parameters at their base value. A more complete analysis will
present the marginal changes in the economic outcome as the input parameter takes on
progressively higher or lower values. When an input has known or reasonably determined
maximum and minimum values, it can be informative to investigate if outcomes are robust
to these alternative input values.
Varying two parameters simultaneously can often provide a richer picture of the
implications of base values and the robustness of the analysis but can be more difficult to
communicate effectively. Analysts should consider using graphs to present these combined
sensitivity analyses by plotting one parameter on the x-axis, the economic outcome on the
y-axis, and treating the second parameter as a shift variable.43 Results of the sensitivity
analysis should be presented clearly and accompanied with descriptive text
Identify switch points. Switch points are defined as those conditions under which the
economic analysis would recommend a different policy decision. For BCA, the switch point
would typically be the input parameter value where estimated net benefits changes sign.
Switch point values for key input parameters can be very informative. For instance, they can
be compared to the available literature to assess whether the values are plausible or well
outside known distributions or observations. While switch points are not tests of
confidence in the statistical sense, they can help provide decision-makers with an
understanding of how robust the analytic conclusions are.
Assess the need for more detailed analysis. Finally, sensitivity analyses may be used as a
screening tool to determine where more extensive effort may be needed. For example, the
plausible range of values for an influential uncertain parameter may be narrowed with
further research or data gathering, which can be used to better characterize the parameter's
uncertainty. If several parameters independently have a large influence on the results of the
analysis when they are varied, then a more sophisticated treatment of uncertainty that
allows for joint consideration of their effects may be necessary. One option is to combine
alternative values for multiple parameters into a scenario that differs from the primary
analysis. It is important that the selected values be consistent with one another and that
choices are explained and well-documented. It is also important to consider that combining
extreme values for multiple inputs (e.g., minimum values) can produce a scenario that is
unlikely so the analysis should include some description of the plausibility of the
combination of values.
43 When the analysis contains many highly uncertain variables, presentation may be facilitated by noting the
uncertainty of each in footnotes and carrying through the central analysis using best point estimates.
5-31 Guidelines for Preparing Economic Analyses | 3rd edition
-------
C Approach - r * ; isiderWhen Data Are Missing
When key data elements are unavailable in an analysis it will not be possible to estimate central
values or perform sensitivity or quantitative analysis around those values. In these cases, it is
important to assess and qualitatively characterize the importance of the missing information in the
analysis. There are also analytic approaches to consider when data are missing.
Break-even analysis. Break-even analysis can be used when one element is missing in an
analysis. Essentially, break-even analysis identifies the switch point value for the missing
element where the net benefits change sign.44 Unlike the case above, however, the switch
point value cannot be associated with any point an underlying distribution. Break-even
analysis may best be explained by example. Suppose a BCA shows that net monetized
benefits are negative, but there is a key health endpoint with an established per-unit
estimate of economic value but without risk estimates that would allow quantification of the
health endpoints. In this case it is possible to estimate the number of cases of the health
endpoint avoided (each valued at the per-unit value estimate) at which overall net benefits
become positive, or where the policy action will "break even."
The same sort of analysis can be performed when analysts lack valuation estimates,
producing a break-even value that should again be assessed for credibility and plausibility.
Break-even analysis can also be used for missing costs when net quantified benefits are
positive, shedding light on what the value of the missing cost estimate would need to be for
net benefits to be negative.
Break-even estimates can be assessed for plausibility either quantitatively or qualitatively.
For example, the break-even unit value estimate for a specific health endpoint may be
compared to values for effects considered to be more or less severe than the endpoint being
evaluated. For the break-even value to be plausible, it should fall between the estimates of
these more and less severe effects. Policy makers will need to determine if the break-even
value is acceptable or reasonable.
Break-even analysis is most useful when there is only one missing value in the analysis, or
when it is applied to a large or important missing value. For example, an analysis missing
risk estimates for two different health endpoints (but with valuation estimates for both),
would need to consider a "break-even frontier" that allows the incidence of both effects to
vary. While it is possible to construct such a frontier, it may be difficult to determine which
points on the frontier are relevant for policy analysis.
Expert elicitation. Expert elicitation is a formal process for obtaining and combining
judgments from experts on missing inputs in the economic analysis.45 The values elicited,
and the uncertainty around these values if characterized in the elicitation, can then be used
in the economic analysis for those missing inputs. Typically, expert elicitations include
multiple experts to capture a range of backgrounds and diversity of knowledge, but
ultimately the responses of these experts are combined into a single estimate or probability
distribution for the input of interest. There are established approaches for the elicitation,
including how to define the target questions, conduct expert interviews and analyze the
44 Boardman et al. (2018) describes determining break-even points under the general subject of sensitivity
analysis and includes empirical examples.
45 OMB Circular A-4 (2023) suggests analyses consider drawing upon expert judgment using Delphi methods, a
form of expert elicitation.
5-32 Guidelines for Preparing Economic Analyses | 3rd edition
-------
responses. Expert elicitations conducted for BCA should follow best practices and carefully
document the elicitation process, results and how those results are applied in the BCA.
Formal expert elicitation can be time-consuming and require substantial resources so it
should be reserved for those cases where its value, in terms of improving the BCA for
decision-making, merits the resources needed. Less formal approaches for drawing upon
expert judgments for missing values may also be useful if those approaches are clearly
identified and described in the economic analysis, including any known limitations. See
Colson and Cooke (2018) for an overview of expert elicitation.
.'vi,-i . *ii videratioii v Related to Uncertainty and Risk
There are additional issues related to uncertainty that may merit consideration, including how to
account for responses to risk information, how to evaluate policies or regulations that provide
information, and how to consider the value of information that may become available later.
Uncertainty may affect private decisions. Households and firms can be expected to
incorporate uncertainty when making decisions such as what actions to take to reduce their
own exposures to risk or what investments to make in response to regulation. As with other
aspects of behavioral responses, to measure benefits and costs of a regulation it is
important to clearly characterize the behavior of firms and households. As described in
Section 5.5.1, analysts should generally assume utility or profit maximization under
uncertainty, taking preferences over uncertainty as given. For households, for example, this
generally means a model of expected utility maximization with whatever degree of risk
aversion best represents the affected households. In practice it also means that existing
estimates of willingness-to-pay will reflect the risk preferences of the populations analyzed.
See Section 5.5.1 for more information on representing economic behavior and behavioral
responses.
Lay and expert risk perceptions. Lay perceptions of risk may differ significantly from
scientific assessments of the same risk. An extensive literature has developed on the topic.46
Because individuals respond according to their own risk perceptions, it is important for the
analyst to be attentive to situations where there is an obvious divergence in these two
measures. In such cases, analyses should clearly state the basis for the economic value
estimates used in their analysis and should also consider describing the known differences
between public risk perceptions and scientific risk assessments. It may also be useful for the
regulation to provide information to the public that may reduce these differences and that
may allay public concerns.
Provision of information. Some policy actions focus on providing information to
individuals on risks to health and welfare. If this information allows them to make better
decisions that improve household welfare, there is an economic benefit to providing this
information. When this is the case, revealed preference approaches can make new
information appear to have a net negative effect on household welfare because households
may undertake new (and costly) activities in response. For example, information on
drinking water quality may lead consumers to buy and use costly filtration systems at home,
which could be misconstrued to mean that providing the information diminished consumer
welfare. An appropriate framework for evaluating the benefits of information provision
46 For a general overview see Renner et al. 2015.
5-33 Guidelines for Preparing Economic Analyses | 3rd edition
-------
under these circumstances is to assess the costs of sub-optimal household decisions under
the less complete information.47 Analysts should carefully consider these issues when they
evaluate policies that focus on information provision.
Option value: Some environmental policies involve irreversible decisions made in the face
of uncertainty. If information that reduces this uncertainty can be expected to develop over
time, then there is a positive value to waiting until this information becomes available.48 In
this case, the value originates from the option to hold off making the decision until
uncertainties are resolved or reduced. An analysis can show the potential costs of making a
decision without this new information. The potential gains from waiting may best be
evaluated in a value-of-information framework where the gains in net benefits from having
better information can be compared to the costs associated with gathering this information,
which includes any forgone benefits due to postponing environmental protection.49
Generally, it is difficult to quantitatively include these option values in an analysis, but the
concept is useful and may be highlighted qualitatively if circumstances warrant Further,
this is an important concept to keep in mind when considering policy approaches. As
described in Section 4.6.5 it may be useful to examine approaches such as voluntary
programs or pilot projects designed to gather information to make a more informed
analysis of the benefits and costs of regulatory approaches.
47 Foster and Just (1989) describes this approach more fully, demonstrating that compensating surplus is an
appropriate measure of willingness-to-pay under these conditions. The authors illustrate this with an empirical
application to food safety.
48 This is sometimes known as quasi-option value, starting with the seminal work of Arrow and Fisher (1974). A
slightly different framing is "real options" analysis following Dixit and Pindyck (1994). These approaches yield
option values that differ slightly but capture the same concept. Traeger (2014) describes the precise relation
between the two and how they can be considered in benefit-cost analysis.
49 Examples of value of information analysis include Marchese et al. (2018) and von Winterfeldt et al. (2020).
5-34 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 5 References
42 U.S.C. 9621: Cleanup standards. Available at: https: //www.law.cornell.edu/uscode/text/42 /9621
(accessed October 21, 2024).
Aldy, J. 2 014. Learning from Experience: An Assessment of the Retrospective Reviews of Agency Rules and the
Evidence for Improving the Design and Implementation of Regulatory Policy. Report for Administrative
Conference of the United States.
Aldy, J., M. Auffhammer, M. Cropper, A. Fraas, and R. Morgenstern. 2022. Looking Back at 50 Years of the
Clean Air Act .Journal of Economic Literature, 60 (l]:l-54.
Allcott, H. and N. Wozny. 2014. Gasoline Prices, Fuel Economy, and the Energy Paradox. Review of Economics
and Statistics, 96(5): 779-795.
Ambec, S., M.A.Cohen, S. Elgie, and P. Lanoie. 2013. The Porter Hypothesis at 20: Can Environmental
Regulation Enhance Innovation and Competitiveness? Review of Environmental Economics and Policy, 7(1): 2
22.
Arrow, K. J. and A. C. Fisher. 1974. Environmental Preservation, Uncertainty and Irreversibility, Quarterly
Journal of Economics, 88(1): 1-9.
Boardman, A.E., D. H. Greenberg, A.R. Vining, and D.L. Weimer. 2018. Cost-Benefit Analysis: Concepts and
Practice. Cambridge: Cambridge University Press.
Brandimarte, Paolo. 2014. Handbook in Monte Carlo Simulation: Applications in Financial Engineering, Risk
Management, and Economics. John Wiley & Sons, Hoboken, NJ.
Brannlund, R. and T. Lundgren. 2009. Environmental Policy without Costs? A Review of the Porter
Hypothesis. International Review of Environmental and Resource Economics, 3(2): 75-117.
Brouhle, K., C. Griffiths, and A. Wolverton. 2005. The Use of Voluntary Approaches for Environmental
Policymaking in the U.S. The Handbook of Environmental Voluntary Agreements, ed. E. Croci. The Netherlands:
Kluwer Academic Publishers.
Busse, M., C. Knittel, and F. Zettelmeyer. 2013. Are Consumers Myopic? Evidence from New and Used Car
Purchases. American Economic Review, 103(1): 220-256.
Colson, A.R. and Cooke, R.M., 2018. Expert Elicitation: Using the Classical Model to Validate Experts'
Judgments. Review of Environmental Economics and Policy, 12(1): 113-132.
Cropper, M., R. Morgenstern, and N. Rivers. 2018. Policy Brief-Facilitating Retrospective Analysis of
Environmental Regulations. Review of Environmental Economics and Policy, 12(2): 359-70.
Croson, R. and N. Treich. 2014. Behavioral Environmental Economics: Promises and Challenges.
Environmental and Resource Economics, 58(3): 335-351.
De Santis, R., and C. Jona Lasinio. 2016. Environmental Policies, Innovation and Productivity in the EU. Global
Economy Journal, 16(4): 615-635.
Dechezlepetre. A., and M. Sato. 2017. The Impacts of Environmental Regulations on Competitiveness. Review
of Environmental Economics and Policy, 11(2): 183-205.
Diamond, P. and J. Hausman. 1994. Contingent Valuation: Is Some Number Better than No Number? Journal of
Economic Perspectives, 8(4): 45-64.
Dixit, A.K. and R.S. Pindyck. 1994. Investment under Uncertainty. Princeton, NJ: Princeton University Press.
Dudley, S., Belzer, R., Blomquist, G., Brennan, T., Carrigan, C., Cordes, J.,... and Hammitt, J. 2017. Consumer's
Guide to Regulatory Impact Analysis: Ten Tips for Being an Informed Policymaker. Journal of Benefit-Cost
Analysis, 8(2): 187-204.
5-35 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Executive Order 12866: Regulatory Planning and Review, October 4,1993. Available at:
https://www.archives.gov/files/federal-register/ executive-orders/pdf/12866.pdf (accessed December 22,
2020).
Farrow, S. 2013. How (not) to Lie with Benefit-Cost Analysis. The Economists' Voice, 10(1): 45-50.
Farrow, S. and A. Rose. 2018. Welfare Analysis: Bridging the Partial and General Equilibrium Divide for Policy
Analysis. Journal ofBenefit Cost Analysis, 9(1): 67-83.
Foster, W. and R. E. Just. 1989. Measuring Welfare Effects of Product Contamination with Consumer
Uncertainty. Journal of Environmental Economics and Management, 17: 266-283.
Fraas, A., E. Kopits, and A. Wolverton. 2023. A Review of Retrospective Cost Analyses. Review of
Environmental Economics and Policy, 17(1): 22-42.
Gayer, T. and W. K. Viscusi. 2016. Determining the Proper Scope of Climate Change Policy Benefits in U.S.
Regulatory Analyses: Domestic versus Global Approaches. Review of Environmental Economics and Policy,
10(2): 245-263.
Gillingham, K.T., S. Houde, and A.A. Van Benthem. 2021. Consumer Myopia in Vehicle Purchases: Evidence
from a Natural Experiment. American Economic Journal: Economic Policy, 13(3): 207-238.
Goulder, L.C. and R.C. Williams III. 2003. The Substantial Bias from Ignoring General Equilibrium Effects in
Estimating Excess Burden, and a Practical Solution. Journal of Political Economy, 111(4): 898-927.
Hahn, R.W. and J.A. Hird, 1991. The Costs and Benefits of Regulation: Review and Synthesis. Yale Journal on
Regulation, 8: 233.
Harberger, A.C. 1964. The Measurement of Waste. American Economic Review, 54(3): 58-76.
Heyes, A.G. and C. Liston-Heyes. 1999. Corporate Lobbying, Regulatory Conduct and the Porter Hypothesis.
Environmental and Resource Economics, 13: 209-218.
Jaffe, A.B. and K. Palmer. 1997. Environmental Regulation and Innovation: A Panel Data Study. The Review of
Economics and Statistics, 79(4): 610-619.
Jaffe, A.B., S.R. Peterson, P.R. Portney, and R.N. Stavins. 1995. Environmental Regulation and the
Competitiveness of U.S. Manufacturing: What Does the Evidence Tell Us? Journal of Economic Literature, 33:
132-163.
Jaffe, J. and R. N. Stavins, 2007. On the Value of Formal Assessment of Uncertainty in Regulatory Analysis.
Regulation & Governance, 1(2): 154-171.
Just, R.E., D.L. Hueth, and A. Schmitz. 2004. The Welfare Economics of Public Policy: A Practical Approach to
Project and Policy Evaluation. Northampton, MA: Edward Elgar Publishing.
Ketcham, J.D., Kuminoff, N.V., and Powers, C.A., 2016. Choice Inconsistencies Among the Elderly: Evidence
from Plan Choice in the Medicare Part D Program: Comment. American Economic Review, 106(12): 3932-61.
Kerr, S. and R. Newell. 2003. Policy-Induced Technology Adoption: Evidence from the U.S. Lead Phasedown.
Journal of Industrial Economics, 51(2): 317-343.
Kling, C.L. and D.J. Phaneuf. 2018. How are Scope and Adding Up Relevant for Benefits Transfer?
Environmental and Resource Economics, 69(3): 483-502.
Knight, F. H. 1921. Risk, Uncertainty, and Profit. Boston, MA: Houghton, Mifflin Co.
Kopp, R., A. Krupnick, and M. Toman. 1997. Cost-Benefit Analysis and Regulatory Reform: An Assessment of
the Science and the Art. Report to the Commission on Risk Assessment and Risk Management.
Leard, B., J. Linn, and Y.C. Zhou. 2023. How Much Do Consumers Value Fuel Economy and Performance?
Evidence from Technology Adoption. Review of Economics and Statistics, 105(1): 158-174.
List, J. 2003. Does Market Experience Eliminate Market Anomalies? Quarterly Journal of Economics, 118(1):
41-71.
5-36 Guidelines for Preparing Economic Analyses | 3rd edition
-------
List, J. 2011. Does Market Experience Eliminate Market Anomalies? The Case of Exogenous Market
Experience. American Economic Review, 101(3): 313-317.
McAfee, P. and T.R. Lewis. 2009 Introduction to Economic Analysis. Saylor Foundation. Available at:
https://open.umn.edu/opentextbooks/textbooks/introduction-to-economic-analysis.
Marchese, D.C., M.E. Bates, J.M. Keisler, M.L. Alcaraz, I. Linkov, and E.A. Olivetti. 2018. Value of information
analysis for life cycle assessment: Uncertain emissions in the green manufacturing of electronic tablets.
Journal of Cleaner Production, 197 (part 1): 1540-1545.
Marten, A.L., R. Garbaccio, and A. Wolverton. 2019. Exploring the General Equilibrium Costs of Sector-Specific
Environmental Regulations. Journal of the Association of Environmental and Resource Economists, 6(6): 1065-
1104.
Martinez-Zarazosa, I., A. Bengochea-Morancho, R. Morales-Lage. 2019. Does Environmental Policy Stringency
Foster Innovation and Productivity in OECD Countries? Energy Policy, 134:110982.
Moral-Benito, E. 2015. Model Averaging in Economics: An Overview. Journal of Economic Surveys, 29(1): 46-
75.
Morgenstern, Richard. 2018. Retrospective Analysis of US Federal Environmental Regulation. Journal of
Benefit Cost Analysis, 9(2): 285-304.
NAS (National Academies of Science). 2017. Valuing Climate Damages: Updating Estimation of the Social Cost
of Carbon Dioxide. Washington, DC: The National Academies Press.
NRC (National Research Council). 2007. Models in Environmental Regulatory Decision Making. Committee on
Models in the Regulatory Decision Process, Board on Environmental Studies and Toxicology, Division on
Earth and Life Studies. Washington, D.C.: National Academies Press.
Newbold, S. C., P.J. Walsh, D.M. Massey, and J. Hewitt. 2018. Using Structural Restrictions to Achieve
Theoretical Consistency in Benefit Transfers. Environmental and Resource Economics, 69(3): 529-553.
Newell, R.G., A.B. Jaffe, and R. Stavins. 1999. The Induced Innovation Hypothesis and Energy-saving
Technological Change. The Quarterly Journal of Economics, 114(3): 941-975.
0MB (Office of Management and Budget). 2023. Circular A-4, Regulatory Analysis, November 9, 2023.
Available at: https: //www.whitehouse.gov/wp-content/uploads/2023 /ll ZCircularA-4.pdf (accessed June
13,2024).
Palmer, K., W.E. Oates, and P.R. Portney. 1995. Tightening Environmental Standards: The Benefit-Cost or the
No-Cost Paradigm? Journal of Economic Perspectives, 9(4): 119-132.
Popp, D. 2003. Pollution Control Innovations and the Clean Air Act of 1990. Journal of Policy Analysis and
Management, 22(4): 641-660.
Porter, M.E. and C. van der Linde. 1995. Toward a New Conception of the Environment-Competitiveness
Relationship. Journal of Economic Perspectives, 9(4): 97-118.
Renner, B., M. Gamp, R. Schmalzle, and H.T. Schupp. 2015. Health Risk Perception. International Encyclopedia
of the Social & Behavioral Sciences, 2nd Ed. 702-709.
Sallee, J., S. West, and W. Fan (2016). Do Consumers Recognize the Value of Fuel Economy? Evidence from
Used Car Prices and Gasoline Price Fluctuations. Journal of Public Economics, 135: 61-73.
Shogren, J. and L. Taylor. 2008. On Behavioral-Environmental Economics. Review of Environmental Economics
and Policy, 2(1): 26-44.
Shogren, J., G. Parkhurst, and P. Banerjee. 2010. Two Cheers and a Qualm for Behavioral Environmental
Economics. Environmental and Resource Economics, 46(2): 235-247.
Traeger, Christian P. 2014. On Option Values in Environmental and Resource Economics. Resource and Energy
Economics, 37: 242-252.
5-37 Guidelines for Preparing Economic Analyses | 3rd edition
-------
U.S. EPA. 1985. Costs and Benefits of Reducing Lead in Gasoline: Final Regulatory Impact Analysis. EPA-230-
05-85-006.
U.S. EPA. 1997. Economic Analysis for the National Emission Standards for Hazardous Air Pollutants for
Source Category: Pulp and Paper Production; Effluent Limitations Guidelines, Pretreatment Standards, and
New Source Performance Standards: Pulp, Paper and Paperboard Category-Phase I. EPA-821-R-97-012.
U.S. EPA. 1998. An SAB Report: Technical Review of the Proposed TSCA Section 403 Regulation (Identification
of Dangerous Levels of Lead). EAP-SAB-EHC-99-004.
U.S. EPA. 2002. Environmental and Economic Benefit Analysis of Final Revisions to the National Pollutant
Discharge Elimination System Regulation and the Effluent Guidelines for Concentrated Animal Feeding
Operations. EPA-821-R-03-003.
U.S. EPA. 2005a Rule to Reduce Interstate Transport of Fine Particulate Matter and Ozone (Clean Air
Interstate Rule); Revisions to Acid Rain Program; Revisions to the NOx SIP Call; Final Rule. Federal Register,
70(91): 25162-25405.
U.S. EPA. 2005b. Standards of Performance from New and Existing Stationary Sources: Electric Utility Steam
Generating Units: Final Rule. Federal Register, 70(95): 28606-28700.
U.S. EPA. 2008. Lead Renovation, Repair, and Painting Program; Final Rule. Federal Register, 73(78): 21691-
21769.
U.S. EPA. 2009. Guidance on the Development, Evaluation, and Application of Environmental Models.
EPA/100/K-09/003.
U.S. EPA. 2014. Retrospective Study of the Costs of EPA Regulations: A Report of Four Case Studies. EPA-240-
F-14-001.
U.S. EPA. 2017. SAB Advice on the Use of Economy-Wide Models in Evaluating Social Costs, Benefits, and
Economic Impacts of Air Regulations. EPA-SAB-17-012.
U.S. EPA. 2019. Economic Analysis for the Navigable Waters Protection Rule: Definition of "Waters of the
United States." EPA-HQ-OW-2018-0149-0004.
von Winterfeldt, D., R.S. Farrow, R.S. John, J. Eyer, A.Z. Rose, and H. Rosoff. 2020. Assessing the Benefits and
Costs of Homeland Security Research: A Risk-Informed Methodology with Applications for the U.S. Coast
Guard. Risk Analysis, 40: 450-475.
Whittington, D. and D. MacRae Jr. 1986. The Issue of Standing in Cost-Benefit Analysis. Journal of Policy
Analysis and Management, 5(4): 665-682.
5-38 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 6 - Discounting Future
Benefits and Oosts
Discounting allows for economically consistent comparisons of benefits and
costs that occur in different time periods. In practice, it is accomplished by
multiplying changes in future consumption (including market and non-market
goods and services] by a discount factor. Discounting reflects that [1] people
prefer consumption today over consumption in the future, and [2] invested
capital is productive and provides greater consumption in the future. Properly
applied, discounting can tell us how much future benefits and costs are worth
today.
Social discounting is the main type of discounting discussed in this chapter.
This is discounting from the broad society-as-a-whole point of view embodied
in benefit-cost analysis [BCA]. Private discounting, on the other hand, is
discounting from the specific, limited perspective of private individuals or
firms. This distinction is important to maintain.
This chapter addresses discounting over the relatively near term, called
intragenerational discounting, as well as discounting over much longer time
horizons, or intergenerational discounting. Intragenerational (a.k.a.,
conventional] discounting applies to contexts that may have decades-long time
frames, but where the timeframe of analysis is within the lifetime of current
generations. Intergenerational discounting addresses very long time horizons
in which the discounted effects will impact generations to come.
This chapter focuses on the most important discounting issues for applied
policy analysis, beginning with practical, basic mechanics and methods for
discounting. It then turns to the theory and foundational logic for discounting
and the different approaches to estimating discount rates. The presentation of
the results should include the full stream of the benefits and costs over the
time horizon of analysis both without discounting and appropriately
discounted. Analysts should present results using both a consumption rate of
interest and, if appropriate, a sensitivity analysis reflecting the shadow price
of capital approach.1
1 This chapter summarizes some key aspects of the core literature on social discounting, but it is not a detailed
review of the vast and varied social discounting literature. Excellent sources for additional information are: hind
Guidelines for Preparing Economic Analyses | 3rd edition
-------
i ,\[i^ riRtliot5^ 1v>j Ilpfuivhig
The most common methods for discounting involve estimating either net present values or
annualized values.2 An alternative method is to estimate a net future value. Net present value,
annualization, and net future value are different ways to express and compare the costs and
benefits of a policy in a consistent manner. These three methods will be discussed below.
^ i i | i-rl V Mil-
The net present value (NPV) of a stream of benefits and costs in the future is the value that those
benefits and costs provide to society today. The NPV at time 0 (the year to which values are
discounted) of a projected stream of current and future benefits and costs is calculated by
multiplying the benefits and costs in each year by a time-dependent weight, or discount factor, d,
and aggregating all of the weighted values. This can be done by discounting the benefits and
subtracting the discounted costs, which is equivalent to discounting the net benefits over all n
years, [n is the number of years in the future until the last year of the time horizon of the analysis)
as shown in the following equation:
NPV = BO + dlBl +... + dn-lBn-1 + dnBn
- CO - dlCl -... - dn-lCri-1 - dnCn
= NBO + dlNBl + d2NB2 +
... + dn-lNBn-1 + dnNBn (1)
where
Bt are the benefits in year t,
Ct are the costs in year t, and
NBt are net benefits, the net difference between benefits and costs (Bt - Ct) in year t.
Alternatively, NPV can be calculated by estimating the present value (PV) of costs and the PV of
benefits separately and then subtract the PV of costs from the PV of benefits:
NPV = (B0 + E?=1 dtBt) - (C0 + Z?=1 dtCt) . (2)
In either case, the discounting weights, dt, are given by:
dt = rJLT7 (3)
1 (1+r)t v J
where r is the discount rate and t is the year.
As shown in equation (1), the benefits and costs should be discounted to the same year to
appropriately calculate net benefits. This is because both future benefits and costs should be
(1982a, b; 1990; 1994), Lyon (1990,1994), Pearce and Turner (1990), Pearce and Ulph (1994), Arrow etal.
(1996), Portney and Weyant (1999), Frederick etal. (2002), Moore et al. (2004), Spackman (2004), Groom etal.
(2005), Cairns (2006), Burgess and Zerbe (2011a), Moore etal. (2013a), Harberger and Jenkins (2015), Li and
Pizer (2021), and Newell et al. (2022,2024)
2 Note that discounting is distinct from inflation, although observed nominal market rates of return reflect
expected inflation. While most of the discussion in this chapter focuses on real discount rates and values, benefits
and costs should also be adjusted for inflation when relevant.
Guidelines for Preparing Economic Analyses \ 3rd edition
-------
evaluated from the perspective of the same year to provide them equal consideration.3 Also, as
discussed in Section 6.1.6.1, the same rate should be used to discount benefits and costs in a given
year. However, in some analyses with very long time horizons there may be reasons to use different
discount rates in different future years, as discussed in Section 6.3.
i i i i: juTiTiiTj ' i i ar versus End-of-Ye i I i .counting
In the NPV equation, Bo, Co, and NBo are the benefits, costs, and net benefits incurred immediately
(when t=0), so they are not multiplied by a discount factor. This makes sense when time is
continuous, but what is "immediate" becomes less clear when time, t, is an entire year. For example,
if a rule is finalized at the beginning of a year and costs and benefits will be realized throughout that
year, are these values "immediate" or should they be discounted one period? If costs and benefits
incurred throughout the year are considered immediate, then they would be Bo and Co in equation
(1) above. This is known as beginning-of-year discounting because all intra-annual effects in the
current year are treated as if they occur at the beginning of the year, when t=0. The alternative is to
treat all intra-year effects in the current year as if they occur at the end of the year, when t= 1, and
discount them back one period. Effects in the next year would then be discounted back two periods.
This is known as end-of-year discounting. The choice between beginning- or end-of-year
discounting does not generally have a large quantitative effect on the analysis. Whichever approach
is adopted should be explicitly stated and applied to both benefits and costs so that the analysis is
internally consistent.4
Tim- r rio 'f [.- >s than One Year
When estimating the NPV, it is important to explicitly state how time periods are designated and
when costs and benefits accrue within each time period. Typically, time periods are in years, but
alternative time periods can be justified if costs or benefits accrue at irregular or non-annual
intervals. To correctly discount intra-year effects, the annual discount rate, r, must be adjusted to an
"effective rate," r, which produces the same result as the annual discount rate if compounded for
one year. The effective discount rate for any non-annual period is:
ft = (1 + r)1 /(# of periods) _ ^ ^
For example, if the annual discount rate is 7% and costs are incurred on a quarterly basis (i.e., there
are four periods in a year), then the effective quarterly discount rate, r, is approximately 1.7%. The
formula for discounting weights, dt, given above, can be used with this effective rate, but t is
measured in quarters rather than years.
3 As discussed in Section 5.2, the analysis should identify the year to which benefits and costs are discounted and
the dollar year used to report them. It is important to identify and distinguish the reporting and dollar years of
the analysis when they differ.
4 Three common Excel functions used for discouning PMT and PV, and NPV - use end-of-year discounting by
default. The PMT and PV functions include a 0/1 "type" argument indicating if the discounting is done at the end
or the beginning of the year. The default is 0, and therefore needs to be changed to 1 to do beginning-of-year
discounting. The NPV function implicitly assumes end-of-year discounting. To use the NPV function to calculate
the net present value using beginning-of-year discounting, the solution to the NPV calculation must be multiplied
by the expression (1 +r), where r is the discount rate. Analyses that use the PMT, PV or NPV functions without
making these adjustments are implicitly assuming end-of-year discounting.
6-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
While the discounting formula can be adjusted to account for intra-annual discounting periods, it
may not be necessary unless exact values are required. The NPV generated by an intra-annual
effective discount rate, r, will be between the NPV using beginning-of-the-year discounting and the
NPV using end-of-the-year discounting using the annual discount rate, r. These NPVs don't usually
differ by much in a typical economic analysis.
mtinuous Discounting
Costs and benefits may also be discounted on a continual basis during the year. In this case, benefits
or costs occurring at the end of a future year (or period), t, are discounted by the weight:
dt = e~ft (5)
Where e is Euler's number, or 2.718, when rounded to three decimal places, and is the base of the
natural logarithm. This is a commonly used expression in economics and finance. Furthermore,
continuous discounting provides a convenient way to represent a discount weight for some
theoretical economic concepts related to discounting. Note equation (5) uses a discount rate
appropriate for continuous discounting, r. As with intra-annual discounting discussed above, the
effective discount rate, r, should produce the same result as the annual discount rate. The effective
discount rate for continuous discounting is:
f = In (1 + r) (6)
In this case, t= 1 represents one year, but the discounting weight is assumed to be applied to every
moment, continuously throughout the year.
6.1.2 Annualiz es
An annualized value is an illustrative cost or benefit that, if incurred every year over the entire time
horizon of the analysis, would produce the same net present value (NPV) as the original time-
varying stream of costs, benefits, or net benefits. In some cases, annualized values are easier to
understand than NPV.
Because the annualized value is constructed to generate the same net present value as the actual
stream of values, comparing annualized values is equivalent to comparing net present values. That
is, one can use either the NPV or the annualized values to determine whether benefits exceed costs
or which option produces the highest net benefits. As with NPV, benefits and costs may be
annualized separately and compared, or the stream of net benefits can be annualized.
The formulas below illustrate the estimation of annualized costs; the formulas are identical for
benefits.5 The exact equation for annualizing depends on whether there are any immediate costs
(i.e., any costs at time zero, t=0).
Annualized costs when there is no cost at t=0 (e.g., no Co in equation (1)) are estimated using the
equation:
Annualized Cost = PVC * , (7)
(l+r)(n+i)_l V- J
5 Variants of these formulas may be common in specific contexts. See, for example, the Equivalent Uniform
Annual Cost approach in the EPA's Air Pollution Control Cost Manual (U.S. EPA 2017).
6-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
where
Annualized Cost = annualized cost accrued at the end of each of n years,
PVC = present value of costs (calculated as in equation (1), above),
r = the discount rate per year, and
n = the length of the time horizon over which costs are annualized.
Annualized costs when there is initial cost at t=0 are estimated using a slightly different equation:
Annualized Cost = PVC * r*^1+^ [8]
(l+r)n-l L J
The annualization approach in equation (7) is generally consistent with end-of-year discounting because
the first cost value, Ci, is discounted one period. Equation (8) is more consistent with beginning-of-year
discounting because there is a value, Co, which is not discounted one period.6 Note that the
numerator expression is the same in both equations, although the PVC is calculated differently
depending upon whether there are costs at t=0. The only difference is the "n+1" and "n" terms in the
denominator of (7) and (8).
Some important caveats are associated with the use of annualized values. First, they are generally
illustrative; the annualized value is not the actual value that will manifest every year. Second, the
annualized value changes with the timeframe of the annualization. This means that the annualized
value will be different for each value of n, even for the same discount rate, r. The longer the
timeframe assumed for the annualization, the lower the annualized value. Third, annualization
formulas assume that the timeframe of n periods begins immediately. If the actual stream of costs
being annualized does not occur immediately, the timeframe for the annualized value and the actual
stream will not be the same.
One special case of equation (7), the annualization formula when there is no cost at t=0, is when
n=°°. In this case, the annualized cost is simply:
Annualized Cost = PVC * r (9)
For example, suppose an action permanently eliminates the use of an environmental amenity (e.g., a
wetland), and the estimated present value of that amenity is $1 million at a discount rate of 2%. The
cost of this policy is the lost value of the amenity in perpetuity -- the period of the analysis is
effectively infinity. The annualized cost of that policy - that is, the cost that if lost every year, forever,
would be equivalent to $1 million in present value today - is $1 million * 2% = $20,000 per year.
The corollary to equation (9) is:
Annualized Cost
Thus, if an environmental amenity is estimated to be worth $20,000 per year, its present value
using a %2 discount rate is $1 million, assuming that it provides benefits into perpetuity.
6 The default PMT function in Excel (with "type" equal to 0) will produce the same answer as equation (7).
Setting the "type" variable to 1 will produce the result from equation (8).
6-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
6.1.3 Net Futui IFV)
Instead of discounting all future values to the present using the NPV, it is possible to estimate the
stream of values from the perspective of some future year, for example, at the end of the last year of
the policy's effects, n. This would be the net future value (NFV). This might be particularly useful
when conducting a retrospective analysis.
The net future value for net benefits (/VBt) is estimated using the following equation:
NFV = aoNBo + aiNBi + C12NB2
+ ... + Cln-lNBn-l+ NBn (H)
Where, as before, NBt are net benefits, (Bt - Ct), in year t. This formula can also be used for either
benefits or costs alone.
In the NFV equation, the accumulation weights, at, are different from the discounting weights in
equation (3) used for NPV, and are given by:
ot=(l + r)CM (12)
where r is again the annual discount rate. The net future value for year n can be expressed in
relation to the net present value for t= 0, as follows:
NFV
NPV = [13]
(1+r)n
The NFV can be modified for intra-annual values using an effective discount rate described in the
NPV section above. It can also be calculated assuming continuous accumulation using the effective
discount rate in equation (6) and accumulation weights:
at = eft (14)
The only difference between equation (14) and equation (5) is the use of r rather than -r in the
exponent.
1" .-f mi paring the
NPV represents the value of a stream of costs and benefits from some point in time (often the
present moment) going forward. NFV represents the value of the stream of costs and benefits at
some future time. Annualization is the calculation of a constant, annual value for costs and benefits
that would produce the same NPV as the actual stream of costs and benefits.
Depending on the circumstances or application of the analysis, one method might have certain
advantages over the others. Discounting to the present to get an NPV is likely to be the most
informative for the standard economic analysis of a policy that will generate future benefits and
costs. NFV may be more appropriate for evaluating the cumulative impacts of regulation or when
conducting a retrospective analysis. The difference between the two is simply the choice of the
reporting or perspective year for the analysis. Annualized values may be used in conjunction with
the NPV to communicate the result or compare options when the costs or benefits are highly
variable over time. It is important to remember, however, that annualized values assume that the
annualization period begins immediately and that the results are sensitive to the annualization
period - the annualized value will be lower the longer the annualization period -- so analysts should
be aware of potentially different annualization periods when comparing annualized values from
one analysis to those from another.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
The choice of discount rate affects the values generated by these discounting methods. For a given
stream of net benefits, the NPV will be lower with higher discount rates, the NFV will be higher with
higher discount rates, and the annualized value may be either higher or lower depending on the
time at which impacts occur and the length of time over which the values are annualized. However,
the ranking of monetized net-benefits among regulatory alternatives is unchanged across these
three methods for any discount rate.
^ i * ' - isitr'ir <~f [-M* IV -1 n* '"In - li «nnrY- r * ill - Rate
Both the size and sign of NPV can be sensitive to the choice of discount rate when there is a
significant difference in the timing of costs and benefits. This is the case for policies that require
large initial outlays or have long delays before benefits are realized, as do many EPA policies. Text
Box 6.1 illustrates how discount rates affect NPV.
In other cases, the discount rate is not likely to affect the sign of the NPV estimate. Specifically, the
NPV will not be affected by the discount rate when:
All effects occur in the same period. In this case, discounting may be unnecessary or
superfluous because net benefits are positive or negative regardless of the discount rate
used.
Costs and benefits of a policy occur consistently over the period of the analysis, or their
relative values do not change over time.
In these cases, whether the NPV is positive does not depend on the discount rate, but the discount
rate can still affect how the present value compares to another policy.
i"138it[i- : in * [ [ tications
Several important analytic components need to be considered when discounting costs and benefits.
''.f/.f C insistent Use of th Pi »count Rate
The same discount rate must be used for both benefits and costs occuring in a given year, as the
discount rate reflects society's intertemporal preferences for trading off consumption over time
independent of the sign of the change in consumption. This is necessary for a consistent
comparison of net-benefits across policy alternatives and helps prevent discounting from being
used to justify a particular policy. A high discount rate reduces the weight given to costs and
benefits in the future and minimizes their impact on the NPV, whereas a low discount rate weights
future impacts more heavily and increases their impact on the NPV. Therefore, almost any policy
can be arbitrarily justified by using separate discount rates for benefits and costs.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 6.1 - Potential Effects of Discounting
To illustrate how different discount rates affect net present value, consider an example where
the benefits of a given program occur 30 years in the future and are valued (in real terms) at
$5 billion at that time. The rate at which the $5 billion future benefits is discounted can
dramatically alter the economic assessment of the policy: $5 billion 30 years in the future
discounted at 1% is worth $3.71 billion in the present, at 3% it is worth $2.06 billion, at 7% it
is worth $657 million and at 10% it is worth only $287 million. In this case, changing the
discount rate from 1% to 10% generates more than an order of magnitude of difference in the
present value of benefits. Longer time horizons will produce even more dramatic effects of
discounting on a policy's NPV. After 100 years, the present value of $5 billion is $260 million
at 3% and only $5.8 million at 7% (see Section 6.3 on intergenerational discounting).
Particularly in the case where costs are incurred in the present and therefore are not affected
by the discount rate, it is easy to see that the choice of the discount rate can determine
whether a policy has positive or negative net benefits.
6.1.6.2 Future Value of Environmental Effects and Uncertainty
There are two issues that are sometimes confounded with social discounting and the choice of
social discount rate, but should be treated separately: how the value of environmental impacts
change over time, and when future benefits and costs may be uncertain. While these issues are
important, particularly in an intergenerational context, they both should be addressed separately in
the economic analysis rather than adjusting the discount rate to account for them.7
First, the future value of environmental effects (i.e., their "current price" in future years) depends
on many factors, including the availability of substitutes and the level of wealth in the future. For
example, the relative price of environmental goods in the future will rise if those environmental
goods become scarcer over time. These changes in relative prices should be applied to future effects
and the associated values discounted, but the discount rate should not be adjusted to incorporate a
change in relative prices.
Second, uncertainty or risk in future benefits and costs resulting from the policy should not be
incorporated into the social discount rate. While it is technically possible to adjust the discount rate
to account for uncertainty, doing so may hide important assumptions and information about the
relative effects of discounting and uncertainty from decision-makers. Uncertainty about future
values should be treated separately when discounting. However, uncertainty about the discount
rate itself is different from uncertainty about future benefits and costs and can affect discounting as
discussed in Section 6.3.3.
6.1.6.3 Placing Effects in Time
Placing effects properly in time is essential for all calculations involving discounting. As discussed
in Section 5.4, analyses should account for implementation schedules and the resulting changes in
emissions or environmental quality, including possible changes in behavior that occur between the
announcement of policy and compliance deadlines. Additionally, a lag may occur between changes
7 See, for example, Moore et al 2017.
6-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
in environmental quality and the corresponding change in welfare. It is the change in welfare which
defines economic value, and not the change in environmental quality itself. The EPA's Science
Advisory Board addressed this issue (U.S. EPA 2001a) for the 2001 Arsenic Rule (U.S. EPA 2001b).
If exposure to arsenic in drinking water is reduced, the number of cancer cases is expected to
decline over time to a lower, steady-state level. How fast this reduction in risk occurs depends on
the "cessation-lag" following reduction in exposure. Whenever values are estimated for future
periods, the analysis should also report those values discounted to the present to allow for a proper
comparison across periods and avoid a potential misunderstanding regarding the magnitude of a
future period's benefits and/or costs relative to those in the current period, all else equal.
i i h" - rio -'f th1-' iialysis
As described in Section 5.4, the guiding principle is that the time horizon should be sufficient to
capture all of the welfare effects from policy alternatives, subject to available resources. This
principle is based on the requirement that BCAs reflect the welfare outcomes of those affected by
the policy. A complete BCA accounts for all welfare changes over the entire time period that an
action is expected to yield benefits and costs.
Analysts should avoid presenting the net benefits or effects of a regulation for a single snapshot
period (e.g., a year), as it is likely incomplete and, therefore, cannot be used to draw clear
conclusions about the overall impact of the regulation. For example, consider the case where a
regulation requires capital expenditures in the first year and has no subsequent costs, but benefits
are realized over many years following the initial investment Presenting the benefits and costs of
the rule in only the first year or only one of the subsequent years would misrepresent the
regulation's expected net benefits.
Previously, n was defined as the final period in which the policy is expected to have impacts. While
a complete BCA should account for impacts expected to occur in all future years, it may be
impractical to do so. One solution is to analyze a time horizon that ends in T
-------
Statutory or other requirements for the policy or the analysis; or
The extent to which benefits and costs are allocated to future generations.
Section 5.4 elaborates on how these first two bullets may influence the time horizon of analysis,
while Section 6.3 elaborates on the third.8
The choice of time horizon for the analysis should be clearly explained and well-documented, and
the analysis should highlight the extent to which the sign of net benefits, or the relative rankings of
policy alternatives, are sensitive to the choice of time horizon. If annualized values are reported
then both annualized benefits and costs should be be reported and the time horizon over which
benefits and costs are annualized should be the same and clearly documented. Furthermore, an
annual value of benefits (or costs) should not be compared to an annualized value of costs (or
benefits) because, as discussed in Section 6.1.2, annual and annualized values represent different
time scales of analysis.
icounting Non-Monetized Effects
A common criticism of discounting for environmental policies is that health impacts such as "lives
saved" or physical impacts such as "improved water quality" are not like money flows. They cannot
be deposited in a bank and withdrawn after earning interest. This criticism does not appreciate that
the valuation approaches are designed to estimate the amount of money that is as valuable to
individuals as the environmental or health effects being examined. If all environmental and health
impacts have been appropriately valued (monetized), then those money-equivalent flows can be
discounted like real money flows over time.
However, some effects cannot always be monetized. In this case, the undiscounted stream of the
non-monetized effects should be presented as they occur over time. As a general matter, these non-
monetized effects should still be discounted in benefit-cost analysis and cost-effectiveness analysis
if they are aggregated over time. This is because they are assumed to hold some value, albeit
unspecified, and discounting assumes that individuals prefer the benefit of that value today over
the future. This is the usual practice in cost-effectiveness analysis (Section 7.5.2.1), where
monetized costs and non-monetized effectiveness measures are both discounted. OMB Circular A-4
(2023) recommends discounting non-monetized health effects.
For some effects, however, the (unknown) marginal value of a change in the non-monetized effect
might be dependent upon the level and timing of that change. That is, marginal values are not
constant over time. For example, suppose there are annual emissions thresholds below which
environmental effects are negligible, but above which lead to major environmental damages. The
economic value of emissions depends upon whether those emissions are above or below this
threshold, and discounting these economic values would be appropriate. If we lack these values,
however, and discount the effects themselves, we are treating all changes as if they had the same
value. Here, it would be preferable to display the undiscounted stream of non-monetized effects
with an appropriate justification and explanation.
,\[i^ i>«f I Pi^^.Fsntm
The goal of social discounting is to compare benefits and costs that occur at different times based
on the rate at which society is willing to make such trade-offs. The analytical and ethical foundation
8 Section 8.2.3.1 provides additional guidance on the appropriate time horizon of analysis specific to accounting
for costs of compliance.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
of the social discounting literature rests on the traditional test of a potential Pareto improvement
in social welfare, whereby those who, on net, benefit from a policy could potentially compensate
those who, on net, experience costs, such that everyone is at least as well off as they were before
(see Chapter 1 and Appendix A). This framework casts the consequences of government policies in
terms of individuals contemplating changes in their own consumption over time.9 In this context,
trade-offs (benefits vs. costs) reflect the preferences of those affected by the policy, and the time
dimension of those trade-offs should reflect the intertemporal preferences of those affected. Thus,
social discounting should seek to mimic the discounting practices of the affected individuals.
Simultaneously, social discounting must reflect social trade-offs in consumption over time, which
may differ from trade-offs from a private, individual perspective.
The literature on discounting often uses a variety of terms to describe identical or very similar key
concepts. For the purposes of the Guidelines, the following fundamental concepts are used in
defining a social discount rate:
The social rate of time preference is the discount rate at which society is willing to trade
consumption in one period (usually year) for consumption in the next period.
Consumption rate of interest is the rate at which an individual is willing to trade
consumption in one period for consumption in the next period. This rate reflects the
individual's rate of time preference and, following the potential Pareto principle, the social
rate of time preference should be based on this individual rate.
The social opportunity cost of capital is the consumption allowed in the next period due
to private investment in the prior period. This is the rate at which society can trade
consumption over time due to productive capital. Benefits and costs should account for
future consumption changes due to changes in private investment
Market interest rates are what we observe in markets for loanable funds. There are
several real market interest rates which, to varying extents and accounting for tax
distortions, can be taken as estimates for the individual rates of time preferences and the
social opportunity cost of capital.
Social discounting is primarily concerned with the relationships among these concepts and how
they are measured.
6.2,1 Consumptirii R/ir- inr-i - r id Social Opportvr,ir
Capital
If capital markets were perfect and complete with no distortions or uncertainties, the market
interest rate would equal both the consumption rate of interest and the social opportunity cost of
capital since it reflects both how individuals value present versus future consumption and how
productive capital can be transformed into future consumption. Following the potential Pareto
principle and valuing future costs and benefits in the same way as the affected individuals, this
market rate would be the appropriate social discount rate.
However, perfect and complete markets do not exist Private sector returns are taxed (often at
multiple levels), capital markets are not perfect, and capital investments often involve private (and
not necessarily social) risks. These factors cause a divergence in the consumption rate of interest
9 The term consumption is broadly defined to include both the use of both private and public goods and services
by households in BCA and includes the intergenerational nature of this change in consumption.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
and the social opportunity cost of capital. That is, there is a divergence between the rates at which
individuals and society can trade consumption over time. Text Box 6.2 illustrates how these rates
can differ.
A large body of economic literature analyzes the implications for social discounting of divergences
between the consumption rate of interest and the social opportunity cost of capital. Most of this
literature is based on the evaluation of public projects, but many of the insights still apply to
regulatory BCA, and the dominant approaches from the literature are briefly outlined here. More
complete recent reviews can be found in Spackman (2004), Burgess and Zerbe (2011a), Moore etal.
(2013a, 2013b), and Harberger and Jenkins (2015). Section 6.2.2 discusses social discounting using
the consumption rate of interest as the social rate of time preference, whereas Sections 6.2.3 and
6.2.4 discuss methods for discounting when investment changes.
6.2,2 Soci *1 R't- <~f Tm)'~ Fi ierence as the Soci *1 L'iscc-.M iv R.'te
If costs and benefits can be represented as changes in consumption profiles over time, then
discounting should be based on the rate at which society is willing to postpone consumption today
for consumption in the future. Thus, the rate at which society is willing to trade current for future
consumption, or the social rate of time preference, is the appropriate discounting concept for
evaluating public policy decisions.
But the social rate of time preference differs from individual rates of time preference. An individual
rate of time preference includes factors such as the probability of death, whereas society can be
presumed to have a longer planning horizon. Additionally, individuals are routinely observed to
have several different types of savings, each possibly yielding different returns, while
simultaneously borrowing at different market interest rates. For these and other reasons, the social
rate of time preference is not directly observable and may not equal any particular market interest
rate. Generally, there are two primary approaches to deriving the social rate of time preference.
6,2,1: f i'f tim^tin^ * > vi >1 I'm- i Time Preferen - J' ing Risk ¦ i - - w sets
One common approach to estimate the social rate of time preference is to use the market rate of
interest from long-term, risk-free assets such as government bonds. The rationale behind this
approach is that this market rate reflects how individuals discount future consumption, and the
government should value policy-related consumption changes as individuals do. In this approach,
the social discount rate should equal the consumption rate of interest found in the market.
In principle, estimates of the consumption rate of interest could be based on after-tax interest rates
consumers face for either saving (i.e., lending) or borrowing. Because individuals have different
marginal tax brackets, different levels of assets, and different opportunities to borrow and invest, the
type of market interest rate that best reflects the consumption rate of interest will differ among
individuals. However, the fact that, on net, individuals generally accumulate assets over their working
lives suggests that the after-tax returns on savings instruments available to the public will provide a
reasonable estimate of the consumption rate of interest for society.
The historical rate of return on long-term government bonds, after-tax and in real terms, is a useful
measure as it is relatively risk-free, maintaining the distinction between risk and social discounting
described in Section 6.1.6. Also, as long-term instruments, they provide more information on how
individuals value future benefits over time frames more relevant for environmental policy analysis.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 6.2 - Social and Consumption Rates of Interest
The following example illustrates how the return on private sector investments may differ from
the consumption rate of interest Suppose a private sector investment for one period is returned
as consumption, the real pre-tax market rate of return on those investments is 5% and that
taxes on capital income amount to 40% of the rate of return. In this case, the private investment
yields a 5% return, 2% is paid in taxes to the government and individuals receive the remaining
3%. From a social perspective, current consumption - if it were instead invested in capital - can
be traded for future consumption at a rate of 5%, with 3% going to individuals and 2% going to
the government But from the individuals' perspective, they are effectively trading consumption
through time at a rate of 3%. Therefore, the consumption rate of interest is 3% and the social
rate of return on private sector investments (also known as the social opportunity cost of
capital) is 5%.
6.2.2.2 Estimating a Social Rate of Time Preference Using the Ramsey
Framework
A second option is to construct the consumption rate of interest as the social rate of time preference
in a framework attributed to Ramsey (1928), which explicitly reflects: (1) preferences for utility in
one period relative to utility in a later period; and (2) the value of additional consumption as
income changes. These factors are combined in the equation:
r = p + rjg (16)
where
r = the consumption rate of interest,
p = the pure rate of time preference,
rj = the elasticity of marginal utility with respect to consumption, and
g = the consumption growth rate.
The pure rate of time preference, p, is the rate at which the representative individual discounts
utility in future periods due to a preference for utility sooner rather than later. The elasticity of
marginal utility with respect to consumption, r], defines the rate at which the well-being from an
additional dollar of consumption declines with the total level of consumption. The consumption
growth vate,g, defines how consumption is expected to grow over time. For example, it may be
expected to increase because incomes are expected to increase over time. Estimating a social rate of
time preference in this framework requires information on each of these arguments. While rj and g
can be derived from data, p is unobservable and must be assumed or calibrated.10 Text Box 6.3
provides a more detailed discussion of the Ramsey equation, and Section 6.3.1 discusses using the
Ramsey framework to guide intergenerational discounting.
10 The Science Advisory Board defined discounting based on a Ramsey equation as the "demand-side" approach,
noting that the value judgments required for the pure rate of time preference make it an inherently subjective
concept (U.S. EPA 2004). However, recent research has developed methodologies to calibrate the pure rate of
time preference using a descriptive approach (Newell et ah, 2022).
6-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 6.3 - The Ramsey Discounting Framework
The Ramsey discounting framework provides an intuitive approach to thinking about, and
potentially calibrating, the social discount rate. It can be derived by considering a representative
individual with utility u(ct) in period t, where ct denotes consumption. The agent is assumed to
make choices to maximize lifetime welfare, / e~ptu(ct)dt, where p is the pure rate of time
preference (i.e., the rate at which the agent discounts utility] and e~pt is the discount factor.
Suppose the agent is considering a one period investment of one dollar in consumption at time t
for additional consumption at time t + 1. The minimum investment rate of return, r, required
for the individual to find the investment desirable is defined by the equation:
= e~P (erJ^L.\ = er-pJ±_ (6.3.11
That is, to be worthwhile the increased utility of consumption in the second period, \ er
discounted back at the pure rate of time preference, p, must be at least equal to the forgone
utility of consumption forgone in the first period to fund the investment, which could be induced
by a regulatory action. The rate r defines the additional return, beyond recovering the initial
investment, required for the agent to be just as well off as before. So, r represents the discount
rate appropriate for comparing a future change in consumption with a change in present
consumption.
If it is assumed, as is common, that the utility function has an iso-elastic form, such that u(ct) =
c1'11
-j^, where r] is the absolute value of the elasticity of margi nal utility, the Ramsey formula can be
recovered. Substituting this utility function into equation (6.3.1), taking the natural log of both
sides of the equation, applying the relationships ln(a) ln(b) = inia/b) and ln(am) = m ln(a),
and solving for r produces:
r = p + r]ln - = p + r>g (6.3.2)
where g is the rate of growth of consumption between t and t + 1.
This definition highlights two reasons that future changes in consumption should be discounted
(as described in section 6.2.2.2).
1. The pure rate of time preference, p, captures the general preference by individuals for
utlity sooner rather than later and measures the rate at which individuals discount their
own future utility.
2. The term, rjg represents that a marginal change in consumption in the future may not
have the same value as a marginal change in consumption today. For example, if
consumption increases over time, the marginal utility of consumption will decrease over
time, implying that a marginal change in future consumption is valued less (and
discounted more) than a marginal change in current consumption.
As shown by Ramsey (1928), in an economy with no taxes, market failures or other distortions,
the social discount rate r, as defined in equation (6.3.2), would be expected to equal the market
interest rate. The market interest rate, in turn, would be equal to the social rate of return on
private investments and the consumption rate of interest. However, distortions and market
failures cause these rates to diverge in practice. As such, r represents the consumption rate of
interest
6-14 Guidelines for Preparing Economic Analyses | 3rd edition
-------
6.2.3 Social Opport ¦< r vi" ipital as the Social Disc*
The social opportunity cost of capital recognizes that the social return to private investments may
exceed the private returns. Therefore, if funding for government projects or capital investments
required to comply with government regulations displace total private investment in the economy
the opportunity cost of those forgone investments may exceed the private returns. In other words,
if a regulation displaces private investments, society will lose the total returns from those forgone
investments, including the tax revenues generated.
Private capital investments might be displaced if public projects are financed with government
debt and government borrowing crowds out private investment. In a regulatory context, private
investment might be displaced if regulated firms cannot pass through capital expenses to
households and the supply of investment capital is relatively fixed. In these cases, demand
pressure in the investment market will tend to raise market interest rates and reduce private
investments that would otherwise have been made.11 A BCA should account for the full social cost
of any declines in private capital investments due the policy being evaluated (and similarly the
social benefits of any increase in private capital investments induced by the policy), as appropriate.
In principle, the social opportunity cost of capital can be estimated by a pre-tax, marginal, risk-free
rate of return on private investments, but this rate is not observed in the marketplace. As a result,
these values are sometimes derived by using National Accounts data to estimate rates of return on
reproducible capital (e.g., Burgess and Zerbe 2011b; Harberger and Jenkins 2015), though there are
some differences in the exact accounts included and their relative weights across these analyses. In
practice, average returns that are likely to be higher than the marginal returns are typically
observed, given that firms will make the most profitable investments first This leads to uncertainty
as to how marginal returns can be estimated. Observed rates also reflect an unknown risk premium
faced in the private sector, which causes them to be higher than a risk-free rate.
In very specific circumstances using the social opportunity cost of capital as the social discount rate
in a BCA for an environmental policy can account for the social costs of displaced capital
investments. In particular, it requires that the policy costs fully crowd out private investments.12
Harberger (1972) recognized this is unlikely to be the case and derived a generalized version of this
approach, assuming that policies displace a mix of consumption and investment In this case, the
social discount rate is a weighted sum of the net pre-tax marginal return to capital (i.e., the social
opportunity cost of capital) and the after-tax marginal return to capital (i.e., the consumption rate
of interest). Sandmo and Dreze (1971), Dreze (1974), and Burgess (1988) extended this approach
to include the marginal cost of foreign financing in an open economy. In practice, this weighted sum
is likely to be closer to the consumption rate of interest than the opportunity cost capital for a
number of reasons. First, the United States is a large, open economy with a high capital mobility.
11 Another justification for using the social opportunity cost of capital argues that the government should not
invest (or compel investment through its policies) in any project that offers a rate of return less than the social
rate of return on private investments. While it is true that social welfare will be improved if the government
invests in projects that have higher values rather than lower ones, it does not follow that rates of return offered
by these alternative projects define the level of the social discount rate. If individuals discount future benefits
using the consumption rate of interest, the correct way to describe a project with a rate of return greater than
the consumption rate is to say that it offers substantial present value net benefits.
12 The terms "displacement" and "crowding out" refers to how total private investment in the economy is
reduced due to new investment in response to the environmental policy. That is, how compliance costs in
response to the policy displace investment that would have occurred without the policy. An environmental policy
has fully crowded out private investment if private investment is reduced by the full compliance cost of policy.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Most regulatory costs are not expected to result in a substantial displacement of capital investment,
which can be funded through an increase in financing by foreign lenders. Second, the benefits of
regulation could induce capital investment (e.g., by increasing productivity or reducing
depreciation), which is unaccounted for in the social opportunity cost of capital approach.13 As
such, the shadow price of capital approach is the analytically preferred approach to account for the
full social cost of any changes in private capital investment expected in response to the policy being
analyzed.
6.2,4 Shade ipital Approach
As noted above, because capital markets are taxed and experience other market distortions, the
consumption rate of interest and the social opportunity cost of capital are not equal. This means
that while individuals are indifferent between consumption and the returns to risk-free private
investment on the margin, society is not. The shadow price of capital approach accounts for this by
adjusting the costs and benefits that affect investment into equivalent consumption impacts (i.e.,
their shadow values) that reflect the social value of altered private investments.14 All impactsthe
costs and benefits that affect consumption and the shadow costs and benefits of affected
investmentsare then discounted using the social rate of time preference that represents how
society trades and values consumption over time.15 Many sources recognize this method as the
preferred analytic approach to social discounting for public projects and policies.16
The shadow price (or social value) of private capital investment captures the perspective that a unit
of private capital produces a stream of social returns at a rate greater than that at which individuals
discount them due to distortions in the capital market noted in the discussion of the social
opportunity cost of capital. This is because a capital investment produces a rate of return for its
owners equal to the consumption rate of interest (post-tax) plus a stream of tax revenues for the
government (generally considered to be used for consumption). Text Box 6.4 illustrates this idea of
the shadow price of capital.
13 This approach has been used by the Federal government for many years and was recommended in previous
EPA Guidelines (2016) and OMB Circular A-4 (2003), but it is technically incorrect and can produce NPV results
substantially different from the shadow price of capital approach. For an example of these potential differences,
see Spackman (2004).
14 A "shadow price" can be viewed as a good's true opportunity cost, which may not equal the market price.
Adjusting the cost and benefits of investment to reflect their consumption equivalent impact is, essentially,
reporting their shadow values, hind (1982a) remains the seminal source for this approach in the social
discounting literature.
15 Because the consumption rate of interest is often used as a proxy for the social rate of time preference, this
method is sometimes known as the "consumption rate of interest - shadow price of capital" approach. However,
as hind (1982b) notes, what is really needed is the social rate of time preference, so more general terminology is
used. Discounting based on the shadow price of capital is referred to as a "supply side" approach by the EPA's
Science Advisory Board (U.S. EPA 2004).
16 See OMB Circular A-4 (2023), Freeman (2003) and the report of the EPA's Advisory Council on Clean Air
Compliance Analysis (U.S. EPA 2004).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 6.4 - Calculating and Applying the Shadow Price of Capital
A highly stylized example illustrates the shadow price of capital concept Suppose that the real
pre-tax annual rate of return on private investments (i.e., the social opportunity cost of capital] is
3.5% and the post-tax consumption rate of interest is 2%. Under these conditions, $1 in private
investment will produce a stream of private consumption of $.02 per year, and tax revenues of
$.015 per year. Further assume that the $1 investment does not depreciate (i.e., it exists in
perpetuity), the annual $0.02 net-of-tax earnings from this investment are consumed each year
and the $.015 annual tax-revenue is also used for consumption in each year. The present value of
the perpetual stream of constant investment income to individuals is the stream of income divided
by the discount rate (Equation (10]), $0.02 / 2% = $1. The present value of the $0,015 per year
stream of tax revenues discounted at 2% is $0,015/0.02 = $0.75. This is the present value of the
additional benefits to society (via the government). Thus, the full social value of this $1 private
investment - the shadow price of capital - is $1,75, greater than the $1 private value that
individuals place on it
This example is a highly stylized case where changes in the productive capital stock persist in
perpetuity and all income from capital assets funds only consumption. A more complete derivation
of the shadow price of capital, as given by Li and Pizer (2021), takes into account depreciation and
the savings rate (i.e., the rate at which individuals invest income):
(1 savings rate)(gross rate of return on capital)
Shadow price of capital = - r-z -r-
[consumption rate of interest + depreciation rate (savings rate)(gross rate of return on capital)\
The gross rate of return on capital is the net rate of return on investments before depreciation (i.e.,
the social opportunity cost of capital plus the depreciation rate). Maintaining the assumptions of a
3.5% social opportunity cost of capital and a 2% consumption rate of interest, and assuming a
depreciation rate of 10% and an equilibrium savings rate of 25% would yield an estimate of 1.17
for the shadow price of capital.
To apply the shadow price of capital estimate in a BCA, one needs additional information about
how much investment is displaced and induced. For example, assume a large public project is
financed with 75% as additional government debt and 25% through increased taxes. Further
supposed that the increase in government debt displaces an equal amount of private investments
and the increase in taxes reduces individuals' current consumption by an equal amount
The shadow price of capital approach would be applied to the cost estimate in the following steps:
1. Separate the costs that displace capital investment from the costs that displace
consumption.
a. $0.75 of every $1 in costs financed through debt displaces investment
b. $0.25 of every $1 in costs financed through taxes displaces consumption.
2. Apply the shadow price of capital (1.17 from the example above) to the $0.75 of costs that
displace private investment. This yields a shadow cost of $0.88 which accounts for the
impact of these costs on investments.
3. Add to this the remaining current cost ($.25) that displaces current consumption, which is
not adjusted for the shadow price of capital.
4. The total social cost of this public project is $1.13 for every $1 spent
The same steps should be followed for the benefits estimate, separating the benefits that induce
capital investment from those that directly increase consumption, to determine the total social
benefits. The total social cost would then be compared to the social benefits of the project
6-17 Guidelines for Preparing Economic Analyses | 3rd edition
-------
When compliance with environmental policies displaces private capital investments (e.g.,
machinery and equipment), the shadow price of capital approach adjusts any capital-displacing
project or policy cost upward by the shadow price of capital (i.e., the effect of displacing capital on
consumption society-wide). This calculation effectively converts changes in private investment into
consumption equivalents, such that all costs and benefits can then be discounted using a social
discount rate equal to the consumption rate of interest. The most complete frameworks for the
shadow price of capital also recognize that while the costs of regulation might displace private
capital, the benefits could induce additional private investments in capital. In principle, a complete
analysis using the shadow price of capital would treat capital adjustments from costs and benefits
in the same fashion.
Policies analyzed in a general equilibrium framework (Chapter 8) will implicitly apply a shadow
price of capital approach. In the case of partial equilibrium analyses, additional steps are necessary
to apply the shadow price of capital approach. The first step is determining whether a policy will
alter private investment flows. Next, the altered private investment flows (positive and negative)
are multiplied by the shadow price of capital to convert them into consumption-equivalent units.
All flows of consumption and consumption equivalents are then discounted using the consumption
rate of interest A simple illustration of this method applied to the costs of a public project is shown
in Text Box 6.4.
timating the Shadow Price of Capital
While the shadow price of capital approach provides a theoretically sound method for measuring
the impact of changes in private capital investment, it is challenging to implement in practice. The
Li and Pizer (2021) specification described in Text Box 6.4, requires estimates of the social rate of
time preference, the social opportunity cost of capital, a depreciation rate, a savings rate, and, in
particular, the extent to which regulatory costs displace private capital investment and benefits
stimulate private capital investment. The first two components can be estimated as described
earlier, and the depreciation rate and savings rate can be estimated from empirical data, but
information on how regulation affects capital formation is more difficult to obtain, making the
approach difficult to implement.17
How policies affect capital investment depends on whether the economy is assumed to be open or
closed to trade and capital flows, and on the magnitude of the policy intervention relative to the
flow of investment capital from abroad. Some argue that early analyses implicitly assumed that
capital flows into the nation were either nonexistent or very insensitive to market interest rates,
known as the "closed economy" assumption.18 However, if an economy has highly mobile capital
flows, including from international sources, that are sensitive to market interest rate changes (the
"open economy" assumption), then total investment in private capital is likely to be less sensitive to
regulatory policy interventions, and there will be little, if any, crowding out19 If there is no
17 In addition to Li and Pizer (2021), Lyon (1990) and Moore etal. (2004) provide reviews of how to calculate
the shadow price of capital and possible settings for the various parameters that determine its magnitude.
Boardman et al. (2011) contains a textbook explanation as well as empirical examples. Depending on the
magnitudes of the various factors, shadow prices from 1 to infinity can result according to Lyon (1990), but the
ratio of the social opportunity cost of capital to the social rate of time preference is an upper bound in the Li and
Pizer (2021) specification.
18 See Lind (1990) for this revision of the shadow price of capital approach.
19 See, for example, Warnock and Warnock (2009).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
crowding out of private investment, then no adjustments using the shadow price of capital are
necessary; benefits and costs should be discounted using the consumption rate of interest alone.
The economic literature is not conclusive on the degree of crowding out and there is limited
empirical evidence of a relationship between the nature and size of projects and capital
displacement This presents challenges to implementing the shadow price of capital approach
outside of a general equilibrium framework.
>?. C."7 [r-k; itir Nr -111 ive Soci L*i icof.mt* R.-n- Estimates
The empirical literature for choosing a social discount rate focuses on estimating the consumption
rate of interest at which individuals trade off consumption through time. Historical real rates of
return on "safe" assets (post-tax), such as U.S. Treasury securities, are normally used to estimate the
consumption rate of interest Some studies and reports have found government borrowing rates range
between 1.5-4%, with long-term interest rates declining for the last two decades.20 Other studies have
expanded this portfolio to include other bonds, stocks and even housing. This generally raises the
range of rates slightly. It should be noted that these rates are realized rates of return, not anticipated,
and they are somewhat sensitive to the choice of time period and the class of assets considered.21
Other economists have constructed a social discount rate by estimating the individual parameters
in the Ramsey equation. These estimates necessarily require judgments about the pure rate of time
preference. Moore et al. (2013a) and Boardman etal. (2011) estimate the social discount rate to be
3.5% under this approach. The Ramsey equation has been used more frequently for
intergenerational discounting, which is addressed in the next section.
Using the social opportunity cost of capital as the social discount rate requires a situation where
private investment is fully crowded out by the costs of environmental policies. This is an unlikely
outcome, but it can be useful for sensitivity analysis and special cases. Estimates of the social
opportunity cost of capital typically range from 4.5% to 8%, depending upon the type of data
used.22
20 Newell and Pizer (2003) find a 200-year average (1798-1999) rate of 4% for long-term (30-year) U.S.
government bonds. According to the U.S. Congressional Budget Office (CBO) (2005), funds continuously
reinvested in 10-year U.S. Treasury notes from 1789 to 2004 would have earned an average inflation-adjusted
return of slightly more than 3% a year. OMB (2003) reported a 30-year average (1973-2002) pre-tax rate for
10-year U.S. Treasury notes of 3.1%. U.S. CBO (2016) estimated that the average real rate for 10-year Treasury
notes was 2.9% between 1990 and 2007. Boardman etal. (2011) suggests 2.71% as the 1953-2001 average real
rate of return on 10-year U.S. Treasury notes. However, the Council of Economic Advisers (CEA 2017) notes a
decades-long downward trend in real rate of return for U.S. Treasury notes. Bauer and Rudebusch (2020,2023)
found that the decline in real interest rates refects a reduction in the equilibrium real interest rate, suggesting
that lower real interest rates are expected to persist. OMB Circular A-4 (2023) states that the more recent 30-
year average (2003-2022) rate for 10-year Treasury marketable securities was 2.0%. U.S. EPA (2023) reported a
1991 -2020 average real rate of return on 10-year Treasury securities of 1.5 to 2.0%, based on the infation
measure used. U.S. CBO (2023) projects a real interest rate on 10-year Treasury notes of 1.5% in 2033, rising to
2.2% by 2053.
21 Ibbotson and Sinquefeld (1984 and annual updates) provide historical rates of return for various assets and
for different holding periods.
22 OMB (2003) estimated a real, pre-tax opportunity cost of capital of 7%. Harberger and jenkins (2015)
estimate an average rate of 8% for "advanced countries."Burgess andZerbe (2011a) estimate a rate of 6% to
Guidelines for Preparing Economic Analyses | 3rd edition
-------
The utility of the shadow price of capital approach hinges on the magnitude of altered capital flows
from the environmental policy. If the policy will substantially displace or induce private investment,
then a shadow price of capital adjustment is necessary before discounting consumption and
consumption equivalents using the consumption rate of interest. Estimates of the shadow price of
capital in the academic literature range from 1.1 to 2.2 (Boardman et al. 2011, Moore et al. 2013a, Li
and Pizer 2021). The economic literature does not provide clear guidance on the likely scale of this
displacement, but it has been suggested that if a policy is relatively small and capital markets fit an
"open economy" model, there is probably little displaced investment.23 Changes in yearly U.S.
government borrowing during the past several decades have been in the many billions of dollars. It
may be reasonable to conclude that EPA programs and policies costing a fraction of these amounts
will not likely result in significant crowding out of U.S. private investments. For these reasons, some
argue that for most environmental regulations, it is sufficient to discount costs and benefits with an
estimate of the consumption rate of interest with sensitivity analysis as appropriate.24
; 11 u ¦ i ¦ 11 ¦ i u Pi^^.Fsntm
Policies designed to address long-term environmental problems such as global climate change,
radioactive waste disposal, groundwater pollution, or biodiversity present unique challenges
because they can involve significant economic effects across generations. Often, costs are imposed
mainly on the current generation to achieve benefits that will accrue primarily to unborn, future
generations. Discounting in this context is generally referred to as intergenerational discounting.
This section discusses the main issues associated with intergenerational social discounting using
the Ramsey discounting framework as a convenient structure for considering how the
"conventional" discounting procedures might need to be modified for policy analysis with very
long, multi-generational time horizons. This discussion presents alternative modeling approaches
to estimate the term structure, or the sequence of discount rates over time, along with important
caveats when using these approaches.
Intergenerational discounting is complicated by at least three factors: (1) the "investment horizon"
is longer than what is reflected in observed market interest rates representative of intertemporal
consumption trade-offs made by the current generation; (2) intergenerational investment horizons
involve greater uncertainty than intragenerational time horizons; and (3) future generations
without a voice in the current policy process are affected. These complications limit the utility of
using observed market rates to evaluate long-term public investments. The leading alternative is to
use model-based approaches to forecast a discount rate representative of expected household
preferences. These models suggest using a social discount rate lower than one based on recently
observed market rates and conditions, especially when uncertainty over the future state of the
world is taken into consideration.
The problem of comparing benefits borne by future generations to costs experienced by the current
generation involves both economic and ethical questions. Therefore, the normative choice of how a
8%, and Moore et al. (2013b) estimate a rate of approximately 5% using the same model but with different
inputs. Using an approach similar to OMB (2003), CEA (2017) estimated real rates of return to capital to be
around 7% based on National Accounts data but noted that approach may be subject to measurement error
leading to an overestimate.
23 hind (1990) first suggested this.
24 See Lesser and Zerbe (1994), Moore et al. (2004), and OMB (2023).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
decision maker should weigh the welfare of present and future generations, along with the
preferences of the current generation regarding future generations, cannot be made on economic
grounds alone. Nevertheless, economics offers important insights concerning intergenerational
discounting, the implications and consequences of alternative discounting methods, and the
systematic consideration of uncertainty.
^ , i ill - hf ''iTr -5 newcii in n Intergenerational Context
The Ramsey framework introduced in Section 6.2.2 is one of the most commonly used approaches
for modeling consumption discount rates.25 It is based on fundamental economic theory and
provides an intuitive organizing framework for thinking about consumption discount rates over
long time horizons. If per capita consumption grows over time as it has since the Industrial
Revolution (Valdes 1999) then future generations will be richer than the current generation. Due
to the diminishing marginal utility of consumption, increases in consumption will be valued less in
future periods than they are today. In a growing economy, changes in future consumption would be
given a lower weight (i.e., discounted at a higher rate) than changes in present consumption in the
Ramsey framework, even setting aside discounting due to the pure rate of time preference, p.
This framework can be viewed in positive terms as a description (or first-order approximation) of
how the economy works in practice. It can also be considered in normative terms to define how
individuals should optimally consume and reinvest economic output over time. As a result, the
individual parameters of the Ramsey equation can be specified using two approaches: the
descriptive (or positive) approach and the prescriptive (or normative) approach.
The descriptive (positive) approach attempts to calibrate the parameters of the Ramsey
equation by using estimates from observed behavior. The resulting consumption discount
rate reflects society's observed preferences for trading off consumption over time and the
best available information on the future growth rate of consumption. Advocates of the
descriptive approach generally call for inferring the discount rate from market rates of
return "because of a lack of justification for choosing a social welfare function that is any
different than what decision-makers [individuals] actually use" (Arrow etal. 1996).
However, this can be difficult to do in practice.
The prescriptive (normative) approach is based on defining a social welfare function that
formalizes the normative judgments that the decision-maker wants to explicitly incorporate
into the policy evaluation. In the case of the Ramsey equation, parameters would then be
chosen to match these desired normative judgments.26 27 The main argument against the
prescriptive approach is that it may not be consistent with individuals' preferences for
inter-temporal trade-offs revealed by their market behavior.
While the Ramsey framework is commonly used and is based on an intuitive description of the
general problem of trading off current and future consumption, it has limitations. Arrow (1996)
25 Text Box 6.3 provides a derivation of the Ramsey framework Key literature on this topic includes Arrow et al.
(1996), Lind (1994), Schelling (1995), Solow (1992), Manne (1994), Toth (1994), Sen (1982), Dasgupta (1982),
Pearceand Ulph (1994), Gollier (2010), and Arrow et al. (2013).
26 Arrow etal. (1996).
27 For instance, there has been a long debate, starting with Ramsey himself, on whether the pure rate of time
preference, which shows a general preference for consumption by the current as opposed to future generations,
should be greater than zero when evaluating public policy decisions.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
contains a detailed discussion of descriptive and prescriptive approaches to discounting over long
time horizons, including examples of rates that emerge under various assumptions about
components of the Ramsey equation.
: Efficiency and Intergeneratior :i Equity
A principal concern when policies span long time horizons is that future generations affected by the
policy are not yet alive. Therefore, they cannot participate in the decision-making process and their
preferences are uncertain. This is not always a severe problem for practical policy analysis. Many
policies impose relatively modest costs and benefits, or have costs and benefits that begin
immediately or occur in the not-too-distant future. In most cases, it suffices to assume future
generations will have preferences like those of present generations. However, for policies where
the costs and benefits are large and distributed asymmetrically over large expanses of time, the
choice of discount rate may involve both efficiency and ethical considerations.
6,3,1 i Lificienc nsiderations
As discussed in Chapter 1 and Appendix A, the BCA efficiency test is grounded in the notion of a
potential Pareto improvement, whereby those who benefit from a policy, on net, could potentially
compensate those who experience costs, on net, such that everyone is at least as well off as they
were before. The potential for this compensation to occur across generations hinges on the interest
rate at which society can transfer wealth across long time horizons. The choice of social discount
rate, therefore, contains an implicit assumption about whether, and at what price, the distribution
of wealth across generations could be adjusted to compensate those who bear costs, on net. Some
have argued that in the U.S. context, the federal government's borrowing rate is a good candidate
for this rate, while others have argued that practical difficulties associated with implementing
intergenerational transfers suggest that the Kaldor-Hicks potential compensation test is limited in
its ability to assess policies affecting multiple generations.28 29 Still others argue that the discount
rate should be below market rates to correct for market distortions, and uncertainties or
inefficiencies in intergenerational transfers of wealth.30 The role of uncertainty is discussed in more
detail below.
6,3,1 siderations
Because future generations cannot participate in decisions made by current generations, social
discounting may raise ethical issues regarding the intertemporal distribution of wealth. This
concern does not suggest forgoing the use of a positive discount rate but has led to suggestions that
the discount rate used in intergenerational contexts should be below market rates to ensure that
generations are treated equally based on ethical principles (e.g., Arrow et al. 1996, Portney and
Weyant 1999).31 One interpretation of this idea is to forgo discounting the utility of future
28 SeeLind (1990) and a summary by Freeman (2003).
29 For more information and theoretical foundations of the Kaldor-Hicks test for potential Pareto
improvements, see Appendix A.
30 Arrow etal. (1996); Weitzman (1998).
31 Another issue is that there are no market rates for intergenerational time periods.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
generations by setting the pure rate of time preference in the Ramsey framework to zero. These
suggestions are for using a prescriptive (i.e., normative) approach for discounting.
lining Di? ¦ vvrt R.-;i-?
Theoretical and empirical support is growing for discount rates that decline over time for
intergenerational discounting (Arrow et al. 2014). That is, the appropriate rate to use in
discounting effects in year 101 to year 100 will be lower than the appropriate rate to use in
discounting effects in year 2 to year 1. Multiple rationales support a declining discount rate, most
notably slowing consumption growth rates and uncertainty about economic growth.
f tional r i ["-> fining Phcoi'in I*->i -s
A slowing of consumption growth rates leads to declining discounting, as is evident from the
Ramsey framework. Using a constant discount rate in BCA is technically correct only if the rate of
economic growth per capita remains fixed over the time horizon of the analysis. In principle, any
changes to income growth, the elasticity of marginal utility of consumption, or the pure rate of time
preference will lead to a discount rate that changes accordingly. If economic growth per capita
changes over time, the discount rate will also fluctuate. In particular, an assumption that the growth
rate is declining systematically over time (perhaps to reflect some physical resource limits) will
lead to a declining discount rate. This is the approach taken in some models of climate change.32
Uncertainty about future consumption growth can also lead to a declining discount rate. The longer
time horizon in an intergenerational policy context implies greater uncertainty about the
investment environment and economic growth over time, and a greater potential for environmental
feedbacks to economic growth (and consumption and welfare). These feedbacks further increase
uncertainty when attempting to estimate the social discount rate. This additional uncertainty
implies effective discount rates lower than those based on observed average market interest rates
(Weitzman 1998, 2001; Newell and Pizer 2003; Arrow etal. 2013; Cropper etal. 2014).33 34
The effect of uncertainty on discount rates is a result of the fact that discounting is a non-linear
operation, such that the average discount factor (i.e., E[e rt]) is not equal to the discount factor
calculated at the average discount rate (i.e., e-£Wt). As an alternative to estimating the average
discount factor, one can calculate the certainty equivalent discount rate schedule, which is the
discount rate schedule that yields the same discount factor in any time period as the average
discount factor across the possible discount rates. Uncertainty about future consumption growth
will cause this certainty equivalent discount rate schedule to decline over time as the potential for
low discount rates will increasingly dominate the expected NPV calculations for benefits and costs
32 See, for example, U.S. EPA (2023).
33 This holds regardless of whether or not the estimated investment effects are predominantly measured in
terms of private capital or consumption.
34 Gollier and Zeckhauser (2005) reach a similar result using a model with decreasing absolute risk aversion.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
far in the future (Weitzman 1998). Text Box 6.5 provides a simple example highlighting how
declining discount rates arise in this fashion.35
6,3,1.1 i poaches to Estimar i' imiir Discou I* u - mi - r Un>, -[flinty
Declining discount rate schedules can be derived from specifications of the Ramsey formula or from
historically estimated stochastic models of interest rates.
If there is uncertainty in the consumption growth rate, then the standard Ramsey formula may
need to be adjusted. Incorporating uncertainty in consumption growth results in a third term
being subtracted from the Ramsey formula to account for the potential of low growth futures
(Gollier 2002; Arrow et al. 2014). If the shocks to consumption growth are independent and
identically distributed, then the precautionary term will cause the discount rate to be lower but
not decline. However, if the shocks are positively correlated over time, then the precautionary
term will grow over time and cause the discount rate to decline (Goiller 2014). If there is
parametric uncertainty regarding the process underlying consumption growth or the other
values in the Ramsey formula, this can also lead to a declining discount rate. However, if the
uncertainty in the growth rate is endogenously incorporated in the benefits or costs
calculations using Monte Carlo simulations, this adjustment is unnecessary.36
The use of historical data to estimate a declining discount rate schedule is shown by Newell and
Pizer (2003). They use historical data on U.S. interest rates and assumptions regarding their future
path to characterize uncertainty and compute a certainty equivalent rate. In this case, uncertainty
in the individual components of the Ramsey equation is not being modeled explicitly. This is
attractive as a descriptive approach because it does not require specifying uncertainty over the
consumption growth rate and parameters of the Ramsey formula, but its results are sensitive to the
selection of a model to represent the stochastic interest rate process (Groom etal. 2007).
Some modelers and government bodies have used fixed step functions for the discount rate term
structure to approximate more rigorously-derived declining discount rate schedules and to reflect
non-constant economic growth, intergeneration equity concerns, and heterogeneity in future
preferences.37 This method acknowledges that a constant discount rate does not adequately reflect
the reality of fluctuating and uncertain growth rates over long time horizons. However, no
empirical evidence suggests the point(s) at which the discount rate declines, so any year selected
for a change in the discount rate will be ad-hoc.
35 While this explanation is motivated by uncertainty over long-term consumption growth, a similar result
arises when there is persistent uncertainty about preferences or heterogeneity in preferences. See Heal and
Millner (2014).
36 For example, see the approach taken in Newell et al. (2022).
37 For instance, in the United Kingdom, the Treasury recommends the use of a 3.5% discount rate for the first 30
years followed by a declining rate over future time periods until it reaches 1 % for 301 years and beyond. The
guidance also requires a lower schedule of rates, starting with 3% for zero to 30 years, where the pure rate of
time preference in the Ramsey framework (the parameter r in our formulation) is set to zero. For details, see
Lowe (2008). Additionally, Weitzman (2001) presents a novel approach to calibrating a fixed step discount rate
schedule based on uncertainty using survey data.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 6.5 - Declining Discount Rates from Uncertainty
The term structure for the certainty equivalent discount rate may decline over time due to
uncertainty about future economic conditions or social preferences. Consider a simple example
where one is attempting to evaluate the net present value of a policy that yields $1 in net benefits
every year, and there is uncertainty as to whether the discount rate is 2% or 4%, with each rate
equally likely. Because discounting is a nonlinear operation, using the average discount rate of
3% will not provide the same result as calculating the expected net present value of the two
equally likely rates. Figure 6.1a presents the present value of this stream of net benefits for time
horizons from 1 year to 300 years. Using the average discount rate of 3% underestimates the
average present value of the payments for long time horizons. However, the plot shows that, the
difference is relatively small over short time horizons.
Figure 6.1a: Net Present Value Figure 6.1b: Certainty Equivalent Discount Rate
As opposed to calculating the average net present value, one could solve for the discount rate
schedule that, when applied to the problem as if there were certainty about the discount rate,
yields the same present value for a particular time horizon as when explicitly accounting for
uncertainty. This discount rate schedule is referred to as the certainty equivalent discount rate.
Figure 6.1b presents the certainty equivalent discount rate for this example. The discount rate
schedule begins close to, but below, the average discount rate of 3% and so for short time
horizons the 3% and certainty equivalent discount rates have approximately equal impact on the
present value. However, as one moves further out in time, the certainty equivalent discount rate
declines and becomes much lower. This effect may be seen in Figure 6.1a. At the 4% discount
rate, after approximately 100 years, future payments do not appreciably affect present value.
However, at the 2% discount rate, extending the time horizon past 100 years appreciably
increases the present value. Therefore, in terms of calculating the average present value it is the
possibility of the discount rate being 2% that matters more (i.e., it dominates). This is the general
effect that causes the certainty equivalent discount rate in Figure 6.1b to decline.
6-25 Guidelines for Preparing Economic Analyses | 3rd edition
-------
~insistency Issues an P immj Discou r w
Another concern regarding declining discount rates is the potential for time inconsistency in policy
recommendations over time (Arrow et al. 2014). Time inconsistency means that a net-beneficial
policy today may not be net-beneficial if evaluated in the future, even when nothing has changed
except for the date of the evaluation. The use of fixed step functions can exacerbate the problem.
Therefore, whether an analysis shows the policy to be net-beneficial will be sensitive to the point in
time the analysis is conducted. Text Box 6.6 provides an illustration of this time consistency
problem.
If the analyst obtains new information between the time the original and updated analysis are
conducted, the results of the analysis may have changed. However, if a fixed declining discount rate
schedule is adopted and not updated between analyses to reflect the arrival of new information,
that could lead to a potential time inconsistency problem (Arrow et al. 2014).
Ion and Challenges
A wide range of potential approaches for calibrating a discount rate or a schedule of declining
discount rates is available for discounting intergenerational costs and benefits. More complex
analysis is justified when the proportion of costs and benefits occurring far out on the time horizon
and the temporal separation of costs and benefits is large. While strong theoretical and empirical
evidence shows that a declining discount rate schedule is appropriate when considering effects
over long time horizons, calibration complications and concerns with time inconsistency remain
notable challenges.
One possible response to such challenges is to select a constant but slightly lower discount rate
when discounting costs and benefits expected to occur far out in the time horizon, reflecting a
certainty equivalent discount rate. Independent of the approach or rate selected, the same discount
rate should be applied to all benefits and costs that occur in the same year for both intra- or
intergenerational consequences to ensure consistency in the analysis (Arrow et al. 2013).
^ i The P.'"!* .-,f Ff r Pi:\r,-Ffntm:j in 11h mi I
This chapter focuses on social discounting, which is discounting from the broad society-as-a-whole
perspective embodied in BCA. By contrast, private discounting is the discounting of expected future
benefits or costs (e.g., revenues or expenditures) from the perspective of private individuals or
firms. Private discount rates reflect the preferences of specific individuals for consumption over
time, as well as the prices that individuals and firms pay to borrow and lend money. These rates
vary among firms, industries, and individuals due to differences in preferences, tax treatments, and
costs of borrowing. Section 6.2.1 describes why market interest rates differ from the consumption
rates of interest
As previously stated, private discount rates should not be used to estimate the NPV of the social net
benefits of policies and projects because the intertemporal preferences of society as a whole (as
measured by the social rate of time preference) are not likely to be equal to private market lending
rates or individual or firm preferences.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 6.6 - Time Inconsistency and Declining Discount Rates
Time inconsistency means that a net-beneficial policy today may not be net-beneficial if evaluated
in the future, even when the only change is the date of the evaluation.
Consider the following stylized example of a declining discount rate used to analyze a policy. The
discount rate schedule is a step function with 3% for benefits and costs that occur one period in
the future and 0% in each period thereafter. The policy will cost $1,000 in the second period from
today and will provide benefits of $1,003 in the third period. If evaluated today, the policy has
positive net benefits ofe 0 03 (e-0 00$l,003 - $1000) = $3.
However, a reevaluation of the policy in the second period would have negative net benefits of e-
0 03 $1,003 - $1000) = $-27, because costs are not discounted while the benefits in period three are
discounted to period two at 3%. Therefore, whether an analysis shows the policy to be net-
beneficial will be sensitive to the point in time the analysis is conducted. This is a time-
inconsistent approach to discounting.
6.4.1 Predicting Private Behaviors and Choices
Private discounting should be used to predict behaviors and choices of individuals and firms in
response to policy, and how investment in the economy and consumption (broadly defined) are
expected to change as a result.38 Individuals and firms can be expected to make decisions based on
their own opportunity costs rather than those of society as a whole. For example, from the
viewpoint of a private firm, the change in a stream of future profits due to the adoption of a
pollution abatement project would be evaluated at the rate at which the firm can borrow. Similarly,
the expected consumption behavior of individuals and households should be modeled consistently
with how they make purchasing decisions. To predict the purchase of durable goods, for example,
private evaluation and perception of the consumer's benefits and costs from using these goods over
time should be used. Failure to account for choices based on appropriate private discount rates will
lead to inconsistencies between the behavior of individuals and firms in the analysis and their
expected behavior in the real world.39 Therefore, private discount rates should be used to evaluate
how firms and individuals will respond to policy.
6.4.2 Treatment of Interest Payments
Any changes in the amount of interest paid for borrowing (e.g., loans) resulting from a potential
regulation should not be included in the calculation of its estimated social benefit or cost Interest
payments do not reflect the use of real resources such as labor, capital, and materials in an
economy. Rather, the interest payment is a transfer between the borrower and lender and would
net out of a social benefit-cost analysis. Private interest rates, in part, reflect the opportunity cost to
society of any changes in the timing of consumption as a result of a regulation, but this opportunity
38 This guidance applies both the regulated sources and any individuals and firms meaningfully affected by the
behavior of the regulated sources.
39 For this same reason, using a social discount rate to model how firms and individuals evaluate private
benefits and costs can lead to misspecif cation of the baseline over time and/or a mistaken projection of their
responses to a policy.
6-27 Guidelines for Preparing Economic Analyses | 3rd edition
-------
cost is already accounted for in social discounting, as discussed above.40 However, interest
payments should be accounted for when evaluating the incidence and economic impacts of a
regulation. For example, if a firm must take out a loan to comply with a regulation, the interest
payment on that loan should be accounted for when estimating the effect of the regulation on the
firm's profits.41 See Chapter 9 for further discussion of how to determine the incidence of a
regulation.
6.4,3 Selecting Private Di
Selecting which discount rate best represents household or firm behavior is a challenge. An
appropriate discount rate may be observed from market behavior, but different households and
firms borrow at different interest rates, and even within a household or firm, borrowing (and
lending) occurs at different rates. For example, firms may borrow at different rates depending on
whether they are financing investments through debt or equity. Therefore, the choice of discount
rate used to represent private behavior should be explained and, if necessary, sensitivity analyses
using different rates should be considered.
The following recommendations are intended as practical and plausible default assumptions rather
than comprehensive and precise estimates of social discount rates that apply in all situations. In
some analyses, there may be compelling reasons to gather data and develop a realistic model with
precise empirical estimates for the factors most relevant to the specific circumstances. In such
cases, these estimates should be presented along with the rationale in the description of the
40 Administrative charges on a loan (e.g., origination fees] may include the cost of preparing and administering
any loans. Changes in these costs, if they can be determined, should be accounted for in a benefit-cost analysis.
41 When evaluating the incidence of a regulation over time, it may also be important to recognize the
annualization of any capital investment. However, when estimating net-benefits, costs should be discounted from
the period they are realized and not necessarily when they are paid for by the regulated source (or other
economic actor). The private amortization schedule of financed costs should not be used.
42 As discussed in the behavioral economics literature, individual behavior is not always consistent with the
conventional discounting framework. For example, households may consume and save different sources of
wealth differently, and therefore are applying different discount rates to those sources of wealth, even when the
sources of wealth are fungible (Thaler 1990). There is also evidence that discount rates for individuals decline
over time, are lower the larger the magnitude of the future value, are higher for gains than for loses and that
individuals may prefer a stream of benefits that increase over time over one that is constant over time despite
each having the same nominal values (Fredrick et al. 2002). Alternative behavioral frameworks have been
proposed that are consistent with these observed patterns of discounting (e.g., Loewenstein and Prelec 1992;
Laibson 1998). Conventional discounting should be used to represent individual, household or firm behavior in
the economic analysis, although alternative discounting frameworks to represent the behavior of individuals or
households may be provided in a sensitivity analysis, provided the alternative framework is well-studied in the
literature in settings comparable to that of the regulation. Care should be taken when applying alternative
discounting models to predict behavior, as observed behavior that at first appears inconsistent with the
conventional framework may actually be consistent with the perceived inconsistency due to omitted
considerations. For example, an individual's discount rate may appear to change over time due to perceived
uncertainty about future outcomes being valued, even though their strict rate of time preference may not be
changing (Fredrick et al. 2002).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
methods and any appropriate peer review. Results based on default assumptions should also be
included for comparison purposes and consistency with OMB guidance, as appropriate. With this
caveat in mind, recommendations for discounting are below.
Display the full time paths of benefits and costs as they are projected to occur over the time
horizon of analysis both without discounting and appropriately discounted.
When determining the net benefits of a regulation, the analysis should compare the
discounted value of the entire time horizon of benefits and costs. It is inappropriate to
characterize the effect of a regulation with only the costs or benefits for a limited period of
time, e.g., a single year, when benefits and costs may occur during other periods. Similarly, it
is inappropriate to compare an annual value to an annualized value.
Calculate the present or annualized value of social benefits and costs using the consumption
rate of interest This is appropriate for situations where all costs and benefits occur as
changes in consumption flows rather than changes in capital stocks (i.e., capital
displacement and inducement effects are negligible). OMB (2023) recommends a real
consumption rate of discount of 2% based on empirical estimates.
To the extent that a regulation is expected to displace or induce short-term or long-term
capital investment, then the shadow price of capital should be applied to the components
of benefits and costs impacting this investment to convert all effects into consumption
equivalents.
o In general, there is uncertainty as to the extent to which private capital is displaced
or induced by regulatory requirements. If the shadow price of capital approach is
not applied explicitly or implicitly using a general equilibrium framework, then
analysts should consider a sensitivity analysis consistent with OMB (2023) to
understand the potential effect of capital investment changes on the discounted
benefits and costs. OMB recommends considering a range of 1.0 to 1.2 as the
shadow cost of capital. The sensitivity analysis should be presented separately and
not part of the primary estimates of benefits, costs, or net benefits, and should be
considered as a check on the robustness of the relative net benefits of the analyzed
options.
If the policy has costs or benefits that extend over a long time horizon (e.g., most benefits
accrue to one generation and most costs accrue to another), then a constant consumption
rate of interest may not be appropriate. The analysis should also present the net benefits
under an additional approach whose rationale is clearly explained. These approaches may
include:
o Calculating the expected present value using a Monte Carlo simulation which
explicitly accounts for uncertainty in the growth rate of consumption and the
correlation between the growth rate and the benefits and costs.43
o Calculating the expected present value of net benefits using a schedule of declining
discount factors (Newell and Pizer 2003, Groom etal. 2007, Hepburn etal. 2009,
OMB 2023).
Regardless of the approach or rate selected, the same discount rate should be applied to all
benefits and costs that occur in the same year to ensure consistency in the analysis, and
43 For example, see Newell et al. (2023).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
benefits and costs should be discounted to the same year when calculating net benefits. In
addition, assumptions that may influence the discount rate (e.g., the gross domestic product
(GDP) growth rate) should be consistent with assumptions made elsewhere in the analysis
when feasible. In cases where this is not possible (e.g., because a valuation estimate has a
discounting assumption embedded in it that cannot be disentangled), the analysis should
clearly explain the limitation, why it cannot be resolved, and its implications for the
analysis.
When discounting future benefits and costs, the following principles should be kept in mind:
Private discount rates should be used to predict the behavior of individuals and firms and to
evaluate economic impacts and incidence, but they should not be used in place of the social
discount rate to assess the social benefits and costs of a policy.
The discount rate should reflect marginal rates of substitution between consumption in
different time periods. It should not be confounded with factors such as uncertainty in
benefits and costs or the value of environmental goods or other commodities in the future
(i.e., the "current price" in future years).
The economic analysis should account for the lag time between a change in regulation and
the resulting welfare impacts. This includes accounting for expected changes in human
health, environmental conditions, ecosystem services, and other related factors.
6-30 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapti jrences
Arrow, K.J., W.R. Cline, K.G. Maler, M. Munasinghe, R. Squitieri, and J.E. Stiglitz. 1996. Intertemporal Equity,
Discounting, and Economic Efficiency. Climate Change 1995: Economic and Social Dimensions of Climate
Change, ed. J.P. Bruce, H. Lee, and E.F. Haites. Cambridge, MA: Cambridge University Press.
Arrow, K., M. Cropper, C. Gollier, B. Groom, G. Heal, R. Newell, W. Nordhaus, R. Pindyck, W. Pizer, P. Portney, T.
Sterner, R. S. J. Tol, M. Weitzman. 2013. Determining Benefits and Costs for Future Generations. Science,
341(26): 349-350.
Arrow, K., M. Cropper, C. Gollier, B. Groom, G. Heal, R. Newell, W. Nordhaus, R. Pindyck, W. Pizer, P. Portney, T.
Sterner, R. Tol, M. Weitzman. 2014. Should Governments Use a Declining Discount Rate in Project Analysis?
Review of Environmental Economics and Policy 8(2): 145-163.
Bauer, M.D. and Rudebusch, G.D., 2020. Interest rates under falling stars. American Economic Review, 110(5)-.
1316-1354.
Bauer, M.D. and Rudebusch, G.D., 2023. The rising cost of climate change: evidence from the bond
market. Review of Economics and Statistics, 105(5): 1255-1270.
Boardman, A.E., D.H. Greenberg, A.R. Vining, and D.L. Weimer. 2011. Cost-Benefit Analysis, Concepts and
Practice, 4th Ed. Upper Saddle River, NJ: Pearson Education, Inc.
Burgess, D.F. 1988. Complementarity and the discount rate for public investment. The Quarterly Journal of
Economics, 103(3): 527-541.
Burgess, D. F. and Zerbe, R. 0. 2011a. Appropriate discounting for benefit-cost analysis. Journal of Benefit-Cost
Analysis, 2(2): 1-20.
Burgess D.F., and Zerbe R.O. 2011b. Calculating the Social Opportunity Cost Discount Rate. Journal of Benefit-
Cost Analysis. 2(3):1-10. doi:l0.2202/2152-2812.1065.
Cairns, J. 2006. Developments in Social Discounting: With Special Reference to Future Health Effects. Resource
and Energy Economics, 28(3): 282-297.
CEA. 2017. Discounting for Public Policy: Theory and Recent Evidence on the Merits of Updating the Discount
Rate. Council of Economic Advisers Issue Brief. January 2017.
Cropper, M.L., M.C. Freeman, B. Groom, and W.A. Pizer. 2014. Declining Discount Rates. American Economic
Review, 104 (5): 538-43.
Dasgupta, P. 1982. Resource Depletion, Research and Development, and the Social Rate of Return. Discounting
for Time and Risk in Energy Policy, ed. R.C. Lind. Washington, DC: Resources for the Future.
Dreze, J.H., 1974. Investment under private ownership: optimality, equilibrium and stability. Allocation under
uncertainty: equilibrium and optimality. Palgrave Macmillan, London.
Frederick, S., G. Lowenstein, and T. O'Donoghue. 2002. Time Discounting and Time Preference: A Critical
Review .Journal of Economic Literature, 40: 351-401.
Freeman III, A.M. 2003. The Measurement of Environmental and Resource Values: Theory and Methods. 2nd Ed.
Washington, DC: Resources for the Future.
Gollier, C. 2002. Time Horizon and the Discount Rate. Journal of Economic Theory 107(2): 463-473.
Gollier, C.2010. Expected Net Present Value, Expected Net Future Value, and the Ramsey Rule. Journal of
Environmental Economics and Management, 59(2): 142-148.
Gollier, C. 2014. Discounting and Growth. American Economic Review, 104(5): 534-537.
Gollier, C. and J. Zeckhauser. 2005. Aggregation of Heterogeneous Time Preferences. Journal of Political
Economy, 113(4): 878-896.
Groom, B., C. Hepburn, P. Koundouri, and D. Pearce. 2005. Declining Discount Rates: The Long and the Short of
It. Environmental and Resource Economics, 32: 445-493.
6-31 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Groom, B., P. Koundouri, E. Panopoulou, and T. Pantelidis. 2007. Discounting the Distant Future: How Much
Does Model Selection Affect the Certainty Equivalent Rate? Journal of Applied Econometrics, 22 (3): 641-656.
Harberger, A.C. 1972. On measuring the social opportunity cost of labor. Project evaluation: collected papers.
Palgrave Macmillan UK.
Harberger, A. C. and G. P. Jenkins. 2015. Musings on the Social Discount Rat e. Journal of Benefit-Cost Analysis,
6(1): 6-32.
Heal, G. and Millner, A. 2014. Agreeing to Disagree on Climate Policy. Proceedings of the National Academy of
Sciences, 111(10): 3695-3698.
Hepburn, C., P. Koundouri, E. Panopoulou, and T. Pantelidis. 2009. Social Discounting Under Uncertainty: A
Cross-Country Comparison .Journal of Environmental Economics and Management, 57: 140-150.
Ibbotson, R.G. and R.A. Sinquefield. 1984. Stocks, Bonds, Bills, and Inflation: The Past and the Future, Financial
Analysts Research Foundation, 1982. Updates published as Stocks, Bonds, Bills and Inflation Yearbook,
Annual, Chicago, IL: R.G. Ibbotson Associates, Inc.
Laibson, D. 1998. Life-Cycle Consumption and Hyperbolic Discount Functions. European Economic Review,
42(3-5): 861-871.
Lesser, J.A. and R.O. Zerbe. 1994. Discounting Procedures for Environmental (And Other) Projects: A
Comment on Kolb and Scheraga .Journal of Policy Analysis and Management, 13(1): 140-156.
Li, Q. and Pizer, W.A., 2021. Use of the consumption discount rate for public policy over the distant
future. Journal of Environmental Economics and Management, 107:102428.
Lind, R.C., ed. 1982a. Discounting for Time and Risk in Energy Policy. Washington, DC: Resources for the
Future.
Lind, R.C. 1982b. A Primer on the Major Issues Relating to the Discount Rate for Evaluating National Energy
Options. Discounting for Time and Risk in Energy Policy, ed. R.C. Lind. Washington, DC: Resources for the
Future.
Lind, R.C. 1990. Reassessing the Government's Discount Rate Policy in Light of New Theory and Data in a
World Economy with a High Degree of Capital Mobility. Journal of Environmental Economics and Management,
18(2): 8-28.
Lind, R.C. 1994. Intergenerational Equity, Discounting, and the Role of Cost-benefit Analysis in Evaluating
Global Climate Policy. Integrative Assessment of Mitigation, Impacts, and Adaptation to Climate, ed. N.
Nakicenovic, W.D. Nordhaus, R. Richels, and F.L. Toth. Laxenburg, Austria: International Institute of Applied
Systems Analysis.
Loewenstein, G. and D. Prelec. 1992. Anomalies in Intertemporal Choice: Evidence and an Interpretation. The
Quarterly Journal of Economics, 107(2): 573-597.
Lowe, J. 2008. Intergenerational wealth transfers and social discounting: Supplementary Green Book
Guidance. HM Treasury.
Lyon, R.M. 1990. Federal Discount Rate Policy, the Shadow Price of Capital, and Challenges for Reforms.
Journal of Environmental Economics and Management, 18(2): 29-50.
Lyon, R.M. 1994. Intergenerational Equity and Discount Rates for Climate Change Analysis. Paper presented
at IPCC Working Group III Workshop on Equity and Social Considerations Related to Climate Change, July 18-
22,1994, Nairobi, Kenya.
Manne, A.S. 1994. The Rate of Time Preference: Implications for the Greenhouse Debate. Integrative
Assessment of Mitigation, Impacts, and Adaptation to Climate Change, ed. N. Nakicenovic, W.D. Nordhaus, R.
Richels, and F.L. Toth. Laxenburg, Austria: International Institute for Applied Systems Analysis (IIASA).
Moore, M.A., A. Boardman, and A.R. Vining. 2013a. More Appropriate Discounting: the Rate of Social Time
Preference and the Value of the Social Discount Rat e. Journal of Benefit-Cost Analysis, 4(1): 1-16.
6-32 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Moore, M.A, A.E. Boardman, and A.R. Vining. 2013b. The Choice of Social Discount Rate and the Opportunity
Cost of Public Funds. Journal of Benefit-Cost Analysis, 4(3): 401-409.
Moore, M.A., Boardman, A.E. and Vining, A.R., 2017. Risk in public sector project appraisal: it mostly does not
matter! Public Works Management & Policy, 22(4]: 301-321.Moore, M.A., A.E. Boardman, A.R. Vining, D.L.
Weimer, and D.H. Greenberg. 2004. Just Give Me a Number! Practical Values for the Social Discount Rate.
Journal of Policy Analysis and Management, 23(4): 789-812.
Newell, R.G. and W.A. Pizer. 2003. Discounting the Distant Future: How Much Do Uncertain Rates Increase
Valuations? Journal of Environmental Economics and Management, 46(1): 52-71.
Newell, R.G., Pizer, W.A. and Prest, B.C., 2022. A discounting rule for the social cost of carbon .Journal of the
Association of Environmental and Resource Economists, 9(5): 1017-1046.
Newell, R.G., Pizer, W.A. and Prest, B.C., 2024. The Shadow Price of Capital: Accounting for Capital
Displacement in Benefit-Cost Analysis. Environmental and Energy Policy and the Economy, 5(1): 49-69.
OMB. 2003. Circular A-4, Regulatory Analysis, September 17, 2003. Available at:
https://obamawhitehouse.archives.gov/omb/circulars a004 a-4 (accessed October 9,2024).
OMB. 2023. Circular A-4, Regulatory Analysis, November 9, 2023. Available at:
https: //www.whitehouse.gov/wp-content/uploads/2023/ll/CircularA-4.pdf (accessed October 9,2024).
Pearce, D.W. and R.K. Turner. 1990. Discounting the Future. Economics of Natural Resources and the
Environment. Baltimore, MD: The Johns Hopkins University Press.
Pearce, D.W. and D. Ulph. 1994. A Social Discount Rate for the United Kingdom. Mimeo No. 95-01, Centre for
Social and Economic Research on the Global Environment, University College London and University of East
Anglia, UK.
Portney, P.R. and J.P. Weyant. 1999. Discounting and Intergenerational Equity. Washington, DC: Resources for
the Future.
Ramsey, F.P. 1928. A Mathematical Theory of Saving. Economics Journal, 38: 543-559.
Sandmo, A. and Dreze, J.H., 1971. Discount rates for public investment in closed and open
economies. Economica, 35(152), pp.395-412.
Schelling, T.C. 1995. Intergenerational Discounting. Energy Policy 23(4/5): 395-401.
Sen, A.K. 1982. Approaches to the Choice of Discount Rates for Social Benefit-cost Analysis. Discounting for
Time and Risk in Energy Policy, ed. R.C. Lind. Washington, DC: Resources for the Future.
Spackman, M. 2004. Time Discounting and the Cost of Capital in Government. Fiscal Studies, 25(4): 467-518.
Solow, R. 1992. An Almost Practical Step toward Sustainability. Paper presented at the Fortieth Anniversary
of Resources for the Future, October 8,1992, in Washington, DC.
Thaler, R.H. 1990. Saving, Fungibility, and Mental Accounts .Journal of Economic Perspectives, 4(1): 193-205.
Toth, F.L. 1994. Discounting in Integrated Assessments of Climate Change. Integrative Assessment of
Mitigation, Impacts, and Adaptation to Climate, ed. N. Nakicenovic, W.D. Nordhaus, R. Richels, and F.L. Toth.
Laxenburg, Austria: International Institute of Applied Systems Analysis (IIASA).
U.S. CBO. 2005. Uncertainty in Analyzing Climate Change: Policy Implications. Available at:
https://www.cbo.gOv/sites/default/files/109th-congress-2005-2006/reports/01-24-climatechange.pdf
(accessed December 22,2020).
U.S. CBO. 2016. An Update to the Budget and Economic Outlook: 2016 to 2026. CBO publication 51908.
August 2016.
U.S. CBO. 2023. The 2023 Long Term Budget Outlook. CBO publication 59014.
U.S. EPA 2001a. Arsenic Rule Benefits Analysis: An SAB Review. EPA-SAB-EC-01-008.
6-33 Guidelines for Preparing Economic Analyses | 3rd edition
-------
U.S. EPA 2001b. National Primary Drinking Water Regulations; Arsenic and Clarifications to Compliance and
New Source Contaminants Monitoring. Final Rule. Federal Register 66(139): 37617.
U.S. EPA. 2004. Review Of The Revised Analytical Plan For EPA's Second Prospective Analysis-Benefits And
Costs Of The Clean Air Act 1990-2020: An Advisory by a Special Panel of the Advisory Council on Clean Air
Compliance Analysis. EPA-SAB-COUNCIL-ADV-04-004.
U.S. EPA. 2016. Guidelines for Preparing Economic Analyses. Report Number 240-R-10-001. Available at:
https://www.epa.gov/environmental-economics/guidelines-preparing-economic-analyses-201 (accessed
October 23, 2024).
U.S. EPA. 2017. EPA Air Pollution and Control Cost Manual. Chapter 2 - Cost Estimation: Concepts and
Methodology Available at: https: //www.epa.gov/economic-and-cost-analysis-air-pollution-regulations/cost-
reports-and-guidance-air-pollution (accessed December 22.2020).
Valdes, B. 1999. Economic Growth: Theory, Empirics and Policy. Northampton, MA: Edward Elgar Publishing.
Warnock, F.E. and V.C. Warnock. 2009. International Capital Flows and U.S. Interest Rates .Journal of
International Money and Finance, 28: 903-919.
Weitzman, M.L. 1998. Why the Far-Distant Future Should Be Discounted at Its Lowest Possible Rate. Journal
of Environmental Economics and Management, 36(3): 201-208.
Weitzman, M.L. 2001. Gamma Discounting. American Economic Review, 91(1): 260-271.
6-34 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 7 - Analyzing Benefits
This chapter provides an overview of the benefits analysis process, discussing
the quantification of benefits first and then their monetization. The aim of a
benefits analysis of an environmental policy is to describe the changes
resulting from that policy and to estimate the social benefits that ensue.
Willingness to pay (WTP] is the preferred measure for benefits.1 WTP
provides a full accounting of individual preferences across trade-offs between
wealth and benefits and is measured in monetary terms to allow the
calculation of net benefits. Net benefits are used to compare policy options
and assess the magnitude of expected net improvements in societal welfare.
When analyzing benefits and costs, the Guidelines assumes the policy under
review improves environmental quality, but at a cost, as noted in Chapter 1.
Benefits, then, are reduced risks to human health and increased welfare from
environmental improvements. This chapter provides tools and methods for
estimating these benefits. However, these same tools and methods are equally
applicable to valuing changes in environmental quality, regardless of the
direction of those changes (e.g., for deregulatory policies where declines in
environmental quality are assessed as a cost].2
Note that a benefits analysis may contain negative elements. For example,
there may be increases in human health risks due to increases in emissions of
a pollutant other than the one being regulated. These risk increases are costs
but may be presented as negative benefits in the benefits analysis, sometimes
described as "disbenefits "or "countervailing risks." Similarly, there may be
negative costs (i.e., benefits] that appear on the cost side of the benefit-cost
ledger. In this way, similar kinds of effects are kept together, which is
appropriate so long as it does not change the conclusion (i.e., the net benefits
of various options are not affected] and the analysis is internally consistent.
This chapter highlights the benefit transfer approach, the chief method for
monetizing benefits in economic analysis of regulatory actions and reviews
available options for more fully incorporating endpoints that are not
1 As described elsewhere in the Guidelines we use "willingness to pay" to refer to both willingness to pay and
willingness to accept compensation concepts. Compensation that falls between the willingness to pay of those
who gain and the willingness to accept of those bearing costs would also be compatible with the potential Pareto
criterion.
2 While this chapter focuses on the anticipated social benefits of regulation, the same approach also applies in a
retrospective setting. See Chapter 5, Text Box 5.1, for more discussion of retrospective analysis.
7-1 Guidelines for Preparing Economic Analyses | 3rd edition
-------
monetized due to lack of existing values or quantification. Chapter 11,
"Presentation of Analysis and Results," presents ways to convey information
on non-monetized benefits to help inform policy-making.3
7.1 The Benefits Analysis Process
Figure 7.1 presents a conceptual model for benefits analysis. After policy options have been
identified, the first step is to identify the changes in environmental contaminants or stressors that
are likely to result from policy options relative to the baseline. These may be measured as changes
in emissions or in concentrations of contaminants, but they can also be considered more broadly.
For example, "stressors" can be the number of hazardous waste sites, and the benefits analysis may
be built upon changes in this metric.
Figure 7.1 - A Conceptual Model for Benefit Analysis
Changes in
Environmental
Contaminants or
Stressors
Changes in
Environmental
Quality
Changes in
Benefit
Endpoints
Valuation of
Benefit
Endpoints
Benefits
C haracterization
Policy Options
Changes in contaminants or stressors often lead to changes in environmental quality such as a
change in ambient air quality. Environmental quality should be interpreted broadly for this
conceptual model, including exposure to contaminants. Often, a great deal of analysis is required to
project how changes in contaminants or stressors affect environmental quality, including modeling
the transport of the pollutants through the environment along a variety of pathways, including
movement through the air, surface water and groundwater; deposition in soils; and ingestion or
uptake by plants and animals (including humans). In many cases, explicit modeling of human intake
or exposure might be another intermediate step in the conceptual model that precedes quantifying
changes in benefit endpoints.
The next step is to identify the benefit endpoints that may be affected by changes in environmental
quality. Benefit endpoints are organized in the Guidelines into broad categories: human health
3 Other methods, such as cost-effectiveness analysis (CEA], can also be used to evaluate policies, CEA does not
require monetization of benefits but rather divides the costs of a policy by a particular effect (e.g., number of
lives saved], CEA can be used to compare proposed policy changes, but unlike BCA, cannot be used to calculate a
single, comprehensive measure of the net effects of a policy, nor can it compare proposed policy changes to the
status quo. While cost effectiveness analysis is not covered extensively in these Guidelines, other methods for
evaluating policies (e.g., distributional analyses] are covered in Chapter 10.
7-2 Guidelines for Preparing Economic Analyses | 3rd edition
-------
improvements, ecological improvements, aesthetic improvements, and reduced materials damages
(Section 7.2], Table 7.1 lists examples of benefit endpoints in each of these categories. Once changes
in benefit endpoints are identified, valuation follows well-defined economic principles (Section 7.2)
using well-established economic methods (Section 7.3). Commonly used methods for each type of
benefit are also described in Table 7.1.
Table 7.1 - Types of Benefits Associated with Environmental Policies:
Categories, Examples
and Commonly Used Valuation Methods
Human Health
Improvement
Examples
Commonly Used Valuation Methods
Mortality risk reductions
Reduced risk of:
Cancer fatality
Acute fatality
Averting behaviors
Hedonics
Stated preference
Morbidity risk reductions
Reduced risk of:
Cancer
Asthma
Cognitive Impairment
Averting behaviors
Cost of illness
Hedonics
Stated preference
Ecological Improvements
Examples
Commonly Used Valuation Methods
Market products
Food; Fuel; Timber; Fish
Production function
Demand analysis for consumer
benefits
Recreation activities and
aesthetics
Wildlife viewing
Fishing and hunting
Boating
Swimming
Hiking
Scenic views
Production function
Averting behaviors
Hedonics
Recreation demand
Stated preference
Valued ecosystem services
Climate moderation
Flood moderation
Groundwater recharge
Sediment trapping
Soil retention
Nutrient cycling
Pollination by wild species
Biodiversity, genetic library
Water filtration
Soil fertilization
Production function
Averting behaviors
Stated preference
Non-use values
Relevant species populations,
communities, or ecosystems
Stated preference
7-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Other Benefits
Examples
Commonly Used Valuation Methods
Aesthetic improvements
Visibility
Taste
Odor
Averting behaviors
Hedonics
Stated preference
Reduced materials damages
Reduced soiling
Reduced corrosion
Averting behaviors
Production / cost functions
Other market and non-market
goods
Reduced fuel expenditures
Reduced infrastructure
expenditures
Enhanced energy security
Demand analysis for consumer
benefits
Production/cost functions
Other methods as needed
Finally, the aggregate value for all benefits, including benefits arising from the primary statutory
objective of the regulation as well as other benefits, provides the basis for characterizing the
benefits of each policy option. Ideally, the benefits analysis would comprehensively assess all
welfare-improving effects all benefit endpoints attributable to a rule or policy decision,
including potential interactions and feedbacks between effects. This may be possible to an extent
with the use of integrated assessment models (IAMs) (see Text Box 7.1 for background on IAMs).
However, the modeling and data required for such a comprehensive assessment make it difficult to
do so in most circumstances.
Benefits analysis need not proceed by enumerating all benefit endpoints separately or follow the
specific sequence described in Figure 7.1, particularly if valuation estimates are linked to effects
further upstream in the model. For example, rather than monetizing enumerated health benefit
endpoints, the hedonic property method (Section 7.3.1.3) may estimate the total value to residents
of changes in the presence of hazardous waste sites - a change in a stressor in Figure 7.1 - by
linking policy changes to changes in property values. This valuation estimate could then be used in
benefits analysis. This method of assessing benefits can be viewed as a reduced form approach to
the modeling.4 Even when viewed as a reduced form approach, however, it is important to think
through the conceptual model to assess whether there are benefit endpoints not reflected in the
reduced form valuation estimate that should be included through additional analysis.
A General Approach to Benefits Estimation
Ultimately, benefits analysis should link policy changes to the value of all benefits that can be
meaningfully attributed to those changes. This is most often done using a pragmatic, general
approach aligned with the conceptual model in Figure 7.1 by tracing policy-related changes through
a set of models to predict changes in specific benefit endpoints, then valuing each endpoint, or
sometimes sets of endpoints, separately. An overall estimate of total benefits is the sum of these
separate components.
4 There are many other ways this type of reduced form approach may be appropriately used, sometimes
including estimates of the benefits per unit of environmental contaminants that are reduced.
7-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 7.1 - Types of Benefits Associated with Environmental Policies:
Categories, Examples and Commonly Used Valuation Methods
Integrated assessment models (IAM) are sometimes used to estimate the benefits of a policy. In
the broadest sense, IAMs are "approaches that integrate knowledge from two or more domains
into a single framework" (Nordhaus 2013), and this class of model has been used in many
disciplines, including earth sciences, biological sciences, environmental engineering, economics
and sociology. In environmental economics, IAMs combine natural processes and economic
systems into a single modeling framework. These models "connect economic activity with
environmental consequences, and ultimately, with valuation" (Keiser and Muller 2017). A full
IAM will capture four components behavior that generates emissions/pollutant loadings,
pollution fate and transport, environmental and human outcomes, and valuation as well as
feedbacks within and across these components. It also aims to capture the importance of these
consequences in a transparent, reproducible way.
IAMs have been used in environmental economics to study stock pollutants, primarily
greenhouse gases (GHGs) (Nordhaus 1993), and flow pollutants, e.g., air pollution (Mendelsohn
1980) and water pollution (Freeman 1979,1982). Current IAMs vary in structure, geographic
resolution and the degree to which they capture feedbacks and valuation of changes in physical
endpoints and regulatory compliance costs, with research often focused on improving these
representations. IAMs have been used to study the interaction between GHG mitigation and
urban and regional air pollution policies (Reilly et al. 2007), the dynamic economic and
ecosystem general equilibrium effects of fisheries management policy (Finnoff and Tschirhart
2008), and linkages in the food-water-energy nexus affecting policy outcomes (Kling et al.
2017). The choice of IAM will depend on the research or policy question.
IAMs are used in BCA in the valuation of changes in GHG emissions. IAMs that combine
representations of climate and economic systems are used to develop monetized estimates of
the damages associated with incremental emissions of carbon dioxide (CO2), denoted as the
social cost of carbon dioxide (SC-CO2), allowing the inclusion in a BCA of social benefits of
actions expected to change these. Specifically, the SC-CO2 is the present value of the stream of
future economic damages associated with an incremental increase (by convention, one metric
ton) in CO2 emissions in a particular year. It is intended to be a comprehensive measure and
includes economic losses due to a wide range of anticipated climate impacts, such as net changes
in agricultural productivity, human health risks, property damages from increased flood
frequencies, the loss of ecosystem services, etc. Analogous metrics estimate the monetary value
of climate impacts associated with other non-CC>2 GHGs, such as methane and nitrous oxide. In
January 2017, the National Academies of Sciences, Engineering, and Medicine issued
recommendations for research and a regularized process for updating the SC-CO2 estimates
used in federal regulatory BCA to ensure that estimates reflect the best available science. Since
the framework used to estimate the social cost of non-CC>2 GHGs is the same as that used for SC-
CO2, the Academies' recommendations also apply to the estimates of the social cost of non-CC>2
GHGs. See U.S. EPA (2023) for a detailed discussion of EPA's implementation of the Academies'
near-term recommendations.
The development of IAMs for use in other aspects of regulatory BCA is an emerging area of
research. For example, the EPA is developing an IAM for broad-scale water quality benefits
analysis by integrating hydrological and water quality modeling (HAWQS/SWAT) with an
economic valuation model (BenSPLASH) (Corona etal. 2020). The Hydrologic and Water Quality
System (HAWQS) is a water quantity and water quality modeling system using the Soil and
Water Assessment Tool (SWAT) as its core engine. The Benefits Spatial Platform for Aggregating
7-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Socioeconomics and H20 Quality (BenSPLASH) is an open-source analytical tool used to
quantify the economic benefits of changes in key water quality parameters and value the surface
water quality benefits of regulatory actions.
In short, the goal is to monetize those benefit endpoints that can be monetized, to quantify those
that can be quantified but not monetized and to provide qualitative characterizations of what
cannot be quantified. The results should then be described in a manner useful for policy makers.5
This general approach can be divided into three steps.
Step 1: Identify relevant benefit endpoints associated with the policy.
Step 2: Quantify significant changes in these benefit endpoints to the extent feasible.
Step 3: Monetize the changes using appropriate valuation methods or by drawing on
values from existing studies.
Each step in this approach is discussed in more detail in the sections that follow. Collaboration with
appropriate experts often will be necessary to execute these steps.6
Step 1: Identify Relevant Benefit Endpoints
The first step is to conduct an initial assessment of the types of benefits associated with the policy
options being considered. This requires evaluation of how conditions and ultimately benefit
endpoints differ between each policy option and baseline conditions (Chapter 5), including the
current and future state of relevant economic and regulatory variables (Section 5.2). The goal for
this step is to enumerate the full set of benefit endpoints and to identify those that should be
further developed for quantification and valuation. In this assessment, analysts should, to the extent
feasible:
Develop an understanding of the changes in environmental contaminants or
stressors resulting from policy options. Initially, the range of policy options being
considered may be very broad. Collaboration among all analysts and policy makers involved
in the policy analysis can help ensure that all potential effects are recognized. It is important
to account for both contaminant or stressor changes directly targeted by the policy options
and those that will occur even if not directly targeted.7
Identify the benefit endpoints likely to be affected by policy options. This step often
requires considering the transport of contaminants through the environment along many
pathways, including movement through the air, surface water and groundwater. Along
these pathways, the pollutants can have detrimental effects on natural resources, such as
affecting oxygen availability in surface water or reducing crop yields. Pollutants can also
have direct or indirect effects on human health, for example, affecting cancer incidence
through direct inhalation or through ingestion of contaminated food. This step is inherently
5 See Chapter 11 for more detail on presenting qualitative, quantified and monetized benefits.
6 A summary of a large-scale benefits exercise that followed these steps is described later in Text Box 7.5.
7 See Chapter 5, Section 5.1 for additional discussion of considering benefits that arise from changes in pollutants
other than those that would be directly regulated by the policy.
7-6 Guidelines for Preparing Economic Analyses | 3rd edition
-------
multi-disciplinary and will include consulting with risk assessors and other experts
involved in the rule or policy, sometimes as part of a formal workgroup (U.S. EPA 2014).
Evaluate the potential changes in benefit endpoints resulting from each policy option.
If policy options differ only in their level of stringency, then each option may have an impact
on all identified endpoints. Where policy options are more complex, however, the options
may have an impact on some endpoints but not on others.
Determine which benefit endpoints warrant further investigation in the overall
benefits analysis using at least the following framing questions:
Which benefit endpoints are likely to be large relative to total benefits or are
otherwise important for informing policy decisions? This determination should be
based on an assessment of the importance of each benefit endpoint to the benefits
analysis, including its potential magnitude, the extent to which it can be quantified
and the extent to which it can be monetized. Preliminary assessments should be
made using the best, readily available quantitative information; however, as a
practical matter, these decisions are often based on professional judgment.
Which benefit endpoints should be included even if they may not be large relative to
total benefits? Some benefit endpoints may not be captured by the first criteria but
are important and informative for other reasons. For instance, benefit endpoints
necessary to evaluate how minority, disadvantaged or susceptible groups are
affected in distributional analyses (Chapter 10) may not be large at a national level
but may be very important at a smaller scale. Benefit endpoints may also be
important because they reflect Agency priorities, are closely related to the
underlying motivation for the rule or are otherwise of particular interest to decision
makers.
Which benefit endpoints are likely to differ across policy options? Analysts should assess
how the effects of each policy option will differ. Benefits categories should be
meaningfully attributed to policy with some degree of confidence, while recognizing
that there will always be uncertainty and that this uncertainty can be characterized in
the benefits analysis. Again, this may be done as part of an interdisciplinary team
working on the rule or policy.
How much uncertainty is associated with the benefit endpoint? All endpoints have some
uncertainty. For example, toxicologic and epidemiologic evidence may be insufficient
to fully determine the likelihood that a contaminant causes a particular health effect
That said, it is important not to limit benefits endpoints at this stage even in the
presence of substantial uncertainty. Highly uncertain benefits may still be important
in the net benefits calculation and should generally be carried through the analysis.
Willingness to pay to avoid a very uncertain probability of a severe effect could be
larger than willingness to pay to avoid a more certain probability of a less serious
effect (McGartland et al. 2017). At a minimum, assessing uncertainty early can inform
what additional analysis is needed to effectively characterize benefits.
What are the costs of undertaking analysis to characterize the benefit endpoint? The
costs of quantifying and monetizing benefit endpoints may be minimal if existing
data and models can be applied. If existing data and models are insufficient, value of
information considerations are important A benefit endpoint may not be worth a
great deal of further investigation if the costs to quantify and monetize it exceed its
Guidelines for Preparing Economic Analyses | 3rd edition
-------
informational value. However, consideration should also be given to other current
and future rulemaking efforts that would rely on this endpoint for benefits.
The outcome of this step can be summarized in a list or matrix that describes the changes expected
from the policy options being considered, defines associated benefit endpoints and identifies the
endpoints that warrant further investigation.
The list of benefit endpoints should be as comprehensive as possible and may be lengthy at first,
encompassing all of those that reasonably can be expected to occur regardless of whether they can
be quantified and/or put in dollar terms. Analysts should preserve and refine this list as the
analysis proceeds. Maintaining the full list of potential effects facilitates later revisions if new
information warrants it. Equally important, benefits that can only be characterized qualitatively
should be presented along with quantitative information in the benefits analysis (see Chapter 11).
n - f , uantify Chair in r'iiiricant Benefit Endpoints
Next, the analysis should quantify changes in the benefit endpoints identified in Step 1 as
warranting further investigation, focusing on changes attributable to each policy option relative to
the baseline. Expertise from a wide array of disciplines in addition to economics is usually needed
in this step, including human health and ecological risk assessment, engineering and natural
sciences. Quantifying endpoints generally requires a function relating changes in emissions,
concentrations and/or exposure to changes in specific ecological services, health effects or risks.
Data are usually needed on the magnitude, duration, frequency and severity of the endpoints. For
example, changes in cancer risks typically come from human health risk assessments, and the
benefits analysis will need information on baseline risks, risk changes associated with each option,
the timing of the risk changes, fatality rates and the size and age distribution of affected
populations. If visibility is the attribute of concern, needed information includes the geographical
areas affected, the baseline visibility and the change in visibility resulting from each policy option.
Sometimes data or modeling constraints will prohibit quantifying significant benefit endpoints. In
these cases, it is useful to quantify changes in environmental stressors or measures of
environmental quality that would lead to benefits. These changes can be informative in the overall
characterization of benefits even if they cannot be aggregated with benefit endpoints.
Analysts should consider the following recommendations when quantifying changes in benefit
endpoints.
Ensure endpoints are appropriate for benefits evaluation. A principal role of the
economist at this stage is to ensure that the endpoints are characterized in ways that are
consistent with principles of economic analysis and the specific models used for benefits
analysis. They should also be characterized in a manner that avoids double-counting. Focus
on the needs of economic analysis is particularly important at the early stages of ecological
or human health risk assessments, and it is generally useful for economists to be part of a
cross-disciplinary team for planning and scoping these assessments.8 The ability to
monetize or even quantify benefits analysis may be limited if effects are described too
broadly, overlap with other benefits categories, cannot be linked to human well-being, or
8 See, for example, the EPA's Framework for Human Health Risk Assessment to Inform Decision Making (U.S. EPA
2014).
7-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
are otherwise incompatible with economic analysis. Text Box 7.2 provides a more detailed
discussion on integrating risk assessment and economics.
Consider how behavior affects benefit endpoints. One area where economists may lend
unique insights at this stage is in assessing how endpoint quantification is affected by
behaviors in the baseline and potential behavioral changes from the policy. These behaviors
often drive, for example, how and how much individuals are exposed to environmental
contaminants. Changes in behavior due to changes in environmental quality (e.g., staying
indoors to avoid detrimental effects of air pollution) can be significant, and economists need
to ensure they are considered in benefits analysis.
Emphasize quantification over qualitative description. Qualitative descriptions are
useful, but benefits endpoints that are not first quantified can generally not be monetized.
The result is that these endpoints have an effective weight of zero in the total benefits
calculation. Even for highly uncertain benefits zero is not usually the best quantitative
weight, and available evidence can often be used to produce some estimate that is more
accurate than assuming these effects do not occur (McGartland etal. 2017, OMB 2023).
: i the Monetary Value of th Endpoints
The next step is to estimate the monetary value to all affected individuals of the quantified benefit
endpoints to obtain the total social benefits of each policy option. This starts with identifying
valuation estimates for quantified benefit endpoints. Importantly, it may not be sufficient to
multiply a change in endpoint by a single value for that endpoint, particularly in the presence of
uncertainty or nonlinearities; valuation must be guided by economic theory (Section 7.2). For
estimating total benefits, it is typical to use a representative agent approach, where values are
calculated for an "average" or representative individual in the relevant population and then
multiplied by the number of individuals in that exposed population.9
When estimating monetary value of effects, analysts should:
Determine which valuation methods are best suited for each endpoint. When possible,
the value estimate should be based on willingness to pay (WTP), but other measures (e.g.,
cost of illness) may be used when there are no available WTP estimates. Valuation methods
are not unique to specific endpoints, and often a given endpoint can be valued through
several methods. Table 7.1 shows general benefit categories, examples of specific benefit
endpoints and associated valuation methods commonly used. Sometimes time and
resources may be available to conduct original research using these methods, but more
often the analysis will need to draw upon existing value estimates in a process called benefit
transfer. Section 7.3 provides details on valuation methods. Benefit transfer is described in
Section 7.4.
9 Though a representative agent approach is often used, models may be available to incorporate heterogeneity.
This can be especially useful for distributional analyses.
7-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 7.2 - Coordinating Economic Analysis and Risk Assessment
Because economists rely on risk assessment outcomes as key inputs into benefits analysis, it is
important to coordinate risk assessment and economic valuation. Health and ecological risk
assessments are designed to support the setting of standards or to rank the severity of different
hazards. However, measures from these assessments can be difficult or impossible to
incorporate into benefits analyses for several reasons. First, the measures may not be
probabilistic expressions of risk, but instead indicate how exposures compare to reference
levels that are not associated with any quantitative level of risk. It may be that the modeled
endpoints cannot be directly related to health outcomes or ecological services that can be valued
using economic methods. Also, risk assessments sometimes focus on outcomes near the tails of
the exposure and/or risk distribution for highly sensitive endpoints, leading to biased benefits
estimates if extrapolated to the general population.
As described in the EPA's Ecological Benefits Assessment Strategic Plan (U.S. EPA 2006) and
Framework for Human Health Risk Assessment (U.S. EPA 2014), coordination between economic
analysis and risk assessment should begin early in the planning process for any risk
assessments, starting with the Planning and Scoping and Problem Formulation stages where a
conceptual model is developed specifying key factors for the assessment including specific
endpoints to be addressed. The EPA's Generic Ecological Endpoints for Ecological Risk Assessment
(U.S. EPA 2016) contains specific guidance to assist ecological risk assessors and economists in
identifying ecological services that are amenable to economic analysis (U.S. EPA 2016).
Throughout the risk assessment process, economists can contribute information and insights
into how behavioral changes may affect realized risk changes. For example, if health outcomes in
a particular risk assessment are such that early medical intervention could reduce the chances
of illness, economists may be able to estimate the probability that individuals will seek
preventive care. Even in cases where the economists' contribution to the risk characterization is
not direct, economists and risk assessors should communicate frequently to ensure that
economic analyses are complete.
Specifically, risk assessors and economists should strive to:
Identify a set of human health and ecological endpoints that are economically meaningful, linked
to human well-being and are monetizable using economic valuation methods. Risk assessors
may be required to model more or different outcomes than if they were attempting to capture
only the most sensitive endpoint This also may require risk assessors and economists to
convert human health or ecological endpoints or indicators measured in laboratory or
epidemiological studies to effects that can be valued in the economic analysis.
Estimate changes in outcome probabilities (human health or ecological) or changes in
continuous outcomes (e.g., IQ) as exposure changes, rather than safety assessment measures
(e.g., reference doses) when possible. For human health, probabilistic dose-response assessment
may be useful for estimating outcome probabilities (WHO 2017, Chiu and Slob 2015).
Work to produce expected or central estimates of risk, rather than bounding estimates as in
safety assessments. Any expected bias in the risk estimates should be clearly described.
Attempt to estimate the timing between changes in emissions or exposures and associated
changes in health and ecological risks or outcomes. For health outcomes these time lags are
referred to as cessation lag (the time between reduced exposure and reduced health risks) or
latency (the time between increases in exposure and increased health risks.)
7-10 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Attempt to characterize the full uncertainty distribution associated with risk estimates. This will
contribute to a better understanding of potential regulatory outcomes and will enable
economists to include risk assessment uncertainty in a broader uncertainty analysis
uncertainty. The EPA's guidance and reference documents on Monte Carlo methods and
probabilistic risk assessment, including the EPA's Policy for Use of Probabilistic Analysis in Risk
Assessments (U.S. EPA 1997e), and the 1997 Guiding Principles for Monte Carlo Analysis (U.S. EPA
1997d) may be of interest
Identify valuation estimates and how they are to be used. Valuation estimates available
for benefits analysis will not always perfectly match the policy context being considered.
Benefit transfer is the exercise of both identifying valuation estimates that sufficiently relate
to the policy context and then transferring the results to the policy analysis. It is important
that this is done in ways that are consistent with economic reasoning and theory and it is not
always sufficient to simply apply a single, fixed value. Section 7.4 contains information on
both general steps for benefit transfer and specific transfer methods to consider.
Describe the source of estimates and confidence in those sources. Valuation estimates
always contain a degree of uncertainty; using them in a context other than the one in which
they were initially estimated can only increase that uncertainty. If many studies of the same
effect have produced comparable values, analysts can have more confidence in using these
estimates in their benefits calculations. In other cases, analysts may have only a single
study, or no directly comparable study, to draw from. In all cases, the benefits analysis
should clearly describe the sources of the valuation estimates and provide a qualitative
discussion of the reliability of those sources.
Avoid double-counting to the extent possible. Double-counting may arise for at least two
reasons. First, different valuation methods often incorporate different subsets of total
benefits, so some types of benefits may be counted twice when aggregating across values.
Second, endpoints may be defined in ways that overlap. For example, a human health
endpoint of avoided "emergency room visits" is likely to overlap with an endpoint of
avoided heart attacks, so valuing these endpoints separately and aggregating them would
introduce double-counting. It is important to avoid double-counting when possible and to
clearly acknowledge any potential overlap when presenting the aggregated results.
Characterize uncertainty. The analysis should include a quantitative uncertainty
assessment when possible using sensitivity analysis or other methods. As with other
aspects of the analysis, the depth and scope of this assessment should be commensurate
with the scale of the benefits analysis. In some cases, it may be sufficient to focus on a few
7-11 Guidelines for Preparing Economic Analyses | 3rd edition
-------
key parameters.10 Important considerations for analysis of uncertainty are provided in
Chapter 5, and principles for presenting information on uncertainty are in Chapter ll.11
The analysis should ultimately present both the aggregate monetized values as well as the value of
each specific benefit endpoint The monetized benefits estimate should be supplemented by
displaying benefits that could be quantified but lack valuation estimates, and a characterization of
benefits that can only be qualitatively described. When data or modeling limitations prevent
quantitative characterization of benefits endpoints, it can be useful to provide quantitative data
related to benefits (e.g., changes in stressors or environmental quality). Chapter 11 discusses the
presentation of information on benefits. When the policy or regulation under consideration is
expected to result in important feedbacks and interactions between various physical and economic
endpoints, analysts should consider whether available integrated approaches for analyzing the
specific policy are more appropriate than quantifying each specific endpoint in a separate analysis.
7.2 Economic Value and Types of Benefits
Economic valuation is based on the traditional economic theory of human behavior and
preferences, centered on the "utility" (or "welfare") that people realize from consumption of goods
and services, both in market and non-market settings. Core to this approach is the principle of
consumer sovereignty, in which values used for benefit-cost analysis (BCA) respect the preferences
individuals have for these goods and services rather than being based, for example, on the
preferences of the analyst or policy maker. Different levels and combinations of goods and services
provide different levels of utility for any one person. Also, because people have different
preferences, utility derived from sets of goods and services will vary across people.
Economic theory suggests that when goods and services are bought and sold in competitive
markets, optimizing consumers maximize their level of utility subject to constraints on their budget
by equating the ratio of the marginal utility (the utility afforded by the last unit purchased) of any
two goods that a person consumes with the ratio of the prices of those goods. If it were otherwise,
that person could reallocate their budget to buy a little more of one good and a little less of the
other good to achieve a higher level of utility.12
Utility is inherently subjective and cannot be measured directly; however, to assign "value" an
operational definition in benefits analysis, it must be expressed in a quantifiable metric. Dollars
10 If the benefits analysis applies statistical relationships to derive a change in health or ecological outcomes in
response to a given change in pollutant exposure (e.g., dose-response functions), the standard errors of the
relevant coefficients can be used to derive confidence intervals to help characterize uncertainty. If the statistical
relationship is characterized by a continuous function, then the confidence interval is proportional to the
magnitude of the change in exposure.
11 Uncertainty in benefits may affect valuation. For example, uncertainty about magnitude of a given risk
reduction, may affect WTPfor that benefit. It may be possible to elicit WTP that reflects uncertainty about risk,
however it may be more pragmatic to apply uncertainty, i.e., a probability of the outcome, to WTP elicited under
certainty. In principle, the difference between these two approaches is likely to be small (SAB, 2024).
12 Behavioral economics studies situations in which individuals' behavior is inconsistent with the standard
economic model (assuming rational choice). The implications of irrational behavior and inconsistent preferences
for welfare analysis are still an emerging area of economics (Just and Just 2016; Shogren and Taylor 2008).
Therefore, our discussion of benefits analysis adheres to the standard economics model of rational, utility-
maximizing behavior and consistent preferences. Chapters 4 (Section 4.4) and 5 (Section 5.5) provide more
discussion of behavioral economics and its implications for environmental policy design.
7-12 Guidelines for Preparing Economic Analyses | 3rd edition
-------
conveniently allow direct comparison of benefits to costs and summing of benefits across different
effects,13 but this choice for the unit of account has no theoretical significance. Table 7.1
summarizes the types of benefits most often associated with environmental protection policies and
provides examples of each benefits types as well as valuation methods often used to monetize the
benefits for each type.
The benefits of an environmental improvement are illustrated graphically in Figure 7.2 which
shows marginal abatement costs (MACs) and marginal damages (MDs) of emissions. Reducing
emissions from eo to ej produces benefits equal to the shaded area under the marginal damages
curve. Many environmental goods and services, such as air quality and biological diversity, are not
traded in markets. The challenge of valuing non-market goods that do not have prices is to relate
them to one or more market goods that do. This can be done either by determining how the non-
market good contributes to the production of one or more market goods (often in combination with
other market good inputs), by observing the trade-offs people make between non-market goods
and market goods, or by asking people directly about the tradeoffs they are willing to make. Section
7.3 provides a discussion of the various revealed and stated preference valuation methods. Of
course, some methods will be more suitable than others in a given scenario for a variety of reasons,
and some will be better able to capture certain types of benefits than others.
The economic valuation of an environmental improvement is the dollar value of the private goods
and services that individuals would be willing to trade for the improvement at prevailing market
prices. The willingness to trade compensation for goods or services can be measured either as
willingness to pay (WTP) or willingness to accept (WTA). WTP is the maximum amount of money an
individual would voluntarily pay to obtain an improvement. WTA is the least amount of money an
individual would accept to forego the improvement14 The key theoretical distinction between WTP
and WTA is their respective reference utility levels. For environmental improvements, WTP uses
the level of utility without the improvement as the reference point while WTA uses the level of
utility with the improvement as the reference point. (Freeman etal. 2014).15
Economists generally expect that the difference between WTP and WTA will be negligible, provided
the values are small relative to household wealth and substitutes are available for the market or
13 Because an individual's utility is unobservable, it cannot be measured in a cardinal sense. However,
economists assume that consumers make choices based on whether one bundle of goods is preferred to another,
but not necessarily by how much. In other words, consumers respond to changes in prices and income by
ordinally ranking consumption bundles using preference relationships. If an individual's preference relationships
display nonsatiation (more is better) and substitutability (if one good in a bundle is decreased, it is possible to
increase another good to make the consumer indifferent) then they can be represented by an ordinal preference
function, or "utility function." While economists cannot observe changes in an individual's utility directly, they
can observe income and consumption decisions at various prices that reflect changes in the ordinal utility
function and can then compute a money-based measure of these utility changes. This money-based measure is
the individual's "willingness to pay" or "willingness to accept" described below. More detail on the development
of this utility function can be found in Just et al. (2005) and Freeman et al. (2014).
14 For simplicity, the discussion in this section is restricted to the case of environmental improvements, but
similar definitions hold for environmental damages. For a more detailed treatment of WTP and WTA and the
closely related concepts of compensating variation, equivalent variation and Hicksian and Marshallian consumer
surplus, seeHanley and Spash (1993), Freeman etal. (2014), Just etal. (2005), and Appendix A of these
Guidelines.
15 See Freeman et al. (2014) for a discussion of how WTP and WTA may be associated with property rights. OMB
Circular A-4 also suggests that WTP and WTA are associated with different views of property rights and notes
associated issues for benefit transfer.
7-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
non-market goods in question (Willig 1976; Hanemann 1991). However, there may be instances in
which income and substitution effects are important (such as for some environmental goods) and
lead to large disparities between WTP and WTA.16 Ultimately, economists use the valuation
estimates to assess policy outcomes by applying the Kaldor-Hicks compensation test (see Appendix A). In
short, the test asks whether hypothetical ly the gainers from a policy could fully compensate the losers
and still be better off and conversely whether the losers could pay the winners to avoid the change
altogether and still be as well off. Since WTP is a consistent measure for this test and to simplify the
presentation, the term WTP is used throughout the remainder of this chapter to refer to the
underlying economic principles behind both WTA and WTP.
Figure 7.2 - Benefits of an Environmental Improvement
Emissions
WTP for environmental quality can also be non-linear. For example, Figure 7.2 illustrates a case in
which marginal damages increase with emissions. When this occurs, it is important to account for
baseline environmental quality when valuing the benefits of incremental improvements. Otherwise,
inconsistent results can occur when estimating the benefits from a series of separate actions. In
addition, sometimes environmental regulations yield relatively small average changes in health or
the environment that may not be noticeable to the public until multiple regulations have achieved a
large aggregate improvement. Just as it is important to account for small average costs imposed by
regulations which can be economically significant when aggregated over a sufficiently large
population it is conceptually correct to account for even very small improvements in
environmental quality. Chapter 5 provides more discussion of analyzing multiple related rules. Text
Box 7.6 in Section 7.4 discusses the issue of estimating multiple improvements in environmental
quality using benefit transfer.
16 For more information see Appendix A and Hanemann (1991 J, Also, Kim etal, 2015, Freeman etal. (2014), and
Horowitz and McConnell (2003) discuss and evaluate various explanations for the disparity between WTP and
WTA, and other studies have estimated the size of the disparity, e.g., Tuncel and Hammitt (2014), and Kniesner,
etal. (2014).
7-14 Guidelines for Preparing Economic Analyses | 3rd edition
-------
The types of benefits that may arise from environmental policies can be classified in multiple ways
(Freeman etal. 2014). As shown in Table 7.1, these Guidelines organize benefits into the following
categories: human health improvements, ecological improvements and other benefits.
In addition, commonly used valuation methods are provided for reference. The list is not meant to
be exhaustive, but rather to provide examples and commonly used methods for estimating values.
The sections below provide a more detailed discussion of each of the benefit categories listed in
Table 7.1.
7,2,1 Human Health Improvements
Human health improvements from environmental policies include effects such as reduced mortality
rates; decreased incidence of non-fatal cancers, chronic conditions and other illnesses; and reduced
adverse reproductive or developmental effects. These categories of outcomes are discussed
separately below.
Generally, it is good practice to fully characterize both the nature of the risk and the affected
populations in benefits analysis, including the age distribution of the affected population. This is
helpful not only to evaluate the best approach for valuing health benefits, but to communicate
clearly with decision makers about who is affected and how they are affected.
vitality
Some U.S. Environmental Protection Agency (EPA) policies will lead to reductions in human
mortality risks due to health conditions such as cancer or cardiovascular disease. In considering the
impact of environmental policy on mortality risk, it is important to remember that environmental
policies do not protect specific, identifiable individuals from death due to environmental causes.
Rather, they generally lead to small reductions in the probability of death for many people.
The value of the mortality risk reductions reflects estimates of individuals' WTP for these small
reductions in the risk of dying. When aggregated over the affected population, this value has
typically been referred to as the "value of statistical life" (VSL) although other terms have been used
(Simon et al. 2019). Regardless of terminology, it is important to recognize that it represents the
tradeoff between wealth or income and small changes in mortality risks and is not the value of life
itself.
For consistency and added transparency across analyses, EPA policy is to apply a single VSL
estimate for the calculation of benefits of mortality risk reductions experienced by all affected
populations associated with all EPA programs and policies. Appendix B describes this
recommended value, its distribution and derivation, and details its application. To reduce public
confusion and misunderstanding, analysts should not use the misleading term "value of life" in
Agency analyses as that term does not accurately describe what the VSL represents.
As discussed in Appendix B, analysts should address the impact of risk and population
characteristics on the VSL qualitatively. In addition, analysts should account for timing
considerations, including:
The effects of latency: delayed manifestation of health or other effects;
Cessation lags: time frame between a reduction in exposure to an environmental
contaminant and the reduced risk to health; and
7-15 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Income growth over time, discounting appropriately where warranted.17
Valuing mortality risk changes in children is particularly challenging. The EPA's Handbook for
Valuing Children's Health Risks (U.S. EPA 2003) provides some information on this topic, including
key benefit transfer issues to consider when using adult-based studies. In addition, OMB's Circular
A-4 advises:
"For regulations where health gains are expected among both children and adults (...) the
monetary values for children should be at least as large as the values for adults (for the
same probabilities and outcomes) unless there is specific and compelling evidence to
suggest otherwise" (OMB 2023, p.51).
Reviews of the literature by Gerking and Dickie (2013) and Robinson et al. (2019) provide support
for this position.
Methods for Valuing Mortality Risk Changes
Because individuals make risk-wealth trade-offs in different contexts, the value of mortality risk
changes can be estimated using a variety of data sources and modeling approaches. The estimate
recommended in Appendix B is derived from a combination of hedonic-wage and stated preference
studies. In the hedonic wage or wage-risk method, value is inferred from the income-risk trade-offs
made by workers for risks faced on the job. Stated preference studies, in which income-risk trade-
offs are solicited directly through surveys, are also used to estimate WTP for reduced mortality
risks. Key considerations in these studies include the extent to which individuals know and
understand the risks involved, and the ability of the study to control for aspects of the actual or
hypothetical transaction that are not risk-related.
There are additional methods that can be used to derive information on risk trade-offs. Averting
behavior studies value risk changes by examining purchases of goods that can affect mortality risk
(e.g., bicycle helmets). However, isolating the portion of the purchase price associated with
mortality risk reductions from other benefits or joint products provided by the good is a
challenging hurdle for this literature. Also, of potential importance is short term avoidance
behavior altering one's activities, including the timing and frequency of activities, to reduce
exposure.18 Another approach is to examine trade-offs between types of risks to estimate relative
preferences for risk reduction. This approach may make the valuation task more manageable for
the respondent but requires multiple steps to obtain a risk-dollar tradeoff.19
Important Considerations
The analyst should keep three important considerations in mind when estimating mortality
benefits (each described in more detail below):
Characterizing and measuring mortality effects;
17 Assumptions about income growth should be consistent throughout the economic analysis. This includes, to
the extent feasible, consistency between income growth assumptions and discounting. See Chapter 6 for more
information on discounting, generally, and Section 6.5 for consistency between income growth and discounting.
18 See Blomquist (2004) for a review of averting behavior studies and Graff Ziffm and Neidell (2013) for a
discussion of short-term averting behavior.
19 See Nielsen, et al. (2019) for an overview and application of risk-risk trade-off method.
7-16 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Heterogeneity in risk and population characteristics; and
The timing of health risk changes.
Characterizing and Measuring Mortality Effects
Although reduced mortality risks associated with an environmental policy are typically small, they
are generally aggregated over the affected population and reported in terms of "statistical lives."
Suppose, for example, that a policy affects 100,000 people and reduces the risk of premature
mortality in the coming year by 1 in 100,000 for each individual. Summing these individual risk
reductions across the entire affected population shows that the policy leads to the equivalent of one
premature fatality averted, or one statistical life "saved," in the coming year.
An alternative metric seeks to capture the remaining life expectancy, or "quantity of life" saved
associated with the risk reductions (Moore and Viscusi 1988) and is typically expressed as
"statistical life years." Looking again at the policy described above for reducing risk in the coming
year, suppose the risks were spread over a population in which each individual had 20 years of
remaining life expectancy. The policy would then "save" 20 statistical life years (1 statistical life x
20 life years). In practice, estimating statistical life years saved requires risk information for
specific subpopulations (e.g., age groups or health status). Statistical life years may be used as an
outcome measure in cost-effectiveness analysis (Institute of Medicine (IOM) 2006). However,
consistent with past Science Advisory Board (SAB) advice, the use of a constant monetized value for
a statistical life year is not supported by the literature and is not recommended for benefits analysis
(U.S. EPA 2007).
Heterogeneity in Risk and Population Characteristics
The WTP to avoid mortality risks can vary both by risk characteristics and by the characteristics of
the affected population. Key risk characteristics include voluntariness (i.e., whether risks are
voluntarily assumed), timing (immediate or delayed), risk source (e.g., natural versus man-made)
and the causative event (e.g., cancer or trauma). Population characteristics include those generally
expected to influence WTP for any good (e.g., income and education) as well as those more closely
related to mortality risks such as baseline risk or remaining life expectancy, health status, risk
aversion and familiarity with the type of risk. The empirical and theoretical literature on the effect
of many of these characteristics on WTP is incomplete or ambiguous. For example, some studies
suggest that older populations are willing to pay less for risk reductions (Viscusi and Aldy 2007);
others find this effect to be small if it exists at all (Alberini et al. 2004). Still others suggest older
populations have higher WTP (Kniesner, Viscusi, and Ziliak 2006; Smith et al. 2004). Similarly,
some studies have found that reductions in fatal cancer risks garner a higher WTP than other kinds
of fatal risks (e.g., Viscusi et al. 2014) while others do not find evidence of a "cancer premium" (e.g.,
Hammittand Haninger 2010). Appendix B contains a more complete discussion of risk and
population characteristics and how they may affect WTP.
Although mortality risk valuation estimates used in economic analyses could reflect these
differences in WTP quantitatively with sufficient empirical evidence, Agency policy is to apply a
single VSL estimate to all populations and mortality risks and to qualitatively describe population
characteristics and risk attributes. One reason for this position is that the empirical evidence in the
literature on the relationship between WTP and the various population and risk characteristics is
inconclusive. In addition, population characteristics become less relevant for applications of VSL in
benefits assessments of national regulations affecting broad spectrums of the population.
7-17 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Timing of Health Risk Changes
Environmental contamination can cause immediate or delayed health effects. If individuals prefer
health improvements earlier in time rather than later, all else equal, then the WTP for reductions in
exposure to environmental pollutants will depend on when the resulting health risk changes will
occur. 20
The effects of timing on the present or annualized value of reduced mortality risk can be considered
in the context of a lifecycle consumption model with uncertain lifetime (Cropper and Sussman
1990; Cropper and Portney 1990; U.S. EPA 2007). In this framework reductions in mortality risk
are represented as a shift in the survival curve the probability an individual survives to future
ages as a function of current age which leads to a corresponding change in life expectancy and
future utility.
If the basis for benefit transfer is a marginal WTP for contemporaneous risk reductions, then
calculating the benefits of a policy with delayed risk reductions requires three steps:
1. Estimating the time path of future mortality risk reductions;
2. Estimating the annual WTP for all future years; and
3. Calculating the present value of these annual WTP amounts.
The first step should account for all the factors that ultimately relate changes in exposure to
changes in mortality risk as defined by shifts in the survival curve.
>rbidity
Morbidity benefits consist of reductions in the risk of non-fatal health effects ranging from mild
outcomes, such as headaches and nausea, to very serious illnesses such as cancer (see Table 7.1).
Non-fatal health effects also include conditions such as birth defects, low birth weight and reduced
cognitive function. Morbidity outcomes need not be so severe to prevent affected individuals from
participating in normal activities but are expected to affect quality of life and labor productivity or
earnings for workers (Graff Zivin and Neidell 2013, 2018). Availability of existing valuation
estimates for morbidity outcomes varies considerably, and the WTP to avoid many health outcomes
do not yet exist
WTP to reduce the risk of experiencing an outcome is the preferred measure of value for morbidity
effects. As described in Freeman et al. (2014), this measure consists of four additive components:
"Averting costs" to reduce the risk of illness;
"Mitigating costs" for treatments such as medical care and medication;
Indirect costs such as reduced earnings from paid work, or lost time maintaining a home
and pursuing leisure activities; and
Monetary equivalent of the disutility of illness (e.g., costs of discomfort, anxiety, pain and
suffering.)
Methods used to estimate WTP vary in the extent to which they capture these components. For
example, cost-of-illness (COI) estimates generally only capture mitigating and indirect costs and
omit averting expenditures and lost utility associated with pain and suffering. Consequently, COI
20 The description here focuses on mortality risk, but the same principles apply to non-fatal health risks.
7-18 Guidelines for Preparing Economic Analyses | 3rd edition
-------
estimates generally understate WTP to reduce a risk or avoid a given health effect Some studies
have estimated that total WTP can be two to four times as large as COI even for minor acute
respiratory illnesses (Alberini and Krupnick 2000). Still, no broadly applicable "scaling factor"
exists that relates COI to WTP.
Methods for Valuing Morbidity
Researchers have developed a variety of methods to value changes in morbidity risks. Some
methods measure the theoretically preferred value of individual WTP to avoid a health effect.
Others (e.g., cost of illness) do not measure WTP but can provide useful data; however, those data
must be interpreted carefully if they are to inform economically meaningful measures. Methods also
differ in the perspective from which values are measured (e.g., before or after the incidence of
morbidity), whether they control for the opportunity to mitigate the illness (e.g., before or after
taking medication) and the degree to which they account for the four components of total WTP set
out above. The three primary methods most often used to value morbidity in an environmental
context are stated preference (Section 7.3.2), averting behavior (Section 7.3.1.4) and COI (Section
7.3.1.5). Hedonic methods (Section 7.3.1.3) are used less frequently to value morbidity from
environmental causes.
Benefits analysis may also be informed by approaches that do not estimate WTP of reduced
morbidity directly. As noted above, risk-risk trade-offs, for example, do not directly estimate dollar
values for risk reductions, but rather, provide rankings of relative risks based on consumer
preferences. Risk-risk trade-offs can be linked to WTP estimates for related risks.21
Other methods for valuing morbidity outcomes suffer from certain methodological limitations
and are therefore generally less useful for policy analysis. For example, health-state indices, such
as Quality Adjusted Life Years (QALYs) are composite metrics that combine information on
quality and quantity of life lived under various scenarios often used for cost-effectiveness (CEA)
or cost-utility analyses (CUA) (see Section 7.5.2.1). While appropriate for use in CEA or CUA,
these measures are consistent with WTP measures only under very strict conditions that
generally do not hold in practice and should not be used for deriving monetary estimates for use
in BCA (Blechrodt and Quiggin 1999; Hammitt 2003; IOM 2006). Another commonly suggested
alternative is jury awards; these also generally should not be used in benefits analysis, for
reasons explained in Text Box 7.3.
Important Considerations
Two factors to consider, in addition to the heterogeneity in risk and population characteristics and
the timing of health risk changes discussed above, when estimating morbidity benefits are:
Characterizing and measuring morbidity effects; and
Third party costs.
21 EPA analyses have used risk-risk trade-offs for non-fatal cancers in conjunction with VSL estimates as one
method to assess the benefits of reduced carcinogens in drinking water (U.S. EPA 2005).
7-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Characterizing and Measuring Morbidity Effects
Key characteristics that will influence the valuation of morbidity effects are their severity,
frequency, duration and symptoms. Severity defines the degree of impairment associated with the
illness. Examples of how researchers have measured severity include daily limitations such as
"restricted activity," "bed disability" and "lost work."22 Severity can also be described using health
state indices that combine multiple health dimensions into a single measure.23 For duration, the
primary distinction is between acute effects and chronic effects. Acute effects are discrete episodes
usually lasting for a limited time period, while chronic effects last much longer and are generally
associated with long-term illnesses. The frequency of effects also can vary widely across illnesses.
Some effects, such as some gastrointestinal illness are one-time events that are unlikely to recur.
Other effects, such as asthma, do recur or can be exacerbated regularly, causing disruptions in
work, school or recreational activities.
For chronic conditions or more serious outcomes, morbidity effects are usually measured in terms
of the number of expected cases of illness. Given the risks faced by each individual and the number
of people exposed to this risk, an estimate of "statistical cases" can be defined analogously to
"statistical lives." In contrast, morbidity effects that are considered acute or mild in nature can be
estimated as the expected number of times a particular symptom associated with an illness occurs
within an affected population over a period of time (e.g., annually), where individual members of
the population may experience the effect more than once. These estimates of "symptom days" may
be used in benefits analysis when appropriate estimates of economic value are available, although a
richer characterization of combinations of symptoms, severity, duration and episode frequency
would be an improvement over much of the existing literature. Some studies have attempted to
address these complexities in a more systematic manner (Cameron and DeShazo 2013).
Third Party Costs
The widespread availability of health insurance and paid sick leave shifts some of the costs of
illness from individuals to others. While this cost-shifting can be addressed explicitly in COI studies,
it may lead to problems in estimating total WTP especially in stated preference studies. If the
researcher does not adequately address these concerns, individuals may mis-state their WTP,
assuming some related costs will be borne by others. Some stated preference studies are designed
to avoid capturing third party or insurance costs in which case the results would be additive to COI.
Regardless, to the extent third party costs represent diversions from other uses in the economy,
they represent real costs to society and should be accounted for in the analysis.
22 As Cropper and Freeman (1991) note, these descriptions are essentially characterizations of a behavioral
response to the illness. Lost workdays, for example, in some cases require a decision on an individual's part not to
go to work due to illness. Such a response may depend upon various socioeconomic factors as well as the physical
effect of the illness.
23 The difference in the indices is intended to reflect the relative difference in disutility associated with
symptoms or illnesses. There are serious questions about the theoretical and empirical consistency between these
"health-related quality of life" index values and WTP measures for improved health outcomes (Hammitt 2002).
Still the inclusion of some aspects of these indices may prove useful in valuation studies (Van Houtven et al.
2006). Examples of economic analyses that have employed some form of health state index include Desvousges et
al. (1998) and Magat et al. (1996).
7-20 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 7.3 - Non-Willingness to Pay Measures
As described earlier, WTP is a valid measure of economic value because it can be used in
potential compensation tests of Kaldor and Hicks. Sometimes, however, other measures are
suggested for use in benefits analysis that are not appropriate. Measures of economic value that
do not measure WTP and cannot be related to changes in utility are not valid for use in benefits
analysis. Three common examples of such values are replacement cost, proxy cost and jury
awards.
Replacement cost. A common consequence of environmental deterioration is damage to
assets. Some analysts suggest that the economic value of the damage is the cost of replacing the
asset In the context of BCA, this is not generally true. It is only true if: (1) damage to the asset is
the only cost incurred; and (2) the least expensive way to achieve the level of satisfaction
realized before the deterioration would be to replace the asset (Freeman, Herriges, and Kling
2014). If the first condition is not met, consideration of replacement costs may be useful but
should be combined with assessments of other costs. If the second condition is not met,
replacement costs might overestimate the consequences. Suppose that water pollution kills fish
in a pond. Replacing those fish with healthy, edible ones might prove extremely expensive; the
pond might need to be dredged and restocked. However, people who are no longer able to catch
fish in the pond might be compensated simply by giving them enough money to purchase
substitutes in the market.
Proxy costs. A closely related concept to replacement cost is the cost of a substitute for the
damaged asset Ideologist H.T. Odum (1996) calculated the number of barrels of petroleum
required to provide the energy to replace the services of wetland ecosystems. However, since
there is no reason to suppose that people would be willing to pay for oil to replace services of
damaged wetlands, this number is economically irrelevant. A similar argument can be made
against the interpretation of "ecological footprints" as an estimate of economic consequences
(Wackernagel and Rees 1996). Dasgupta (2002) interprets these approaches as single-factor
theories of value, fallacies that were disproved in general by Samuelson's (1951) "non-
substitution theorem."
Jury awards. Attempts are sometimes made to value environmental improvements using jury
awards. Using jury awards in this way may prove problematic for several reasons. First, cases
only go to trial if both sides prefer the expected value of an adjudicatory outcome to the
certainty of a pre-trial settlement. Cases that go to juries are "atypical" by definition. Second,
since adjudication does not always occur and can never be infallible, jury awards often do, and
arguably should (Shavell 1979), embody "punitive" as well as "compensatory" elements. Juries
make examples of guilty defendants to try to deter others from committing similar offenses. For
this reason, jury awards may overstate typical damages. Finally, jury awards are ex-post
measures based on a known outcome, not the probability of experiencing an adverse event
These estimates are not appropriate for application to ex-ante evaluation of the value
associated with a statistical probability.
7-21 Guidelines for Preparing Economic Analyses | 3rd edition
-------
C. C >| h - nefits
Many EPA policies will produce ecological benefits by enhancing the delivery of ecosystem services,
defined here as "direct or indirect contributions that ecosystems make to the well-being of human
populations" (Thompson et al. 2009). Examples of EPA policies affecting ecosystem services
include: reducing acid precipitation that may acidify forests and freshwater ecosystems; controlling
pesticides and other environmental contaminants that affect pollinators such as bees, as well as
predators of pests; reducing nutrient pollution from municipal wastewater treatment plants, septic
systems, fertilizer and manure runoff, and atmospheric deposition that may lead to changes in the
composition and attributes of receiving water bodies.
In each of these examples, environmental regulation may not directly affect goods or services in the
household utility function. Instead, they affect ecological inputs into the processes that generate
such goods and services. The valuation of ecosystem services is not fundamentally different than
the valuation of other productive assets on which our economy depends (Polasky 2008; Barbier
2012, OMB 2024). The relevant endpoints are goods or services that enter the household utility
function directly. Ecosystem services that contribute to the production of those assets, but are not
directly valued by households, should be recognized as inputs in an ecological production function
(EPF) and monetization should value their marginal product Making the distinction between final
ecosystem services and ecological inputs and identifying the relevant linkages is a challenging task
facing analysts of environmental policy (Boyd and Banzhaf 2007).
7,2,2 t L"> ologic M R ti -.n Rjfm tions
An ecological production function is a description of how ecosystems combine inputs to produce
ecosystem services that consumers enjoy directly or are used in the production of goods or services
that are enjoyed by consumers. The natural science literature provides guidance for some cases on
the form of the ecological production function (MacArthur and Wilson 1967; Kingsland 1985) and
numerous examples (Hamel etal. 2015; Reddy etal 2015; Kremen et al. 2007; Jaramillo etal. 2010).
Knowledge of the relevant ecological linkages is essential to predicting the effects of environmental
policies on ecosystem service provision and an economic analyst will likely benefit from
collaborating with ecologists or other natural scientists to predict the effects of the proposed policy.
The Agency's Ecological Benefits Assessment Strategic Plan describes an interdisciplinary approach
for conducting ecological benefits assessments (U.S. EPA 2006). To familiarize themselves with
"benefit relevant indicators" of ecological endpoints that may be affected by policy measures,
analysts may also wish to consult National Ecosystem Services Partnership (2016) or National
Ecosystem Services Classification System (U.S. EPA 2015a).
There are several sources an analyst might consult for potentially useful ecological production
functions. In addition to searching the scientific literature on the topic of interest, some large-scale
research ventures maintain suites of models of pollination, storm protection, pollution treatment,
groundwater recharge and other phenomena. The Natural Capital Project, for example, maintains
19 models of Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST). Another suite of
such models is the Artificial Intelligence for Ecosystem Services platform (ARIES). Bagstad etal.
(2012) review and compare InVEST and a number of other models that might be used to model the
generation of ecosystem services. In addition, the EPA's Office of Research and Development has
compiled lists of potentially useful models in its Ecoservice Models Library (ESML 2024).
Using an "off-the-shelf" ecological production function in benefits analysis may have the
disadvantage of not being tailored to the specific application, but the advantages could include:
7-22 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Some EPFs have been extensively peer-reviewed and found to be of practical value;
Using an existing EPF may save considerable time and effort;
In some instances, it may be possible to calibrate the parameters of an ecological production
function, even if the parameters are unknown, using readily available data or summary
statistics for the policy case (e.g., Massey et al. 2017);
Alternatively, it may be possible to develop useful bounding results that hold regardless of
particular parameter values (e.g., Simpson etal. 1996; Simpson 2016).
Even when using established and tested ecological production functions, there is no substitute for
substantive familiarity with the subject matter. The analyst should understand the logic behind the
form of the ecological production function and consult with an expert on the subject (often a
biologist or other natural scientist) before deciding to adopt one in the work. Another consideration
is that arguments of the production function should be relevant to the problem the analyst is
addressing. Many ecosystem service models relate habitat area to the provision of a service; this
may not inform questions of how pollutants affect the provision of the service.
Ecological production functions sometimes exhibit what may seem to be counterintuitive effects.
Consider as an example, nutrients (primarily reactive nitrogen and phosphorus) from municipal,
agricultural or other sources that may enter water bodies. The term "eutrophication" describes the
consequences of excessive delivery of nutrients. Marine biologists have documented that, in some
circumstances, increased nutrient availability may enhance some desirable endpoints, such as the
support of larger populations of fish caught by commercial and recreational fishers (e.g., Breitburg
et al. 2009). In another example, bees that carry pollen between orange groves are a necessary
component of orange productions, however, carrying pollen between different groves may at times
hybridize fruits resulting in a lower value crop (Sagoff 2011). These examples are unusual, but they
underscore a point: preserving systems in, or restoring them to, a more "natural" state may not
always enhance the value of the services they provide.
7,2,2 J F -merit" Using Ecological Production Functions
Knowing only the ecological production function is generally insufficient to conduct economic
valuation. While ecological production functions are analogous to production functions that are a
staple of textbook microeconomics, they often differ in one important respect: the inputs and
outputs of ecological production functions are often not traded in markets (U.S. EPA 2009b).
Consequently, rather than being able to observe prices, we must infer them using the tools of
nonmarket valuation. Massey et al. (2006), Newbold and Massey (2010), Smith and Crowder
(2011), and Finnoff and Tschirhart (2011) exemplify how these linkages can be made for
commercial fisheries and recreational anglers.
Monetizing the ecological benefits of environmental regulations using ecological production
functions proceeds in three phases (Bateman, 2012). The first is to project changes in the ecological
inputs caused by the regulation. This phase may require its own extensive modeling effort such as
hydrological models that predict the effect of land use changes on nutrient and sediment loadings to
lakes, rivers and streams. The second phase employs the ecological production function to project
how the changes in those inputs affect the provision of final ecosystem services. To use the nutrient
and sediment pollution example again, this would require a model of aquatic ecosystems to project
changes in environmental goods that people value such as fish to catch and water amenities like
clarity and odor. Finally, changes in final ecosystem services are valued using nonmarket valuation
methods. Revealed and stated preference approaches to nonmarket valuation are described in
detail in Sections 7.3.1 and 7.3.2. When resource constraints prevent an original nonmarket
7-23 Guidelines for Preparing Economic Analyses | 3rd edition
-------
valuation study, benefit transfer can be used to apply values estimated in other contexts; See
section 7.4 for a detailed discussion with caveats.
7,2,2.' K:- nerir- K/tii ration When tl - Dlogical Production Function is Not
Known
When ecological production functions are not known, it may be easier and/or more defensible to
infer ecosystem service values from other relationships. If only the changes in the ecological inputs
are known, these can be used in revealed preference approaches to valuation by observing their
impact on complementary market behaviors.
Fundamental results in economics establish that these production relationships may be
equivalently expressed as profit functions and that profits may be capitalized into the value of
assets such as advantageously located property. As such, hedonic valuation methods are frequently
proposed for ecosystem service valuation (see, for example, Swinton et al. 2007; Bishop and
Timmins 2018). Several researchers have conducted hedonic property value studies to estimate
values of assets such as forest cover (Kim and Johnson 2002; Tyrvainen and Miettinen 2000;
Mansfield etal. 2005; Sander etal. 2010), wetlands (Tapsuwan etal. 2009; Mahan etal. 2000;
Woodward and Wui 2001; Bin and Polasky 2005), or other varieties of "open space" (Sander and
Polasky 2009; Cho etal. 2006; Irwin and Bockstael 2002; Irwin 2002; Thorsnes 2002).
The estimation of recreational demand, or, more generally, locational choice models (e.g., Kuminoff
et al. 2013) are based on similar underlying principles: choices of where to visit or live are made to
maximize utility (or profits) and the ecological attributes of an area affect such choices (McConnell
1990; Parsons 1991; Phaneuf etal. 2008). Hedonic price, recreational demand or locational choice
models may be regarded as "reduced form" representations of ecological production from which
the analyst can infer the values individuals ascribe to ecosystem services by observing the choices
they make, provided that the analyst can adequately control for potentially confounding factors.
These approaches are discussed further in Sections 7.3.1.2 and 7.3.1.3.
" C, 'sYh -1 E vnefii'.:
Other types of potential benefits from environmental policies include aesthetic improvements,
reduced material damages and other benefits resulting from changes that occur in response to a
regulation.
Aesthetic improvements include effects such as the improved taste and odor of tap water resulting
from water treatment requirements and enhanced visibility resulting from reduced air pollution.
Increased visibility due to improved air quality can be divided into two types of benefits: residential
visibility benefits and recreational visibility benefits. Improvements in residential visibility are
generally assumed to only benefit residents living in the areas in which the improvements are
occurring, while all households in the United States are usually assumed to derive some benefit
from improvements in visibility in areas such as national parks. The benefits received, however,
may decrease with the distance from the recreational area in which the improvements occur.24
Reduced materials damages include welfare impacts that arise from changes in the provision of
service flows from human-made capital assets such as buildings, roads and bridges. Materials
24 Boyle et al (2016) estimate WTPfor shifts in the distribution of visibility in national parks.
7-24 Guidelines for Preparing Economic Analyses | 3rd edition
-------
damages can include changes in both the quantity and quality of such assets. Benefits from reduced
material damages typically involve cost savings from reduced maintenance or restoration of soiled
or corroded buildings, machinery or monuments.
Some positive welfare effects resulting from regulation do not fit into the previous categories.
Specific examples include lower consumer expenditures on fuel or electricity from regulations that
improve vehicle fuel economy or appliance energy efficiency, or reduced infrastructure
expenditures from regulations that encourage green infrastructure for stormwater management
Whether these effects are presented as cost savings or benefits is not important for the calculation
of net benefits. Section 5.5.2 discusses issues for analysts to consider when they estimate that a
regulation that strengthens environmental protection results in net private cost savings, which
would not typically be expected.
Methods and Previoi dies
Changes in the stock and quality of human-made capital assets are assessed in a manner similar to
their "natural capital" counterparts. Analytically, the valuation of reduced materials damages
parallels the methods for valuing the tangible end products from managed ecosystems such as
agriculture or forestry. For example, effects from changes in air quality on the provision of the
service flows from physical resources are handled in a fashion similar to the effects from changes in
air quality on crops or commercial timber stocks. The most common empirical applications involve
air pollution damages and the soiling of structures and other property.
Linking changes in environmental quality with the provision of service flows from materials can be
difficult because of the limited scientific understanding of the physical effects, the timing of the
effects and the behavioral responses of producers and consumers. An analysis of reduced materials
damages often begins with an environmental fate and transport model to determine the direct
effects of the policy on the stocks and flows of pollutants in the environment Then stressor-
response functions are used to relate local concentrations of pollutants to corrosion, soiling or
other physical damages that affect the production (inputs) or consumption (outputs) of the
material service flows. The market response to these impacts serves as the basis for the final stage
of the assessment, in which some type of structural or reduced-form economic model that relates
averting or mitigating expenditures to pollution levels is used to value the physical impacts. The
degree to which behavioral adjustments are considered when measuring the market response is
important, and models that incorporate behavioral responses are preferred to those that do not
Adams and Crocker (1991) provide a detailed discussion of this and other features of materials
damages benefits assessment. Also see the EPA's benefits analysis of household soiling for an
example that employs a reduced-form economic model relating defensive expenditures to ambient
pollution (U.S. EPA 1997f).
When other benefits result from changes in market goods, then demand analysis of the affected
market can be a useful approach. Non-market valuation approaches such as those discussed in the
remainder of this chapter may be required to measure welfare effects from changes in non-market
goods.
For goods bought and sold in undistorted markets, the market price indicates the marginal social
value of an extra unit of the good. Often, there are no markets for environmental goods. While some
natural products are sold in private markets, such as timber and fish, the analyst's concern will
7-25 Guidelines for Preparing Economic Analyses | 3rd edition
-------
typically be with non-market inputs, which are, by definition, not traded in markets.25 To overcome
this lack of market data, economists have developed a number of methods to value environmental
quality changes. Most of these methods can be broadly categorized as either revealed preference or
stated preference methods.
In cases where markets for environmental goods do not exist, WTP can often be inferred from
choices people make in related markets. Specifically, because environmental quality is often a
characteristic or component of a private good or service, it is sometimes possible to disentangle the
value a consumer places on environmental quality from the overall value of a good. Methods that
employ this general approach are referred to as revealed preference methods because values are
estimated using data gathered from observed choices that--when combined with several important
auxiliary assumptions (individuals have complete and stable preferences, are expected utility
maximizers, have all relevant information, etc.) reveal the preferences of individuals. Revealed
preference methods include production or cost functions, travel cost models, hedonic pricing
models and averting behavior models. This section also discusses COI methods, which are
sometimes used to value human health effects when estimates of WTP are unavailable.
In situations where no markets for environmental or related goods exist to infer WTP, economists
sometimes rely on survey techniques to gather choice data from hypothetical markets. The
methods that use this type of data are referred to as stated preference methods because they rely on
choice data that are stated in response to hypothetical situations, rather than on choice behavior
observed in the real world. Stated preference methods include contingent valuation, conjoint
analysis, and contingent ranking.
As noted in Chapter 8, although there is an expanding body of work that uses computable general
equilibrium (CGE) models to evaluate nonmarket goods (Smith etal. 2004; Carbone and Smith
2008), CGE models often lack a credible way to represent environmental externalities or the
benefits that accrue to society from mitigating them. As such, a CGE model's economic welfare
measure is typically incomplete and not a suitable means at present to capture the benefits of a
regulation.
As a general matter, revealed preference methods have the advantage that they are based on actual
tradeoffs and decisions made by individuals. Stated preference methods sometimes have the
advantage that the choice question can be tailored to obtain values that more closely align with the
needs of benefits analysis. Each of these revealed and stated preference methods is discussed in
detail below, starting with an overview of the method, a description of its general application to
environmental benefits analysis, and a discussion of issues involved in interpreting and
understanding valuation studies. The discussion concludes with a separate overview of benefit-
transfer methods.
It is important to keep in mind that research on all of these methods is ongoing. The limitations and
qualifications described here are meant to characterize the state of the science at the time these
25 There are examples in which environmental goods have been traded in markets. The Clean Air Act
Amendments ofl 990, for example, initiated a market in sulfur dioxide (SO2). Similarly, studies in the
computational economics literature have performed ex-ante calibrated analyses by constructing hypothetical
markets for environmental goods. The non-market good valuation is determined through shadow prices in
imposed regulation-induced scarcity in the environmental good. However, prices in such markets are determined
by ex-ante policy driven quantity constraints, and not through empirically based statistical methods by
considerations of marginal utilities or marginal products.
7-26 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Guidelines were written. Analysts should work with the National Center for Environmental
Economics (NCEE) to determine the usefulness of additional resources as they become available. In
practice, analyses will often need to draw upon values from multiple methods to value benefits.
Text Box 7.4 briefly describes original valuation studies using multiple methods conducted by the
Agency to estimate benefits of improved water quality in the Chesapeake Bay using many of the
methods discussed here.
'V-. f [ ealed Preference Methods and the * t of Illness
A variety of revealed preference methods for valuing environmental changes have been
developed and are widely used by economists. While these methods all use observable data
to estimate or infer value, they each have their own set of advantages and limitations. The
following common types of revealed preference methods are discussed in this section:
Production or cost functions;
Travel cost models;
Hedonic models; and
Averting behavior models.
This section also discusses the cost of illness (COI) approach to valuation. It is worth noting that
estimation approaches can span more than one method. For example, the random utility
maximization framework in discrete choice models is commonly applied to travel cost models.26
i i H '11 ion an r reactions
Discrete changes in environmental circumstances generally cause both consumer and producer
effects, and it is common practice to separate the welfare effects brought about by changes in
environmental circumstances into consumer surplus and producer surplus.27 Marginal changes can
be evaluated by considering the production side of the market alone.
Econom ndations of Production and Cost Functions
Inputs to production contribute to welfare indirectly. The marginal contribution of a productive
input is calculated by multiplying the marginal product of the input by the marginal utility obtained
from the consumption good, in which the input is employed in production. The marginal value of a
consumption good is recorded in its price. While marginal products are rarely observed, the need to
observe them is obviated when both inputs and outputs are sold in private markets because prices
can be observed.28 Environmental goods and services are typically not traded in private markets,
and therefore the values of environmental inputs must be estimated indirectly.
26 For an example of random utility maximization (RUM) and how it is applied to a travel cost approach, please
see the travel cost applications later in this revealed preference section. RUM is also commonly applied in stated
preference methods and its various estimation approaches.
27 See Appendix A for more detail.
28 The suitability of prices for welfare analysis depends on the structure of the market this is discussed in the
section titled, "Considerations In evaluating and understanding production and cost functions," within Section
7.3.1.1.
7-27 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 7.4 - Benefits Analysis of the Chesapeake Bay TMDL
In 2010, the EPA established the Chesapeake Bay (CB) total maximum daily load (TMDL), a
comprehensive "pollution diet" to restore clean water to the Bay and area streams, creeks and
rivers. The EPA's National Center for Environmental Economics (NCEE) was tasked with
assessing the TMDL's benefits in a multi-faceted analysis of recreational and aesthetic
amenities. NCEE began by conducting a scoping exercise to inform which detailed benefits
analyses to pursue. Categories yielding small benefits in previous analyses were shelved (U.S.
EPA 2002, 2009b). The EPA's CB Program Office provided data on water clarity and pollutant
loadings with and without the TMDL. NCEE engaged external experts on CB fisheries and water
quality to obtain their best professional judgments of potential stock size changes relative to
current water quality conditions, holding all other influences constant. NCEE used an extension
of the U.S. Geological Service's SPARROW (SPAtially Referenced Regression on Watershed
attributes) model to predict nutrient loadings and chlorophyll in lakes (Moore et al. 2011).
Summaries appear below. Estimates are not additive across studies; overlaps may exist among
homeowners, recreators and respondents.
Hedonic property value analysis: Walsh et al. (2017) used spatially explicit water quality data
paired with economic, geographic and demographic variables to analyze the value of water
clarity to home buyers using over 200,000 property sales in Maryland. Klemick et al. (2018)
then used meta-analysis to synthesize the value of clarity improvements in Maryland and to
transfer the results to properties in Delaware, Virginia and the District of Columbia. Together,
they found that predicted water clarity improvements from the TMDL would result in a 0.7-
1.3% increase in property value for waterfront homes. Properties farther from the water had
smaller effects. Total near-waterfront property values could increase by about $458 to $802
million from water clarity improvements, which is equivalent to an annualized value of $14 to
$56 million at discount rates of 3 and 7%.
Market analysis: Like many fresh goods, fish and shellfish are highly perishable; producers
cannot easily adjust supply in the short run to respond to changes in demand. Moore and
Griffiths (2018) developed a two-stage inverse demand model to describe how prices respond
to supply changes in other commodity groups. The model allowed NCEE to estimate consumer
welfare impacts of an increase in CB fish and shellfish harvests while allowing other areas'
harvests to act as substitutes. The estimated annual value of expected harvest improvements is
$14.2 million.
Fishing model: NCEE estimated benefits to recreational anglers using a linked participation
and site-choice recreation demand model. The model relied on historic catch rate data from the
Marine Recreational Fisheries Statistics Survey, an intercept survey that uses weights based on
historic visitation frequencies at each intercept site. The data were used to estimate a random
utility site-choice model and trip counts from respondent zip codes were used to estimate a
participation model conditional on the inclusive value of all sites as estimated by the site-choice
model. The resulting estimates of recreational fishing benefits range between $5.7 and $67.6
million per year.
Other recreation demand: NCEE used a recreation demand model to estimate the benefits
from other outdoor recreation activities using data on total visitor counts to national and state
parks in Maryland, Virginia and Delaware, supplemented with survey data on the number of
recreation trips taken to the CB area. The marginal effects of water quality on recreators' site
choices were estimated in a second-stage regression, using estimates of site-specific constants
from the first-stage site-choice model as the dependent variable and measures of average water
7-28 Guidelines for Preparing Economic Analyses | 3rd edition
-------
quality conditions and other fixed site attributes as explanatory variables. The estimated annual
outdoor recreation benefits (exclusive of recreational fishing) range from $120 to $321 million.
Stated preference survey: Moore, et al. (2018) conducted an SP survey linking forecasted
water quality changes to ecological endpoints to estimate use and nonuse values for aesthetic
and ecological improvements in the CB and watershed lakes. The survey estimated WTP for
improvements in water clarity; populations of three CB species (striped bass, blue crab, and
oysters); and the condition of freshwater lakes in the CB Watershed. They found that benefits to
watershed lakes and nonuse values account for a large proportion of total WTP and would
significantly affect the benefit-cost ratio of pollution reduction programs. Estimated benefits
from the projected environmental improvements after the TMDL range from $4.47 billion to
$7.79 billion per year.
Production possibilities can be represented in three equivalent ways:
As a production function relating output to inputs;
As a cost function relating production expenses to output and to input prices; and
As a profit function relating earnings to the prices of both output and inputs.29
The value of a marginal change in some environmental condition can be represented as a marginal
change in the value of production, as a marginal change in the cost of production, or as a marginal
change in the profitability of production.30 It should be noted, however, that problems of data
availability and reliability often arise. Such problems may motivate the choice among these
conceptually equivalent approaches, or favor of another approach.
Note that derivation of values on the margin does not require any detailed understanding of
consumer demand conditions. To evaluate marginal effects via the production function approach,
the analyst needs to know the price of output and the marginal product of the environmental input
To derive the equivalent measure using a cost function approach, the analyst needs to know the
derivative of the cost function with respect to the environmental input In the profit function
approach, the analyst needs to know the derivative of the profit function with respect to the
environmental input
In the statements above, note the emphasis that marginal effects are being estimated. Estimating
the net benefits of larger, non-marginal changes represent a greater challenge to the analyst In
general, this requires consideration of changes in both producer and consumer surplus.
Links Between Production and Hedonic and Other Models
A fourth way to estimate environmental effects on production possibilities is through the
profitability of enterprises engaged in production. The value of a fixed asset, such as a parcel of
land, is related to the stream of earnings that can be achieved by employing it in its most profitable
use. Its rental value is, therefore, equal to the profits that can be earned from it over the period of
use. The purchase price of the land parcel is equal to the expected discounted present value of the
29 Varian (1992) describes the relationships among these functions.
30 For a review of statistical procedures used for estimating production, cost, and profit functions see Berndt
(1991).
7-29 Guidelines for Preparing Economic Analyses | 3rd edition
-------
stream of earnings that can be realized from its use over time. Therefore, the production, cost and
profit function approaches described above are also equivalent to inferences drawn from the effects
of environmental conditions on asset values. This fourth approach is known as "hedonic pricing,"
and will be discussed in detail in
Section 7.3.1.3. It is introduced now to show that production, cost, or profit function approaches are
generally equivalent to hedonic approaches.
"Production" as a term is broad in meaning and application, especially with regard to hedonic
pricing. While businesses produce goods and services in their industrial facilities, one might also
say that developers "produce" housing services when they build residences. Therefore, hedonic
pricing approaches can measure the value of the environment in "production," whether they are
focusing on commercial or residential properties. Similarly, households may "produce" their health
status by combining inputs such as air and water filtration systems and medical services along with
whatever environmental circumstances they face. Or they "produce" recreational opportunities by
combining "travel services" from private vehicles, their own time, recreational equipment
purchases and the attributes of their destination. Much of what is discussed elsewhere in this
section is associated with this type of production analysis. This is not to say that estimation of
production, cost or profit functions is necessarily the best way to approach such problems, but
rather, that all of these approaches are conceptually consistent
General Application of Production and Cost Functions
Empirical applications of production and cost function approaches are diverse. Among other topics,
the empirical literature has addressed the effects of air quality changes on agriculture and
commercial timber industries. It also has assessed the effects of water quality changes on water
supply treatment costs and on the production costs of industry processors, irrigation operations
and commercial fisheries.31 Production, cost or profit functions have found interesting applications
to the estimation of some ecological benefits. Probabilistic models of new product discovery from
among diverse collections of natural organisms can also be regarded as a type of "production."
Finally, work in ecology points to "productive" relationships among natural systems that may yield
insights to economists as well.
Considerations in Evaluating and Understanding Production and Cost Functions
The analyst should consider the following factors when estimating the values of environmental
inputs into production:
Data requirements and implications. Estimating production, cost or profit functions requires
data on all inputs and/or their prices. Omitted variable bias is likely to arise absent such
information and may motivate the choice of one form over another. Economists have typically
preferred to estimate cost or profit functions. Data on prices are often more complete than are data
on quantities, and prices are typically uncorrelated with unobserved conditions of production,
whereas input quantities are not.
The model for estimation. Standard practice involves the estimation of more flexible functional
forms (i.e., functions that can be regarded as second-order approximations to any production
31 See, for example, Price and Heberling (2018) and the studies reviewed therein on source water quality.
7-30 Guidelines for Preparing Economic Analyses | 3rd edition
-------
technology). The translog and generalized Leontief specifications are examples.32 Estimation often
will be more efficient if a system of equations is estimated (e.g., simultaneous estimation of a cost
function and its associated factor demand equations), although data limitations may impose
constraints.
Market imperfections. Analysts should consider the impact of market imperfections and tax
distortions. Most analyses assume perfectly competitive behavior on the part of producers and input
suppliers and assume an absence of other distortions. When these assumptions do not hold, the
interpretation of welfare results becomes more challenging. While there is an extensive literature on the
regulation of externalities under imperfect competition, originating with Buchanan (1969), analysts
should exercise caution and restraint in attempting to correct for departures from competitive behavior.
The issues can become quite complex and there is typically no direct evidence of the magnitude of the
departures. In many circumstances it might reasonably be argued that departures from perfect
competition are not of much practical concern (Oates and Strassman 1984).
Perhaps a more pressing concern in many instances will be the wedge between private and social
welfare consequences that arise with taxation. An increase in the value of production occasioned by
environmental improvement typically will be split between private producers and the general public
through tax collection. The issues here also can become quite complex (see Parry etal. 1997), with
interactions among taxes leading to sometimes surprising implications. While it is difficult to give
general advice, analysts may wish to alert policy makers to the possibility that the benefits of
environmental improvements in production may accrue to different constituencies.
ivel Costs
Recreational values associated with an environmental improvement constitute a potentially large
class of use benefits (see Table 7.1 for examples). However, measuring these values is complicated
by the fact that the full benefits of recreation activities are rarely reflected in the price to access
them. Travel cost models address this problem by inferring the value of changes in environmental
quality through observing the trade-offs recreators make between environmental quality and the
cost of visiting sites. A heuristic example is choosing between visiting a nearby recreation location
with low environmental quality versus a more distant location with higher environmental quality.
The outcome of the decision of whether to incur the additional travel cost to visit the location with
higher environmental quality reveals information about the recreator's WTP for environmental
quality. Travel cost models are often referred to as recreation demand models because they are
most often used to value the availability or quality of recreational opportunities.
Econom ndation of Travel Cost Models
Travel cost models of recreation demand focus on the choice of the number of trips to a given site
or set of sites that a traveler makes for recreational purposes. Because there is generally no explicit
market or price for recreation trips, travel cost models rely on the assumption that the "price" of a
recreational trip is equal to the cost of visiting the site. These costs include both participants'
monetary cost and opportunity cost of time. Monetary costs include all travel expenses. For
example, when modeling day trips taken primarily in private automobiles, travel expenses would
32 See Coelli, etal (2005) for more details on the properties and estimation of a range of production functions.
7-31 Guidelines for Preparing Economic Analyses | 3rd edition
-------
include roundtrip travel distance in miles multiplied by an estimate of the average cost per mile of
operating a vehicle, plus any tolls, parking and admission fees.
A participant's opportunity cost of time for a recreational day trip is the value of the participant's
time spent traveling to and from the recreation site plus the time spent recreating since it is time
that could have been devoted to other activities. In most cases, onsite recreational time is assumed
to be constant across a recreator's choice alternatives and is, therefore, not included in the estimate
of travel costs.33 Although estimates of the opportunity cost of time ranging from zero to more than
100% of wage rates can be found in the literature, time spent traveling for recreational purposes is
generally valued at some fraction of an individual's full wage rate.34 The fraction of the wage rate
used is important because it directly affects estimates of willingness to pay. As the fraction of the
wage rate assumed to represent the opportunity cost of time rises, it causes total travel cost
estimates to rise, which in turn cause estimates of willingness to pay to also rise.
Most commonly in the recreation demand literature, researchers have used one-third of a person's
annual hourly wage as an estimate of participants' hourly opportunity cost of time, although
estimates of two-thirds of the full wage rate can also be found (Parsons 2003b; English, Leggett, and
McConnell 2015; Phaneuf and Requate 2017). Within that range, the U.S. Department of
Transportation (DOT) guidance recommends valuing recreational travel at 50% of the hourly
median household income for local travel and 70% for intercity travel (U.S. DOT 2016). A number of
researchers have also developed methods for estimating recreaters' opportunity cost of time
endogenously, although no one method has yet been fully embraced in the literature.35 Unless
compelling reasons for deviating from the standard wage rate assumptions are present, analysts
should generally rely on the standard one-third of the wage rate opportunity cost assumption when
estimating recreation travel in original studies. Conducting analyses using one-half of the wage rate
can also be justifiable in some cases when done in addition to the one third assumption as a way to
check the sensitivity of estimates to opportunity cost assumptions.
Even among studies that use the same fraction of the wage rate to estimate the opportunity cost of
time, care must still be taken in comparing estimates across studies. First, researchers in the
literature vary in their use of personal or household income in calculating opportunity costs.
Household income tends to be greater on average than personal income resulting in larger
opportunity cost estimates. Second, when researchers do not have recreators' self-reported
incomes they have often used population median or average income levels. Average income is
generally higher than median income because higher incomes in the tail of the distribution tend to
pull the average up. Lastly, in cases where household income is used, opportunity cost estimates
will depend on whether costs are assumed to accrue to adults and children or only to adults. The
literature is not clear on the preferred choices specification of opportunity costs, so the analyst
must use best professional judgement to decide what is best on a case-by-case basis.
33 If onsite time is assumed to be an additional choice variable, then estimation will require a model that
accounts for the decision of how long to recreate at a site. Examples of models investigating onsite time include
Bell and Leeworthy (1990); McConnell (1992); Larson (1993); Herman and Kim (1999); and Landry and
McConnell (2007).
34 For an in-depth discussion of the valuation of time in different contexts, see U.S. EPA (2020).
35 For examples, see McConnell and Strand (1981); Smith, Desvousges, and McGivney (1983); Bockstael et al.
(1987); McConnell (1992); McKean etal. (1995), Feather and Shaw (1999), Palmquist et al. (2010), Fezzi etal.
(2014); and Larson and Lew (2014).
7-32 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Hourly opportunity costs are multiplied by round trip travel time and time on-site to calculate a
person's full opportunity cost of time. Total travel costs are the sum of monetary travel costs and
full opportunity costs. Following the law of demand, as the cost of a trip increases the quantity of
trips demanded generally falls, all else equal. This means that participants are more likely to visit a
closer site than a site farther away.
While travel costs are the driving force of the model, they do not completely determine a
participant's choice of sites to visit. Site characteristics, such as parking, restrooms or boat ramps;
participant characteristics, such as age, income, experience and work status; and environmental
quality also can affect demand for sites. Changes in the measures of environmental quality are
generally the focus of economic analyses done in support of the regulatory decision-making
process. The identification and specification of the appropriate site and participant characteristics
are generally determined by a combination of data availability, statistical tests and the researcher's
best judgment. Ultimately, every recreation demand study strikes a compromise in defining sites
and choice sets, balancing data needs and availability, costs and time.36
General Application by Type of Travel Cost Model
Travel cost models can logically be divided into two groups: single-site models and multiple-site
models. Apart from the number of sites they address, the two types of models differ in several ways.
The basic features of both model types are discussed below.
Single-site models. Single-site travel cost models examine recreators' choices of how many trips
to make to a specific site over a fixed period of time (generally a season or year). It is expected that
the number of trips taken will increase as the cost of visiting the site decreases and/or as the
benefits realized from visiting increase. The price of close substitute sites could also affect
demand. Income and other participant characteristics act as demand curve shifters. For example,
avid outdoor recreators (fishermen or birders for example) may be more likely to take more trips
than non-avid recreators, all else equal. Most current single-site travel cost models are estimated
using count data models because the dependent variable (number of trips taken to a site) is a non-
negative integer. See Haab and McConnell (2003) and Parsons (2003a) for detailed discussions
and examples of recreation demand count data models.
Single-site models are most commonly used to estimate the value of a change in access to a site,
particularly site closures (e.g., the closure of a lake due to unhealthy water quality). The lost access
value due to a site closure is given by the area under the demand curve between a participant's
current price and the price at which trip demand falls to zero.37 Estimating the value of a change in
the cost of a site visit, for example the addition or increase of an admission fee, is another common
application of the model. Although it is possible with alternative data and model structures, single
site models are not generally used for valuing changes in site quality.
A weakness of the single-site model is its inability to deal with large numbers of substitute sites. If,
as is often the case, a policy affects several recreation sites in a region, then traditional single-site
models are required for each site. In cases with large numbers of sites, defining the appropriate
substitute sites for each participant and estimating individual models for each site can impose
overwhelming data collection and computational costs. Because of these difficulties, most
36 For a comprehensive treatment of the theoretical and econometric properties of recreation demand models,
see Phaneufand Smith (2005). Best practices are discussed in Lupi et al. 2020.
37 The price at which trip demand falls to zero is commonly called the choke price.
7-33 Guidelines for Preparing Economic Analyses | 3rd edition
-------
researchers have opted to refrain from using single-site models when examining situations with
large numbers of substitute sites.38
Multiple-site models. Multiple-site models examine a recreator's choice of which site to visit from
a set of available sites (known as the choice set) on a given choice occasion and, in some cases, can
also examine how many trips to make to each specific site over a fixed period of time. Compared to
the single-site model, the strength of multiple-site models lies in their ability to account for the
availability and characteristics of substitute sites. By examining how recreators trade the differing
levels of each site characteristic and travel costs when choosing among sites, it is possible to place a
per trip (or choice occasion) dollar value on site attributes or site availability for single sites or
multiple sites simultaneously.
The two most common multiple-site models are the random utility maximization (RUM) travel cost
model and Kuhn-Tucker (KT) system of demand models. Both models may be described by a
similar utility theoretic foundation, but they differ in important ways. In particular, the RUM model
is a choice occasion model while the KT model is a model of seasonal demand.
Random utility maximization models.39 In a RUM model each alternative in the recreator's
choice set is assumed to provide the recreator with a given level of utility, and on any given choice
occasion the recreator is assumed to choose the alternative that provides the highest level of utility
on that choice occasion.40 The attributes of each of the available alternatives, such as the amenities,
environmental quality and the travel costs, are assumed to affect the utility of choosing each
alternative. Because people generally do not choose to recreate at every opportunity, a non-
participation option is often included as a potential alternative.41 From the researcher's
perspective, the observable components of utility enter the recreator's assumed utility function.
The unobservable portions of utility are captured by an error term whose assumed distribution
gives rise to different model structures. Assuming that error terms have a type 1 extreme value
distribution leads to the closed-form logit probability expression and allows for maximum
likelihood estimation of utility function parameters. Using these estimated parameters, it is then
possible to estimate WTP for a given change in sites' quality or availability.
38 Researchers have developed methods to extend the single-site travel cost model to multiple sites. These
variations usually involve estimating a system of demand equations. See Bockstael, McConnell, and Strand
(1991) and Shonkwiler (1999)for more discussion and other examples of extensions of the single-site model.
39 English et al. (2018) is a highly scrutinized RUM model conducted for the damage assessment following the
Deepwater Horizon spill and demonstrates a number of sensitivity analyses. Additionally, the public archive for
the case contains a wealth of information. See www.doi.gov/deepwaterhorizon/adminrecord under the heading
"5.10 Lost Human Use;" see Section 5.10.4 for technical reports discussing issues surrounding RUM estimation.
40 While the standard logit recreation demand model treats each choice occasion as an independent event, the
model can also be generalized to account for repeated choices by an individual.
41 In a standard nested logit RUM model, recreators are commonly assumed to first decide whether or not to
take a trip, and then conditional on taking a trip, to next choose which site to visit. By not including a non-
participation option, the researcher in effect assumes that the recreator has already decided to take a trip, or in
other words, that the utility of taking a trip is higher than the utility of doing something else for that choice
occasion. In other words, models lacking a participation decision only estimate the recreation values of the
segment of the population that participates in recreation activities (i.e., recreators), while models that allow for
non-participation incorporate the recreation values of the whole population (i.e., recreators and non-recreators
combined). Because of this, recreation demand models without participation decisions tend to predict larger per
person welfare changes than models allowing non-participation.
7-34 Guidelines for Preparing Economic Analyses | 3rd edition
-------
However, because the RUM model examines recreation decisions on a choice occasion level, it is
less-suited for predicting the number of trips over a time period and measuring seasonal welfare
changes. A number of approaches have been used to link the RUM model's estimates of values per
choice occasion to estimates of seasonal participation rates. See Parsons, Jakus, and Tomasi (1999)
for a detailed discussion.
The nested logit and mixed logit models are extensions of the basic logit model. The nested logit
model groups similar alternatives into nests where alternatives within a nest are more similar to
each other than they are to alternatives outside of the nest. In very general terms, recreators are
first assumed to choose a nest and then, conditional on the choice of nest, they then choose an
alternative within that nest Nesting similar alternatives allows for more realistic substitution
patterns among sites than is possible with a basic logit The mixed logit is a random parameter logit
model that allows for even more flexible substitution patterns by estimating the variation in
preferences (or correlation in errors) across the sample. If preferences do not vary across the
sample, then the mixed logit collapses to a basic logit42
The Kuhn-Tucker (KT) model. The KT model is a seasonal demand model that estimates
recreators' choice of which sites to visit (like a multiple-site model) and how often to visit them over a
season (like a single-site model). The model is built on the theory that people maximize their
seasonal utility subject to their budget constraint by purchasing the quantities of recreation and
other goods that give them the greatest overall utility. Similar to the RUM model, the researcher
begins by specifying the recreator's utility function. Taking the derivative of this utility function
with respect to the number of trips taken, subject to a budget and non-negative trip constraint,
yields the "Kuhn-Tucker" conditions. The KT conditions show that trips will be purchased up to the
point that the marginal rate of substitution between trips and other spending is equal to the ratio of
their prices. In cases where the price of a trip exceeds its marginal value none will be purchased.
Given assumptions on the form of the utility function and the distribution of the error term,
probability expressions can be derived, and parameter estimates may then be recovered. While
recent applications have shown that the KT model can accommodate a large number of substitute
sites (e.g., von Haefen et al. 2004), the model is computationally intensive compared to RUM
models.43
Considerations in Evaluating and Understanding Recreation Demand Studies
Definition of a site and the choice set. The definition of what constitutes a unique site has been
shown to have a significant effect on estimation results. Ideally, one could estimate a recreation
demand model in which sites are defined as specific points such as exact fishing location, campsites,
etc. The more exact the site definition, the more exact the measure of travel costs and site
attributes, and therefore WTP, that can be calculated. However, in situations with many potential
alternatives, the large data requirements may be cost and time prohibitive, estimation may be
problematic, and aggregation may be required. The method of aggregation has been shown to have
a significant effect on estimated values. The direction of the effect will depend on the situation
being evaluated and the method of aggregation chosen (Parsons and Needleman 1992; Feather
1994; Kaoru, Smith, and Liu 1995; Parsons, Plantinga, and Boyle 2000).
42 See Train (1998) and Train (2009) for detailed descriptions of the nested and mixed logit models.
43 For a basic application of the KT model see Phaneufand Siderelis (2003). For more advanced treatments of
the models see Phaneuf Kling, and Herriges (2000), and von Haefen and Phaneuf(2005).
7-35 Guidelines for Preparing Economic Analyses | 3rd edition
-------
In addition to the definition of what constitutes a site, the number of sites included in a recreator's
choice set can have a significant effect on estimated values. When defining choice sets, the most
common practice in the literature has been to include all possible alternatives available to the
recreator. In many cases availability has been defined by location within a given distance or travel
time.44 This strategy has been criticized on the grounds that people may not know about all possible
sites, or even if they do know they exist they may not seriously consider them as alternatives. In
response to this, a number of researchers have suggested methods that either restrict choice sets to
include only those sites that the recreators seriously consider visiting (e.g., Peters et al. 1995, and
Haab and Hicks 1997) or that weight seriously-considered alternatives more heavily than less-
seriously-considered alternatives (e.g., Parsons, Massey, andTomasi 2000).
Multiple-site or multipurpose trips. Recreation demand models assume that the particular
recreation activity being studied is the sole purpose for a given trip. If a trip has more than one
purpose, it likely violates the travel cost model's central assumption that the "price" of a visit is
equal to the travel cost.45 The common strategy for dealing with multipurpose trips is simply to
exclude them from the data used in estimation.46 See Mendelsohn etal. (1992) and Parsons (2003b)
for further discussion.
Day trips versus multi-day trips. The recreation demand literature has focused almost exclusively
on single-day trip recreation choices. Adding the option to stay longer than one day adds another
choice variable in the estimation, thereby greatly increasing estimation difficulty. Also, as trip
length increases multipurpose trips become increasingly more likely, again casting doubt on the
assumption that a trip's travel costs represent the "price" of one single activity. A few researchers
have estimated models that allow for varying trip length. The most common strategy has been to
estimate a nested logit model in which each choice nest represents a different trip length option.
See Kaoru (1995), Shaw and Ozog (1999), and English etal. (2018) for examples. The few multi-day
trip models in the literature find that the per-day value of multi-day trips is generally less than the
value of a single-day trip, which suggests that estimating the value of multi-day trips by multiplying
a value estimated for single-day trips value by the number of days of will overestimate the multi-
day trip value.
donic Models
Hedonic pricing models use statistical methods to measure the contribution of a good's
characteristics to its price. These models are applicable to goods that can be thought of as
"bundling" together many attributes that vary in quantity and quality. Houses differ in size, layout,
location and exposure to environmental contaminants. Labor hours can be thought of as "goods"
differing in attributes like safety risks and supervisory nature that should be reflected in wages.
Hedonic pricing models use variation in prices of such goods to estimate the value of these
attributes.
44 Parsons and Hauber (1998) explore the implication of this strategy by expanding the choice set
geographically and find that beyond some threshold the effect of additional sites is negligible.
45 Parsons and Wilson (1997) suggest including a dummy variable to account for differences in multipurpose
trips.
46 Excluding any type or class of trip (like multiple-site or multipurpose) will produce an underestimate of the
population's total use value of a site. The amount by which benefits will be underestimated will depend on the
number and type of trips excluded.
7-36 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Econorr ndations of Hedonic Models
Hedonic pricing studies estimate economic value by weighing the advantages against the costs of
different choices. A standard assumption underlying hedonic pricing models is that markets are in
equilibrium, which means that no individual can improve her welfare by choosing a different home
or job. For example, if an individual changed location she might move to a larger house, or one in a
cleaner environment. However, to receive such amenities, the individual must pay for a more
expensive house and incur transaction costs to move. The more the individual spends on her house,
the less the individual has to spend on food, clothing transportation, and all the other things desired
or needed. So, if the difference in prices paid to live in a cleaner neighborhood is observable, then
that price difference can be interpreted as the WTP for a better environment
One key requirement in conducting a hedonic pricing study is that the available options differ in
measurable ways. To see why, suppose that all locations in a city's housing market are equally
polluted or that all jobs in a labor market expose workers to the same risks over a given period of
time. The premium that homeowners place on environmental quality or that workers place on lower
occupational risks could not be measured in this case. A hedonic pricing study requires a comparison
to purchases of more expensive houses in less polluted neighborhoods, or to wages in lower-paying
but safer jobs. However, there is also a practical limit on the heterogeneity of the sample. Workers in
different countries earn very different wages and face very different job risks, but this does not mean
it is possible to value the difference in job risks by reference to international differences in wages.
This is because: (1) there are many other factors that differ between widely separated markets; and
(2) people are not very mobile between disparate sites. Comparing wages or home prices across
decades or other long periods of time raises similar concerns if preferences change over time. For
these reasons it is important to exercise care in defining the spatial and temporal market in which
choices are made.
A related issue is that only environmental attributes or health risks that market participants are
aware of and understand can be valued using hedonic pricing methods (and revealed preference
methods more generally). If homeowners are unable to recognize differences in health outcomes,
visibility and other consequences of differences in environmental quality at different locations, or if
workers are unaware of differences in risks at different jobs, then a hedonic pricing study would
not be suitable for estimating the values for those attributes. For example, groundwater
contamination such as from leaking underground storage tanks can be difficult for
homeowners to detect (Zabel and Guignet 2012). In contrast, stated preference surveys can directly
ask respondents how they value changes in specific environmental commodities or health risks.
General Application by Type of Hedonic Pricing Study
Hedonic wage studies, also known as wage-risk or compensating wage studies, are based on the
premise that individuals make trade-offs between wages and occupational risks of death or injury.
Hedonic wage studies can assess the value to workers of changes in workplace morbidity and
mortality risks, which may then be applied to environmental risks using benefit-transfer
techniques. Viscusi (2013) provides an overview of the method. Most current hedonic wage studies
begin with estimation of the risk, calculated as workplace fatalities per worker. The Bureau of
Labor Statistics (BLS) Census of Fatal Occupational Injuries (CFOI) is the most common source for
workplace risk information, a complete record of U.S. workplace fatalities since 1992. CFOI reports
these fatalities by three-digit occupation and four-digit industry classifications, as well as the
7-37 Guidelines for Preparing Economic Analyses | 3rd edition
-------
circumstances of the fatal events.47 Typically, these data are used to construct the number of
annual fatalities within categories such as a given industry, occupation or industry-occupation cell.
This is the numerator for the annual risk rate for the hedonic wage study. Other data sources, most
commonly from the Current Population Survey also conducted by BLS, are used to estimate the
number of workers in these categories, providing the denominator for the annual risk rate, as well
as characteristics of workers, including wage rate. The estimating equation then uses the wage rate
as the dependent variable, usually in a linear or semi-log specification, and the coefficient on the
risk measure provides the basis for the implicit wage-risk tradeoff for mortality risk valuation.
There are questions about the applicability of hedonic wage study results to environmental benefits
assessment For example, hedonic wage estimates are derived from populations that are working
age and able to work, and they reflect preferences of those who have chosen relatively risky
professions. These characteristics may differ from populations affected by environmental
contaminants. There is also a difference in risk context between fatal workplace accidents and
environment-related mortality from, for example, cancer. Still, hedonic wage studies have been
widely used to estimate the value of fatal risk reductions, because they provide revealed-preference
information on how people trade off risks for money.48 Historically, the EPA has used a VSL
estimate primarily derived from hedonic wage studies. For more information on the Agency's VSL
estimate, see Appendix B.
Hedonic property value studies measure the contributions of various characteristics to property
prices. These studies are typically conducted using residential housing data, but they have also been
applied to commercial and industrial property, agricultural land and vacant land. Bartik (1988) and
Palmquist (1988,1991) provide detailed discussions of benefits assessment using hedonic
methods. Bishop etal. (2020) review best practices for using hedonic property models to measure
WTP for environmental amenities. Property value studies require large amounts of data. Market
data on individual housing units' prices and other attributes are strongly preferred to aggregate
data such as census tract median home values and characteristics.
Hedonic property value studies have examined the effects of air quality (e.g., Smith and Huang
1995, Bishop andTimmins 2018), water quality (e.g., Leggett and Bockstael 2000; Walsh etal.
2017; Guignetetal. 2022), natural amenities (e.g., Landry and Hindsley 2011; Guignetet al. 2017),
and land contamination (e.g., Messer et al. 2006; Guignet 2013; Walsh and Mui 2017) on property
values. As discussed in Section 7.1, the hedonic property approach can value changes in stressors,
contaminant releases or media concentrations, or other intermediate endpoints linked to
environmental benefits. The type of environmental amenity included in the analysis is often driven
by data availability.
Other hedonic studies. Applicability of the hedonic pricing method is not limited to property and
labor markets. For example, hedonic pricing methods can be combined with travel cost methods to
examine the implicit price of recreation site characteristics (Brown and Mendelsohn 1984; Phaneuf
etal. 2008; Kovacs 2012). Hedonic analysis of light-duty vehicle prices has been used to examine
the value of fuel economy and other features (Espey and Nair 2005; Fan and Rubin 2010).
47 More information on the CFOI data is available at: http://www.bls.gov/iif/oshfatl.htm.
48 For example, the EPA's SAB has recognized the limitations of these estimates for use in estimating the benefits
of reduced cancer incidence from environmental exposure. Despite these limitations, however, the SAB concluded
that these estimates were the best available at the time (U.S. EPA 2000).
7-38 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Considerations in Evaluating and Understanding Hedonic Pricing Studies
There are numerous statistical issues associated with applying hedonic pricing models to value
changes in environmental quality and health risks. Below, we highlight the issues of identification
and causality, spatial correlation, defining and measuring the environmental amenity, and
interpretation of the estimates.
Identification and causality. A common challenge in hedonic pricing studies is establishing the
direction of causality between the independent variable of interest (environmental quality or safety
risks) and the dependent variable (e.g., home prices or wages). People choose among different
houses not only because they can trade off differences in environmental amenities against price, but
also because of other characteristics, like curb appeal, school quality and crime. If these other
characteristics are not included in the hedonic regression and are correlated with environmental
quality, then the analysis may not identify the causal impact of environmental quality on prices. In
this situation, endogeneity or omitted variable bias would lead to incorrect estimates of the value of
environmental quality to home buyers (Taylor, Phaneuf, Liu 2016). Similarly, if the risk of
accidental death is correlated with the risk of serious, nonfatal injuries, the premium estimated in a
hedonic wage equation would overstate WTP for reductions in mortality if these other risks were
omitted from the regression.
Approaches to identify causal effects in the hedonic property value literature include repeat sales
models, which can identify the effect of changes in environmental quality over time using the
sample of homes that sold multiple times during the study period, and quasi-experimental
approaches, which rely on "natural experiments" in which environmental quality varies for reasons
that are exogenous to home prices. Quasi-experimental approaches include instrumental variables,
regression discontinuity, matching and difference-in-difference models (e.g., Greenstone and Gayer
2009; Greenstone and Gallagher 2008; Gamper-Rabindran and Timmins 2013). Spatial fixed effects
denoting discrete geographic units such as Census tracts or counties can also help control for
difficult-to-measure local characteristics, but environmental quality must vary within this spatial
unit for these models to yield useful valuation estimates (Abbott and Klaiber 2010). While fixed
effects alone may not mitigate omitted variable bias if unobserved characteristics correlated with
environmental quality also vary within the Census tract or other spatial unit, research has found
that a combination of spatial fixed effects, quasi-experimental identification and temporal controls
can greatly reduce bias (Kuminoff et al. 2010). There may also be spatial correlation in the
dependent variable or the error term of the model if home prices are directly affected by the prices
of nearby homes (for example, due to the home appraisal process). Spatial econometrics techniques
allow analysts to account for some of these sources of dependence, reducing bias and improving the
consistency or efficiency of parameter estimates (Anselin 2001). However, incorrect specification of
the structure of the spatial correlation can also bias parameter estimates (Gibbons and Overman
2012). Spatial fixed effects and geographic clustering of standard errors are also useful approaches
to address spatial correlation of property characteristics (Bishop et al. 2020).
Defining and measuring the environmental amenity. Another important issue is the way that
the environmental amenity or health risk included in a hedonic model is defined and measured. The
ideal measure is an indicator that market participants value and that can be linked to a change in
environmental policy, but such measures are not always available. For instance, available water
quality indicators may not fully reflect water quality or ecosystem health (Griffiths et al. 2012).
Water clarity has been shown to positively affect property prices (Michael etal. 2000; Gibbs etal.
2002; Walsh etal. 2017; Guignetetal. 2022), but it is not always a good indicator of ecosystem
health (Shaw, Mechenich, and Klessig 2004). Furthermore, data on water clarity may contain errors
7-33 Guidelines for Preparing Economic Analyses | 3rd edition
-------
because clarity cannot be accurately measured under cloud cover (Olmanson, Bauer, and Brezonik
2008). If water clarity is measured with error or is not a good proxy for home buyers' perceptions
of water quality, then measurement error could produce valuation estimates that are biased toward
zero due to attenuation (Greene 2000), though empirical research has found that objective
measures of water clarity have higher predictive power than individuals' subjective measure of
water clarity (Poor etal. 2001).
Interpretation of the estimates. Understanding how to interpret hedonic model estimates is
important. As with many results in economics, hedonic pricing models are best suited to the
valuation of small, or marginal, changes in attributes. Under such circumstances, the slope of the
hedonic price function can be interpreted as WTP for a small change. Hedonic price functions
typically reflect equilibria between consumer demands and producer supplies for fixed levels of the
attributes being evaluated. Public policy, however, is sometimes geared to larger, discrete changes
in attributes. When this is the case, using hedonic model estimates to calculate benefits is more
complicated. Palmquist (1991) describes conditions under which exact welfare measures can be
calculated for discrete changes. See Freeman etal. (2014) and Ekeland, Heckman, and Nesheim
(2004) for treatments.
Studies that compare prices before and after a change in environmental quality using repeat sales
and quasi-experimental approaches raise particular challenges for interpretation (Klaiber and
Smith 2013). These approaches are sometimes called "capitalization" rather than hedonic studies
because they estimate the extent to which changes in amenities are capitalized into prices over
time. The capitalization effect only equals WTP if WTP remains stable over the study time horizon
(Kuminoff and Pope 2010). If marginal WTP for environmental quality is increasing (decreasing)
over time in the study area, then the capitalization estimate will tend to overestimate
(underestimate) the benefits of cleanup.
In property value studies, if gentrification or re-sorting occurs such that people with a higher WTP
move to neighborhoods with improving environmental quality, pre- and post-cleanup housing
prices reflect the preferences of two distinct groups of people.49 In addition, the capitalization
estimate from repeat sales and quasi-experimental models represents the average rather than
marginal change in property values that occurs in response to a change in an amenity (Parmeter
and Pope 2013). If residents do not re-sort, their preferences and incomes are not changing over
time, and WTP is linear in environmental quality, then a capitalization estimate can be interpreted
as a measure of WTP. These conditions are less likely to hold in a study that examines a large
change in environmental quality over a relatively long timespan. For example, Parmeter and Pope
(2013) argue that Chay and Greenstone's (2005) quasi-experimental study of the housing price
effects of improvements in air quality in nonattainment counties after passage of the 1970 Clean Air
Act Amendments provides a capitalization rather than a WTP estimate because of the 10-year
timespan of the study and the non-marginal reduction in air pollution. However, the assumption
that a capitalization estimate provides a good approximation of WTP might be reasonable for
studies covering relatively short periods of time and examining small changes in environmental
quality.
49 Residential sorting models provide another alternative to hedonic and capitalization studies in the property
value literature. These models derive estimates of WTP explicitly accounting for residential sorting behavior and
resulting changes in a variety of neighborhood amenities (e.g., Klaiber and Phaneuf2010; Kuminoff and Jarrah
2010).
7-40 Guidelines for Preparing Economic Analyses | 3rd edition
-------
iviors
The averting behavior method infers values for environmental quality from observations of actions
people take to avoid or mitigate the increased health risks or other undesirable consequences of
reductions in environmental quality. Examples of such defensive actions can include the purchase
and use of air filters, the activity of boiling water prior to drinking it and the purchase of preventive
medical care or treatment By analyzing the expenditures associated with these averting behaviors,
economists can attempt to estimate the value individuals place on small changes in risk or
environmental quality. Dickie (2017) provides a detailed overview of the approach.
Econorr ndations of A verting Behavior Methods
Averting behavior methods can be best understood from the perspective of a household production
framework. Households can be thought of as producing health outcomes by combining an
exogenous level of environmental quality with inputs such as purchases of goods that involve
protection against health and safety risks (Freeman et al. 2014; Dickie 2017). To the extentthat
averting behaviors are available, the model assumes that a person will continue to take protective
action as long as the expected benefit exceeds the cost of doing so. If there is a continuous
relationship between defensive actions and reductions in health risks, then the individual will
continue to avert until the marginal cost just equals the individual's marginal WTP for these
reductions. Thus, the value of a small change in health risks can be estimated from two primary
pieces of information: (1) The cost of the averting behavior or good, and (2) its effectiveness, as
perceived by the individual, in offsetting the loss in environmental quality.
One approach to estimation is to use observable expenditures on averting and mitigating activities
to generate values that may be interpreted as a lower bound on WTP. As noted earlier in Section
7.2.1.2, WTP for small changes in environmental quality can be expressed as the sum of the values
of four components: changes in averting expenditures, changes in mitigating expenditures, lost time
and the loss of utility from pain and suffering. The first three terms of this expression are
observable, in principle, and can be approximated by calculating changes in these costs after a
change in environmental quality. The resulting estimate can be interpreted as a lower bound on
WTP that may be used in benefits analysis (Shogren and Crocker 1991; Quiggin 1992).
General Application of Averting Behavior Method
Although the first applications of the averting behavior method estimated the benefits of reduced
soiling of materials from environmental quality changes (Harford 1984), recent research has
primarily focused on health risk changes. Conceptually, the averting behavior method can provide
WTP estimates for a variety of other environmental benefits such as damages to ecological systems
and materials.
Some averting behavior studies focus on behaviors that prevent or mitigate the impact of specific
symptoms (e.g., shortness of breath or headaches), while others have examined averting
expenditures in response to specific episodes of contamination (e.g., groundwater contamination).
Because many contaminants can produce similar symptoms, studies that estimate values for
symptoms may be more amenable to benefit transfer than those are episode-specific and do not
value specific symptoms or illnesses. The latter could potentially be more useful, however, for
assessing the benefits of a regulation expected to reduce the probability of similar contamination
episodes.
7-41 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Considerations in Evaluating and Understanding Averting Behavior Studies
Perceived versus actual risks. As in other revealed preference methods, analysts should
remember that consumers base their actions on perceived benefits from their behaviors. Many
averting behavior studies explicitly acknowledge that their estimates rest on consistency between
the consumer's perception of risk reduction and actual risk reduction. While there is some evidence
that consumers are rational regarding risk for example, consumer risk-reduction expenditures
increase as risk increases there is also evidence that there are predictable differences between
consumers' perceptions and actual risks. For example, individuals tend to overestimate risks that
are very small or that are novel or unfamiliar (Renner et al. 2015). Thus, averting behavior studies
can produce biased WTP estimates for a given change in objective risk. Surveys may be necessary to
determine the benefits individuals perceive they are receiving when engaging in defensive
activities. These perceived benefits can then be used as the object of the valuation estimates. For
example, if surveys reveal that perceived risks are lower than expert risk estimates, then WTP for
risk reduction can be estimated with the lower, perceived risk (Blomquist 2004).
Data requirements and implications. Data needed for averting behavior studies include
information detailing the effect(s) being averted (e.g., specific illnesses, exposure to environmental
contaminants); actions taken to avert or mitigate damages; the costs of those behaviors and
activities; and other variables that affect health outcomes, like age, health status or chronic
conditions. Another significant challenge in many averting behavior applications is that output level
(e.g., health) is unobserved and may change when averting actions are taken, which complicates
calculation of WTP.
Separability of joint effects. Analysts should exercise caution in interpreting the results of studies
that focus on goods in which there may be significant joint benefits (or costs). Many defensive
behaviors not only avert or mitigate environmental damages, but also provide other benefits. For
example, air conditioners obviously provide cooling in addition to air filtering, and bottled water
may not only reduce health risks, but also taste better. Conversely, it also is possible that the
averting behavior may have negative effects on utility. For example, wearing helmets when riding
bicycles or motorcycles may be uncomfortable. Failure to account for these "joint" benefits and
costs associated with averting behaviors will result in biased estimates of WTP.
Modeling assumptions. Restrictive assumptions are sometimes needed to make averting behavior
models tractable. Analysts drawing upon averting behavior studies will need to review and assess
the implications of these assumptions for the valuation estimates.
ess
A frequent alternative to WTP estimates is the avoided COI, which estimates the resource costs
associated with an adverse health effect (an "illness"). Though WTP is the preferred valuation
measure for BCA, the COI method is straightforward to implement and explain to policy makers,
and has been widely applied, particularly in health economics or studies of the burden of disease
(e.g., Trasande et al. 2016). COI estimates for many illnesses are readily available from existing
studies. COI estimates are usually less expensive to develop than WTP estimates using stated or
revealed preference approaches, so it may be feasible to develop new COI estimates for a given
benefits analysis even with typical time and resource constraints. Jo (2014) and Tarricone (2006)
provide overviews of the method.
7-42 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Econom ndations of CO I Studies
Relating cost of illness to WTP. The COI method does not estimate WTP but is based on
estimating the market value of goods and services used to treat illness and the lost productivity due
to that illness. It does not incorporate any disutility from illness, the costs of averting behaviors
taken to avoid the illness, or risk preferences that would be inherent in estimates of WTP to reduce
risks of illness. Because of these limitations COI is best viewed as a proxy for WTP when WTP
estimates are not available, and is generally considered to be a lower bound on WTP, especially for
more serious illnesses.50 Available comparisons of COI and total WTP estimates suggest that the
difference can be large but varies greatly across health effects; COI estimates cannot be simply
"scaled up" to approximate WTP.
In some cases, COI may be additive to a WTP estimate that did not account for certain costs. COI
estimates capture medical expenses passed on to third parties such as health insurance companies
and hospitals, whereas WTP estimates generally do not. COI estimates can capture the value of lost
productivity, something that may be overlooked in WTP estimates especially when derived from
consumers or employees covered by sick leave. In practice, because there is a risk of double-
counting when adding COI and WTP, doing so requires a careful evaluation of the studies in
question, what they each do and do not include, and how they can be appropriately added together.
Types of costs considered in COI studies. COI studies generally distinguish between direct costs
(costs related to medical treatment) and indirect costs (costs related to lost productivity). Many COI
studies estimate both direct and indirect costs, but some may focus solely on costs of treatment
while other studies, broadly categorized as COI, only reflect lost productivity.
Direct Costs are those related to treatment and care for the illness. These costs include the
value of goods and services spent for items such as physicians' services, testing,
hospitalization, medications and medical devices. But it also includes the value of household
expenditures, transportation, accommodation and other resources spent on care for the
illness. COI studies may not capture all of these costs. For example, studies relying solely on
databases of medical expenditures might not capture the costs of household expenditures.
Indirect costs refer to productivity losses associated with the illness, most often measured
by the human capital approach where earnings reflect the value of productive time.51 That
is, assuming the wage equals the value of marginal product Losses to productivity,
therefore, are a social cost and can be measured by the wage rate.52 Lost productivity may
be focused on the short-term, e.g., for illnesses where the losses are associated with a loss of
workdays or, for more serious illnesses a permanent loss of income.
In principle, indirect costs should also consider the costs of lost home productivity and the
value of leisure, but this is not always done in COI studies. Lost productivity for home health
care, e.g., the time spent by members of the household in caring for family members or
accompanying patients to medical appointments, should also be included in indirect costs.
50 However, any particular COI estimate is not necessarily going to be lower than WTP for a given health
condition. Depending on the design of the studies, WTP could reflect avoidance costs that are lower than the cost
of illness once the illness has been contracted.
51 For an in-depth discussion of the valuation of time in different contexts, see U.S. EPA (2020).
52 The EPA has a similar approach for cost analysis that is also based on the opportunity cost of time; see U.S.
EPA "Handbook on Valuing Changes in Time Use Induced by Regulatory Requirements and Other EPA Actions"
[U.S. EPA 2020).
7-43 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Note that the human capital approach applies not just to lost work time at a given wage in
what would be considered a traditional COI study but for any impact on productivity
associated with adverse health effects. For example, lowered IQ - an effect associated with
exposures to many pollutants - has been related to labor participation and lower lifetime
earnings, a loss of human capital (Salkever 1995; Lin etal. 2018). This relationship can be
useful in economic analyses to value the benefits of avoiding IQ losses. Additionally,
exposure to ozone has been linked to loss of productivity among agricultural workers
(Crocker and Horst 1981; Graff Zivin and Neidell 2012).53
General Application by Type of COI Study
Prevalence-based estimates. Prevalence-based COI estimates are derived from the costs faced by
all individuals who have a sickness in a specified time period. For example, an estimate of the total
number of individuals who currently have asthma, as diagnosed by a physician, reflects the current
prevalence of physician-diagnosed asthma. Prevalence-based COI estimates for asthma include all
direct and indirect costs associated with asthma within a given time period, such as a year.
Prevalence-based COI estimates are a measure of the financial burden of a disease but will still
generally underestimate total WTP for avoiding the disease altogether. They are most applicable for
valuation of policies that reduce or eliminate morbidity associated with existing cases of illness.
Incidence-based estimates. By contrast, incidence-based COI estimates reflect expected costs for
new cases of an illness in a given time period. For example, the number of individuals who receive a
new diagnosis of asthma from a physician in a year reflects the annual incidence of physician-
diagnosed asthma. Incidence-based COI estimates reflect the expected value of direct medical
expenditures and lost income and productivity associated with a disease from the time of diagnosis
until recovery or death. Because these expenses can occur over an extended time period, incidence-
based estimates should be discounted to the year the illness is diagnosed and expressed in present
value terms. Incidence-based COI estimates are most applicable for valuation of policies that reduce
the expected number of new cases of disease, which is often the case for environmental regulations.
Bottom-up, top-down and econometric approaches. There are three primary methods for
estimating COI for a given health condition. The "bottom up" approach constructs a typical profile
of treatment for the condition and then uses unit costs to estimate total treatment costs over time,
usually based on databases of medical expenditures. The "top-down" approach, on the other hand,
typically starts with aggregate expenditures across a number of illnesses and then attributes these
expenditures across that set of illnesses. Finally, the econometric approach to COI typically uses
data on total costs for a given sample over a given time period and then econometrically estimates
the difference in costs between those with and without a given health condition. The difference
provides an estimate of the cost of treatment for the illness. Bottom-up or econometric approaches
are generally best-suited for benefits analysis.
Considerations in Evaluating and Understanding COI Studies
Technological change. Medical treatment technologies and methods are constantly changing, and
this could push the true cost estimate for a given illness either higher or lower. When using
53 For examples of how productivity estimates have been used in economic analyses, see the primary benefits
analysis for the 2011 Transport Rule as well as the supplemental benefits analysis for the 2015 Ozone NAAQS.
7-44 Guidelines for Preparing Economic Analyses | 3rd edition
-------
previous COI studies, the analyst should be sure to research whether and how the generally
accepted treatment has changed from the time of the study.
Measuring health care costs. The COI literature uses a variety of methods to measure health care
costs.54 One important distinction is between medical expenditures and medical charges.
Expenditures are the better indicator of social costs because they better represent actual resources
used by healthcare providers rather than the "list price," which is often discounted. Studies that rely
on medical charges may use them as-is or try to approximate expenditures using hospital-specific
cost-to-charge ratios. For benefits analysis adjusted charges are better than unadjusted charges, but
studies that use expenditure data are even more preferred.
Measuring the value of lost productivity. The value of lost productivity in many studies may only
reflect persons in the work force, omitting the productivity costs of those persons not involved in
paid jobs. Homemakers' household upkeep and childcare services, retired persons' volunteering
efforts and students' time in school all directly or indirectly contribute to the productivity of
society. In cases where an affected individual requires a caregivers' assistance, e.g., when children,
elderly or impaired individuals are affected, the caregiver may also incur time away from work and
lost productivity. The value of lost leisure time to an individual and their family is not included in
most COI studies.55 A second set of considerations is the choice of wage rate in the study, which will
reflect the study population and may not match the wage rate of the population in the policy case.
erence
The distinguishing feature of stated preference methods compared to revealed preference methods
is that stated preference methods rely on people's responses to hypothetical questions while
revealed preference methods rely on observations of actual choices. Stated preference methods use
surveys that ask respondents to consider one or a series of hypothetical scenarios that describe a
potential change in a non-market good. The advantages of stated preference methods include their
ability to estimate non-use values and to incorporate hypothetical scenarios that closely
correspond to a policy case. The main disadvantage of stated preference methods is that they may
be subject to systematic biases that are difficult to test for and correct For this reason, OMB
Circular A-4 advises,
"If both revealed-preference and stated-preference studies that are directly applicable to
regulatory analysis are available, you should consider both kinds of evidence and compare
or combine the findings when feasible. If the results diverge significantly, you should, when
feasible, compare the overall quality of the two bodies of evidence. Other things equal,
revealed preference data are preferable to stated preference data because revealed
preference data are based on actual decisions." (p.37)
The Report of the NOAA Panel on Contingent Valuation is often cited as an early source of
recommendations for best practices for stated preference studies. Often referred to as the "NOAA
Blue Ribbon Panel," this panel, comprised of five distinguished economists including two Nobel
Laureates, deliberated on the usefulness of stated preference studies for policy analysis for the
National Oceanic and Atmospheric Administration (NOAA) (Arrow et al. 1993). The panel focused
54 See Onukwugha, et al. 2016 for a review of methods and their prevalence in the COI literature.
55 For an in-depth discussion of the valuation of time in different contexts, see U.S. EPA 2020.
7-45 Guidelines for Preparing Economic Analyses | 3rd edition
-------
on a rather narrow application of stated preference the use of contingent valuation to estimate
non-use values for litigation in the United States. In the years since, stated preference research has
advanced significantly and its applications have expanded to more diverse contexts. In 2012, The
Journal of Economic Perspectives published a symposium on contingent valuation including an
assessment of the state of the science, prompting additional discussion in a 2013 Applied Economic
Perspectives and Policy paper (Haab et al. 2013). More recently, Johnston et al. (2017) published an
updated set of guidelines that reflects contemporary stated preference research, changes in survey
methods and technology and the transfer of primary estimates to different policy scenarios.
'. I. f D;on 'Pith r ([ndatiori of Stated Preference Methods
The role of non-use value in BCA has been well established since the 1990s (see Kopp 1992, and
Bishop and Welsh 1992 for early discussions of non-use value and welfare theory). Further,
ignoring non-use value in environmental regulatory analysis can lead to large omissions in benefits
estimation and a misallocation of resources. A regulatory analysis should carefully consider when
non-use values might be substantial and, given stated preference is the only valuation approach
that captures them, what studies are available to draw from and how to evaluate the validity of
their results.
The responses elicited from stated preference surveys, if truthful, unbiased and well-informed, are
either direct expressions of WTP or can be used to estimate WTP for the good in question. However,
the caveats listed above are paramount. While many environmental economists believe that
respondents can provide truthful answers to hypothetical questions and therefore view stated
preference methods as useful and reliable if conducted properly, a non-trivial fraction of
economists are more skeptical of the results elicited from stated preference surveys. Due to this
skepticism, it is important to employ validity and reliability tests of stated preference results before
applying them to policy decisions.
If the analyst decides to conduct a stated preference survey or use stated preference results in a
benefit transfer exercise, then a number of survey design issues should be considered. Stated
preference researchers have attempted to develop methods to make individuals' choices in stated
preference studies as consistent as possible with market transactions or consequential referenda.
Reasonable consistency with the framework of market transactions is a guiding criterion for ensuring
the validity of stated preference value estimates. Three components of market transactions need to be
constructed in stated preference surveys: the commodity, the payment and the scenario (Fischoff and
Furby 1988).
Stated preference studies need to carefully define the commodity to be valued, including
characteristics of the commodity such as the timing of provision, certainty of provision and
availability of substitutes and complements. The definition of the commodity generally involves
identifying and characterizing attributes of the commodity that are relevant to respondents.
Commodity definition also includes defining or explaining baseline or current conditions, property
rights in the baseline, the policy scenarios, as well as the source of the change in the environmental
commodity.56
56 Depending on the scenario, the description of the commodity may produce strong reactions in respondents
and could introduce bias. In these cases, the detail with which the commodity of the change is specified needs to
7-46 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Respondents also must be informed about the transaction context, including the method, timing,
and duration of payment. The transaction must not be coerced, and the individual should be aware
of her budget constraint The payment vehicle should be described as a credible and binding
commitment should the respondent decide to purchase the good (Carson et al. 1997). The timing
and duration of a payment involves individuals implicitly discounting payments and calculating
expected utility for future events. The transaction context and the commodity definition should
describe and account for these temporal issues.
The hypothetical scenario(s) should be described to minimize potential strategic behavior such as
"free riding" or "yea-saying." In the case of free riding, respondents will underbid their true WTP for
a good if they believe it will be provided regardless of their response. In the case of yea-saying,
respondents pledge amounts greater than their true WTP with the expectation that they will not be
made to pay for the good but believing that their response could influence whether the good will be
provided. Incentive-compatible choice scenarios and attribute-based response formats have been
shown to mitigate strategic responses. Both are discussed below.
It is recognized in both the experimental economics and the survey methodology literatures that
different survey formats can elicit different responses. Changing the wording or order of questions
can also influence responses. Therefore, the researcher should provide a justification for her
choice of survey format and include a discussion of the ramifications of that choice.
7,3,2.1 v neral Application by Type of Stated Preferer* - idy
Two main types of stated preference survey format are currently used: direct WTP questions and
stated choice questions. Stated choice questions can be either dichotomous choice questions or
multi-attribute choice experiments. Because survey formats are still evolving, and many different
approaches have been used in the literature, no definitive recommendations are offered here
regarding selection of the survey format Rather, the following sections describe some of the most
commonly used formats and discuss some of their known and suspected strengths and weaknesses.
Researchers should select a format that suits their topic, and should use focus groups, pretests and
statistical validity tests to address known and suspected weaknesses in the selected approach.
Direct/Open-Ei estions
Direct/open-ended WTP questions ask respondents to indicate their maximum WTP for the specific
quantity or quality changes of a good or service that has been described to them. An important
advantage of open-ended stated preference questions is that the answers provide direct, individual-
specific estimates of WTP which requires a smaller sample size and simpler estimation approach.
While these advantages could lower the cost of the study, early stated preference studies found that
some respondents had difficulty answering open-ended WTP questions and non-response rates to
such questions were high. Such problems are more common when the respondent is not familiar
with the good or with the idea of exchanging a direct dollar payment for the good. An example of a
stated preference study using open-ended questions is Brown et al. (1996).
be balanced against the ultimate goals of the survey. Regardless, the commodity needs to be specified with
enough detail to make the scenario credible.
7-47 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Various modifications of the direct/open-ended WTP question format have been developed to help
respondents arrive at their maximum WTP estimate. In iterative bidding respondents are asked if
they would pay some initial amount, and then the amount is changed up or down depending on
whether the respondent says "yes" or "no" to the first amount. This continues until a maximum
WTP is determined for that respondent Iterative bidding has been shown to suffer from "starting
point bias," wherein respondents' maximum WTP estimates are systematically related to the dollar
starting point in the iterative bidding process (Rowe and Chestnut 1983; Boyle et al. 1988;
Whitehead 2002). A payment card is a list of dollar amounts from which respondents can choose,
allowing respondents an opportunity to look over a range of dollar amounts while they consider
their maximum WTP. Mitchell and Carson (1989) and Rowe etal. (1996) discuss concerns that the
range and intervals of the dollar amounts used in payment card methods may influence
respondents' WTP answers.
Stated Choice Questions
While direct/open-ended WTP questions are efficient in principle, researchers have generally
turned to other stated preference techniques in recent years. This is largely due to difficulties
respondents face in answering direct WTP questions and the lack of easily implemented procedures
to mitigate these difficulties. Researchers also have noted that direct WTP questions with various
forms of follow-up bidding may not be "incentive compatible." That is, the respondents' best
strategy in answering these questions is not necessarily to be truthful (Freeman et al. 2014).
In contrast to direct/open-ended WTP questions, stated choice questions ask respondents to
choose a single preferred option or to rank options from three or more choices. When analyzing the
data, the dependent variable will be continuous for open-ended WTP formats and discrete for
stated choice formats.57 In principle, stated choice questions can be distinguished along three
dimensions:
The number of alternatives each respondent can choose from in each choice scenario
surveys may offer only two alternatives (e.g., yes/no, or "live in area A or area B"); two
alternatives with an additional option to choose "don't know" or "don't care;" or multiple
alternatives (e.g., "choose option A, B or C").
The number of attributes varied across alternatives in each choice question (other than price)
alternatives may be distinguished by variation in a single attribute (e.g., mortality risk) or
multiple attributes (e.g., mortality risk, length and severity of illness, source of risk, etc.).
The number of choice scenarios an individual is asked to evaluate through the survey.
Any particular stated choice survey design could combine these dimensions in any given way. For
example, a survey may offer two options to choose from in each choice scenario, vary several
attributes across the two options, and present each respondent with multiple choice scenarios. The
statistical strategy for estimating WTP is largely determined by the survey format adopted, as
described below.
The earliest stated choice questions were simple yes/no questions. These were often called
referendum questions because they were often posed as, "Would you vote for ..., if the cost to you
57 Some researchers use the term "contingent valuation" to refer to direct WTP and dichotomous
choice/referendum formats and "stated preference" to refer to other stated choice formats. In these Guidelines,
the term "stated preference" is used to refer to all valuation studies based on hypothetical choices (including
open-ended WTP and stated choice formats), as distinguished from "revealed preference."
7-48 Guidelines for Preparing Economic Analyses | 3rd edition
-------
were $X?" However, these questions are not always posed as a vote decision and are now
commonly called dichotomous choice questions.
Stated preference researchers have also adapted a choice question approach used in the marketing
literature called conjoint analysis. These are more complex choice questions in which the
respondent is asked repeatedly to pick her preferred option from a list of two or more options. Each
option represents a package of product attributes. Including cost as an attribute and varying it
across options allows researchers to estimate marginal WTP for each attribute of the good. Holmes
and Adamowicz (2003) refer to this as attribute-based stated choice.
Dichotomous choice WTP questions. Dichotomous choice questions present respondents with a
specified environmental change costing a specific dollar amount and then ask whether they would
be willing to pay that amount for the change. The primary advantage of dichotomous choice WTP
questions is that they are easier to answer and less prone to manipulation than direct WTP
questions, because the respondent is not required to determine their exact WTP, only whether it is
above or below the stated amount Sample mean and median WTP values can be derived from
analysis of the frequencies of the yes/no responses to each dollar amount Bishop and Heberlein
(1979), Hanemann (1984), and Cameron and James (1987) describe the necessary statistical
procedures for analyzing dichotomous choice responses using logit or Probit models. Dichotomous
choice responses will reveal an interval containing WTP, and in the case of a "yes" response this
interval will be unbounded from above. As a result, significantly larger sample sizes are needed for
dichotomous choice questions to obtain the same degree of statistical efficiency in the sample
means as direct/open-ended responses that reveal point-values for WTP (Cameron and James
1987).
To increase the estimation efficiency of dichotomous choice questions, some studies have used
what is called a double-bounded approach. In double-bounded questions the respondent is asked
whether she would be willing to pay a second amount, higher if she said yes to the first amount, and
lower if she said no to the first amount.58 Sometimes multiple follow-up questions are used to try to
narrow the interval around WTP even further. These begin to resemble iterative bidding style
questions if many follow-up questions are asked. Similar to starting point bias in iterative bidding
questions, the analyses of double-bounded dichotomous choice question results suggest that the
second responses may not be independent of the first responses (Cameron and Quiggin 1994,1998;
and Kanninen 1995).
Multi-attribute choice questions. In multi-attribute choice questions (also known as conjoint
analysis), respondents are presented with alternative choices that are characterized by different
combinations of attributes and prices. Multi-attribute choice questions ask respondents to choose
the most preferred alternative (a partial ranking) from multiple alternative goods (i.e., a choice set),
in which the alternatives within a choice set are differentiated by their attributes including price
(Johnson et al. 1995 and Roe et al. 1996). The analysis takes advantage of the differences in the
attribute levels across the choice options to determine how respondents value marginal changes in
each of the attributes. To measure WTP, a cost (e.g. a tax or measure of travel costs) is included in
multi-attribute choice questions as one of the attributes of each alternative. This price and the
mechanism by which the price would be paid need to be explained clearly and plausibly, as with
any payment mechanism in a stated preference study. Boyle and Ozdemir (2009) examine the
58 Alberini (1995) illustrated an analysis approach for deriving WTP estimates from such responses and
demonstrated the increased efficiency of double-bounded questions. The same study showed that the most
efficient range of dollar amounts in a dichotomous choice study design was one that covered the mid-range of
the distribution and did not extend very far into the tails at either end.
7-43 Guidelines for Preparing Economic Analyses | 3rd edition
-------
impact of question design choices, such as the ordering of attributes and the number of alternatives
in a single question, on the mean WTP estimate.
There are many desirable aspects of multi-attribute choice questions, including the nature of the
choice being made. To choose the most preferred alternative from some set of alternatives is a
common decision experience in posted-price markets, especially when one of the attributes of the
alternatives is a price. One can argue that such a decision encourages respondents to concentrate
on the trade-offs between attributes rather than taking a position for or against an initiative or
policy. This type of repeated decision process may also diffuse the strong emotions often associated
with environmental goods, thereby reducing the likelihood of yea-saying or of rejecting the premise
of having to pay for an environmental improvement Presenting repeated choices also gives the
respondent some practice with the question format, which may improve the overall accuracy of her
responses, and gives her repeated opportunities to express support for a program without always
selecting the highest price option.59 One challenge of attribute-based methods is representing the
environmental changes in small number of separable attributes. Extensive focus group research is
required to choose the most salient attributes and find the best way to convey those changes to
respondents. To estimate marginal values, the attributes must be able to change independent of one
another without respondent rejecting the scenario.
7.3.1 ' 1 »nsideratior ting Stated Preference Results
Survey mode. The mode used to administer a survey is an important component of survey
research design because it is the mechanism by which information is conveyed to respondents, and
likewise determines the way in which individuals can provide responses for analysis. Common
survey modes include telephone, in-person, mail and electronic surveys administered by computer
or smart phone. Telephone surveys are primarily conducted with a trained interviewer using
random digit dialing (RDD) to contact households. In-person surveys are conducted in a variety of
ways, including door-to-door, intercepts at public locations and via telephone recruiting to a central
facility. Mail surveys are conducted by providing written survey materials for respondents to self-
administer. Electronic surveys are computerized and can be self-administered at a central facility,
at home via the internet or on a smart phone wherever internet access is available. As technology
and society have changed, so has the preference for one mode over the other. With the influx of
market research, telemarketing, telephone scams and the abandonment of landlines, the telephone
has become a less effective way to administer surveys. Similarly, response rates to mail surveys
have declined substantially (Stedman etal. 2019). With increased prevalence of smartphone
technology, internet access and email accessibility, computerized surveys have emerged as an
expedient means of survey administration. Researchers may also choose to combine modes using
one for recruiting and the other for survey administration. With every survey mode mentioned,
there are inherent biases. These biases are generally classified as social desirability bias, sample
frame bias, avidity bias, and non-response bias. See Maguire (2009), Loomis and King (1994),
Mannesto and Loomis (1991), Lindberg et al. (1997), and Ethier et al. (2000) for a discussion of
different biases in survey mode.
Framing effects. An important issue regarding survey formats is whether information provided in
the questions influences the respondents' answers in one way or another. For example, Cameron
59 Some applications of multi-attribute survey formats include Layton and Brown (2000), Boyle et al. (2001),
Morey et al. (2002), and Moore et al. (2018). Studies that investigate the effects of multi-attribute choice
question design parameters include Johnson et al. (2000) and Adamowicz et al. (1997).
7-50 Guidelines for Preparing Economic Analyses | 3rd edition
-------
and Huppert (1991) and Cooper and Loomis (1992) find that mean WTP estimates based on
dichotomous choice questions may be sensitive to the ranges and intervals of dollar amounts
included in the WTP questions. Kanninen and Kristrom (1993) show that the sensitivity of mean
WTP to bid values can be caused by model misspecification, failure to include bid values that cover
the middle of the distribution, or inclusion of bids from the extreme tails of the distribution.
Selection of payment vehicle. The payment vehicle in a stated preference study refers to the
method by which individuals or households would pay for the good described in a particular survey
instrument Examples include increases in electricity prices, changes in cost of living, a one-time
tax, or a donation to a special fund. It is imperative that the payment vehicle is compulsory and does
not introduce strategic responses or bias. Voluntary payment vehicles such as donations can be
subject to free riding behavior and cause respondents to overstate their willingness to pay.
Incentive Compatibility. A survey instrument is incentive compatible when respondents are motivated
to answer truthfully and do not use their responses to try to influence a particular outcome. Incentive
compatibility is driven primarily by the consequentiality of the survey and the question format used.
Consequentiality requires that survey participants believe there Is a positive probability that the survey
outcome will have actual consequences. Establishing a link between survey responses and actual
outcomes described by the scenario mitigates several types of bias associated with stated preference
valuation including hypothetical bias and yea-saying (Cummings and Taylor 1999; Carson and Groves
2007; Landry and List 2007; Vossler and Evans 2009; Herriges etal. 2010; Vossler and Watson
2013). An incentive compatible question format will reduce strategic behavior by respondents. Single
binary choice formats meet this criterion with the least assumptions though other formats may be
incentive compatible under strict conditions (Johnston etal. 2017).
Treatment of "don't know" or neutral responses. Based on recommendations from the NOAA
Blue Ribbon panel (Arrow et al. 1993), many surveys have included "don't know" or "no
preference" options. The Contemporary Guidance for Stated Preference (Johnston etal. 2017)
recommends including a no-answer option for sensitive topics but not necessarily for all
applications. There have been questions about how such responses should enter the empirical
analysis. Examining referendum-style dichotomous choice questions, Carson etal. (1998) found
that when those who chose not to vote were coded as "no" responses, the mean WTP values were
the same as when the "would not vote" option was not offered. Offering the "would not vote" option
did not change the percentage of respondents saying "yes". Thus, they recommend that if a "would
not vote" option is included, it should be coded as a "no" vote, a practice that has become
widespread. Stated preference studies should always be explicit about how they treat "don't know,"
"would not vote," or other neutral responses.
Reliability, in general terms, means consistency or repeatability. If a method is used numerous
times to measure the same commodity, then the method is considered more reliable if the
variability of the results is lower than an alternative.
Test-retest approach. Possibly the most widely applied approach for assessing reliability
in stated preference studies has been the test-retest approach. Test-retest assesses the
variability of a measure between differenttime periods. Loomis (1989), Teisl etal. (1995),
McConnell et al. (1998), and Hoban and Whitehead (1999) all provide examples of the test-
retest method for reliability.
Meta-analysis of stated preference survey results for the same good also may provide
evidence of reliability. Meta-analysis evaluates multiple studies as though each was
constructed to measure the same phenomenon. Meta-analysis attempts to sort out the
effects of differences in the valuation approach used in different surveys, along with other
factors influencing the elicited value. For example, Boyle et al. (1994) use meta-analysis to
7-51 Guidelines for Preparing Economic Analyses | 3rd edition
-------
evaluate eight studies conducted to measure values for groundwater protection (also see
Section 7.4).
Validity tests seek to assess whether WTP estimates from stated preference methods behave as a
theoretically correct WTP should. Three types of validity discussed below are: content validity,
criterion validity and convergent validity.
Content validity. Content validity refers to the extent to which the estimate captures the
concept being evaluated. Content validity is largely a subjective evaluation of whether a
study has been designed and executed in a way that incorporates the essential
characteristics of the WTP concept
To evaluate a survey instrument, analysts look for features that researchers should have
incorporated into the survey scenario. First, survey should clearly define the environmental
change being valued. The description should include a careful exposition of the conditions
in the baseline case and how these would be expected to change over time if no action were
taken. Next, the action or policy change should be described, including an illustration of how
and when it would affect aspects of the environment that people might care about. Boyd and
Banzahf (2007), and Boyd and Krupnick (2013) put a finer point on this concept and
advocate developing the valuation scenario based on "ecological endpoints" rather than
intermediate goods that are less clearly associated with outcomes of interest. For example,
if respondents ultimately care about the survival of a certain species, it is more sensible to
structure questions to ask about WTP for the species' survival than to ask about
degradation of habitat, as respondents are unlikely to know the relationship between
habitat attributes and species survival. Respondent attitudes about the provider and the
implied property rights of the survey scenario can be used to evaluate the appropriateness
of features related to the payment mechanism (Fischhoff and Furby 1988). Survey
questions that probe for respondent comprehension and acceptance of the commodity
scenario can offer important indications about the validity of the results (Bishop et al.
1997).
Criterion validity. Criterion validity assesses whether stated preference results relate to
other measures that are considered to be closer to the concept being assessed (or WTP).
Ideally, one would compare results from a stated preference study of use values (the
measure) with those from actual market data (the criterion). Another approach would be to
estimate a sample of individuals' WTP for a commodity using a stated preference survey
and then later give the same sample of individuals or a different random sample of
individuals drawn from the same population a real opportunity to buy the good (see
Mitchell and Carson 1989; Carson et al. 1987; Kealy et al. 1990; Brown et al. 1996; and
Champ etal. 1997 for examples.)
When unable to conduct such comparisons, sensitivity to scope and income has been used
to assess criterion validity. "Scope tests" are concerned with how WTP responds to changes
in the amount of the referenced good provided in the valuation scenario (Smith and
Osborne 1996; Rollins and Lyke 1998; Heberlein etal. 2005). If the referenced good is
indeed a "normal good," utility theory implies that WTP should increase with the provision
of the good. For the same reason one would expect WTP to exhibit positive income elasticity
(McFadden 1994 and Schlapfer 2006). Neither test is necessary or sufficient to establish
criterion validity (Heberlein et al. 2005), but either can serve as a useful proxy when an
alternate measure of WTP for the same good is unavailable. Diamond (1996) suggests that
stronger scope tests can be conducted by comparing departures from strict "adding up" of
WTP for partial changes and relating them to the income elasticity of WTP. Other
7-52 Guidelines for Preparing Economic Analyses | 3rd edition
-------
researchers, however, argue that the Diamond test may not be practicable and imposes a
specific structure on the preference function which may not be appropriate (Carson et al.
2001).
Convergent validity. Convergent validity examines the relationship between different measures of
a concept60 This differs from criterion validity in that one of the measures is not taken as a
criterion upon which to judge the other measure. The measure of interest and the other measure
are judged together to assess consistency with one another. If they differ in a systematic way (e.g.,
one is usually larger than another for the same good), it is not clear which one is more correct. If
stated preference estimates are being compared with revealed preference estimates, care should be
taken that the same values are being captured by both approaches. Stated preference estimates
often include non-use values whereas revealed preference estimates do not capture that portion of
total economic value.
Hypothetical bias occurs when the responses to hypothetical stated preference questions are
systematically different than what individuals would pay if the transactions were to actually occur.
Widely cited as one of the most common problems with the stated preference method (List and
Gallet 2001 and Murphy and Allen 2005), researchers have made advances in techniques to
minimize such bias. These techniques include the use of "cheap talk" methods to directly tell
respondents about the potential for hypothetical bias (Cummings and Taylor 1999 and List 2001);
calibrating hypothetical values (List and Shogren 1998 and Blomquist et al. 2009); and allowing
respondents to express uncertainty in their responses and restricting the set of positive responses
to those about which the respondent was most certain (Vossler etal. 2003). Several studies have
shown that attribute-based choice experiments reduce hypothetical bias in the bid amounts and the
marginal value of attributes relative to other elicitation methods (Carlsson and Martinsson 2001:
Murphy and Allen 2005; List etal. 2006).
Tests for hypothetical bias often involve a comparison of actual payments and responses to
hypothetical scenarios that use the same solicitation approach. The actual payments typically occur
in one of three scenarios. Market transactions are the most common (Cummings et al. 1995 and List
and Shogren 1998) but generally involve payments for private goods while most stated preference
applications are concerned with public or quasi-public goods. Simulated markets can be used to
solicit actual donations for public good provision (Champ et al. 1997). However, donation
solicitations are subject to free riding, so while it may be possible to test for hypothetical bias using
this approach, both the actual and hypothetical payment scenarios lack incentive compatibility and
may not represent total WTP. In rare instances comparisons have been made between actual
referenda for public good provision and hypothetical responses to the same scenario, but the
conditions for a valid comparison of this sort are exceedingly difficult to satisfy (Johnston 2006).
Some experiments have, however, simulated public good provision in various ways and with
varying levels of success (Carson et al. 2001; Landry and List 2007; Vossler and Evans 2009;
Vossler etal. 2012).
Non-response bias is introduced when non-respondents would have answered questions
systematically differently than those who did answer. Non-response bias can take two forms: item
non-response and survey non-response.
60 Mitchell and Carson (1989) define convergent validity and theoretical validity as two types of construct
validity. Construct validity examines the degree to which the measure is related to other measures as predicted
by theory.
7-53 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Item non-response bias occurs when respondents who agreed to take the survey do not
answer all the choice questions in the survey. Information available about respondents from
other questions they answered can support an assessment of potential item non-response
bias for the WTP questions that were unanswered. The key issue is whether there were
systematic differences in potential WTP-related characteristics of those who answered the
WTP questions and those who did not Characteristics of interest include income, gender,
age, expressed attitudes and opinions about the good or service, and information reported
on current use or familiarity with the good or service. Statistically significant differences
may indicate the potential for item non-response bias, while finding no such differences
suggests that the chance of significant non-response bias is lower. However, the results of
this comparison are only suggestive because respondents and non-respondents may only
differ in their preference for the good in question (McClelland et al. 1991).
Survey non-response bias is created when those who refuse to take the survey have
preferences that are systematically different from the preferences of those who do respond.
Although it is generally thought that surveys with high response rates are less likely to
suffer from survey non-response bias, it is not a guarantee.61 For survey non-respondents,
there may be no available data to determine how they might systematically differ from
those who responded to the survey. The most common approach is to examine the relevant
measurable characteristics of the respondent group, such as income, resource use, gender,
age, etc., and to compare them to the characteristics of the study population. Similarity in
mean characteristics across the two groups suggests that the respondents are
representative of the study population and that non-response bias is expected to be
minimal.
A second way to evaluate potential survey non-response bias is to conduct a short follow-up
survey with non-respondents. This can sometimes be accomplished through interviews
conducted during the recruiting phase. Such follow-ups typically ask a few questions about
attitudes and opinions on the topic of the study as well as collecting basic socioeconomic
information. Questions need to match those in the full survey closely enough to compare
non-respondents to respondents. The follow-up must be very brief, or response rates will
be low (OMB 2006).
\ Revealed and " rv- - Ki - ^rence Data
Instead of looking at revealed preference and stated preference data as two separate methods for
estimating environmental benefits, many researchers have used them in combination. The practice
has been in use much longer in the marketing and transportation literature and many of the lessons
learned by those researchers are now being employed in environmental economics. In theory, the
strengths of each data type should help overcome some of the weaknesses of the other. As
described by Whitehead et al. (2008) in an assessment of the state of the science, the advantages of
combining revealed preference and stated preference data include:
61 Note that OMB's Guidance on Agency Survey and Statistical Collections (OMB 2006) has fairly strict
requirements for response rates and their calculation for Agency-sponsored surveys, recommending that "ICRs
for surveys with expected response rates of 80 percent or higher need complete descriptions of the basis of the
estimated response rate...ICRs for surveys with expected response rates lower than 80 percent need complete
descriptions of how the expected response rate was determined, a detailed description of steps that will be taken
to maximize the response rate...and a description of plans to evaluate non-response bias" (pp. 60-70).
7-54 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Helping to ground the hypothetical stated preference data with real world behavior,
potentially decreasing any hypothetical bias;
Providing the ability to test the validity of both data sources;62
Increasing the range of historical stated preference data to include conditions not observed in
the past and thereby reducing the need to make predictions outside of the sample;
Increasing the sample size;
Extending the size of the market or population to include larger segments than captured by
either method alone; and
Exploiting the flexibility of stated preference experimental design to overcome revealed
preference data's potential multicollinearity and endogeneity problems fvon Haefen and
Phaneuf 2008).
The different strategies for combining revealed preference and stated preference data can be
grouped into three categories. The first two methods rely on joint estimation. If the revealed
preference and stated preference data have similar dependent and independent variables and the
same assumed error structures, then they can simply be pooled together and treated as additional
observations (Adamowicz et al. 1994; Boxall, Englin, and Adamowicz 2003; and Morgan, Massey,
and Huth 2009). If the revealed preference and stated preference data sources cannot be pooled, it
is sometimes possible to use them in a jointly estimated mixed model that relies on a utility
theoretic specification of the underlying WTP function (Huang, et al. 1997; Kling 1997; Eom and
Larson 2006, Jeon and Herriges 2016, Whitehead and Lew 2020; Hindsley etal. 2022). If the data
cannot be combined in estimation, it can still be useful to estimate results separately and then use
them to test for convergent validity between the two data sources (Carson et al. 1996 and Schlapfer
etal. 2004).
As noted at the outset of this chapter, benefit transfer is the approach most often used by the
Agency for monetizing benefits in economic analysis. Benefit transfer refers to the use of estimated
values of environmental quality changes from primary studies to the evaluation of similar changes
that are of interest to the analyst (Freeman et al. 2014). The case under consideration for a new
policy is referred to as the "policy case." Cases from which estimates are obtained are referred to as
"study cases." A benefit transfer study identifies stated preference or revealed preference study
cases that sufficiently relate to the policy context and transfers their results to the policy case.
Benefit transfer is necessary when it is infeasible to conduct an original study focused directly on
the policy case and is the most common approach for completing a BCA at the EPA. Given the time
and analytical resource constraints under which most regulatory analysis activities are conducted,
conducting new revealed or stated preference studies that are tailor-made to examine all of the
(sometimes numerous) endpoints changed by the policy or regulation in question is near
impossible (Newbold et al. 2018a). Because original studies are time consuming and expensive,
benefit transfer can reduce both the time and financial resources required to develop estimates of a
proposed policy's benefits. Benefit transfer might also be useful as a scoping exercise to predict the
62 Herriges, Kling, and Phaneuf (2004) point out that revealed preference may not always be valid for
estimating WTP for quality changes when weak complementarity cannot be assured.
7-55 Guidelines for Preparing Economic Analyses | 3rd edition
-------
approximate magnitude of benefits that might then be more precisely estimated with an original
study.
While there is no universally accepted single approach for conducting benefit transfer, there are
some generalized steps involved in the process. These steps are described below.
Figure 7.3 - Steps for Conducting Benefit Transfer
Step 1: Describe
the Policy Case
Step 2: Select
Study Cases
\
Step 4: Report
Results
Step 1: Describe the Policy Case
The first step in a benefit-transfer study is to clearly describe the policy case with respect to the
baseline so that its characteristics and consequences are well-understood. Are human health risks
reduced by the policy intervention? Are ecological benefits expected (e.g., increases in populations
of species of concern)? It is also important to identify to the extent possible the population affected
by the proposed policy and to describe their demographic and socioeconomic characteristics (e.g.,
users of a particular set of recreation sites, children living in urban areas or older adults across the
United States). Information on the affected population is generally required to translate per person
(or per household) values to an aggregate benefits estimate.
Step 2: Select Study Cases
A benefit-transfer study is only as good as the study cases from which it is derived, and it is
therefore crucial that studies be carefully selected. First, the analyst should identify potentially
relevant studies conducting a comprehensive literature search. Online searchable databases
summarizing valuation research may be especially helpful at this stage.63 Because peer-reviewed
academic journals may be more likely to publish work based on methodological contributions,
analysts should be aware of the potential for publication bias. Some studies of interest may be
found in government reports, working papers, dissertations, unpublished research and other "gray
literature" (Rosenberger and Stanley 2006; Johnston and Rosenberger 2010; Johnston et al.
2015).64 While including studies from the gray literature may help mitigate publication bias, use of
valuation estimates that are not published in peer-reviewed journals may necessitate subsequent
peer review in some form (i.e., formal peer review or a less formal peer input). See U.S. EPA (2015a)
for more guidance (in particular, Sections 3.5.7 and 3.5.8).
63 For example, the Environmental Valuation Reference Inventory (EVRI) is maintained by Environment Canada
and managed by a cross-county working group. EVRI contains summaries of over 4,000 studies that can be
referenced according to keyword, study type, region and environmental asset. EVRI also provides a bibliography
on benefit transfer. See www.evri.ca for more information.
64 Newer, unpublished research may also be on the cutting edge of methods.
7-56 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Next, the analyst should develop an explicit set of selection criteria to evaluate each of the
potentially relevant studies for quality and applicability to the policy case. The quality of the value
estimates in the study cases will in large part determine the quality of the benefit transfer. As a first
step, the analyst should review studies according to the criteria listed for each methodology in the
previous sections in this chapter. Results from study cases must be valid as well as relevant.
Concerns about the quality of the studies, as opposed to their relevance, will generally hinge on the
methods used. Valuation approaches commonly used in the past may now be regarded as
unacceptable for use in benefits analysis. Studies based on inappropriate methods or reporting
obsolete results should be removed from consideration.
It is unlikely that any single study will match perfectly with the policy case; however, study cases
potentially suitable for use in benefit transfer should be similar to the policy case in their: (1)
definition of the environmental commodity being valued (include scale and presence of
substitutes); (2) baseline and extent of environmental changes; and (3) characteristics of affected
populations.65 Analysts should avoid using benefit transfer in cases where the policy or study case
have large differences in context-especially concerning goods with unique attributes (such as a
national park), where the valuation estimate is ex ante and the policy case is ex post-especially if
the policy introduces a significant change in the attributes of the good, or where the magnitude of
the change or improvement across the two cases differs substantially (OMB 2023).66 It is crucial to
remember that economic value is determined on the margin and depends upon how scarce
something is relative to the demand for it at the time and place it is provided (Simpson 2017).
The analyst should determine whether adjustments should and can be made for important
differences between each study and policy case. For example, some case studies will be based on
stated preference methods while others may be based on revealed preference methods. The ability
of the analyst to make these adjustments will depend, in part, on both the number of value
estimates for suitably similar study sites and the method used to combine these estimates. These
methods are now discussed in turn.
Step 3: Transfer Values
There are several approaches for transferring values from study cases to the policy case. These
include unit value transfers and function transfers, and they may use techniques from meta-
analysis if multiple studies are available. Transfers may also be structural or non-structural
(referring to a utility-theoretic structure). Each of these approaches is typically used to develop per
person or per household value estimates that are then aggregated over the affected population to
estimate total benefits. In general, when reporting transfer results, researchers should provide
information on the background of the problem, the strategy for selecting studies, analytic methods used,
results, discussion and conclusions.
Unit value transfers are the simplest of the benefit-transfer approaches. They take a point
estimate of WTP for a unit change in the environmental resource from a study case or cases
and apply it directly (or simply adjusted) to the policy case. The point estimate may be a
65 In some cases, the transfer method itself may inform the choice of study cases to include. Meta-analysis
approaches (discussed below) can facilitate some forms of statistical validity testing. For example, Moeltner
(2015, 2019) uses Bayesian methods in a meta-analysis to identify optimal pooling of studies.
66 OMB Circular A-4 provides other guidance for benefit transfer. Analysts can also consider whether a function
transfer that includes the adding-up condition (Text Box 7.6) can account for differing magnitudes across policy
cases and study cases.
7-57 Guidelines for Preparing Economic Analyses | 3rd edition
-------
single estimated value from a single case study, but it can also be the average of a small
number of estimates from a few case studies. For example, a study may have found a WTP of
$20 per household for a one-unit increase on a water quality scale. A simple unit value
transfer would estimate total benefits for the policy case by multiplying $20 by the number
of units by which the policy is expected to increase water quality and by the number of
households who will benefit from the change. This approach can be useful for developing
preliminary, order-of-magnitude estimates of benefits, but it should be possible to base final
benefit estimates on more information than a single point estimate from a single study. If
multiple studies are available, the mean or median WTP value may provide a useful point
estimate, though analysts should consider weighting estimates by inverse variance or
sample size to give more weight to more precise estimates when calculating the mean
(Nelson and Kennedy 2009; Nelson etal. 2013). Point estimates reported in study cases are
typically functions of several variables, and simply transferring a summary estimate
without controlling for differences among these variables can yield inaccurate results.
Therefore, unit values are often adjusted to account for these differences, e.g., changes in
income over time (Boardman etal. 2018). It is important to recognize that unit value
transfer assumes that the original good, as well as the characteristics and tastes of the
population of beneficiaries, are the same as the policy good.
Function transfers use information on other factors that influence WTP to adjust the unit
value for quantifiable differences between the study case and the policy case. This is
accomplished by transferring the estimated function upon which the value estimate in the
study case is based to the policy case. This approach implicitly assumes that the population
of beneficiaries to which the values are being transferred has potentially different
characteristics but similar tastes as the original one and allows the analyst to adjust for these
different characteristics. Generally, benefit function transfers maybe preferable to unit value
transfers as they incorporate information relevant to the policy scenario (OMB 2023;
Johnston and Rosenberger 2010).
To implement a function transfer, suppose that in the hypothetical example above the $20
unit value was the result of averaging the results of an estimated WTP function over all
individuals in the study case sample, where the WTP function included income, the baseline
water quality level and the change in the water quality level for each household. A function
transfer would estimate total benefits for the policy case by:
1. Applying the WTP function to a random sample of households affected in the policy
case using each household's observed levels of income, baseline water quality and
water quality change;
2. Averaging the resulting WTP estimates; and
3. Multiplying this average WTP by the total number of households affected in the policy
case.
If the WTP function is nonlinear and statistics on average income, baseline water quality,
and water quality changes are used in the transfer instead of household level values, then
bias will result. Feather and Hellerstein (1997) provide an example of a function transfer
that attempts to correct for such bias. Although function transfers can adjust and
compensate for small differences between the case and policy study populations, they are
subject to the same basic usage rules governing unit value transfers. Function transfers
should only be used if the case and policy studies are evaluating sufficiently similar
environmental goods, change in environmental levels, and affected populations.
7-58 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Meta-analysis uses results from multiple valuation estimates in a new unit or function
transfer.67 Meta-analysis is an umbrella term for a suite of techniques that synthesize the
results of empirical research. This could include a simple ranking of results, a meta-analytic
average or other central tendency estimate, or a multivariate regression. The advantage of
these meta-analytic methods is that they incorporate and can potentially control for more
information than transfers based on a single estimate. This approach has been widely used in
environmental economics (see Rolfe etal. 2015, Johnston etal. 2018, Guignetetal. 2022, and
Newbold etal. 2018a).
There are several guidelines for meta-analyses that outline protocols that should be followed in
conducting or evaluating a study. See Bergstrom and Taylor (2006); Nelson and Kennedy (2009);
Nelson etal. (2013); Nelson (2015): and Boyle etal. (2013, 2015) for more information. Some
may choose to follow a systematic review protocol as described in Text Box 7.5. The EPA's Peer
Review Handbook (U.S. EPA 2015b) recommends that meta-analyses used in regulatory analysis
should generally be peer-reviewed. Boyle and Wooldridge (2018) emphasize that the purpose of
a meta-analysis for benefit transfer is prediction and the purpose of a traditional meta-analysis is
to summarize a literature. This latter paper provides a number of technical suggestions to
"provide the best econometric prediction of value for a benefit-transfer application."
Structural benefit transfer involves deriving a benefit transfer function from an assumed
form of the direct or indirect utility function and calibrating or estimating the form of the
transfer function using insights from economic theory.68 The advantages of structural
transfer functions are that they can accommodate different types of economic value
measures (e.g., WTP, WTA or consumer surplus) and can be constructed to satisfy certain
theoretical consistency conditions (e.g., WTP bounded by income). Using a structural
benefit transfer or preference calibration approach is one way to ensure that that the
adding-up condition holds (see Text Box 7.6). However, Johnston etal. (2018) discuss the
tradeoff between theory and accuracy of the transfer in structural benefit transfer. They
conclude that core concepts such as diminishing marginal utility are necessary but there can
be a trade-off between empirical accuracy of transfers and imposing a specific functional
form to satisfy stronger theoretical restrictions; and there is not a consensus in the
empirical literature on the appropriate balance.
67 A typical meta-analysis combines estimates from many studies, but meta-analyses that combine multiple
estimates from one study or more than one application of the same protocol are also common. This latter type is
often referred to as an internal meta-analysis. See Text Box 7.5 for an example of a study that applied an internal
meta-analysis.
68 See Smith and Pattanayak (2002) and Smith, Pattanayak, and van Houtven (2006) for descriptions on the
method. See Newbold etal. (2018b), discussed in Text Box 7.6, for an example of a functional form of a meta-
analysis being based on theory.
7-53 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 7.5 - Systematic Review Approaches and Benefit Transfer
EPA risk assessments have increasingly adopted systematic review approaches as
recommended by the National Research Council (NRC) (NRC 2014), and these approaches may
provide a useful model for identifying and evaluating literature for a benefit transfer or meta-
analysis. The Institute of Medicine (IOM) has defined systematic review as "a scientific
investigation that focuses on a specific question and uses explicit, prespecified scientific
methods to identify, select, assess, and summarize the findings of similar but separate studies"
(IOM 2011). A key element in conducting a systematic review is preparation of a protocol which
details in advance the methods that will be used in conducting the review. Major advantages of
systematic review include improved documentation and transparency, as well as minimization
of potential bias in how the review is conducted (NRC 2014).
The steps in conducting a systematic review, as outlined by the NRC (2014), are:
Problem Formulation: define the study question (roughly equivalent to "describe the
policy case");
Develop a protocol for conducting the systematic review: the protocol defines the
methods to be used (e.g., search strategy, inclusion/exclusion criteria study evaluation
criteria);
Evidence Identification: conduct the literature search and screen the literature search
results (apply the search strategy and the inclusion/exclusion criteria from the protocol
to identify relevant studies);
Evidence Evaluation: evaluate quality of studies by applying the criteria specified in the
protocol to each included study; and
Evidence Integration: develop conclusions from the included studies to answer the
study question.
Which benefit transfer method to choose is notalways obvious. Boyle etal. (2013) note
there is no consensus on which method works best There have been numerous studies
comparing the (convergent) validity and reliability of transfers (see Rosenberger 2015 and
Kaul et al. 2013 for summaries). Some general lessons are "that function transfers tend to
be more accurate than value transfers; transfers of values for environmental quantity
changes tend to be more accurate than those for quality changes; geographic similarity
between sites improves the accuracy of transfers, especially for value transfers; combining
information from multiple studies improves the accuracy of transfers; and that transfers
based on stated preference valuation formats with more options per question, such as
choice experiments, have larger transfer errors than methods with fewer choices per
question, such as contingent valuation surveys" (Newbold etal. 2018a). Few studies test the
validity and reliability of meta-analytic transfers, however. Johnston et al. (2018) describe
how increasingly complex methods may not always be worthwhile, noting that more
flexible transfer functions tend to outperform unit value transfers but value transfers may
outperform other types of transfers when the sites are very similar. Function transfers that
adjust estimates for a few key variables (e.g., income elasticity) may have lower transfer
error than complex function transfers that control for numerous characteristics. The benefit
transfer literature is large and diverse. The EPA will continue to monitor it and update these
recommendations as necessary.
7-60 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 7.6 - The Adding-Up Condition in Benefit Transfer
When a benefit transfer function is estimated in a meta-analysis or in some other fashion, it is
important to carefully consider the form of the estimating equation used to relate study
characteristics to willingness-to-pay. For consistency in regulatory accounting, a willingness-to-
pay function must satisfy a basic adding-up condition (Kling and Phaneuf 2018). In other words,
WTP for good X, plus WTP for good Y given good X, must equal WTP for X and Y together.
A benefit transfer function that violates the adding-up condition can lead to inconsistent policy
evaluations. For example, an omnibus policy that appears to have lower net benefits than a set
of component policies that, when combined, yield the same water quality outcomes and have
the same total cost as the omnibus policy does not satisfy the adding up condition. In this case,
the policy change appears to have larger benefits if it is broken up into several smaller policy
changes. A benefit transfer function that violates the adding-up condition also could lead to
inconsistent policy rankings, since independently evaluating the provision of goods X and Y
could pass a benefit-cost test while evaluating the provision of both goods X and Y together
could fail a benefit-cost test.
Analysts who use meta-analysis to estimate a benefit transfer function or apply a benefit
transfer function developed in a previous study should ensure that the resulting willingness-to-
pay function satisfies the adding-up condition. If the function fails to satisfy the adding-up
condition, the analyst should consider re-estimating the benefit transfer function using a
different functional form that does satisfy the adding-up condition. One way to ensure that a
benefit transfer function complies with the adding-up condition is to use a "structural benefit
transfer" or "preference calibration" approach, as described in the main text
Newbold et al. (2018b) examine existing valuation studies and document violations of the
adding-up condition and impacts on benefit-cost results because of these violations. They
further describe a structural meta-analytic model that meets the adding-up condition and
compare it to a non-structural model that does not. They find that the nonstructural model
produces much larger benefits estimates than the structural model and that the violations of the
adding-up condition are severe in the non-structural model.
Step 4: Report the Results
In addition to reporting the final benefit estimates from the transfer exercise, the analyst should
clearly describe all key judgments and assumptions, including the criteria used to select study cases
and the choice of the transfer approach. The uncertainty in the final benefit estimate should be
quantified and reported when possible. Any limitations should also be discussed69 (see Chapter 11
on Presentation of Analysis and Results).
7.5 Accommodating Benefits that Cannot Be Quantified
and/or Monetized
It often will not be possible to quantify and value every significant benefit or endpoint for all policy
options. For example, it often is not possible to quantify the various ecosystem changes that may
69 See Stanley et al. (2013) for additional recommendations for reporting on meta-analyses.
7-61 Guidelines for Preparing Economic Analyses | 3rd edition
-------
result from an environmental policy. While Chapter 11 discusses how to present these benefits to
provide a fuller accounting of all effects, this section discusses what analysts can do to incorporate
these endpoints more fully into the analysis.
" * i ilitatr Discussions
When there are potentially important effects that cannot be quantified, the analyst should include a
qualitative discussion of the benefit endpoints. The discussion should explain why a quantitative
analysis was not possible and the reasons for believing that these non-quantified effects may be
important for decision making. Chapter 11 discusses how to describe benefit categories that are
quantified in physical terms but not monetized.
" * C v I r^rnatrv i ialytical Approaches
Alternative analyses exist that can support benefits valuation when robust value estimates and/or
risk estimates are lacking. These analyses, including cost-effectiveness, break-even and bounding
analysis, can provide decision makers with some useful information. However, analysts should
remember that because these alternatives do not estimate the net benefits of a policy or regulation,
they fall short of BCA in their ability to identify an economically efficient policy. This shortcoming
and any others should be discussed when presenting results from these analyses to decision
makers.
"WI i i iCi i -» nivene ualysis
Cost-effectiveness analysis (CEA) is most useful when outcome measures from a policy are not in
dollar terms, for example, the number of expected premature mortalities avoided, or "lives saved."
Cost-effectiveness is calculated by dividing the annualized cost of the option by the annual non-
monetary outcome measure, resulting in a ratio of cost per unit (e.g., dollar per life saved). Because
the outcome is a ratio, it is more sensitive to how benefits and costs are characterized. Whereas a
net benefits outcome (benefits minus costs) is robust to whether negative benefits are counted as
costs, the same is not true for a ratio. This is one reason net benefits are generally preferred to
benefit-cost ratios.
The economically preferred option from CEA is clear when all options achieve the same result (e.g.,
the same number of tons reduced): the option with the lowest cost per unit is the most cost-
effective option. More typically, however, options vary not only in costs but also in the outcomes
they produce, and there is no generally accepted criterion defining the economically preferred
option in this case. Each cost-effectiveness ratio represents a different trade-off between the
outcome measure and costs, but there is no information on which, if any, of the options is efficient.
Still, cost-effectiveness on outcomes can inform decision-making in the absence of monetized
benefits.
Because cost-effectiveness is defined by cost-per-unit, CEA requires a single outcome measure. It is
not possible to perform a CEA where there are two separate outcomes, which is often the case for
environmental regulations. For example, if a program reduces both hydrocarbon and nitrogen
oxide emissions, it is probably not possible to develop a cost per ton of hydrocarbons reduced and a
separate cost per ton of nitrogen oxides, because the same costs produce both outcomes. For health
and safety regulations, however, there are a number of measures that integrate disparate health
outcomes into a single metric for cost-effectiveness calculations. These metrics were largely
7-62 Guidelines for Preparing Economic Analyses | 3rd edition
-------
developed for comparing public health or medical interventions. The most common metric is a
quality-adjusted life year (QALY), which combines health-related quality of life with longevity. Cost-
effectiveness using QALYs is sometimes referred to as "cost-utility analysis" (CUA) because the
health-related quality of life component is based on stated preferences about the impact of different
health conditions.
The application of QALYs to regulatory analysis has been evaluated in detail by the IOM (IOM
2006). It is important that cost-effectiveness analysis using QALYs be distinct from BCA. Converting
QALYs to a monetary value using a "cost per QALY" is not fully consistent with utility theory
underlying BCA (IOM 2006; Hammitt 2002), but is an approach suggested for consideration in OMB
Circular A-4 (2023). When there is a BCA, cost-effectiveness analysis should be considered a
complement that provides a different perspective on the trade-offs of a regulatory action.
"WI I eak-Even an i unding Analysis
Chapter 5 describes several approaches for analyzing and characterizing uncertainty. Two methods
that can be particularly useful for benefits analysis with missing information are break-even and
bounding analysis. For example, analysts who have per unit estimates of economic value, but lack
risk estimates can use break-even analysis to estimate the number of cases (each valued at the per
unit value estimate) at which overall net benefits become positive, or where the policy action will
break even.70 This estimate then can be assessed for plausibility either quantitatively or
qualitatively. The same sort of analysis can be performed when analysts lack valuation estimates,
producing a break-even value that should again be assessed for credibility and plausibility. Policy
makers will need to determine if any break-even value is acceptable or reasonable. Bounding
analysis can help when analysts lack value estimates for a particular endpoint Reducing the risk of
health effects that are more severe and of longer duration should be valued more highly than those
that are less severe and of shorter duration, all else equal. If robust valuation estimates are
available for effects that are unambiguously "worse" and others that are unambiguously "not as
bad," then one can use these estimates as the upper and lower bounds on the value of the effect of
concern.
70 Circular A-4 (OMB 2023) refers to these values as "switch points" in its discussion of sensitivity analysis.
Section 5.4.4 on uncertainty analysis also contains related discussions on switch points.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 7 References
Abbott, J.K. and H.A. Klaiber. 2010. An Embarrassment of Riches: Confronting Omitted Variable Bias and
Multi-Scale Capitalization in Hedonic Price Models. The Review of Economics and Statistics, 93(4): 1331-1342.
Adamowicz, W.L., J. Louviere, and M. Williams. 1994. Combining Revealed and Stated Preference Methods for
Valuing Environmental Amenities. Journal of Environmental Economics and Management, 26: 271-292.
Adamowicz, W.L., J. Swait, P. Boxall, J. Louviere, and M. Williams. 1997. Perceptions versus Objective
Measures of Environmental Quality in Combined Revealed and Stated Preference Models of Environmental
Valuation. Journal of Environmental Economics and Management, 32: 65-84.
Adams, R.M. and T.D. Crocker. 1991. Materials Damages. Measuring the Demand for Environmental Quality, ed.
J. Braden and C. Kolstad. North Holland: Elsevier Science Publishers. 271-302.
Alberini, A. 1995. Optimal Designs for Discrete Choice Contingent Valuation Surveys: Single-Bound, Double-
Bound, and Bivariate Models. Journal of Environmental Economics and Management, 28(3): 287-306.
Alberini, A., M. Cropper, A. Krupnick, and N.B. Simon. 2004. Does the Value of a Statistical Life Vary with Age
and Health Status? Evidence from the United States and Canada. Journal of Environmental Economics and
Management, 48: 769-792.
Alberini, A. and A. Krupnick. 2000. Cost-of-Illness and WTP Estimates of the Benefits of Improved Air Quality:
Evidence from Taiwan. Land Economics, 76(1): 37-53.
Anselin, L. 2001. Spatial Econometrics. v4 Companion to Theoretical Econometrics, ed. B.H. Baltagi. Oxford, UK:
Blackwell Publishing Ltd.
Arrow, K.J., etal. 1993. Report of the NOAA Panel on Contingent Valuation. Federal Register, 58(10): 4601-
4614. Available at:
https://www.researchgate.net/publication/235737401 Report of the NOAA panel on Contingent Valuatio
n (accessed December 23, 2020).
Bagstad, K.J., D. Semmens, R. Winthrop, D. Jaworski, and J. Larson. 2012. Ecosystem Services Valuation to
Support Decision Making on Public Lands: A Case Study for the San Pedro River, Arizona. USGS Scientific
Investigations Report, 5251. Available at: https://puhs.er.usgs.gov/puhlication/sir20125251 (accessed
December 23, 2020).
Barbier, E.B. 2012. Progress and Challenges in Valuing Coastal and Marine Ecosystem Services. Review of
Environmental Economics and Policy, 6(1): 1-19.
Bartik, T.J. 1988. Measuring the Benefits of Amenity improvements in Hedonic Price Models. Land Economics,
64: 172-183.
Bateman, I.J., 2012. National Ecosystem Assessment: the UK Experience. Review of Environment, Energy and
Economics.
Bell, F.W., and V.R. Leeworthy. 1990. Recreational Demand by Tourists for Saltwater Days. Journal of
Environmental Economics and Management, 18(3): 189-205.
Bergstrom, J.C. and L.O. Taylor. 2006. Using Meta-Analysis for Benefits Transfer: Theory and Practice.
Ecological Economics, 60: 351-360.
Berman, M.D. and H.J. Kim, 1999. Endogenous on-site time in the recreation demand model. Land Economics:
603-619.
Berndt, E.R. 1991. The Practice of Econometrics: Classic and Contemporary. Boston: Addison-Wesley.
Bin, 0. and S. Polasky. 2005. Evidence on the Amenity Value of Wetlands in a Rural Setting. Journal of
Agricultural and Applied Economics, 37(3): 589-602.
Bishop, Kelly C., N.V. Kuminoff, H.S. Banzhaf, K.J. Boyle, K. von Gravenitz, J.C. Popek, V.K. Smith, and C.D.
Timmins. 2020. Review of Environmental Economics and Policy, 14(2): 260-281.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Bishop, R.C., P.A. Champ, T.C. Brown, and D.W. McCollum. 1997. Non-Use Values: Theory and Empirical
Applications. In Determining the Value of Non-Market Goods: Economic, Psychological, and Policy Relevant
Aspects of Contingent Valuation Methods, ed. R. Kopp, W. Pommerehne, and N. Schwarz. Boston: Kluwer. 59-
82.
Bishop, R.C. andT. Heberlein. 1979. Measuring Values of Extra-Market Goods: Are Indirect Measures Biased?
American Journal of Agricultural Economics, 61: 926-930.
Bishop, R.C. and Welsh, M.P., 1992. Existence Values in Benefit-cost Analysis and Damage Assessment. Land
Economics, 405-417.
Bishop, K.C. and C. Timmins. 2018. Using Panel Data to Easily Estimate Hedonic Demand Functions. Journal of
the Association of Environmental and Resource Economists, 5(3): 517-543.
Blomquist, G. 2004. Self-protection and Averting Behavior, Values of Statistical Lives, and Benefit Cost
Analysis of Environmental Policy. Review of Economics of the Household, 2: 89-110.
Blomquist, G.C., K. Blumenschein, and M. Johannesson. 2009. Eliciting Willingness to Pay without Bias Using
Follow-Up Certainty Statements: Comparisons between Probably/Definitely and a 10-Point Certainty Scale.
Environmental and Resource Economics, 43(4): 473-502.
Boardman, A.E., D.H. Greenberg, A.R. Vining, and D.L. Weimer. 2018. Cost-Benefit Analysis: Concepts and
Practice, 5th Ed Cambridge, United Kingdom: Cambridge University Press.
Bockstael, N.E., I.E. Strand, and W.M. Hanemann. 1987. Time and the Recreation Demand Model. American
Journal of Agricultural Economics, 69: 293-302.
Bockstael, N.E., K.E. McConnell, and I.E. Strand. 1991. Recreation, in Environmental Health Effects. Measuring
the Demand for Environmental Quality, ed. J. Braden and C. Kolstad. North Holland: Elsevier Science
Publishers.
Boxall, P.C., J. Englin, and W.L. Adamowicz. 2003. Valuing Aboriginal Artifacts: A Combined Revealed-Stated
Preference Approach .Journal of Environmental Economics and Management, 45: 213-230.
Boyd, J. and S. Banzhaf. 2007. What are Ecosystem Services? The Need for Standardized Environmental
Accounting Units. Ecological Economics, 63(2/3): 616-626.
Boyd, J. and A. Krupnick. 2013. Using Ecological Production Theory to Define and Select Environmental
Commodities for Nonmarket Valuation. Agricultural and Resource Economic Review, 42 (1): 1-32.
Boyle, K.J., T.P. Holmes, M.F. Teisl, and B. Roe. 2001. A Comparison of Conjoint Analysis Response Formats.
American Journal of Agricultural Economics, 83(2): 441-454.
Boyle, K.J., S. Kaul, and C.F. Parmeter. 2015. Meta-analysis: Econometric Advances and New Perspectives
Toward Data Synthesis and Robustness. Benefit Transfer of Environmental and Resource Values, 383-418.
Boyle, K.J. and S. Ozdemir. 2009. Convergent Validity of Attribute-Based, Choice Questions in Stated-
Preference Studies. Journal of Environmental and Resource Economics, 42: 247-264.
Boyle, K. J., C. C. Parmeter, B. Boehlert, and R. Paterson. 2013. Due Diligence in Meta-Analyses to Support
Benefit Transfers. Environmental and Resource Economics, 55(3): 357-386.
Boyle, K., R. Paterson, R. Carson, C. Leggett, B. Kanninen, J. Molenar, J. Neumann. 2016. Valuing Shifts in the
Distribution of Visibility in National Parks and Wilderness Areas in the United States .Journal of
Environmental Management, 173:10-22.
Boyle, K.J., G.L. Poe, and J.C. Bergstrom. 1994. What Do We Know About Groundwater Values? Preliminary
Implications from a Meta-Analysis of Contingent-Valuation Studies. American Journal of Agricultural
Economics, 76:1055-1061.
Boyle, K.J., M.P. Welsh, and R.C. Bishop. 1988. Validation of Empirical Measures of Welfare Change: A
Comparison of Nonmarket Techniques: Comment and Extension. Land Economics, 64(1): 94-98.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Boyle, K.J. and J. Wooldridge. 2018. Understanding Error Structures and Exploiting Panel Data in Meta-
analytic Benefit Transfers. Environmental and Resource Economics, 69: 609-635.
Breitburg, D. L., D. W. Hondorp, L. A. Davias, and R. J. Diaz. 2009. Hypoxia, Nitrogen, and Fisheries:
Integrating Effects Across Local and Global Landscapes. Annual Review of Marine Science, 1: 329-349.
Brown, G.J. and R. Mendelsohn. 1984. The Hedonic Travel Cost Method. Review of Economics and Statistics, 66:
427-33.
Brown, T.C., P.A. Champ, R.C. Bishop, and D.W. McCollum. 1996. Which Response Format Reveals the Truth
about Donations to a Public Good? Land Economics, 72(2): 152-166.
Buchanan, Jr., J.M. 1969. External Diseconomies, Corrective Taxes, and Market Structure. American Economic
Review, 59:174-177.
Cameron, T.A. and J.R. DeShazo. 2013. Demand for Health Risk Reductions. Journal of Environmental
Economics and Management, 65(1): 87-109.
Cameron, T.A. and D.D. Huppert. 1991. Referendum Contingent Valuation Estimates: Sensitivity to the
Assignment of Offered Values. Journal of American Statistics Association, 86(416): 910-918.
Cameron, T.A. and M.D. James. 1987. Efficient Estimation Methods for Use with 'Closed-Ended' Contingent
Valuation Survey Data. Review of Economics and Statistics, 69: 269-276.
Cameron, T.A. and J. Quiggin. 1994. Estimation Using Contingent Valuation Data from a 'Dichotomous Choice
with Follow-Up' Questionnaire. Journal of Environmental Economics and Management, 27: 218-234.
Cameron, T.A. and J. Quiggin. 1998. Estimation Using Contingent Valuation Data from a 'Dichotomous Choice
with Follow-Up' Questionnaire: Reply. Journal of Environmental Economics and Management, 35(2): 195-199.
Carbone, J.C. and V.K. Smith. 2008. Evaluating Policy Interventions with General Equilibrium Externalities.
Journal of Public Economics. 92:1254-1274.
Carlsson, F. and P. Martinsson. 2001. Do Hypothetical and Actual Marginal Willingness to Pay Differ in Choice
Experiments?: Application to the Valuation of the Environment. Journal of Environmental Economics and
Management, 41(2): 179-192.
Carson, Richard T., and Theodore Groves. 2007. Incentive and informational properties of preference
questions. Environmental and Resource Economics, 37:181-210.
Carson, R.T., N.E. Flores, and W.M. Hanemann. 1998. Sequencing and Valuing Public Goods. Journal of
Environmental Economics and Management, 36(3): 314-323.
Carson, R.T., N. Flores, and N. Meade. 2001. Contingent Valuation: Controversies and Evidence. Environmental
and Resource Economics, 19(2): 173-210.
Carson, R.T., W.M. Hanemann, and R.C. Mitchell. 1987. The Use of Simulated Political Markets to Value Public
Goods. Discussion Paper 87-7, Department of Economics, University of California, San Diego, March.
Carson, R.T., W.D. Shaw, S.E. Ragland, and J.L. Wright. 1996. Using Actual and Contingent Behavior Data with
Differing Levels of Time Aggregation to Model Recreation Demand. Journal of Agricultural and Resource
Economics, 21:130-149.
Carson, R.T., Groves, T., and Machina, M.J., 1997. Stated Preference Questions: Context and Optimal Response.
National Science Foundation Preference Elicitation Symposium, University of California, Berkeley.
Champ, P.A., R.C. Bishop, T.C. Brown, and D.W. McCollum. 1997. Using Donation Mechanisms to Value Nonuse
Benefits from Public Goods .Journal of Environmental Economics and Management, 33(2): 151-162.
Chay, K. and M. Greenstone. 2005. Does Air Quality Matter? Evidence from the Housing Market. Jo urn al of
Political Economy, 113(2): 376-424.
Chiu, W.A. and Slob, W. 2015. A Unified Probabilistic Framework for Dose-Response Assessment of Human
Health Effects. Environmental Health Perspectives. DOI: http: //dx.doi.org/10.1289 /ehp.14Q9385
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Cho, S.H., Bowker, J.M. and Park, W.M., 2006. Measuring the Contribution of Water and Green Space Amenities
to Housing Values: An Application and Comparison of Spatially Weighted Hedonic Models. Journal of
Agricultural and Resource Economics-. 485-507.
Coelli, T.J., D.S.P. Rao, C.J. O'Donnell, G.E. Battese. An Introduction to Efficiency and Productivity Analysis, 2nd
Edition. 2005. Springer Science+Business Media, Inc.
Cooper, J. and J. Loomis. 1992. Sensitivity of Willingness-to-Pay Estimates to Bid Design in Dichotomous
Choice Contingent Valuation Models. Land Economics, 68(2): 211-224.
Corona J., T. Doley, C. Griffiths, M. Massey, C. Moore, S. Muela, B. Rashleigh, W. Wheeler, S.D. Whitlock, J.
Hewitt. 2020. An Integrated Assessment Model for Valuing Water Quality Changes in the U.S. Land Economics,
96(4):478-492.
Crocker, T. D., & Horst, R. L. 1981. Hours of Work, Labor Productivity, and Environmental Conditions: A Case
Study. The Review of Economics and Statistics, 63(3): 361-368.
Cropper, M.L. and A.M. Freeman. 1991. "Environmental Health Effects" in Measuring the Demand for
Environmental Quality, ed. J. Braden and C. Kolstad. Amsterdam, the Netherlands: Elsevier Science Publishers.
Cropper, M.L. and F.G. Sussman. 1990. Valuing Future Risks to Life. Journal of Environmental Economics and
Management, 19:160-174.
Cropper, M.L. and P.R. Portney. 1990. Discounting and the Evaluation of Lifesaving Programs. Journal of Risk
and Uncertainty, 3: 369-379.
Cummings, R.G. and L.O. Taylor. 1999. Unbiased Value Estimates for Environmental Goods: A Cheap Talk
Design for the Contingent Valuation Method. American Economic Review, 89(30): 649-665.
Cummings, R.G., D.S. Brookshire, and W.D. Schulze. 1986. Valuing Environmental Goods: An Assessment of the
Contingent Valuation Method. Totowa, NJ: Rowman and Allanheld.
Cummings, R.G., G.W. Harrison, and E. Rutstro. 1995. Homegrown Values and Hypothetical Surveys: Do
Dichotomous Choice Questions Elicit Real Economic Commitments? The American Economic Review, 85(1):
260-266.
Dasgupta, P. 2002. Modern Economics and Its Critics. Fact and Fiction in Economics: Models, Realism and
Social Construction, ed. U. Maki. Cambridge: Cambridge University Press.
Desvousges, W.H., F.R. Johnson, and H.S. Banzhaf. 1998. Environmental Policy Analysis with Limited
Information: Principles and Applications of the Transfer Method. Northampton, MA: Edward Elgar Publishing.
Diamond, P. 1996. Testing the Internal Consistency of Contingent Valuation Surveys. Journal of Environmental
Economics and Management, 30(3): 337-347.
Dickie, M. 2017. "Averting Behavior Methods," Chapter 8 in A Primer on Non-Market Valuation, 2nd Ed., ed.
P.A. Champ, K.J. Boyle, and T.C. Brown. Springer Science+Business Media.
Ekeland, I., J. Heckman, and L. Nesheim. 2004. Identification and Estimation of Hedonic Models. Journal of
Political Economy, 112: S60-S109.
English, E., C. Leggett, and K. McConnell. 2015. Value of Travel Time and Income Imputation. Memorandum
E2, DWH-AR0056732. Available at: https: //www.fws.gov/doiddata/dwh-ar-documents/940/DWH-
AR0056732.pdf (accessed December 23,2020).
English, E. R.H. von Haefen, J. Herriges, C. Leggett, F. Lupi, K. McConnell, M. Welsh, A. Domanski, and N. Meade.
2018. Estimating the Value of Lost Recreation Days from the Deepwater Horizon Oil Spill. Journal of
Environmental Economics and Management, 91: 26-45.
Eom, Y.S. and D.M. Larson. 2006. Improving Environmental Valuation Estimates through Consistent Use of
Revealed and Stated Preference Information. Journal of Environmental Economics and Management, 32: 209-
221.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
ESML, EcoService Models Library. 2024. Retrieved from https://www.epa.gov/eco-research/ecoservice-
models-librarv (accessed October 15,2024).
Espey, M., and S. Nair (2005). Automobile Fuel Economy: What is it Worth? Contemporary Economic Policy,
23(3): 317-323.
Ethier, R.G., G.L. Poe, W.D. Schulze, and J. Clark. 2000. A Comparison of Hypothetical Phone and Mail
Contingent Valuation Responses for Green-Pricing Electricity Programs. Land Economics, 76(1): 54-67.
Fan, Q., and J. Rubin. 2010. Two-Stage Hedonic Price Model for Light-Duty Vehicles: Consumers Valuations of
Automotive Fuel Economy in Maine. Transportation Research Record, 2157: 119-128.
Feather, P.M. 1994. Sampling and Aggregation Issues in Random Utility Model Estimation. American Journal of
Agricultural Economics, 76: 772-780.
Feather, P. and D. Hellerstein. 1997. Calibrating Benefits Function Transfer to Assess the Conservation
Reserve Program. American Journal of Agricultural Economics, 79(1): 151-162.
Feather, P. and W.D. Shaw. 1999. Estimating the Cost of Leisure Time for Recreation Demand Models. Journal
or Environmental Economics and Management, 38: 49-65.
Fezzi, C., I. J. Bateman, and S. Ferrini. 2014. Using Revealed Preferences to Estimate the Value of Travel Time
to Recreation Sites .Journal of Environmental Economics and Management, 67(1): 58-70.
Finnoff, D., and J. Tschirhart. 2008. Linking Dynamic Ecological and Economic General Equilibrium Models.
Resource and Energy Economics, 30(2): 91-114.
Finnoff D., and J. Tschirhart. 2011. Inserting Ecological Detail into Economic Analysis: Agricultural Nutrient
Loading of an Estuary Fishery. Sustainability, 3:1688-1722.
Fischhoff, B., and L. Furby. 1988. Measuring Values: A Conceptual Framework for Interpreting Transactions
with Special Reference to Contingent Valuation of Visibility. Journal of Risk and Uncertainty, 1:147-184.
Freeman III, A.M. 1979. The Benefits of Air and Water Pollution Control: A Review and Synthesis of Recent
Estimates. Report prepared for the Council on Environmental Quality. Washington, DC.
Freeman III, A.M. 1982. Air and Water Pollution Control: A Benefit-Cost Assessment. Hoboken, NJ: John Wiley &
Sons.
Freeman III, A.M., J. Herriges, and C. Kling. 2014. The Measurement of Environmental and Resource Values:
Theory and Methods. 3rd Ed. Washington, DC: Resources for the Future.
Gamper-Rabindran, S., and Timmins, C., 2013. Does Cleanup of Hazardous Waste Sites Raise Housing Values?
Evidence of Spatially Localized Benefits. Journal of Environmental Economics and Management, 65(3): 345-
360.
Gerking, Shelby, and M. Dickie. 2013. Valuing Reductions in Environmental Risks to Children's Health. Annual
Review of Resource Economics, 5(1): 245-260.
Gibbons, S., and H.G. Overman. 2012. Mostly Pointless Spatial Econometrics? Journal of Regional Science, 52:
172-191. https://doi.Org/l 0.1111 /i.l467-9787.2012.00760.x
Gibbs, J. P., J. M. Halstead, K. J. Boyle, and J.C. Huang. 2002. An Hedonic Analysis of the Effects of Lake Water
Clarity on New Hampshire Lakefront Properties. Agricultural and Resource Economics Review, 31(1): 39-46.
Graff Zivin, J.S., and M.J. Neidell. 2012. The Impact of Pollution on Worker Productivity. American Economic
Review 102, 3652-3673.
Graff Zivin, S. J., and M. J. Neidell. 2013. Environment, Health, and Human Capital. Journal of Economic
Literature, 51(3): 689-730.
Graff Zivin, S. J., and M.J. Neidell. 2018. Air Pollution's Hidden Impacts. Science, 359(6371): 39-40.
Greene, W. 2000. Econometric Analysis. 4th Ed. Englewood Cliffs: Prentice Hall.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Greenstone, M., and J. Gallagher. 2008. Does Hazardous Waste Matter? Evidence from the Housing Market and
the Superfund Program. The Quarterly Journal of Economics, 123(3): 951-1003.
Greenstone, M., and T. Gayer. 2009. Quasi-experimental and Experimental Approaches to Environmental
Economics. Journal of Environmental Economics and Management, 57(1): 21-44.
Griffiths, C., H. Klemick, M. Massey, C. Moore, S. Newbold, D. Simpson, P. Walsh, and W. Wheeler. 2012. U.S.
Environmental Protection Agency Valuation of Surface Water Quality Improvements. Review of Environmental
Economics and Policy, 6(1): 130-146.
Guignet, D., 2013. What Do Property Values Really Tell Us? A Hedonic Study of Underground Storage
Tanks. Land Economics, 89(2): 211-226.
Guignet, D., Griffiths, C., Klemick, H. and Walsh, P. J., 2017. The Implicit Price of Aquatic Grasses. Marine
Resource Economics, 32(1): 21-41.
Guignet, D, Heberling, M.T., Papenfus, M., Griot, 0., and Holland. B., 2022. Properly Values and Water Quality:
A Nationwide Meta-analysis and the Implications for Benefit Transfer. Land Economics, 98(2): 191-218.
Haab, T.C. and R.L. Hicks. 1997. Accounting for Choice Set Endogeneity in Random Utility Models of
Recreation Demand .Journal of Environmental Economics and Management, 34:127-47.
Haab, T. C., M.G. Interis, D.R. Petrolia, and J.C. Whitehead. 2013. From Hopeless to Curious? Thoughts on
Hausman's "Dubious to Hopeless" Critique of Contingent Valuation. Applied Economic Perspectives and Policy,
35(4): 593-612. http://www.istor.org/stable/43695804
Haab, T.C. and K.E. McConnell. 2003. Valuing Environmental and Natural Resources: The Econometrics of Non-
Market Valuation. Cheltenham, UK: Edward Elgar Publishing.
Hamel, P., Chaplin-Kramer, R., Sim, S. and Mueller, C., 2015. A New Approach to Modeling the Sediment
Retention Service (InVEST 3.0): Case study of the Cape Fear catchment, North Carolina, USA. Science of the
Total Environment, 524:166-177.
Hammitt, J.K. 2002. QALYs versus WTP. Risk Analysis, 22(5): 985-1001.
Hammitt, J.K. 2003. Valuing Health: Quality Adjusted Life-Years or Willingness to Pay? Risk in Perspective,
11(1): 1-6.
Hammitt, J.K. and K. Haninger, 2010. Valuing Fatal Risks to Children and Adults: Effects of Disease, Latency,
and RiskAversion. Journal of Risk and Uncertainty, 40(1): 57-83.
Hanemann, W.M. 1984. Welfare Evaluations in Contingent Valuation Experiments with Discrete Responses.
American Journal of Agricultural Economics, 66: 332-341.
Hanemann, W.M. 1991. Willingness to Pay and Willingness to Accept: How Much Can They Differ? American
Economic Review, 81(3): 635-647.
Hanley, N., and C.L. Spash. 1993. Cost-Benefit Analysis and the Environment. E. Elgar.
Harford, J.D. 1984. Averting Behavior and the Benefits of Reduced Soiling. Journal of Environmental Economics
and Management, 11: 296-302.
Havranek, T., Stanley, T.D., Doucouliagos, H., Bom, P., Geyer-Klingeberg, J., Iwasaki, I., Reed, W.R., Rost, K. and
van Aert, R.C.M. 2020. Reporting Guidelines for Meta-Analysis in Economics. Journal of Economic Surveys,
34(3): 469-47
Heberlein, T.A., M.A. Wilson, R.C. Bishop, and N.C. Schaeffer. 2005. Rethinking the Scope Test as a Criterion for
Validity in Contingent Valuation. Journal of Environmental Economics and Management, 50(1): 1-22.
Herriges, J.A., C.L. Kling, and D.J. Phaneuf. 2004. What's the Use? Welfare Estimates from Revealed Preference
Models when Weak Complementarity Does Not Hold. Journal of Environmental Economics and Management,
47: 55-70.
Hindsley, P., Landry, C. E., Schnier, K., Whitehead, J. C., & Zarei, M. 2022. Joint Estimation of Angler Revealed
Preference Site Selection and Stated Preference Choice Experiment Recreation Data Considering Attribute
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Non-attendance. Journal of Environmental Economics and Policy, 12(1): 44-62.
https://doi.org/10.1080/21606544.2022.207784Q
Hoban, J.C. and T.J. Whitehead. 1999. Testing for Temporal Reliability in Contingent Valuation with Time for
Changes in Factors Affecting Demand. Land Economics 75(3): 453-465.
Holmes, T. and W. Adamowicz. 2003. Attribute Based Methods. A Primer on the Economic Valuation of the
Environment, ed. P. Champ, T. Brown and K. Boyle. Kluwer: 171-219.
Horowitz, J., and K. McConnell. 2003. Willingness to accept, willingness to pay and the income effect .Journal
of Economic Behavior & Organization, 51(4): 537-545.
Huang, J.C., T.C. Haab, and J.C. Whitehead. 1997. Willingness to Pay for Quality Improvements: Should
Revealed and Stated Preference Data be Combined? Journal of Environmental Economics and Management, 34:
240-255.
Institute of Medicine (IOM). 2006. Valuing Health for Regulatory Cost-Effectiveness Analysis, ed. Wilhelmine
Miller, Lisa A. Robinson, and Robert S. Lawrence. Washington, DC: The National Academies Press.
Institute of Medicine (IOM). 2011. Finding What Works in Health Care: Standards for Systematic Reviews.
Washington, DC: National Academies Press.
Irwin, E.G. and N. Bockstael. 2002. Interacting Agents, Spatial Externalities, and the Endogenous Evolution of
Residential Land Use Patterns. Journal of Economic Geography, 2: 31-54.
Irwin, E. 2002. The Effects of Open Space on Residential Property Values. Land Economics, 78(4): 465-481.
Jaramillo, J., Chabi-Olaye, A., and Borgemeister, C. 2010. Temperature-dependent Development and
Emergence Pattern of Hypothenemus hampei (Coleoptera: Curculionidae: Scolytinae) from Coffee Berries.
Journal of Economic Entomology, 103(4): 1159-1165.
Jeon, H., and Herriges, J.A. 2016. Combining Revealed Preference Data with Stated Preference Data: A Latent
Class Approach. Environmental and Resource Economics, 68:1053-1086. https://doi.org/10.1007/slQ640-
016-0060-0
Jo, C. 2014. Cost-of-illness Studies: Concepts, Scopes, and Methods. Clinical and Molecular Hepatology, 20:
327-337.
Johnson, F.R., M. Ruby, S. Banzhaf, and W.H. Desvousges. 2000. Willingness to Pay for Improved Respiratory
and Cardiovascular Health: a Multiple-format, Stated-preference Approach. Health Economics, 9(4): 295-317.
Johnson, F.R., W.H. Desvousges, E.E. Fries, and L.L. Wood. 1995. Conjoint Analysis of Individual and Aggregate
Environmental Preferences. Triangle Economic Research Technical Working Paper No. T-9502.
Johnston, R.J. 2006. Is Hypothetical Bias Universal? Validating Contingent Valuation Responses using a
Binding Public Referendum .Journal of Environmental Economics and Management, 52(1): 469-481.
Johnston, R.J., K.J. Boyle, W. Adamowicz, J. Bennett, R. Brouwer, T.A. Cameron, W.M. Hanemann, N. Hanley, M.
Ryan, R. Scarpa, R., R. and Tourangeau. 2017. Contemporary Guidance for Stated Preference Studies. Journal
of the Association of Environmental and Resource Economists, 4(2): 319-405.
Johnston, R.J., J. Rolfe, R.S. Rosenberger, and R. Brouwer. 2015. "Introduction to Benefit Transfer Methods" in
Benefit Transfer of Environmental and Resource Values. The Netherlands: Springer Netherlands.
Johnston, R.J., J. Rolfe and E. Zawojska. 2018. Benefit Transfer of Environmental and Resource Values:
Progress, Prospects and Challenges. International Review of Environmental and Resource Economics, 12(2-3):
177-266.
Johnston R.J. and R.S. Rosenberger. 2010. Methods, Trends and Controversies in Contemporary Benefit
Transfer. Journal of Economic Surveys, 24(3): 479-510.
Just, R.E., D.L. Hueth, and A. Schmitz. 2005. Welfare Economics of Public Policy: A Practical Approach to Project
and Policy Evaluation. Northampton, MA: Edward Elgar Publishing.
7-70 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Just, D.R. and R.E. Just 2016. Empirical Identification of Behavioral Choice Models under Risk, American
Journal of Agricultural Economics, 98(4): 1181-1194.
Kanninen, B.J. 1995. Bias in Discrete Response Contingent Valuation. Journal of Environmental Economics and
Management, 28(1): 114-125.
Kanninen, B.J. and B. Kristrom. 1993. Sensitivity of Willingness-to-Pay Estimates to Bid Design in
Dichotomous Choice Contingent Valuation Models: A Comment. Land Economics, 69(2): 199-202.
Kaoru, Y. 1995. Measuring Marine Recreation Benefits of Water Quality Improvements by the Nested Random
Utility Model. Resource and Energy Economics, 17(2): 119-136.
Kaoru, Y., V.K. Smith, and J.L. Liu. 1995. Using Random Utility Models to Estimate the Recreational Value of
Estuarine Resources. American Journal of Agricultural Economics, 77:141-151.
Kaul, S., Boyle, K. J., Kuminoff, N. V., Parmeter, C. F. & Pope, J. C. 2013. What Can We Learn from Benefit
Transfer Errors? Evidence from 20 Years of Research on Convergent Validity. Journal of Environmental
Economics and Management, 66(1): 90-104.
Kealy, M.J., M. Montgomery, and J.F. Dovidio. 1990. Reliability and Predictive Validity of Contingent Values:
Does the Nature of the Good Matter? Journal of Environmental Economics and Management, 19(3): 244-263.
Keiser, D., and N. Muller. 2017. Air and Water: Integrated Assessment Models for Multiple Media. Annual
Review of Resource Economics, 9:165 84.
Kim, Y.S. and Johnson, R.L., 2002. The Impact of Forests and Forest Management on Neighboring Property
Values. Society &NaturalResources, 15(10): 887-901.
Kim, Younjun, Catherine L. Kling, and Jinhua Zhao. 2015. Understanding Behavioral Explanations of the WTP-
WTA Divergence Through a Neoclassical Lens: Implications for Environmental Policy. Annual Review of
Resource Economics. Annual Reviews. 7(1): 169-187.
Kingsland SE. 1985. Modeling Nature: Episodes in the History of Population Ecology. Chicago: The University of
Chicago Press.
Klemick, H., C. Griffiths, D. Guignet, and P. Walsh. 2018. Improving Water Quality in an Iconic Estuary: An
Internal Meta-analysis of Property Value Impacts Around the Chesapeake Bay. Environmental and Resource
Economics, 69(2): 265-292.
Klaiber, H.A., and D.J. Phaneuf. 2010. Valuing open space in a residential sorting model of the Twin Cities.
Journal of Environmental Economics and Management, 60: 57-77.
Klaiber, H.A. and V.K. Smith. 2013. Quasi Experiments, Hedonic Models, and Estimating Trade-offs for Local
Amenities. Land Economics, 89(3): 413-431.
Kling, C.L. 1997. The Gains from Combining Travel Cost and Contingent Valuation Data to Value Nonmarket
Goods. Land Economics, 73: 428-439.
Kling, C.L., R.W. Arritt, G. Calhoun, and D.A. Keiser. 2017. Integrated Assessment Models of the Food, Energy,
and Water Nexus: A Review and an Outline of Research Needs. Annual Review of Resource Economics, 9: 143-
163.
Kling, C.L. and D. Phaneuf. 2018. How are scope and adding up relevant for benefits transfer? Environmental
and Resource Economics, 69: 483-502.
Kniesner, T.J, W.K. Viscusi, and J.P. Ziliak. 2006. Life-Cycle Consumption and the Age-Adjusted Value of Life.
Contributions to Economic Analysis & Policy, 5(1): Article 4.
Kniesner, T.J., W.K. Viscusi, and J.P. Ziliak. 2014. Willingness to Accept Equals Willingness to Pay for Labor
Market Estimates of the Value of a Statistical Life. Journal of Risk and Uncertainty, 48(3): 187-205.
Kopp, R.J. 1992. Why Existence Values Shouldbe used in Cost-benefit Analysis. Journal of Policy Analysis and
Management, 11(1): 123-130.
7-71 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Kremen, C., Williams, N.M., Aizen, M.A., Gemmill-Herren, B., LeBuhn, G., Minckley, R., Packer, L., Potts, S.G.,
Roulston, T.A., Steffan-Dewenter, I. and Vazquez, D.P., 2007. Pollination and Other Ecosystem Services
Produced by Mobile Organisms: A Conceptual Framework for the Effects of Land-use Change. Ecology Letters,
10(4): 299-314.
Kovacs, K.F., 2012. Integrating Property Value and Local Recreation Models to Value Ecosystem Services from
Regional Parks. Landscape and Urban Planning, i05(2-4): 79-90.
Kuminoff, N.V. and A.S. Jarrah. 2010. A New Approach to Computing Hedonic Equilibria and Investigating the
Properties of Locational Sorting Models .Journal of Urban Economics, 67(3): 322-35.
Kuminoff, N.V., C.F. Parmeter, and J.C. Pope. 2010. Which Hedonic Models Can We Trust to Recover Marginal
Willingness to Pay for Environmental Amenities? Journal of Environmental Economics and Management, 60
(l):145-60.
Kuminoff, N. and J. Pope. 2010. Hedonic Equilibria, Land Value Capitalization, and the Willingness to Pay for
Public Goods. Available at: https://www.semanticscholar.org/paper/Hedonic-Equilihria%2C-Land-Value-
Capitalization%2C-and-Kuminoff-Pope/e415575e741fbdab622ed34a8f658c3bd41418f9 (accessed
December 23, 2020).
Kuminoff, N.V., V.K. Smith, and C. Timmins. 2013. The New Economics of Equilibrium Sorting and Policy
Evaluation using Housing Markets. Journal of Economic Literature, 51(4): 1007-62
Landry, C.E. and Hindsley, P. 2011. Valuing Beach Quality with Hedonic Property Models. Land Economics,
87(1): 92-108.
Landry, Craig E., and John A. List. 2007. Using ex ante approaches to obtain credible signals for value in
contingent markets: Evidence from the field. American Journal of Agricultural Economics, 89(2): 420-429.
Landry, C., and K. McConnell. 2007. Hedonic OnsightCost Model of Recreation Demand. Land Economics, 83:
253 -267.
Larson, D.M. 1993. Joint Recreation Choices and Implied Values of Time. Land Economics, 69(3): 270-286.
Larson, D. M. and D. K. Lew, 2014. The Opportunity Cost of Travel Time as a Noisv Wage Fraction. American
fournal of Agricultural Economics. 96(2): 420-437.
Layton, D.F. and G. Brown. 2000. Heterogeneous Preferences Regarding Global Climate Change. The Review of
Economics and Statistics, 82(4): 616-624.
Leggett, C.G. and N.R. Bockstael. 2000. Evidence of the Effects of Water Quality on Residential Land Prices.
Journal of Environmental Economics and Management, 39:121-144.
Lin, D., R. Lutter, and C.J. Ruhm. 2018. Cognitive Performance and Labour Market Outcomes. Labour
Economics, 51:121-135.
Lindberg, K., R.L. Johnson, and R.P. Berrens. 1997. Contingent Valuation of Rural Tourism Development with
Tests of Scope and Mode Stability. Journal of Agricultural and Resource Economics, 22(1): 44-60.
List, J.A. 2001. Do Explicit Warnings Eliminate the Hypothetical Bias in Elicitation Procedures? Evidence from
Field Auctions for Sportscards. American Economic Review, 91(5): 1498-1507.
List, J.A. and C.A. Gallet. 2001. What Experimental Protocol Influence Disparities between Actual and
Hypothetical Stated Values? Environmental and Resource Economics, 20: 241-254.
List, J.A. and J. Shogren. 1998. Calibration of Difference Between Actual and Hypothetical Valuations in a Field
Experiment. Journal of Economic Behavior and Organization, 37:193-205.
List, J.A., P. Sinha, and M.H. Taylor. 2006. Using Choice Experiments to Value Non-Market Goods and Services:
Evidence from Field Experiments. Advances in Economic Analysis & Policy 5(2): 1-39.
Loomis, J.B. 1989. Test-Retest Reliability of the Contingent Valuation Method: A Comparison of General
Population and Visitor Response. American Journal of Agricultural Economics, 71(1): 76.
7-72 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Loomis, J. and M. King. 1994. Comparison of Mail and Telephone-Mail Contingent Valuation Surveys .Journal
of Environmental Management, 41: 309-324.
Lupi, F., D.J. Phaneuf, and R.H. von Haefen. 2020. Best Practices for Implementing Recreation Demand Models.
Review of Environmental Economics and Policy, 14:2: 302-323.
MacArthur, R.H. and Wilson, E.O., 1967. The Theory of Island Biogeography. Princeton: Princeton Univ. Press:
203.
Magat, A.M., W.K. Viscusi, and J. Huber. 1996. A Reference Lottery Metric for Valuing Health. Management
Science, 42(8): 1118-1130.
Maguire, K.B. 2009. Does Mode Matter? A Comparison of Telephone, Mail and In-Person Treatments in
Contingent Valuation Experiments. Journal of Environmental Management, 90(11): 3528-3533.
Mahan, B.L., S. Polasky, and R.M. Adams. 2000. Valuing Urban Wetlands: a Property Price Approach. Journal of
Land Economics, 76:100-113.
Mannesto, G. and J. Loomis. 1991. Evaluation of Mail and In-person Contingent Valuation Surveys: Results of a
Study of Recreational Boaters .Journal of Environmental Management, 32(1): 177-190.
Mansfield, C., S.K. Pattanayak, W. McDow, R. McDonald, and P. Halpin. 2005. Shades of Green: Measuring the
Value of Urban Forests in the Housing Market. Journal of Forest Economics, 11(3): 177-199.
Massey, D.M., C. Moore, S.C. Newbold, T. Ihde, and H. Townsend. 2017. Commercial Fishing and Outdoor
Recreation Benefits of Water Quality Improvements in the Chesapeake Bay. NCEE Working Paper 2017-02.
Massey, D.M., S. Newbold, and B. Genter. 2006. Valuing Water Quality Changes Using a Bioeconomic Model of
a Coastal Recreational Fishery. Journal of Environmental Economics and Management, 52: 482-500.
McClelland, G.H., W.D. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L. Deck, and M. Thayer. 1991.
Valuing Eastern Visibility: A Field Test of the Contingent Valuation Method. U.S. EPA Cooperative Agreement
#CR-815183-01-3. Draft report prepared for the U.S. Environmental Protection Agency, Washington, DC
McConnell, K.E. 1990. Double Counting in Hedonic and Travel Cost Models. Land Economics, 66(2): 121-27.
McConnell, K.E. 1992. On-site Time in the Demand for Recreation. American Journal of Agricultural Economics,
74: 918-25.
McConnell, K.E. and I. Strand. 1981. Measuring the Cost of Time in Recreation Demand Analysis: An
Application to Sportfishing. American Journal of Agricultural Economics, 63:153-156.
McConnell, K.E., I.E. Strand, and S. Valdes. 1998. Testing Temporal Reliability and Carry-over Effect: The Role
of Correlated Responses in Test-retest Reliability Studies. Environmental & Resource Economics, 12(3): 357.
McFadden, D. 1994. Contingent Valuation and Social Choice. American Journal of Agricultural Economics,
76(4): 689-708.
McGartland, A., R. Revesz, D.A. Axelrad, C. Dockins, P. Sutton, T.J. Woodruff. 2017. Estimating the health
benefits of environmental regulations. Science, 357:457-458.
McKean, J. R., D. M. Johnson, and R. G. Walsh (1995), Valuing Time in Travel Cost Demand Analysis: An
Empirical Investigation. Land Economics, 71: 96-105.
Mendelsohn RO. 1980. An Economic Analysis of Air Pollution from Coal-fired Power Plants. Journal of
Environmental Economics and Management, 7: 30-43.
Mendelsohn, R., J. Hoff, G. Peterson, and R. Johnson. 1992. Measuring Recreation Values with Multiple
Destination Trips. American Journal of Agricultural Economics, 74: 926-933.
Messer, K.D., W.D. Schulze, K.F. Hackett, T.A. Cameron, and G.H. McClelland. 2006. Can Stigma Explain Large
Property Value Losses? The Psychology and Economics of Superfund. Environmental and Resource Economics,
33: 299-324.
7-73 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Michael, H. J., K.J. Boyle, and R. Bouchard. (2000). Does the Measurement of Environmental Quality Affect
Implicit Prices Estimated from Hedonic Models? Land Economics-. 283-298.
Mitchell, R.C. and R.T. Carson. 1989. Using Surveys to Value Public Goods: The Contingent Valuation Method.
Washington, DC: Resources for the Future.
Moeltner K. 2015. "Optimal Scope and Bayesian Model Search," in Benefit Transfer of Environmental and
Resource Values: The Economics of Non-Market Goods and Resources (14), ed. Johnston R., Rolfe J., Rosenberger
R., Brouwer R. Springer, Dordrecht.
Moeltner, K. 2019. Bayesian Nonlinear Meta Regression for Benefit Transfer. Journal of Environmental
Economics and Management, 93: 44-62.
Moore, C. and Griffiths, C., 2018. Welfare Analysis in a Two-stage Inverse Demand Model: An Application to
Harvest Changes in the Chesapeake Bay. Empirical Economics, 55(3): 1181-1206.
Moore, C., D. Guignet, C. Dockins, K.B. Maguire, and N.B. Simon. 2018. Valuing Ecological Improvements in the
Chesapeake Bay and the Importance of Ancillary Benefits. Journal of Benefit-Cost Analysis, 9(1): 1-26.
Moore, M.J. and W.K. Viscusi. 1988. The Quantity-Adjusted Value of Life. Economic Inquiry, 26(3): 369-388.
Moore, R., C. Johnston, R. Smith, and B. Milstead. 2011. Source and Delivery of Nutrients to Receiving Waters
in the Northeastern and Mid-Atlantic Regions of the United States .Journal of the American Water Resources
Association, 47(5): 965-990.
Morey, E., T. Buchanan, and D. Waldman. 2002. Estimating the Benefits and Costs to Mountainbikers of
Changes in Access Fees, Trail Characteristics and Site Closures: Choice Experiments and Benefits Transfer.
Journal of Environmental Economics and Management, 64: 411-422.
Morgan, O.A., D.M. Massey, and W.L. Huth. 2009. Diving Demand for Large Ship Artificial Reefs. Marine
Resource Economics, 24: 43-59.
Murphy, J.J. and P.G. Allen. 2005. A Meta-analysis of Hypothetical Bias in Stated Preference Valuation.
Environmental and Resource Economics, 30: 313-325.
National Academies of Sciences, Engineering, and Medicine. 2017. Valuing Climate Damages: Updating
Estimation of the Social Cost of Carbon Dioxide. National Academies Press. Washington, DC Available at
https://www.nap.edu/catalog/24651 /valuing-climate-damages-updating-estimation-of-the-social-cost-of
(accessed May 30,2017).
National Ecosystem Services Partnership. 2016. Federal Resource Management and Ecosystem Services
Guidebook. 2nd ed. Durham: National Ecosystem Services Partnership, Duke University,
https://nespguidebook.com
National Research Council (NRC). 2014. Review ofEPA's Integrated Risk Information System (IRIS) Process.
Washington, DC: The National Academies Press.
Nelson JP. 2 015. Meta-analysis: Statistical Methods. Benefit Transfer of Environmental and Resource Values: A
Guide for Researchers and Practitioners. Dordrecht: Springer.
Nelson, J.P., R. Johnston, R. Rosenberger et al. 2013. Meta-analysis: Statistical Methods. Benefit Transfer of
Environmental and Resources Values. New York: Springer.
Nelson J.P. and P. Kennedy. 2009. The Use (and abuse) of Meta-analysis in Environmental and Natural
Resource Economics: An Assessment. Environmental Resource Economics, 42(3): 345-377.
Nielsen, J.S., S. Chilton, S., and H. Metcalf. 2019. Improving the Risk-risk Trade-off Method for Use in Safety
Project Appraisal Responses. Environmental Economics and Policy Studies, 21: 61-86.
Newbold, S. and D. Massey. 2010. Recreation Demand Estimation and Valuation in Spatially Connected
Systems. Resource and Energy Economics, 32(2): 222-240.
7-74 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Newbold, S., R.D. Simpson, D.M. Massey, M. Heberling, W. Wheeler, J. Corona, and J. Hewitt. 2018a. Benefit
transfer challenges: Perspectives from U.S. Practitioners. Environmental and Resource Economics, 69: 467-
481.
Newbold, S.C., Walsh, P.J., Massey, D.M., and Hewitt, J. 2018b. Using structural restrictions to achieve
theoretical consistency in benefit transfers. Environmental and Resource Economics, 69: 529-553.
Nordhaus, W.D. 1993. Rolling the 'DICE': An Optimal Transition Path for Controlling Greenhouse Gases.
Resource and Energy Economics, 15: 27-50.
Nordhaus, W.D. 2013. "Integrated economic and climate modeling," in Handbook of Computable General
Equilibrium Modeling, ed. Peter B. Dixon and Dale W. Jorgenson. Amsterdam: North Holland, Elsevier. 1069-
1131.
Oates, W.E. and D.L. Strassman. 1984. Effluent Fees and Market Structure. Journal of Public Economics, 24: 29-
46.
Odum, H.T. 1996. Environmental Accounting: Energy and Environmental Decision Making. New York: John
Wiley & Sons.
OMB. 2006. Guidance on Agency Surveys and Statistical Information Collections: Questions and Answers
When Designing Surveys for Information Collections. Available at:
https://obamawhitehouse.archives.gov/sites/default/files/omb/inforeg/pmc survey guidance 2006.pdf
(accessed December 23,2020).
OMB. 2023. Circular A-4, Regulatory Analysis, November 9, 2023. Available at:
https://www.whitehouse.gov/wp-content/uploads/2023/ll ZCircularA-4.pdf (accessed June 13, 2024).
OMB. 2024. Guidance for Assessing Changes In Environmental and Ecosystem Services In Benefit-Cost
Analysis. February 28, 2024. Available at: https:IIwww.whitehouse.gov/wp-
content/uploads /2024/02 /ESGui dance.pdf (accessed June 14,2024).
Olmanson, L. G., M. E. Bauer, and P. L Brezonik. (2008). A 20-year Landsat water clarity census of Minnesota's
10,000 lakes. Remote Sensing of Environment, 112(11): 4086-4097.
Onukwugha, E., J. McRae, A. Kravetz, S. Varga, R. Khairnar, and C.D. Mullins. 2016. Cost-of-illness Studies: An
Updated Review of Current Methods. PharmacoEconomics, 34(1): 43-58
Palmquist, R.B. 1988. Welfare Measurement for Environmental Improvements Using the Hedonic Model: The
Case of Nonparametric Marginal Prices .Journal of Environmental Economics and Management, 15: 297-312.
Palmquist, R.B. 1991. "Hedonic Methods," in Measuring the Demand for Environmental Quality, ed. J. Braden
and C. Kolstad. Amsterdam, The Netherlands: Elsevier Science Publishers.
Palmquist, R.B., D.J. Phaneuf, and V. K. Smith. 2010. Short Run Constraints and the Increasing Marginal Value
of Time in Recreation." Environmental Resource Economics, 46:19-41.
Parmeter, C. and J. Pope. 2013. Quasi-Experiments and Hedonic Property Value Methods. Handbook on
Experimental Economics and the Environment, ed. J. A. List and M. K. Price. Cheltenham, UK: Edward Elgar
Publishing.
Parry, I.W.H., L.H. Goulder, and D. Burtraw. 1997. Revenue-Raising vs. Other Approaches to Environmental
Protection: The Critical Significance of Pre-Existing Tax Distortions. RAND Journal of Economics, 28: 708-731.
Parsons, George. 1991. A Note on the Choice Residential Location in Travel Cost Demand Models. Land
Economics, 67: 360-364.
Parsons, G.R. 2003a. A Bibliography of Revealed Preference Random Utility Models in Recreation Demand.
Available online at http:// www.ocean.udel.edu/cms /gparsons /rumbib.pdf (accessed March 11, 2011).
Parsons, G.R. 2003b. The Travel Cost Model. In A Primer on Non-Market Valuation, ed. K. Boyle and G.
Peterson. London: Kluwer Academic Publishers.
7-75 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Parsons, G.R. and A.B. Hauber. 1998. Spatial Boundaries and Choice Set Definition in a Random Utility Model
of Recreation Demand. Land Economics, 74: 32-48.
Parsons, G.R. and M.S. Needelman. 1992. Site Aggregation in a Random Utility Model of Recreation. Land
Economics, 68: 418-433.
Parsons, G.R., and A.J. Wilson. 1997. Incidental and Joint Consumption in Recreation Demand. Agricultural and
Resource Economics Review, 26(1): 1-6.
Parsons, G.R., D.M. Massey, and T. Tomasi. 2000. Familiar and Favorite Sites in a Random Utility Model of
Beach Recreation. Marine Resource Economics, 14: 299-315.
Parsons, G.R., P.M. Jakus, and T. Tomasi. 1999. A Comparison of Welfare Estimates from Four Models for
Linking Seasonal Recreational Trips to Multinomial Logit Models of Site Choice. Journal of Environmental
Economics and Management, 38:143-157.
Parsons, G.R., A.J. Plantinga, and K.J. Boyle. 2000. Narrow Choice Sets in a Random Utility Model of Recreation
Demand. Land Economics, 76: 86-99.
Peters, T., W.L. Adamowicz, and P.C. Boxall. 1995. Influence of Choice Set Considerations in Modeling the
Benefits from Improved Water Quality. Water Resources Research, 31(7): 1781-1787.
Phaneuf, D.J., C.L. Kling, and J.A. Herriges. 2000. Estimation and Welfare Calculations in a Generalized Corner
Solution Model with an Application to Recreation Demand. Review of Economics and Statistics, 82: 83-92.
Phaneuf, D.J. andT. Requate. 2017. A Course in Environmental Economics. New York: Cambridge University
Press.
Phaneuf, D.J. and C. Siderelis. 2003. An Application of the Kuhn-Tucker Model to the Demand for Water Trail
Trips in North Carolina. Marine Resource Economics, 18: 1-14.
Phaneuf, D.J. and V.K. Smith. 2005. "Recreation Demand Models," in Handbook of Environmental Economics, 2,
ed. K. Maler and J. Vincent. North-Holland.
Phaneuf D.J., V.K. Smith, R.B. Palmquist, and J.C. Pope. 2008. Integrating Property Value and Local Recreation
Models to Value Ecosystems Services in Urban Watersheds. Land Economics, 84(3): 361-381.
Polasky, S. 2008. What's Nature Done for You Lately: Measuring the Value of Ecosystem Services. Choices,
23(2): 42-46.
Poor, P. J., K.J. Boyle, L.O. Taylor, and R. Bouchard. 2001. Objective versus Subjective Measures of Water Clarity
in Hedonic Property Value Models. Land Economics, 77(4): 482-493.
Price, J.I., M.T. Heberling. 2018. The Effects of Source Water Quality on Drinking Water Treatment Costs: A
Review and Synthesis of Empirical Literature. Ecological Economics, 151:195-209.
Quiggin, J. 1992. Risk, Self-Protection and Ex Ante Economic Value Some Positive Results. Journal of
Environmental Economics and Management, 23(1): 40-53.
Reddy, P.V., J.M. Schumacher, and J.C. Engwerda. 2015. "Optimal Management with Hybrid Dynamicsthe
Shallow Lake Problem," in Mathematical Control Theory. Springer: 111-136.
Reilly, J.S., Paltsev, B. Felzer, X. Wang, D. Kicklighter, J. Melillo, R. Prinn, M. Sarofim, and C. Wang. 2007. Global
Economic Effects of Changes in Crops, Pasture, and Forests due to Changing Climate, Carbon Dioxide, and
Ozone. Energy Policy, 35(11): 5370-83.
Renner, B., M. Gamp, R. Schmalzle, and H.T. Schupp. 2015. Health Risk Perception. International Encyclopedia
of the Social & Behavioral Sciences, 2:_702-709.
Robinson, L., Raich, W., Hammitt, J., and O'Keeffe, L. 2019. Valuing Children's Fatality Risk Reductions. Journal
of Benefit-Cost Analysis: 1-22.
Roe, B., Boyle, K.J., and M.F. Teisl. 1996. Using Conjoint Analysis to Derive Estimates of Compensating
Variation. Journal of Environmental Economics and Management, 31(2): 145-159.
7-76 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Rolfe J., R. Brouwer, and R.J. Johnston. 2015. Benefit Transfer of Environmental and Resource Values, 14. ed.
Johnston R., Rolfe J., Rosenberger R., Brouwer R. The Netherlands: Springer, Dordrecht.
Rollins, K. and A. Lyke. 1998. The Case for Diminishing Marginal Existence Values. Journal of Environmental
Economics and Management, 36(3): 324-334.
Rosenberger, Randall. 2015. Benefit Transfer Validity and Reliability. Benefit Transfer of Environmental and
Resource Values, 307-326.
Rosenberger, R. S. and T. D. Stanley. 2006. Measurement, Generalization, and Publication: Sources of Error in
Benefit Transfers and Their Management. Ecological Economics, 60 (2): 372-378.
Rowe, R.D. and L.G. Chestnut. 1983. Valuing Environmental Commodities: Revisited. Land Economics, 59: 404-
410.
Rowe, R.D., W.D. Schulze, and W.S. Breffle. 1996. A Test for Payment Card Biases. Journal of Environmental
Economics and Management, 31(2): 178-185.
Sagoff, M. 2011. The Quantification and Valuation of Ecosystem Services. Ecological Economics, 70 (3): 497-
502.
Salkever, D.S. 1995. Updated Estimates of Earnings Benefits from Reduced Exposure of Children to
Environmental Lead. Environmental Research, 70(1): 1-6.
Samuelson, P.A. 1951. Abstract of a Theorem Concerning Substitutability in Open Leontief Models. In Activity
Analysis of Production and Allocation, ed. T. Koopmans. New York, NY: John Wiley & Sons, Inc.
Sander, H.A. and S. Polasky. 2009. The Value of Views and Open Space: Estimates from a Hedonic Pricing
Model for Ramsey County, Minnesota, USA. Land Use Policy, 26(3): 837-845.
Sander, H., S. Polasky, and R.G. Haight. 2010. The Value of Urban Tree Cover: A Hedonic Property Price Model
in Ramsey and Dakota Counties, Minnesota, USA. Ecological Economics, 69:1646-1656.
Schlapfer, F. 2006. Survey Protocol and Income Effects in the Contingent Valuation of Public Goods: A Meta-
analysis. Ecological Economics, 57(3): 415-429.
Schlapfer, F., A. Roschewitz, and N. Hanley. 2004. Validation of Stated Preferences for Public Goods: A
Comparison of Contingent Valuation Survey Responses and Voting Behavior. Ecological Economics, 51:1-16.
Shavell, S. 1979. Risk Sharing and Incentives in the Principal and Agent Relationship. Bell Journal of
Economics, 10: 55-73.
Shaw, B., C. Mechenich, and L. Klessig. 2004. Understanding Lake Data. UW Extension. Madison, WI: Board of
Regents of the University of Wisconsin System. Available at: https: / /www.uwsp.edu/cnr-
ap/UWEXLakes/Documents/uwexlakes/understandingLakeData.pdf (accessed December 23. 2020).
Shaw, W.D. and M.T. Ozog. 1999. Modeling Overnight Recreation Trip Choice: Application of a Repeated
Nested Multinomial Logit Model. Environmental and Resource Economics, 13(4): 397-414.
Shogren, J.F. and L.O. Taylor. 2008. On Behavioral-Environmental Economics. Review of Environmental
Economics and Policy. 2(1): 26-44.
Shogren, J.F. and T.D. Crocker. 1991. Risk, Self-Protection, and Ex Ante Economic Value. Journal of
Environmental Economics and Management, 20(1): 1-15.
Shonkwiler, J.S. 1999. Recreation Demand Systems for Multiple Site Count Data Travel Cost Models. Valuing
the Environment Using Recreation Demand Models, ed. J.A. Herriges and C.L. Kling. Edward Elgar Publishing.
253-269.
Simon, Nathalie B., Chris Dockins, Kelly B. Maguire, Stephen C. Newbold, Alan J. Krupnick, and Laura 0. Taylor.
2019. Policy BriefWhat's in a Name? A Search for Alternatives to "VSL." Review of Environmental Economics
and Policy, 13(1): 155-161.
Simpson, R.D., 2016. Putting a price on ecosystem services. Issues in Science and Technology, 32(4): 57.
7-77 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Simpson, R. D. 2017. The Simple but Not-too-simple Economic Valuation of Ecosystem Services: Principles
and an Example. Journal of Environmental Economics and Policy, 6(1): 96-106.
Simpson, R.D., R.A. Sedjo, and J.W. Reid. 1996. Valuing Biodiversity for Use in Pharmaceutical Research.
Journal of Political Economy, 104:163-185.
Smith M.D. and Crowder L.B. 2011. Valuing Ecosystem Services with Fishery Rents: a Lumped-parameter
Approach to Hypoxia in the Neuse River Estuary. Sustainability, 3: 2229-2267.
Smith, V.K. and J. Huang. 1995. Can Markets Value Air Quality? A Meta-Analysis of Hedonic Property Value
Models .Journal of Political Economy, 103(11): 209-227.
Smith, V.K. and L.L. Osborne. 1996. Do Contingent Valuation Estimates Pass a "Scope" Test? A Meta-analysis.
Journal of Environmental Economics and Management, 31(3): 287-301.
Smith, V.K. and S. Pattanayak. 2002. Is Meta-Analysis a Noah's Ark for Non-Market Valuation? Environmental
and Resource Economics, 22(1-2): 271-296.
Smith, V.K., W. Desvousges, and M. McGivney. 1983. The Opportunity Cost of Travel Time in Recreation
Demand Models. Land Economics, 59: 259-278.
Smith, V.K., S.K. Pattanayak, and G.L. van Houtven. 2006. Structural Benefits Transfer: An Example Using VSL
Estimates. Ecological Economics, 60: 361-371.
Smith, V.K., M.F. Evans, H. Kim, and D.H. Taylor. 2004. Do the Near Elderly Value Mortality Risks Differently?
Review of Economics and Statistics, 86(1): 423-429.
Stanley, T.D, H. Doucouliagos, M. Giles, J.H. Heckenmeyer, R.J. Johnston, P. Laroche, J.P. Nelson, M. Paldam, J.
Poot, G. Pugh, R.S. Rosenberger, K. Rost. 2013. Meta-analysis of Economics Research Reporting Guidelines.
Journal of Economic Surveys, 27(2): 390-394.
Stedman, R.C., N. A. Connelly, T.A. Heberlein, D.J. Decker, and S.B. Allred. 2019. The End of the (Research)
World As We Know It? Understanding and Coping with Declining Response Rates to Mail Surveys. Society &
Natural Resources, 32 (10): 1139-54. doi:10.1080/08941920.2019.1587127.
Swinton, S.M., Lupi, F., Robertson, G.P. and Hamilton, S.K. 2007. Ecosystem services and agriculture:
cultivating agricultural ecosystems for diverse benefits. Ecological Economics, 64(2): 245-252.
Tapsuwan, S., Ingram, G., Burton, M. and Brennan, D., 2009. Capitalized amenity value of urban wetlands: a
hedonic property price approach to urban wetlands in Perth, Western Australia. Australian Journal of
Agricultural and Resource Economics, 53(4): 527-545.
Tarricone, R. 2006. Cost-of-illness Analysis: What Room in Health Economics? Health Policy, 77:51-63.
Taylor, L.O., Phaneuf, D.J. and Liu, X., 2016. Disentangling Property Value Impacts of Environmental
Contamination from Locally Undesirable Land Uses: Implications for Measuring Post-cleanup Stigma. Journal
of Urban Economics, 93: 85-98.
Teisl, M., K.J. Boyle, D.W. McCollum, and S.D. Reiling. 1995. Test-Test Reliability of Contingent Valuation with
Independent Sample Pretest and Posttest Control Groups. American Journal of Agricultural Economics, 77:
613-619.
Thompson, B.H., R.E. Paradise, P.L McCarty, K. Segerson, W.L. Ascher, D.C. McKenna, G. Biddinger, A. Bostrom,
A., D.J. Evans, J. Boyd, and R. Costanza. 2009. Valuing the Protection of Ecological Systems and Services; a
report of the EPA Science Advisory Board.
Thorsnes, P.G. 2002. The Value of a Suburban Forest Reserve: Estimates from Sale of Vacant Residential
Building Lots. Land Economics, 78(3): 426-441.
Train, K.E. 1998. Recreation Demand Models with Taste Differences Over People. Land Economics, 74: 230-
239.
Train, K.E. 2009. Discrete Choice Methods with Simulation. Cambridge: Cambridge University Press.
7-78 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Trasande, L., R. T. Zoeller, U. Hass, A. Kortenkamp, P. Grandjean, J. P. Myers, J. DiGangi, P. M. Hunt, R. Rudel, S.
Sathyanarayana, M. Bellanger, R. Hauser, J. Legler, N. E. Skakkebaek and J. J. Heindel. 2016. Burden of Disease
and Costs of Exposure to Endocrine Disrupting Chemicals in the European Union: an Updated Analysis.
Andrology, 4: 565-572.
Tuncel, T. and J. Hammitt. 2014. A New Meta-Analysis on the WTP/WTA Disparity. Journal of Environmental
Economics and Management, 68:175-187.
Tyrvainen, L. and A. Miettinen. 2000. Property Prices and Urban Forest Amenities. Journal of Environmental
Economics and Management, 39(2): 205-223.
U.S. DOT. 2016. The Value of Time Savings: Departmental Guidelines for Conducting Economic Evaluations,
Revision 2 (2016 Update). Available at: https: //www.transportation.gov/office-policy/transportation-
policy/revised-departmental-guidance-valuation-travel-time-economic (accessed April 1, 2020)
U.S. EPA. 1997d. Guiding Principles for Monte Carlo Analysis. Risk Assessment Forum, EPA/630/R-97/001.
March. Available at: http://cfpuh.epa.gov/ncea/raf/recordisplay.cfm?deid=29596 (accessed January 31,
2011).
U.S. EPA. 1997e. Policy for use of probabilistic analysis in risk assessment. Memorandum of Fred Hansen,
Deputy Administrator, May 15,1997. Available at: http: //www.epa.gov/spc/pdfs/probpol.pdf (accessed
February 28, 2011).
U.S. EPA. 1997f. Regulatory Impact Analyses for the Particulate Matter and Ozone National Ambient Air
Quality Standards and Proposed Regional Haze Rule, prepared by Innovative Strategies and Economics
Group, OAQPS, Research Triangle Park, NC, July 16.
U.S. EPA. 2000. SAB Report on EPA's White Paper Valuing the Benefits of Fatal Cancer Risk Reduction. EPA-
SAB-EEAC-00-013.
U.S. EPA. 2002. Environmental and Economic Benefit Analysis of Final Revisions to the National Pollutant
Discharge Elimination System Regulation and the Effluent Guidelines for Concentrated Animal Feeding
Operations. EPA-821-R-03-003.
U.S. EPA. 2003. Children's Health Valuation Handbook. Office of Children's Health Protection and Office of
Policy, Economics, and Innovation, EPA 100-R01-002. Available at:
http://yosemite.epa.gov/ee/epa/eed.nsf/wehpages/HandhookChildrensHealthValuation.html (accessed
December 23, 2020).
U.S. EPA. 2005. Economic Analysis for the Final Stage 2 Disinfectants and Disinfection Byproducts Rule.
December 2005. EPA 815-R-05-010. Available at:
https: //nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P100500X.TXT (accessed December 23, 2020).
U.S. EPA. 2006. Ecological Benefits Assessment Strategic Plan. EPA-240-R-06-001. Office of the Administrator,
Washington, DC. Available at: https://nepis.epa.gov/Hxe/ZvPUR I,.cgi?Dockev= PI 00N07W.TXT (accessed
December 23, 2020).
U.S. EPA. 2007. SAB Advisory on EPA's Issues in Valuing Mortality Risk Reduction. Available at:
http://yo Semite, epa.gov/sab /sabproduct.nsf/4128007E7876B8F0852 573760058A978/$File/sab-08-
001.pdf (accessed December 23, 2020).
U.S. EPA. 2009a. Valuing the Protection of Ecological Systems and Services: A Report of the EPA Science
Advisory Board.
U.S. EPA. 2009b. Environmental Impact and Benefits Assessment for Final Effluent Guidelines and Standards
for the Construction and Development Category. EPA-821-R-09-012. Available at:
https://www.epa.gov/sites/production/files/2015-Q6/documents/cd envir-benefits-assessment 2009.pdf
(accessed December 23,2020).
U.S. EPA. 2011. Benefits and Costs of the Clean Air Act, 1990 to 2020: Second Prospective Study.
U.S. EPA. 2014. Framework for Human Health Risk Assessment to Inform Decision Making. EPA/100/R-
14/001.
7-73 Guidelines for Preparing Economic Analyses | 3rd edition
-------
U.S. EPA. 2015a. National Ecosystem Services Classification System (NESCS): Framework Design and Policy
Application. EPA-800-R-15-002. United States Environmental Protection Agency.
U.S. EPA. 2015b. Peer Review Handbook. 4th Ed. EPA/100/B-15/001. Available at:
https://www.epa.gov/osa/peer-review-handbook-4th-edition-2015 (accessed December 23. 20201
U.S. EPA. 2016. Generic Ecological Assessment Endpoints (GEAEs) For Ecological Risk Assessment: Second
Edition with Generic Ecosystem Services Endpoints Added. EPA/100/F15/005.
U.S. EPA. 2020. Handbook on Valuing Changes in Time Use Induced by Regulatory Requirements and Other EPA
Actions. EPA/236/B15/001.
U.S. EPA. 2023. Report on the Social Cost of Greenhouse Gases: Estimates Incorporating Recent Scientific
Advances. Available at: EPA Report on the Social Cost of Greenhouse Gases: Estimates Incorporating Recent
Scientific Advances.
U.S. EPA. 2024. SAB Report: Review of BenMAP and Benefits Methods. EPA-SAB-24-003.
Van Houtven, G., J. Powers, A. Jessup, and J.C. Yang. 2006. Valuing Avoided Morbidity Using Meta-regression
Analysis: What Can Health Status Measures and QALYs Tell us about WTP? Health Economics, 15: 775-795.
Varian, H. 1992. Microeconomic Analysis. New York: W.W. Norton & Co., Inc.
Viscusi, W.K. 2013. Using Data from the Census of Fatal Occupational Injuries to Estimate the "Value of a
Statistical Life." US Bureau of Census Monthly Labor Review, October 2013.
Viscusi, W.K., J. Huber, and J. Bell. 2014. Assessing Whether There is a Cancer Premium for the Value of a
Statistical Life. Health Economics, 23(4): 384-396.
Viscusi, W.K. and J. Aldy. 2007. Age Differences in the Value of Statistical Life: Revealed Preference Evidence.
Review of Environmental Economics and Policy, 1(2): 241-260.
von Haefen, R.H. and D.J. Phaneuf. 2005. Continuous Demand System Approaches to Nonmarket Valuation. In
Applications of Simulation Methods in Environmental & Resource Economics, ed. R. Scarpa and A. Alberini.
Dordrecht: Springer.
von Haefen, R.H. and D.J. Phaneuf. 2008. Identifying Demand Parameters in the Presence of Unobservables: A
Combined Revealed and Stated Preference Approach. Journal of Environmental Economics and Management,
56(1): 19-32.
von Haefen, R.H., D.J. Phaneuf, and G.R. Parsons. 2004. Estimation and Welfare Analysis with Large Demand
Systems. Journal of Business and Economic Statistics, 22(2): 194-205.
Vossler, C.A., R.G. Ethier, G.L. Poe, and M.P. Welsh. 2003. Payment Certainty in Discrete Choice Contingent
Valuation Responses: Results from a Field Validity Test. Southern Economic Journal, 69(4): 886-902.
Vossler, Christian A., and Mary F. Evans. "Bridging the gap between the field and the lab: Environmental
goods, policy maker input, and consequentiality." Journal of Environmental Economics and Management 58.3
(2009): 338-345.
Vossler, Christian A., Maurice Doyon, and Daniel Rondeau. "Truth in consequentiality: theory and field
evidence on discrete choice experiments." American Economic Journal: Microeconomics 4.4 (2012): 145-171.
Vossler, Christian A., and Sharon B. Watson. "Understanding the consequences of consequentiality: Testing
the validity of stated preferences in the field." Journal of Economic Behavior & Organization 86 (2013): 137-
147.
Wackernagel, M. and W. Rees. 1996. Our Ecological Footprint: Reducing Human Impact on Earth. Gabriola
Island, BC. New Society Publishers.
Walsh, P., Griffiths, C., Guignet, D. and Klemick, H., 2017. Modeling the Property Price Impact of Water Quality
in 14 Chesapeake Bay Counties. Ecological Economics, 135:103-113.
Walsh, P. and P. Mui, 2017. Contaminated Sites and Information in Hedonic Models: An Analysis of a NJ
Property Disclosure Law. Resource and Energy Economics, 50:1-14.
7-80 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Whitehead, J.C. 2002. Incentive Incompatibility and Starting-Point Bias in Iterative Valuation Questions. Land
Economics, 78(2): 285-297.
Whitehead, J.C., Lew, D.K. 2020. Estimating Recreation Benefits Through Joint Estimation of Revealed and
Stated Preference Discrete Choice Data. Empirical Economics, 58: 2009-2029.
https://doi.org/10.1007/s00181-019-Q1646-z
Whitehead, J.C., S.K. Pattanayak, G.L. Van Houtven, and B.R. Gelso. 2008. Combining Revealed and Stated
Preference Data to Estimate the Nonmarket Value of Ecological Services: An Assessment of the State of the
Science. Journal of Economic Surveys, 22: 872-908.
Willig, R.D. 1976. Consumer's Surplus without Apology. The American Economic Review, 66(4): 589-597.
Woodward, R.T. and Y. Wui. 2001. The Economic Value of Wetland Services: A Meta-analysis. Ecological
Economics, 37(2): 257-270.
World Health Organization (WHO). 2017. Guidance on Evaluating and Expressing Uncertainty In Hazard
Assessment, 2nd edition. Available at:
http: //www.who.int/ipcs/methods/harmonization/areas/hazard assessment/en/ (including APROBA
spreadsheet)
Zabel, J.E. and D. Guignet. 2012. A Hedonic Analysis of the Impact of LUST Sites on House Prices. Resource and
Energy Economics, 34(4): 549-564.
7-81 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 8 - Analyzing Costs
As discussed in Chapter 1, the distinction between what is labeled a benefit or
a cost in regulatory analysis is, to some extent, arbitrary so long as the
analysis is internally consistent. The compliance cost of a regulatory action
might be considered a benefit of a deregulatory action, and vice versa.
Likewise, reduced fuel expenditures that accrue to a firm or consumer due to
regulation have been described as both negative costs and positive private
benefits. On which side of the ledger specific categories fall is somewhat
immaterial. What is important is that analysts make every effort to account for
all costs and benefits to illuminate key differences across the options under
consideration with regards to net benefits. As noted above, these Guidelines
are framed from the perspective of a policy that improves environmental quality,
but at a cost. This chapter discusses methods and modeling approaches for
estimating a regulation's costs, which are typically reflected in market
outcomes, for use in benefit-cost analysis [BCA]. For a discussion of methods
and modeling approaches for estimating a regulation's benefits, see Chapter 7.
Estimating the costs of regulation involves a series of decisions. The analyst
must determine the scope of the analysis to appropriately capture the range of
anticipated effects from the policy.1 Analysts must determine the types of
costs that are likely to occur within a specific regulatory context and choose
the most defensible way to measure them based on the best available data and
methods. Both the scope of the analysis and how costs are measured will
affect the choice of economic model.2 Models vary in their ability to capture
certain costs; whether they are static or dynamic; their level of geographic and
sectoral detail; and their scope; among others. After selecting one or more
economic models, analysts face a series of implementation decisions, such as
how to best parameterize the model and how to account for uncertainty.
1 Several executive and legislative mandates require that different aspects of costs be considered in a regulatory
analysis. For instance, Executive Order (EO) 12866 specifies that an assessment of the costs of a regulation
should include "any adverse effects on the efficient functioning of the economy and private sector (including
productivity, employment, and competitivenessThe Unfunded Mandates Reform Act (UMRAJ of1995 requires
that cost estimates account for indirect and implicit costs on state and local governments. Many of these "costs"
are categorized as economic impacts and therefore discussed in Chapter 9.
2 A model is a "simplification of reality that is constructed to gain insights into select attributes of a particular
physical, biologic, economic, or social system" (National Research Council (NRC) 2009) and the simplifications
necessary to tractably model complex systems will introduce uncertainty.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
. i The >"f I
While estimating the costs of regulation is often portrayed as relatively straightforward
particularly compared to estimating benefits it must be guided by economic theory. As such, the
appropriate measure of cost to use in a BCA is social cost. Social cost represents the total burden
that a regulation will impose on society, defined as the sum of all opportunity costs incurred as a
result of the regulation. These opportunity costs consist of the value to society of the goods and
services no longer produced and consumed as resources are reallocated away from other activities
towards activities such as pollution abatement To be complete, an estimate of social cost should
include the opportunity costs of current and future consumption and leisure that will be forgone as
a result of the regulation (e.g., effects in the future could occur because of effects on capital
investment).3 For example, errors can easily occur if the analyst confuses transfers with costs or
ignores pre-existing regulation or taxes in the affected market
The social cost of a regulation is generally not the same as its effects on gross domestic product
(GDP) or other broad measures of economic activity.4 See Section 8.2.2.1 for more discussion.
Likewise, social cost is distinct from but includes the cost of compliance borne by the regulated
entity. The compliance cost is the private cost that a regulated entity incurs to reduce or prevent
pollution to comply with an environmental regulation for instance, through the installation and
operation of pollution abatement equipment
To estimate social cost, analysts may use one or some combination of compliance cost, partial
equilibrium and/or general equilibrium approaches. A compliance cost approach assesses the costs
of abatement and other actions taken to comply (e.g., monitoring, testing, reporting and
recordkeeping requirements) for the directly regulated sector(s).5 A partial equilibrium approach
models the supply and demand responses of the regulated sector(s) to these compliance costs and
may be extended to consider a small number of related sectors (e.g., markets that supply
intermediate goods to the regulated sector(s), markets for substitute or complementary products,
or markets that supply abatement equipment or other services to comply with requirements).
When broader economy-wide impacts are expected due to a regulation, a compliance cost or partial
equilibrium approach will miss these impacts. In this case, a general equilibrium approach is
needed to more fully estimate social cost Regardless of the approach taken, it is expected that
analysts will need to estimate compliance costs associated with abatement requirements, as they
are a necessary input for generating estimates that rely on a partial or general equilibrium
approach. Models that utilize each of these three approaches to estimate social cost are discussed in
Section 8.3.
3 For a more detailed treatment of the material in this section, see Pizer and Kopp (2005).
4 GDP is defined as the sum of the value (price times quantity) of all market goods and services produced in the
economy and is equal to either Consumption (C) + Investment (I) + Government (G) + (Exports (X) - Imports
(M)), or Labor (L) + Capital (K) + Taxes (T).
5 The term direct cost is sometimes used to refer to the costs incurred by regulated entities to comply with the
regulation. The term indirect cost is sometimes used to refer to costs incurred in related markets or experienced
by consumers or government not subject to the regulation, often transmitted through changes in the prices of
goods or services produced by the regulated sector.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
npliance Cost Approach
A compliance cost approach estimates the direct compliance expenditures incurred by regulated
entities (e.g., individual emitting units or facilities) when installing and operating abatement
technologies or processes to comply with a regulation, conditional on a given level of output. It does
not attempt to estimate welfare impacts associated with a change in the amount of production or
use of inputs but generally assumes that regulated sources are cost-minimizing in their compliance
behavior. Its primary advantage is the ability to generate highly detailed and, when data are
available, relatively specific information on compliance options and their associated costs that
reflect the heterogeneity of regulated entities. This detailed information can be very useful, as many
stakeholders are keenly interested in understanding the anticipated cost of meeting regulatory
requirements. Furthermore, reporting detailed assessments of how regulated entities are expected
to respond to a regulation can generate useful public comments that further the U.S. Environmental
Protection Agency's (EPA's) understanding of available compliance options and their costs.
A compliance cost approach typically does not account for other producer or consumer behavioral
changes that may result from a new regulation.6 However, it can still provide a reasonable estimate
of social cost when changes in a regulated sector's outputs and input mix (aside from direct
compliance activities) are expected to be minimal. However, when significant changes on the
producer or consumer side are expected to occur or profit maximizing firm behavior differs from
cost minimizing behavior due to market imperfections, a compliance cost approach may
substantially misestimate the social cost of a regulation.7 Likewise, a compliance cost approach does
not capture supply side responses, such as changes in the composition of goods produced by the
industry or changes in product quality, and the associated changes in consumer and producer
welfare that result. A key question for analysts is whether it is worth expending additional
resources to expand beyond a compliance cost approach to capture other potentially substantial
costs within the sector itself, related sectors or in the overall economy.
i C K "iti-i[ E:;F.Ji[ih[Approach
In contrast to a compliance cost approach, a partial equilibrium approach to cost estimation
accounts for market changes in the regulated sector. Market responses to the regulation may
include reduced industry output or higher prices as firms pass on some costs directly to consumers.
The goal of a partial equilibrium approach is to measure the net change in consumer and producer
surplus relative to the pre-regulatory equilibrium.8
6 A compliance cost approach does not imply that the costs are ultimately borne by producers. Rather, these
costs may be passed through to consumers (see Chapter 9). The key assumption with a compliance cost approach
is that there are no significant changes in markets except for the compliance activity.
7 The degree to which a demand response influences social cost depends on a variety of factors, such as the
magnitude of the price change, the price elasticity of demand for output of the regulated sector and the degree of
competition in the market. An elasticity is a measure of how responsive a firm or consumer is to a change in
price. In the case of demand, it is the percentage change in the quantity of the product that is demanded by
consumers divided by the percentage change in the product's price. See Appendix A for more discussion of
elasticities.
8 Consumer surplus is the sum of consumers' net benefits i.e., what they are willing to spend on a good or
service over and above market price. Thus, it is the area under the market demand (marginal benefit) curve but
8-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
In theory, in the absence of market distortions (e.g., pre-existing taxes), the social cost of a
regulation can be assessed with a partial equilibrium approach of the regulated market (Just et al.
2005; Harberger 1964).9 While a policy may have effects in many other markets, market clearing
conditions effectively cancel out these effects with regard to aggregate welfare (Farrow and Rose
2018).10 Thus, a partial equilibrium approach is sufficient for estimating social cost when the
analyst expects that a regulation will result in appreciable changes in market activities, but the
effects will be confined primarily to a single market or a small number of markets. The use of a
partial equilibrium approach assumes that the effects of the regulation in all other markets, outside
of those being modeled, will be minimal. See Appendix A.4.3 for more detailed discussion.
Figures 8.1 and 8.2 illustrate how social cost can be defined in partial equilibrium. Figure 8.1 shows
a competitive market before the imposition of an environmental regulation. The shaded area below
the demand curve and above the equilibrium price line is consumer surplus. The area above the
supply curve and below the price line is producer surplus.11 The sum of these two areas defines the
total welfare generated in this market (i.e., the net benefits to society from producing and
consuming the good or service represented in this market). For simplicity, total welfare as depicted
ignores the negative pollution externality arising in this market, which the regulation is designed to
correct.12
In this market, the imposition of a new environmental regulation raises firms' production costs.
Each unit of output is now more costly to produce because of expenditures incurred to comply. As a
result, firms will respond by reducing their level of output. For the industry, this will appear as an
upward shift in the supply curve. This is shown in Figure 8.2 as a movement from S0 to Si. The effect
on the market of the shift in the supply curve is to increase the equilibrium price from P0 to Pi and to
decrease the equilibrium output from Qo to Qh holding all else constant As seen by comparing
above market price. Producer surplus is producers' revenues minus the variable cost of production. Thus, it is the
area above the market supply (marginal cost) curve but below market price. See Appendix A.
9 As defined in Chapter 5, market distortions are factors such as pre-existing taxes, externalities, trade barriers,
federal, state or local regulations or imperfectly competitive markets that move consumers or firms away from
the economically efficient outcome. These factors should be accounted for in the baseline and analyzed when
they interact with the policy under consideration.
10 In theory, impacts in undistorted related markets are "pecuniary" and do not need to be included if the social
costs have been correctly measured in the primary market, but pecuniary effects are important to consider in
inefficient related markets (Boardman et al. 2011). It is also likely that most regulations will result in winners
and losers. Economic impact analysis evaluates how different groups are impacted by a regulation. See Chapter
9.
11 Producer surplus may be interpreted as the profits plus the fixed cost of producers. Profits equal total
revenues, price multiplied by quantity of total output, minus the total costs that vary with the level of production
(i.e., variable costs). These costs equal the area under the supply curve and exclude costs that do not vary with
production (i.e., fixed costs). Over time, the share of costs attributable to investments that are fixed declines and
the supply curve becomes more elastic. See Section 8.2.3.2.
12 Appendix A presents a graphical representation of how to account for this externality. Reduction of the
negative externality would be quantified in the benefits portion of an analysis. The supply curve in Figure 8.1
corresponds to the marginal private cost (MPC) curve described in Figure A.5.
8-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Figures 8.1 and 8.2, the overall effect on welfare is a decline in both producer and consumer
surplus.13
In the long run (i.e., when all costs are variable), compliance costs in this market equal the area
between the old and new supply curves, bounded by the new equilibrium output, Q^.14 Useful
insights about the total costs of the regulation can be derived from Figures 8.1 and 8.2. First, when
consumers are price sensitive as reflected in the downward sloping demand curve a higher
price causes them to reduce consumption of the good. If costs are estimated ex-ante and this price
sensitive behavior is not taken into account (i.e., the cost estimate is based on the original level of
output, Q0), realized compliance costs will be misestimated. Extending the vertical dotted line in
Figure 8.2 from the original equilibrium quantity to the new supply curve (Si) and estimating costs
assuming this quantity illustrates this point A second insight is that realized compliance costs are
only part of the total social costs of a regulation. The black triangle shown in Figure 8.2 is an
additional, real cost arising from the regulation. It reflects the forgone net benefit (or opportunity
cost) from the reduction in output
Figure 8.1 - Competitive Market Before Regulation
Consumer Surplus
Producer Surplus
13 The figure depicts an equal distribution of welfare between consumers and producers in both the old and new
equilibria. Depending on the elasticities of supply and demand, this may not be the case. The elasticities
determine the magnitude of the price and quantity changes induced by the cost increase as well as the
distribution of costs.
14 In the long run, costs are variable and are fully represented in the movement of the supply curve. In the short
or medium run, fixed costs may not affect the supply curve although they could contribute to compliance costs.
See Tietenberg (2002).
8-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Figure 8.2 - Competitive Market After Regulation
P
Pi
Po
Consumer Surplus
Producer Surplus
Compliance Costs
Opportunity Cost of
Reduced Output
D
0
Qi Qo
Q
Under the assumption that impacts outside this market are not significant, the social cost of the
regulation is equal to the sum of the compliance cost and the opportunity cost of reducing output
shown in Figure 8.2. It is exactly equal to the reduction in producer and consumer surplus from the
pre-regulation equilibrium shown in Figure 8.1.
When the effects of a regulation are expected to impact a limited number of markets beyond the
regulated sector, it still may be sufficient to use a partial equilibrium approach to estimate social
cost. A multi-market approach extends a single-market, partial equilibrium representation of the
directly regulated sector to include closely related markets. These may include the upstream
suppliers of major inputs to the regulated sector (including pollution abatement equipment or
services), downstream producers who use the regulated sector's output as an input and producers
of substitute or complimentary products. Vertically or horizontally related markets will be affected
by changes in the equilibrium price and quantity in the regulated sector. As a consequence, they
will experience equilibrium adjustments of their own that can be analyzed in a similar fashion.15
The preceding discussion describes the use of a partial equilibrium approach when the regulated
market is perfectly competitive and both producers and consumers are price takers (i.e., their
behavior does not meaningfully affect prices). In many cases, however, some form of imperfect
competition (e.g., market power) may better characterize the regulated market or closely related
markets. Firms in imperfectly competitive markets will adjust differently to the imposition of a new
15 Just et al (2005) detail methods for evaluating partial equilibrium welfare changes across multiple related
markets (see also Bullock 1993). Estimating welfare is only possible when the relevant relationships among the
sectors (e.g., cross-price elasticities) are correctly specified. Pizer and Kopp (2005) and Kokoski and Smith
(1987) provide additional discussion of when these methods are suitable for estimating social cost.
8-6 Guidelines for Preparing Economic Analyses | 3rd edition
-------
regulation, which can alter the estimate of social cost16 If the regulated markets or closely related
markets are imperfectly competitive, this may significantly influence compliance behavior and
costs, in which case the market structure should be reflected in the analysis.17
Figures 8.3 and 8.4 demonstrate how imperfect competition can result in additional social costs.
Figure 8.3 begins with the case of a monopolistic firm. Unlike the case of perfect competition, the
price of the good is not set equal to its marginal cost.18 Firms with market power can instead set
price equal to their marginal revenue to maximize profit19 However, this results in less of the good
being produced (shown as Qo) than is socially optimal (i.e., what would occur under perfect
competition), Qo*. The welfare loss from the lower-than-optimal level of production is the black
triangle.
Figure 8.3 - Monopoly Market Before Regulation
16 For further discussion of the welfare effects of environmental regulation in the context of imperfectly
competitive markets, see Chapter 6 ofBaumol and Oates (1988), Requate (2006), and Chapter 6 ofPhaneufand
Requate (2017). See Ryan (2012), Ferris etal. (2014), and Wolverton, etal. (2019) for examples where
accounting for the way market structure affected firm decision-making would have potentially led to a different
estimate of a regulation's costs. Kellogg and Reguant (2021) provide additional examples in the energy sector.
17 Section 8.2.3.6 describes how environmental regulations may create conditions that lead to imperfectly
competitive markets.
18 For expositional purposes we continue to label the marginal cost curve as the supply curve. However, in the
case of a monopoly, this curve no longer identifies the amount that will be produced at a given market price
(Mankiw 1998). However, as in the case of perfect competition, it still identifies the amount that will be produced
given the marginal revenue received.
19 To maximize its profit, the monopolist produces QO units such that the additional revenue it receives from
selling one more unit of the good (its marginal revenue) equals the additional cost of producing that unit of the
good (its marginal cost). MRO represents the additional revenue that the firm receives for each additional unit of
production.
8-7 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Figure 8.4 - Monopoly Market After Regulation
P
Pi
Po
~ Consumer Surplus
~ Producer Surplus
Cost of Market Power
¦ Compliance Cost
n Opportunity Cost of
Reduced Output
D
0
Qi Qo
Q
Figure 8.4 shows what occurs when the monopolist is subject to an environmental regulation. As in
the case of perfect competition, the regulation causes the supply curve to shift upward from S0 to Si.
The equilibrium price rises from P0 to Pi and output falls from Qo to Qi. By further restricting output,
the exercise of market power exacerbates the welfare loss from the regulation. The opportunity
cost of this reduced production has now increased. The additional opportunity cost of the reduced
output as a result of the regulation is greater than in a perfectly competitive market. Imperfectly
competitive input markets may also exacerbate the social cost of a regulation (e.g., Busse and
Keohane 2007).
Existing regulations may distort the behavior of regulated sources and other market participants in
response to an environmental regulation and, in turn, result in additional social costs. A well-known
example is in the context of electric utilities where some states regulate investment and retail
prices to assure that producers do not exercise market power. However, this results in two
potential distortions. First, retail prices in these states are lower than would occur in a competitive
market, which means more electricity is consumed than is economically efficient.20 Second, the
ability to only pass along some types of costs and not others to consumers results in more capital-
intensive production (including pollution abatement) than is economically efficient (e.g., Parry
2005; Burtraw and Palmer 2008; Fowlie 2010). Thus, it is important for analysts to be mindful of
the potential for interactions with existing market regulations and, when feasible, account for them
when modeling compliance behavior of market participants.
20 In some regulated electricity markets, prices are set equal to the average cost of production, which is lower
than the additional (marginal) cost of production. In these circumstances, the production and consumption of
electricity is greater than is economically efficient because the additional benefit of production exceeds the cost
to produce it.
8-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
i , * - tieral bjm ? niroach
A general equilibrium approach to cost estimation concurrently considers the effect of a regulation
across all sectors in the economy. It is structured around the assumption that, for some discrete
period of time, an economy can be characterized by a set of equilibrium conditions in which supply
equals demand in all markets. When the imposition of a regulation alters conditions in one market,
a general equilibrium approach will determine a new set of prices for all markets that will return
the economy to equilibrium. These prices in turn determine the outputs and consumption of goods
and services in the new equilibrium. In addition, a new set of prices and demands for the factors of
production (labor, capital and land), the returns to which compose the income of businesses and
households, will be determined in general equilibrium. The social cost of the regulation can then be
estimated by comparing the value of variables in the pre-regulation "baseline" equilibrium with
those in the post-regulation, simulated equilibrium.21
When the imposition of an environmental regulation is expected to have appreciable effects in
markets beyond those that are directly subject to the regulation, a partial equilibrium approach
may be insufficient to adequately estimate social cost A general equilibrium approach, which
captures linkages between markets across the entire economy, is most likely to add value when
both cross-price effects and pre-existing distortions (e.g., taxes, regulations, market power in
other markets) are expected to be significant (U.S. EPA 20 1 7).22 23
Consider as an example a regulation that imposes emission limits on the electric utility sector. In
the long run, we expect at least some, if not all, compliance costs are passed through to consumers
as increases in the electricity price. Because electricity is used as an input in the production of many
goods, the prices of these products may also increase to reflect the increase in their marginal cost of
production. Increases in prices may cause households to alter their choices. For example, their
consumption of energy-intensive goods and services may decrease relative to other goods.
Furthermore, the number of hours they are willing to work may change in part because when goods
become more expensive, households can afford less with the same income; thus, their real wage has
declined.24 On the margin, they respond by changing the number of hours worked. When an
environmental regulation affects the real wage such that individuals opt to work fewer hours, it can
exacerbate pre-existing tax distortions in the labor market (Goulder et al. 1997). The impacts of a
regulation also may interact with pre-existing distortions in other markets, which may cause
21 Computable general equilibrium (CGE) models are discussed in Section 8.3.3. Hazilla and Kopp (1990),
Jorgenson and Wilcoxen (1990), U.S. EPA (1997), U.S. EPA (2011), and Marten etal. (2019) use CGE models to
estimate the social cost of environmental regulations.
22 Cross-price effects are measured by elasticities. For example, the cross-price elasticity of demand is defined as
the percentage change in the quantity of product X demanded by consumers in response to a change in the price
of product Y. If two markets are unrelated, then the cross-price effect is expected to be near or equal to zero. If
they are substitutes, then the cross-price elasticity is positive. If they are complements, then it is negative.
23 The previous section shows how the social cost of a regulation can be estimated in a single market using
partial equilibrium analysis. The example demonstrates how a regulation may cause an opportunity cost in that
market. Distortions with similar opportunity costs already exist in many, if not most, markets as a result of taxes,
regulations and other distortions. When the imposition of a regulation causes a new distortion in one market, it
may interact with pre-existing distortions in other markets, which may cause additional impacts on welfare.
24 In general equilibrium analysis, all prices and wages are real, i.e., measured relative to a numeraire, a specific
single price or weighted average of prices such as the gross domestic product (GDP) deflator. Here, the consumer
price level rises relative to the numeraire. The result is a fall in the real wage the nominal wage divided by the
consumer price level.
8-9 Guidelines for Preparing Economic Analyses | 3rd edition
-------
additional impacts on welfare.25 In cases such as these, a general equilibrium approach is capable of
identifying the nature and magnitude of the costs of complying with a regulation as they flow
through the economy, including changes in substitution among factors of production, trade
patterns, endogenous demands and even intertemporal consumption. These effects are partially or
wholly missed by compliance cost and partial equilibrium approaches.
Figure 8.5 illustrates how a regulation can interact with pre-existing tax distortions in the labor
market. A pre-existing tax equal to a share of the gross (pre-tax) wage (Wg°) causes the net (after-
tax) wage [Wn°) to be lower than the gross (pre-tax) wage by the amount of the tax. With this tax
distortion, the quantity of labor supplied is Lo and there is an opportunity cost of reduced labor
supplied. When a new regulation is imposed in another market, raising production costs will
increase the price level and may lower labor supply. This is shown in Figure 8.6 as a decrease in the
net wage to Wn1 and a decrease in the amount of labor supplied to L\. The opportunity cost of the
labor tax along with the increased distortion as the difference between the gross and net wage has
increased.26 The interaction between the effect of a regulation and the distortion from a tax is
especially pronounced in the labor market.27 Similar interactions are likely to occur in other
markets with significant pre-existing distortions (e.g., capital markets). In cases where they are
likely to have a significant impact, analysts should incorporate these distortions into models used to
estimate social cost28
25 See Text Box 8.3 for a discussion of interactions that could also affect benefits estimation.
26 Recall in this example that the tax is a share of the gross wage, so as the gross wage goes up, the distortion
from the tax also increases. Alternatively, if the amount of tax does not depend on the gross wage, it may need to
be increased to maintain government revenues, which is a common assumption in general equilibrium analysis.
However, it is not necessary for the tax level to change for the new regulation to exacerbate the pre-existing tax
distortion in the labor market.
27 The labor tax distortion affects individual labor supply decisions at the margin. While full-time workers may
not change (or be able to change) hours worked in response to a fall in the real wage, part-time workers those
in households with more than one full-time worker or potential retirees may be more likely to adjust the
number of hours they work Parry (2003) discusses the theoretical and empirical basis for this depiction of the
labor market.
28 Economists have long recognized the "tax interaction effect" (Ballard and Fullerton 1992), and a rich body of
work has focused on them in the context of environmental regulation (Goulder; 2000; Parry and Bento 2000;
Murray et al. 2005; and Bento and Jacobsen 2007). If an environmental regulation raises revenue through a tax
on pollution or another revenue-raising provision, and the revenue is used to reduce pre-existing distortions such
as taxes on wages, the tax-interaction effect may be offset. This is known as the "revenue recycling effect." The
offset may be partial, complete, and in some cases, the overall efficiency of the tax system may actually improve.
The net result is an empirical matter, depending on the nature of the full set of interactions across the economy
and how the revenue is raised. One offsetting factor is that society also incurs a welfare loss from raising
revenues through taxes due to the difference between the value of an additional dollar raised by the government
and the value of that dollar to a private individual (i.e., the marginal cost of public funds).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Figure 8.5 - Labor Market with Pre-Existing Distortion Before Regulation
Figure 8.6 - Labor Market with Pre-Existing Distortion After Regulation
8.2 Estimating Social Cost
When estimating social cost, the objective is to measure the incremental cost for each regulatory
option under consideration. Incremental cost is defined as the additional cost associated with a new
8-11 Guidelines for Preparing Economic Analyses | 3rd edition
-------
requirement relative to a baseline.29 Often when specifying a baseline from which to measure costs,
the analyst needs to first identify what abatement activities are already in place or anticipated as a
result of existing regulations. The costs associated with previously installed abatement controls are
not counted toward the cost of the rule, as these occurred prior to the regulation under
consideration and are therefore in the baseline.30 Similarly, costs that have not yet been realized
but will be incurred to comply with existing regulations should not be counted towards the cost of
the regulation under consideration. Identifying which abatement controls are already in use also
aids the analyst in identifying what additional abatement control options may be available to
further reduce emissions.31
It is important that analysts derive the most defensible central estimates of the compliance costs
associated with identified abatement strategies, as they are the building block for developing social
cost estimates. Social cost estimates should continue to rely on central assumptions and inputs that
are well-supported by standard engineering practice and the published scientific literature. In
addition, analysts should ensure that: (1) the information supporting cost estimation is appropriate
for its intended use; (2) the scientific and technical procedures, measures, methods and/or models
employed to generate the information are reasonable for, and consistent with, the intended
application; and (3) the data, assumptions, methods, quality assurance, sponsoring organizations
and analyses employed to generate the information are well-documented.32 33 As previously
discussed, analysts are advised to focus quantification on categories of costs that are expected to
have a large influence on the net benefits and relative ranking of the options under consideration.
Analysts also should describe the process for quantifying the different types of costs that underlie
the aggregate cost estimate and present them in a disaggregated and informative manner.
i * -Mi;¦[ liai - * t iivvii11'ition
Recall that compliance costs are the additional costs that regulated entities incur to reduce or
prevent pollution to comply with the regulation. There are a variety of different types of compliance
costs including but not limited to the following.
29 While this chapter focuses on the anticipated social costs of regulation, the same approach also applies in a
retrospective setting (see Chapter 5).
30 See Chapter 5 for a detailed discussion of establishing the baseline. Other issues relevant to specifying a
defensible baseline for cost estimation include, for example, ensuring consistency in key assumptions across costs
and benefits; and treatment of anticipatory actions to meet regulatory requirements.
31 Note that the expected abatement strategies that underpin estimates of social costs also affect the expected
change in the level and exposure to the environmental contaminant and therefore the benefits of the regulation.
32 At times, EPA uses externally derived (e.g., contractor, industry association or advocacy group) cost estimates
for its regulatory analyses. Any cost estimate produced by an external source and used by the EPA in its analyses
should meet these criteria.
33 Some statutes require the EPA to choose a regulatory option that is demonstrably affordable. In this case,
analysts should continue to rely on the most defensible central estimate of costs. Estimating an upper bound
(instead of a central estimate) of the compliance cost associated with the chosen option to demonstrate
affordability will bias the net benefits of the regulation downward and/or could result in artificially low levels of
stringency.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Treatment/Capture: The cost of any method, technique or process designed to remove
pollutants, after their generation in the production process, from air emissions, water
discharges or solid waste.
Recycling: The cost of on-site or off-site processing of waste for an alternative use.
Disposal: The cost of the final placement, destruction, or disposition of waste after
pollution treatment/capture and/or recycling has occurred.
Prevention: The cost of preventing pollution from being generated or contamination from
occurring during the production process.
Entities that directly incur compliance costs to meet regulatory requirements may include firms,
households, and government agencies. For example, firms normally incur costs to purchase and
operate pollution control equipment; households may incur the costs of periodic inspections of
pollution control equipment on vehicles; and government agencies may implement, administer,
monitor, and enforce a regulation. In the case of product standards, compliance costs include the
incremental cost of designing and manufacturing the compliant product relative to the already
existing noncompliant product
It is relatively straightforward to infer a value for compliance costs when an explicit monetary
payment (e.g., purchasing pollution abatement equipment) is made. Compliance costs for which
monetary values are not readily available are often more difficult to quantify. For example, the
value and length of time households spend on vehicle inspections may be uncertain. Guidance on
how to value time spent on such activities is discussed in Section 8.2.4. Compliance costs are also
more difficult to quantify when, instead of installing abatement equipment, firms modify
production processes to prevent emissions. Regardless of the ease with which compliance costs can
be estimated or what terminology is used to characterize them, if the compliance activities require
resources that are redirected from other activities relative to the baseline, the value of those
resources should be accounted for in compliance costs.
A compliance cost estimate reasonably approximates the social cost of a regulation when the value
of the resources used for compliance generally reflect their social opportunity cost and prices or
other producer and consumer behavior are not expected to change significantly as a result of a
regulation. Determining whether compliance activities change prices or behavior requires: (1)
estimates of supply and demand conditions, (2) an assessment of how compliance costs affect
production costs, and (3) evidence of whether producers in the sector significantly change their
level of production relative to one another. When compliance costs are used to estimate social
costs, the analysis should provide evidence that justifies their choice.
It is common to refer to different categories of compliance costs, such as fixed or variable costs, as a
way to systematically identify the costs that may result from a regulation. In practice, these
categories of compliance costs may not be entirely distinct Table 8.1 summarizes the main cost
categories discussed in this and subsequent sections.
ced Costs
Fixed costs do not change with the level of production or abatement over a specific time period,
often referred to as the short run. They are typically one-time costs, or costs that only occur once
over the time horizon of the analysis, such as the installation of pollution control equipment.
However, fixed costs may also refer to recurring costs that are independent of the level of
production or abatement over a given time period (note that in the long run, virtually all fixed costs
are variable.) Two common categories of fixed costs are described below.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Capital costs are costs related to the installation or retrofit of structures or equipment. These
expenditures include materials and labor used for equipment installation and startup. Once
equipment is installed, capital costs generally do not change with the level of abatement or
production. Capital costs may also include changes to the production process.
Table 8.1 - Types of Costs Associated with Environmental Policies:
Categories, Examples and Commonly Used Approaches to Quantification
Compliance Costs
Examples
Common Approach to
Quantification34
Fixed costs
Capital costs;
Research and Development investments
Compliance cost
Variables costs
Operating costs;
Monitoring, reporting and recordkeeping costs;
Transaction costs
Compliance cost
Other Opportunity Costs
Examples
Common Approach to
Quantification
Reduced output in regulated
markets
Higher product prices cause reduced quantity
demanded
Partial equilibrium
Changes in product quality
Trade-offs with reliability or longevity
Partial equilibrium
Changes in behavior in the
regulated or final goods markets
Changes in investment (e.g., delayed adoption);
Rebound effects
Partial equilibrium
Transition costs
Search costs for new jobs;
Costs of initially scarce new equipment
Partial equilibrium
Economy-Wide Costs
Examples
Common Approach to
Quantification
Interactions with pre-existing
distortions in related markets
Tax interaction effect;
Trade barriers
General equilibrium
Macroeconomic feedbacks
Capital-induced growth effects
General equilibrium
Research and development (R&D) costs are incurred to develop new products, processes, or
techniques. These costs are in addition to capital costs and should be accounted for when
estimating compliance costs. Similarly, if a supplier to the regulated entity is expected to incur R&D
expenditures in response to the regulation those costs should also be included in the estimate of the
social costs.
In the case where a supplier incurs the R&D expense, care should be taken to avoid double-
counting; in the long run, the supplier will reflect these R&D costs in the price charged to the
34 While opportunity and economy-wide costs may be quantified without explicit partial or general equilibrium
modeling, such modeling may be necessary to avoid the double-counting that would occur if these costs were
added directly to cost estimates from a compliance cost approach.
8-14 Guidelines for Preparing Economic Analyses | 3rd edition
-------
regulated entity. If social costs are estimated using the prices regulated entities pay, accounting for
these additional R&D investments by the supplier, or any other resources reflected in prices
charged to the regulated entity, do not need to be added to the estimate of social costs.
R&D costs incurred in the past by a regulated entity or supplier should not be counted as a cost of
the regulation, as these costs are sunk.35'36 If past R&D costs are reflected in the market prices of
inputs sold by the supplier (for instance, due to market power) then the cost associated with these
past R&D expenditures should also be excluded from the social cost estimate when possible.
ble Costs
Variable costs change with the level of production or abatement37 They are the sum of the marginal
cost for each unit that is produced. Common categories of variables costs are described below.
Operating costs are recurring expenditures associated with the operation and maintenance of
equipment, including salaries and wages, energy inputs, materials, and supplies, purchased services
and maintenance or repair of equipment associated with pollution abatement or waste
management In general, operating costs increase with the level of abatement or the amount of
production or use.
Monitoring, reporting, and recordkeeping costs are incurred to demonstrate or assure
compliance with a regulation. They may be incurred by regulated entities or regulators and
generally reflect the use of resources that should be accounted for when estimating the cost of a
regulation. While these types of costs are identified here as variable costs, some may also be a fixed
cost, such as the installation of pollution monitoring equipment.
Transactions costs are the costs incurred when buying or selling a good or service. They may
include the costs of searching out a buyer or seller, bargaining and enforcing contracts.
Transaction costs reflect the use of real resources (e.g., time, equipment) and should be included in
an estimate of social costs.
8.2,2 Social Cost Estimation
Social costs most often differ from compliance costs because the imposition of the regulation causes
changes in behavior beyond just the compliance activity required by the regulation. The most
straightforward example of an opportunity cost not accounted for by a compliance cost approach is
the value to producers and consumers of reduced output in the regulated market as demand
35 "Sunk costs" are costs that have already been incurred and cannot be reversed (e.g., existing investments in
pollution control equipment or previous use of labor). For a deregulatory action, those costs that are sunk should
not be accounted for in the benefits of that action while avoided future compliance costs should be accounted for.
Care should be taken when identifying whether certain costs are sunk as certain investments and other fixed
costs may actually be, in part, reversible, in which case there is an opportunity cost of continuing to use those
resources. For example, there may be scrappage value of pollution abatement equipment
36 Note, however, that anticipatory actions such as planning and designing for future R&D activities that
are initiated in expectation of promulgation (for instance, in response to the proposal) may still be attributable
to the regulation.
37 Use of some resources, especially energy, also can cause negative environmental or other externalities.
Techniques for non-market valuation can be applied even when impacts are counted on the cost side of the
ledger in a benefit-cost analysis (BCA) (see Chapter 7).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
responds to higher product prices. However, social costs can differ from compliance costs even
without additional behavioral responses. For example, differences may occur when environmental
regulations apply to the performance of a final good for instance, a vehicle or cleaning product
and the installation of an abatement technology can sometimes also result in changes in other
product attributes valued by consumers. While these changes may be positive or negative,
examples in the literature mainly focus on trade-offs between emissions and attributes such as
decreased performance or reliability or reduced safety due to material substitution (e.g., Klemick et
al. 2015; Klier and Linn 2016).
Other behavioral changes in the regulated or final good markets are also not captured by a
compliance cost approach. For example, the design of the regulation itself may result in changes in
investment behavior to avoid the costs of new requirements. In particular, vintage-differentiated
regulations that impose more stringent abatement requirements on new emission sources can
result in delays in the adoption of new, cleaner equipment and potentially increase investment in
older, dirtier equipment (e.g., Gruenspecht 1982; Nelson et al. 1993; Jacobsen and Bentham 2015).
While this behavior lowers the cost of complying with the regulation, it also weakens its overall
stringency. In other cases, the design of the regulation can lead to changes in the utilization of the
regulated product. For example, the literature has estimated a rebound effect from energy
efficiency and vehicle fuel economy standards; because these regulations make it cheaper to
consume energy or fuel on a per-unit basis, demand for these services and therefore emissions
from them increase relative to the case without the rebound (Gillingham and Rapson 2 016). 38
As already discussed, economy-wide costs also may arise when the regulation interacts with pre-
existing distortions such as taxes, other regulations, trade barriers or market power to move
private behavior further away from the economically efficient outcome.
When compliance costs do not fully represent all the opportunity costs of a regulation, partial or
general equilibrium analytic approaches can be used to estimate social cost In some cases, the
analyst can construct or use available partial or general equilibrium modeling tools to credibly
estimate expected changes in the social cost of regulation. When a model is not available, it may still
be possible to estimate the social cost associated with anticipated behavioral change by applying
findings from the peer-reviewed literature. Analysts should justify their choice of estimates and
explain their applicability to the specific context of the rule. If a range of credible estimates are
available, analysts should reflect that range in the analysis and discuss key factors or sources of
uncertainty that influence the estimates. Regardless of the origin of the estimate be it modeled,
empirically estimated, or taken from published studies it is also important that the underlying
behavioral assumptions are consistent with the rest of the BCA.
8.2.2.1 Meas cial Cost
It is possible to estimate the social cost of a regulation by adding up the net change in consumer and
producer surplus in all affected markets. Consumer's equivalent variation (EV) and compensating
38 As Section 5.2 mentions, behavioral economics can have implications for benefits and costs of a regulation.
For example, if consumers mis-optimize or are loss averse, they may not adopt energy-saving technologies for
which private benefits of adoption appear to exceed private costs. This raises the possibility that a regulation
could yield positive private net benefits to consumers or firms. See Section 7.2 for more discussion.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
variation (CV) are other measures that have been utilized.39 As households are the ultimate
beneficiaries of government and investment expenditures, the EV and CV measures focus on
changes in consumer welfare rather than on changes in demand.40
The social cost of a regulation is generally not the same as a change in GDP or aggregate
consumption (U.S. EPA 2017). As measures of social cost, changes in GDP and aggregate
consumption both miss potentially important regulatory effects, such as impacts on leisure demand
or the demand for nonmarket goods.41 GDP is also comprised of more than just changes in
consumption as it is a measure of total economic output.42 For instance, a regulation that requires
firms to install new capital in a given year will see an increase in investment. However, capital also
affects the availability of goods and services that can be consumed over a much longer time period.
As a result, GDP effectively double counts the new capital installed since investment and
consumption are both components of GDP.43
8.2.2 ifers
Environmental regulations may also affect transfers. Transfers are shifts in money or resources
from one part of the economy (e.g., a group of individuals, firms, or institutions) to another in a way
that does not affect the total resources that are available to society. In other words, the loss to one
part of the economy is exactly offset by the gain to another. Since social cost represents the total
burden that a regulation imposes on the economy, it nets out transfers.44'45 Examples of transfers
39 Both EV and CV are monetary measures of the change in household utility brought about by changes in prices
and incomes resulting from the imposition of a regulation. Appendix A describes the relationship between
consumer surplus, equivalent variation and compensating variation. EV and CV are particularly well-suited for
partial and general equilibrium analysis because both modeling frameworks require the explicit
characterization of consumer preferences. Calculating EV and CV requires only pre- and post-policy price and
utility levels.
40 EV and CV can also provide a complete welfare metric (incorporating both benefits and costs) if non-market
goods are explicitly accounted for in consumer utility functions. However, these metrics are often only used to
assess the social cost of a regulation because traditional economic models do notyet incorporate non-separable
benefits, or explicit linkages between environmental quality and economic costs (see Text Box 8.3).
41 SeePaltsev etal. (2009), U.S. EPA (2011), and Paltsev and Capros (2013) for examples of how these measures
differ in specific policy contexts.
42 It is also the case that transfer payments, which are excluded from BCA, are subsumed within the government
spending category of GDP. In addition, while changes to trade patterns due to a regulation may be reflected in
both GDP and welfare, they are not necessarily equivalent measures (Paltsev et ah, 2009; Paltsev and Capros
2013).
43 Several reasons exist for why GDP is not the preferred measure of social cost and overall welfare in general.
For instance, GDP does not include non-market environmental costs or benefits. An example of where it is a
potentially misleading metric is when improvements in environmental quality from a regulation also lead to
reductions in hospital visits that reduce GDP. GDP is also a flow measure of expenditures and does not account
for changes in the capital stock. An example of when this is potentially important is if pollution damages
buildings, then the expenditures on maintenance and repair would increase GDP at the expense of returning the
stock of capital to its original state.
44 Transfers are important for understanding how a regulation affects the private cost of a regulation for
different groups. Thus, they are included in an economic impact analysis. See Chapter 9.
45 An exception is when one group has economic standing in the analysis, and the other does not. See Chapter 5.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
include payments for most taxes and subsidies received, as well as higher revenues for producers in
imperfectly competitive markets due to higher prices.46 However, it is important to note that not all
taxes or subsidies should necessarily be excluded from estimates of social costs under the
assumption that they are transfers. For example, the opportunity cost of a firm employing labor in a
compliance activity is inclusive of payroll taxes since they are a form of compensation (e.g.,
insurance against old age) and not a transfer.
While transfers should be excluded from an estimate of social cost, the conditions leading to the
transfer may create additional costs that should be accounted for in a partial or general equilibrium
framework. For example, when existing taxes are already distorting behavior in a socially inefficient
manner for instance, by changing the decision of how much to work the change in behavior
induced by the regulation can cause the welfare loss associated with these distortions to also
change. These additional changes in welfare due to interactions between the environmental policy
and pre-existing tax distortion should be included in an estimate of social costs.
r Time
After the imposition of a new environmental regulation, the economy moves to a new long-run
equilibrium set of prices and quantities that allow all markets to clear. Since compliance costs
represent permanent additions to the cost of production for a firm, effects in closely related sectors
are incurred in the new equilibrium.
However, in some contexts it is possible that firms and/or consumers may incur additional short-
term costs during the period when the economy is adjusting to the new equilibrium. These are
known as transition costs. Examples include costs to train workers to use new equipment, search
costs as some workers seek employment in other sectors, and additional costs associated with
initially limited availability of new monitoring or abatement equipment. It is also possible that at
least some factors of production are fixed initially, limiting the ability of firms to respond quickly to
new regulatory requirements. For instance, contractual or technological constraints may prevent
firms from fully adjusting their input mix or output decisions until those contracts expire or
technology is ready to be replaced. Similarly, the number of firms may change over time depending
on the cost of new firms to enter. If these types of adjustment costs are substantial, a sole focus on
long run costs may underestimate the total social cost of regulation.
Thus, it is important to consider both short- and long-run effects when measuring costs over time.
In addition, analysts must make choices about the time horizon of the analysis, the use of a static
versus a dynamic framework, discounting and technical change, employment effects and effects on
market structure.
8.2.C i Tint - E i rizon
The time horizon for calculating producer and consumer adjustments to a new regulation should be
considered carefully. The analyst should strive to estimate the present value of all future costs of a
regulation (see Chapter 6). If the analyst is only able to estimate a regulation's costs for one or a few
representative future years, the analyst must take care to ensure that the year(s) selected are truly
46 For example, taxes are generally thought of as transfers between households or firms and government such
that an offsetting change in government revenue and household income due changes in behavior induced by the
regulation are not social costs. Regulations may also create scarce compliance assets, such as allowances in cap-
and-trade systems. Generally, the gratis receipt of or any payments for allowances are also a transfer (see, for
example, Burtraw and Evans 2009).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
representative, that no important transitional costs are effectively dismissed by assumption and
that no one-time costs are assumed to be ongoing.
In the short run, at least some factors of production and consumer demand are fixed. If costs are
evaluated over a short period of time, then contractual or technological constraints can prevent
firms from responding quickly to increased compliance costs by adjusting their input mix or output
decisions. In the long run, by contrast, all factors of production are variable. Firms can adjust any of
their factors of production in response to a new regulation and can even change their production
processes. Similarly, consumers, including producers in other sectors, may not be able to adjust
demand for the output of the regulated sector in the short run, but have more flexibility in the long
run. The time horizon for the analysis should be long enough to capture any flexibility the
regulation provides firms in their compliance approach.
However, if transition costs seem likely, analysts should also consider presenting evidence that sheds
light on the length of the transition period and the magnitude of these costs. In some cases, regulatory
requirements are phased in gradually over time, either explicitly through graduated compliance dates
or requirements or implicitly through characteristics like vintage differentiation (i.e., varying regulatory
requirements based on the age of the plant). For example, consider a regulation that enacts more
stringent requirements on new sources of a pollutant Selecting a time period of analysis that is
early in the program when only a few new sources of production are affected may not accurately
capture a future year in which most sources of production are new for the purposes of the
regulation. A regulation also may influence the rate at which old sources are replaced by new ones.
Chapter 5 contains additional guidance on how to determine the most appropriate time horizon for
analysis.
8.2.« mics
One key decision for the analyst is whether to assume that economic conditions are invariant over
time (i.e., static) or attempt to account for expected future changes in prices and economic activity
(i.e., dynamic). Costs that are estimated at a given point in time or for a selection of distinct points
in time and compared to the baseline are static. They provide snapshots of costs faced by firms,
government and households but do not allow behavioral changes from one time period to affect
responses in another time period. A dynamic framework, one that explicitly captures trade-offs
across time periods, allows for this possibility.47
In most cases, a regulation will continue to have economic impacts after its initial implementation.
If these intertemporal impacts are likely to be significant, they should be included in the estimation
of social cost. Pizer and Kopp (2005) note that static productivity losses from environmental
regulations are amplified over time due to their effect on capital accumulation (a lower capital stock
over time reduces economic output and therefore welfare). A static model would miss this effect. In
some cases, the potential effect of a regulation on long-term growth may be significantly larger than
its effect on the regulated sector alone.48 In addition to these capital-induced growth effects, the
47 Note that a comparative static framework compares snapshots of key economic outcomes before and after a
change in an exogenous factor such as a regulatory requirement.
48 Pizer and Kopp (2005) estimated that the "additional cost of this accumulation effect on welfare can be as
much as 40 percent above the static cost that ignores changes in capital stock " Hazilla and Kopp (1990) and
Jorgenson and Wilcoxen (1990) also showed that this effect is potentially significant. It is important to note,
however, that this conclusion is based on studies of large-scale changes in environmental regulation (i.e., the
welfare effects of the 1972 Clean Water Act and 1977 Clean Air Act Amendment).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
evaluation of costs in a dynamic framework may be important when a proposed regulation is
expected to affect product quality, productivity, innovation and/or changes in markets indirectly
affected by the environmental policy. Dynamic effects also impact net levels of measured consumer
and producer surplus over time. See Section 8.4 for how a regulation's potential dynamic impacts
affect model choice.
Conceptually, a dynamic framework allows the analyst to specify the process by which the economy
moves between equilibria in response to a regulation across time. In practice, however, economists
have more experience characterizing long-run equilibria than the pathways between them. While
shorter run equilibria can be approximated by treating some factors of production as fixed (e.g.,
labor or capital), very near-term transitional costs are typically ignored in the modeling approaches
discussed in Section 8.3. 49
8.2J icounting
Costs that occur over time must be properly and consistently discounted to allow for legitimate
comparisons with benefits.50 Procedures for social discounting in economic analyses are reviewed
in considerable detail in Chapter 6.
There are two applications of discounting that are closely related to the modeling of social costs.
First, when modeling firms' behavior, the analyst should use a discount rate that reflects the
industry's cost of capital, just as a firm would. The social cost of the regulation, on the other hand, is
calculated using the social discount rate, the same discount rate used for estimating the benefits of
the regulation. Section 6.4 provides additional details on the choice of discount rate when modeling
behavior, such as firms' compliance decisions. Second, when a dynamic general equilibrium model
is used to estimate social costs, any displacement of investment due to the regulation has already
been accounted for and the social cost estimates should only be compared to present value
estimates of benefits discounted at the consumption discount rate.
8.2.C T ¦ hni> al Change and
Estimating the social cost of an environmental regulation over a relatively long time horizon
requires assumptions about future technological change. Jaffe et al. (2002) lay out a conceptual
framework for understanding how technological change in response to environmental regulation
may affect the relationship between inputs and output, ultimately reducing the unit costs of
production. It is possible for an environmental regulation to change overall productivity in a sector
over time in one of two ways: 1) the sector is more productive than before, but inputs are used in
the same proportion as before to produce output (i.e., unbiased technical change) or 2) the
regulation affects the growth rate of one or more inputs over time in a way that changes the relative
productivity of the inputs to production (i.e., biased technical change).
49 Dynamic stochastic general equilibrium (DSGE) models represent business cycles within an economywide
framework via random autocorrelated productivity shocks. While built on rely on highly aggregate
representations of the economy, they may offer general insights into the role of modeling uncertainty and how
regulations interact with short-run dynamics (Annicchiarico et al. 2021, U.S. EPA 2017).
50 It is equally important to properly discount cost estimates of different regulatory approaches to facilitate
valid comparisons.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Compliance with an environmental regulation may result in the adoption of existing technology,
improvement, or application of existing technology to a new use and/or development of entirely
new technologies or processes (Sue Wing 2006). Whether it is more appropriate to capture these
compliance responses as affecting overall productivity or the relative productivity of one or more
inputs is an empirical matter.
Despite its importance as a determinant of economic welfare, the process of technical change is not
well-understood. Different approaches to environmental regulation present widely differing
incentives for how compliance is achieved and the relative role of technological innovation (e.g.,
Fischer et al. 2003).51 As a result, the same environmental end may be achieved at significantly
different costs, depending on the pace and direction of technical change.
The empirical economics literature also has observed that variable costs of production or
environmental abatement often decline over time with cumulative experience. Ferioli etal. (2009)
note that just the act of deploying a new technology can result in substantial process improvements
that translate into cost reductions. Building on this empirical observation, log-linear or S-shaped
learning curves that related the scale of production to per unit costs of production have often been
used to represent how costs decline over time with experience. However, explanations for why this
occurs vary (e.g., workers learn from mistakes and determine shortcuts; ad hoc processes become
standardized), and substantial uncertainty remains regarding how learning occurs over time for a
specific technology (Yeh and Rubin 2012). For instance, what learning rate is appropriate for a new
versus a mature technology? To what extent does the learning rate change over time or remain
relatively constant? Do costs always decline or increase in some cases?52
The EPA's Advisory Council on Clean Air Compliance Analysis stressed the importance of relying on
sector-specific empirical data to inform assumptions regarding learning effects whenever
possible.53 When no data are available or the evidence is outdated, it recommended the use of a
single default learning rate for transparency reasons. Sensitivity analysis is also recommended to
better understand the influence that learning curves can have on costs (U.S. EPA 2007).54 55 Given
uncertainty regarding how and when learning curves should be applied, analysts also should
51 For instance, the realized costs of Title IV of the 1990 Clean Air Act Amendment's Sulfur Dioxide (S02)
Allowance Trading program are considerably lower than initial predictions, in part due to incentives to innovate
in response to the policy (e.g., Bellas and Lange 2011; Frey 2013; Chan et al. 2018). See Chapter 4 for a discussion
of how different regulatory approaches may affect innovation.
52 Yeh and Rubin (2012) point to examples where technologies that were tested on a smaller scale or a
controlled setting actually experienced increased costs upon deployment due to unexpected performance or
reliability challenges.
53 OMB's Circular A-4 recommends that a cost analysis incorporate credible changes in technology over time,
noting evidence from the literature that variable costs of deploying new technologies or existing technologies in
new applications decrease over time (OMB 2023).
54 A useful description of the calculations used to identify a learning curve are found in van derZwaan and Rabl
(2004). The U.S EPA (2016b) reviews learning rates in the published literature for manufacturing and electric
utilities with a specific focus on the production of transportation-related goods (e.g., cars, ships, trucks). Grubb et
al. (2021) report estimated learning rates for energy and related technologies. Note that the empirical estimates
in the literature represent a biased sample, since they only represent technology that has been successfully
deployed (Sagarand van derZwaan 2006).
55 Note that cost decreases associated with technological change and learning may have additional costs
associated with them such as training costs.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
discuss and justify their assumptions. See Section 5.5.3 for additional discussion of technical change
and learning.
8.2.C > -cial Cost an- ' L'ntpl jyment
Recall that compliance costs include the value of labor for activities such as the installation and
maintenance of abatement technologies as well as monitoring, recordkeeping, and reporting. The
social cost of a regulation also includes the value of lost output associated with the reallocation of
resources (including labor) away from production of output and toward pollution abatement, the
value of induced changes in consumption and the deadweight loss from changes in the use of time
(i.e., due to pre-existing tax distortions in the labor market). Employment effects more generally,
such as those driven by labor-leisure choice, are not part of social costs under two commonly held
assumptions: if the economy is at full-employment (i.e., every worker who wants a job at the
prevailing wage has one) and with de minimus transition costs (Ferris and McGartland 2014).
Typically, a regulation reallocates workers among economic activities increasing employment in
some industries and decreasing it in others rather than affect the general employment level
(Arrow et. al 1996).56 For these reasons, employment effects should be characterized in the
economic impact analysis, as explained in Chapter 9.
8.2.; tiveness
As discussed in Section 8.1.2, imperfect competition in baseline market conditions may influence
the social cost of a regulation. Introducing an environmental regulation may also create conditions
that affect the size and market structure of industry, which may then allow firms to exercise market
power.57 Analysts should assess any expected changes to market structure as a result of the
regulation and, in particular, whether it will lead to imperfect competition and impact social cost
Environmental regulations can potentially affect the number of producers and the market structure
of the regulated sector by raising production costs, modifying economies of scale, or affecting
barriers to entry. For example, spatial heterogeneity in the stringency of environmental regulations
or compliance costs, and in turn their effect on production costs, can lead to market consolidation at
existing firms (e.g., Gray and Shadbegian 2002). Market structure can also be affected by the impact
of compliance activities and abatement technologies on the minimum efficient scale for firms in the
industry.58 Positive economies of scale for abatement technologies can lead to reduced entry and
greater exit (Millimet et al. 2009). Similarly, larger firms in the industry may have a competitive
advantage in the presence of economies of scale (Dean et al. 2000). Differences in product offerings
by firms may also affect market structure. If some firms subject to new product standards already
have compliant products, they will have a distinct advantage over others. Regional differences in
56 This does not mean, of course, that specific individual workers are not harmed by a policy (e.g., if they lose
their jobs).
57 The focus of this section is on how these changes may affect market structure, including reducing competition
(i.e., increase market power), and consequently affect the social cost of a regulation. These consequences of a
regulation may also affect the composition and distribution of costs within a sector and closely related markets.
See Chapter 9 for discussion of these market impacts.
58 Note, however, that it is theoretically ambiguous as to whether a reduction in output will be accompanied by
a net reduction in the number of firms in a regulated market for common regulatory designs such as
performance standards (e.g., Requate 1997; Lahiri and Ono 2007).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
regulatory requirements may also lead to product differentiation, which can then create or increase
market concentration (e.g. Brown et al. 2008; Chakravorty et al. 2008). Regulations can also create
barriers to entry either due to vintage-differentiated standards, whereby new entrants have stricter
standards, or through the control of patents on abatement technologies held by incumbents who
innovated as a result of the regulation. By decreasing opportunities for entry or the number of firms
in a sector, it is possible that incumbent or otherwise advantaged firms may now be able to charge
higher prices, which in turn further reduces competition and, all else equal, increases the social cost
of the regulation (see Figure 8.3 which shows the additional cost associated with reduced output
due to market power).
The effects of imperfect competition on the social cost of regulation may also increase over time as
markets adjust. For example, Fowlie et al. (2016) evaluate the social cost of a cap-and-trade
program in the cement sector, where firms have significant market power in local markets. They
show that, in the short run before production methods and the number of firms adjust, the
additional social cost of the regulation due to imperfectly competitive markets is relatively small.
However, in the longer run as firms change their production processes and firms exit the market,
the effect of imperfect competition on the social cost of the regulation is much higher. It is therefore
important that analysts evaluate potential differences between short and long run effects of a
regulation on market power.
8,2,4 Valuing Time
Compliance with environmental regulations changes the use of productive resources, including
people's time. Often, these changes occur at the workplace where labor is required to undertake
pollution control activities. Less often, time outside of the workplace is also affected; for example,
product bans might cause consumers to switch to substitutes that occupy more time. Changes in
time use can affect the social costs of a regulation. The EPA has produced a separate document on
how to value work time and nonwork time in regulatory analyses. We summarize the
recommendations below, but analysts should consult U.S. EPA (2020) for a detailed discussion.
The opportunity cost of worktime is determined by the value of the marginal product that would
have occurred absent the regulation. As a proxy for this opportunity cost, analysts should use the
employer's cost of employing a worker, consisting of the wage, fringe benefits and any overhead
costs.59 The value of work time will vary based on the industries and occupations affected by the
regulation. If overhead data are not available, U.S. EPA (2020) recommends that analysts use a
default multiplier applied to wages plus fringe benefits for the value of worktime.60 U.S. EPA (2020)
provides links to data sources for wages, benefits and overhead rates that normally are included
when valuing worktime.
Non-work time includes time spent on leisure, household production or other unpaid activities. Its
opportunity cost may vary by the types of activities forgone; the utility derived from the activity
59 Overhead costs are employer costs associated with labor, but not paid directly to workers, such as the value of
personnel services and training activities. For more information on how wages, fringe benefits and overhead
costs are defined, see Section 2.1 of U.S. EPA (2020).
60 The default multiplier in U.S. EPA (2020) reflects multiplier values used in prior analyses based on industry
Source: U.S. EPA (2020). The Bureau of Labor Statistics (BLS) Occupational Employment Statistics (OES) are
available at https://www.bls.gov/oes/tables.htm, while the Employer Costs for Employee Compensation and
occupation-specific benefit and overhead rates affected by EPA regulations.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
that occupies time; whether workers have a continuous choice over their hours of paid work; the
socioeconomic characteristics of affected individuals; and more. As a proxy for the opportunity cost
of non-work time, analysts should add the value of voluntary fringe benefits to the wage net of any
taxes paid by workers to federal, state and local governments on earned income.61 Table 8.2
summarizes the recommended approach and data sources for estimating work time and non-work
time.
In unusual circumstances, analysts may have access to information that allows an alternative
approach to estimating the value of work time or nonwork time. If utilized, analysts should explain
why the alternative is preferred to the approach recommended here and in U.S. EPA (2020).
Table 8.2 Estimating the Value of Work and Non-Work Time
Type of time
affected
Displaced activity
Estimation
approach
Data sources
Work time:
Tasks completed
while working for
pay
Other market
work in the same
industry and
occupation as
workers asked to
complete the
required tasks
Employer costs of
labor = Wages +
Fringe benefits +
Overhead costs
BLS OES or ECEC data on wages and
fringe benefits
For overhead costs, use industry
specific data as available
If overhead is not available, use the
recommended multiplier to obtain a
fully loaded wage
Non-work time:
Tasks completed
outside of paid
work time
Other nonmarket
activities such as
leisure and
nonmarket work
Individual
valuation of time =
(Wages -
Taxes on earned
income) +
Voluntary fringe
benefits
BLS OES or ECEC data on wages and
voluntary benefits
Adjust wage estimates using Census
CPS data on median household income
before and after taxes to estimate
average income tax rate
Source: U.S. EPA (2020). The Bureau of Labor Statistics (BLS) Occupational Employment Statistics (OES) are
available at https://www.bls.Qov/oes/tables.htm, while the Employer Costs for Employee Compensation.
8.2.5 Compliance Assumptions
In most cases, analysts should develop baseline and policy scenarios that assume full compliance
with existing and newly enacted (but not yet implemented) regulations. Assuming full compliance
focuses the analysis on the incremental effects of the new regulatory action without double-
counting benefits and costs already accounted for in previous regulatory analyses. That said, it is
important to determine whether specific policy options are more likely to result in compliance
issues or may be more difficult to enforce. In such cases, it is important to evaluate these effects
(e.g., options that require monitoring and reporting may have higher costs, but compliance is easier
61 Voluntary fringe benefits are the categories of employer-paid benefits that are not legally required and
include paid leave, supplemental pay (e.g., for overtime), insurance and retirement and savings plans.
8-24 Guidelines for Preparing Economic Analyses | 3rd edition
-------
to verify) and explore whether alternative options would result in improved compliance and/or
easier enforcement.
Assumptions about compliance behavior in the baseline and policy scenarios should be clearly
explained in the analysis. When compliance rates are uncertain or expected to vary across policy
options, analysts should explore the sensitivity of the results to these assumptions. See Section
5.4.2 in Chapter 5 for a more in-depth discussion.
Models Used in lire -"«f Bv
ion
Several types of models have been used to estimate the social costs of environmental regulation.
They range from models that estimate costs in a single industry (or part of an industry) to models
that estimate costs for the entire U.S. economy. In this section, we focus on three main model types:
compliance cost models, partial equilibrium (PE) models, and computable general equilibrium
(CGE) models. Input-output and input-output econometric models should not be used to estimate
social cost; however, these approaches and their limitations are also described. Analysts are
encouraged to consult with the National Center for Environmental Economics (NCEE) early in the
rulemaking process for help in identifying the most appropriate approaches for estimating the costs
of a specific regulation.
In practice, some models are simple enough to be implemented in a spreadsheet. Others consist of
systems of hundreds or even thousands of equations that require specialized software. Many
models are data intensive.62 63 Given model complexity, a simple model that captures key economic
features may be useful to identify which aspects of the regulatory options under consideration
likely matter from a cost perspective and therefore warrant further investigation in a more complex
model. Likewise, the use of a simple analytic general equilibrium approach is a less resource-
intensive way to build intuition within an internally consistent framework before utilizing a CGE
model (U.S. EPA 2017). Analysts should rely on a model that is "no more complicated than
necessary to inform the regulatory decision" (U.S. EPA 2009).
In some cases, use of more than one type of model may be warranted. Specifically, a more
aggregate CGE analysis may complement the cost estimates of a detailed compliance cost or PE
model (U.S. EPA 2017). For example, direct cost estimates from a compliance cost model can be
used as an input into a PE or CGE model. In some cases, models also can be linked to combine the
sectoral detail of a PE approach with the economy-wide features of a CGE model. Text Box 8.1
discusses linking models. Table 8.3 summarizes key attributes by model type.
When selecting a model, it is important to evaluate whether it is the most appropriate for the
question at hand (i.e., fit for purpose) and does a reasonable job of approximating the market(s)
62 Data requirements for these models vary, though advances in computing power, data availability and more
user-friendly software packages continually reduce the barriers to sophisticated model-based analysis. Refer to
Chapter 9 for a discussion of the public and private data sources that can be used for cost estimation.
63 Analysts should take great care in ensuring the quality of a model's data and specifications. See Section 8.4 for
a discussion of approaches to parameter selection, and ways to address parameter and model uncertainty.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
and behavioral responses of interest Most model types involve tradeoffs between different
strengths and weaknesses. Below are several factors that may be helpful in choosing a model.64
Types of impacts being investigated. Models differ in their abilities to estimate different
types of costs.
Geographic scope of expected impacts. Some models are well-suited for examining
regional or local impacts but may not capture the full range of costs at the national level and
vice versa.
Sectoral scope of expected impacts. Some models are highly aggregate and lack the detail
necessary to capture important aspects of compliance behavior within a single sector.
Likewise, highly detailed sector models often do not capture effects on other sectors and
may not adequately capture demand response.
Expected magnitude of impacts. A model well-suited for estimating the cost of a
regulation with large effects may have difficulty estimating the cost of a regulation with
relatively smaller expected effects, and vice versa.
Expected importance of interactions and feedbacks with other sectors. When
regulations are expected to have substantial effects on the broader economy, it is important
to choose a model that can capture those effects.
Other criteria for ensuring a specific model are appropriate and of sufficient quality to analyze the
effects of a regulation are discussed in Text Box 5.2. For example, chosen models should be subject
to credible and objective peer review to ensure consistency with scientific and economic theory
before being used in regulatory analysis. Comprehensive documentation of model components, and
as possible, underlying data sources should be publicly available.
Usually, some combination of the above factors will determine the most appropriate type of model
for a specific application. Analysts should present a reasoned discussion of the factors that inform
their model choice. Analysts should describe the main upstream and downstream sectors affected,
whether close substitutes to the regulated good are available, the extent to which the goods affected
are substitutes or complements to leisure, and the existence of pre-existing distortions in affected
sectors (e.g., subsidies, imperfect competition, other regulations, or externalities). Evidence from
the literature such as supply and demand elasticities that indicate market responsiveness (e.g., of
consumers, input markets, substitutes, and complements) will aid the analyst in justifying model
choice. Ultimately, models need to be supported by the data: for example, a single-market PE
analysis requires demand and supply elasticities, while a multi-market or CGE analysis requires
cross-price elasticities.
64 This list of factors is informed by Industrial Economics, Inc. (IEc 2005).
8-26 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 8.1 - Linking Models
CGE models are aggregate representations of the economy that allow an analyst to capture the
interactions of producers and consumers as changes in prices and quantities in the regulated
sector percolate through the rest of the economy. These economy-wide interactions are
captured through exogenously specified elasticities of substitution that approximate detailed
demand and supply responses from the policy. There may however, be instances when CGE
models do not have sufficient detail to quantify how regulated entities may respond to a
regulation, such as the types of compliance methods that are available.
Partial equilibrium and compliance cost approaches typically do not suffer from a lack of detail.
They often have technology-rich representations that reflect the range of salient characteristics
for regulated sources as well as installation and operation costs for each individual compliance
technology. However, often demand and supply are specified in a very simple way and
interactions with other potentially affected markets are not considered.
Much could be gained by linking these two modeling approaches in a coherent and sensible way
to take advantage of the technological detail of compliance cost or PE models and the
theoretically consistent economic structure of CGE models (Bohringer and Rutherford 2008).
There are a number of studies, many in the energy context, that have leveraged such linkages
(e.g., Cai and Arora 2015; Rausch and Karplus 2014; Kiuila and Rutherford 2013; Lanz and
Rausch 2011; Sue Wing 2006; Schafer and Jacoby 2005; McFarland et al. 2004). The Science
Advisory Board (SAB) (U.S. EPA 2017) recommended that EPA make linking more aggregate
CGE models to more detailed models of households, industries or sectors a research priority. It
signaled a clear preference for two-way linkages between models: the CGE model simulates
prices and investment for use as inputs to the compliance cost or PE model, while the
compliance cost or PE model computes technology capacities and output supplies that are used
as inputs by the CGE model. The two models are run in an alternating fashion until convergence.
It is important to note that, in practice, any linking exercise is dependent on the information
available from the sector model and the representation of relevant sectors and markets in the
CGE model. As the information and available models may differ significantly across regulatory
analyses, any application of linking also may present unique challenges and considerations.
To link a compliance cost model with a CGE model, the accounting of outputs and inputs
between the two models needs to be sufficiently aligned. To do this, it is important to
disaggregate compliance costs into the factors (e.g., labor, capital, energy, materials) that
correspond to the inputs to the sector's production function as specified in the CGE model.
However, this is often not a straightforward exercise. For instance, the fixed cost of a compliance
method may include both the capital used for a compliance technology and the labor to install it
Likewise, variable costs may include materials as well as labor for maintenance. However, in
both cases the shares of the compliance cost from the specific inputs are rarely available. It is
also a challenge to aggregate compliance cost information up to the sector level for the purpose
of linking to the CGE model. A compliance cost model often provides information on the
expected compliance behavior and cost for each affected entity. The CGE model usually
represents a sector with a single representative firm.
Many of the challenges of linking compliance cost models to CGE models also apply to linking
CGE models with PE models. However, because PE models may have their own sets of
assumptions on baseline forecasts, elasticities of demand and supply, functional forms, and/or
technical change that may differ from the underlying assumptions in the CGE model, there may
be additional complexities that must be grappled with and reconciled in some way to ensure
that the linkage remains feasible and produces sensible results.
8-27 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Table 8.3 Summary of Key Attributes by Model Type
Attributes
Sector-Specific
Compliance Cost
Sector-Specific
Partial
Equilibrium
Economy-Wide
CGE
Significant industry detail; rich set of
technologies
Sometimes
None
Account for facility or market constraints
Sometimes
Model changes in regulated producer
behavior (e.g., input and process changes)
Sometimes
Represent interactions and feedbacks
between sectors
None
Limited or none
Model demand side response
None
Limited
Can directly estimate welfare effects
None
Sometimes
8.3.1 Compliance Cost Models
Compliance cost models are used to estimate the direct costs of compliance with a regulation.
Estimates by engineers and other experts are used to produce algorithms that characterize the
changes in costs resulting from the adoption of various compliance options and are usually
determined for individual facilities or for categories of model facilities with varying baseline
characteristics. To estimate the control costs of a regulation for an entire sector, disaggregated data
that adequately reflect the industry's heterogeneity are used as an input into the model. The
disaggregated cost estimates are then aggregated to the industry sector level. These models are
most informative when the data are available to capture heterogeneity across facilities, both in
terms of individual characteristics (e.g., facility age and production technology, input costs) and
compliance options.
The structure of compliance cost models can vary depending on the scope of an analysis. For
instance, compliance cost models may include many of the categories of costs previously described
in Section 8.2.1 (e.g., capital costs, operating and maintenance expenditures, monitoring,
measurement, and reporting costs). Moreover, some compliance cost models are designed to allow
the integrated estimation of control costs for multiple pollutants and multiple regulations. Some
models account for cost changes over time, including technical change and learning. While most
compliance cost models are for facilities within a specific industry, they may also be models of
households.
While precise estimates of compliance costs are an important component of any analysis, recall
that, in cases where the regulation is not expected to significantly affect market supply and demand
in the regulated market, compliance costs can be considered a reasonable approximation of social
cost. Compliance cost models usually focus on the supply side because regulations are typically
imposed on producers. In circumstances where producer and consumer behavior are appreciably
affected, these models are not able to provide estimates of changes in industry prices and output
resulting from the imposition of a regulation.
8-28 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Advantages
Compliance cost models often contain significant industry detail and can provide relatively
precise estimates of the costs incurred by regulated sources (or categories of regulated
sources) when complying with a regulation.
Once constructed, compliance cost models often require fewer resources to implement and
are relatively straightforward to use and easy to interpret
Limitations
As they usually focus on the supply side and do not capture changes in production among
affected sources, compliance cost models can only provide estimates of social cost in certain
cases.
Compliance cost models are often limited to estimating the costs of complying with
regulatory requirements for a single industry.
Linear xiels
Often linear programming models are used in the analysis of EPA regulations to estimate
compliance costs. Linear programming models minimize (or maximize) a linear objective function
by choosing a set of decision variables, subject to a set of linear constraints. In the EPA's regulatory
context, the objective function is usually to minimize compliance costs incurred by the regulated
sources. The decision variables represent the production and compliance choices available to the
regulated entities. The constraints may include available technologies, productive capacities, fuel
supplies and regulations on emissions.
Although linear programming models can be constructed to examine multiple sectors or even
economy-wide effects, they are commonly focused on a single sector. For the regulated sector, a
linear programming model can incorporate a large number of technologies and compliance options,
such as end-of-pipe controls, fuel switching and changes in plant operations. Similarly, the model's
constraints can include multiple regulations that require simultaneous compliance. The objective
function usually includes the fixed and variable costs of each compliance option.
In addition to compliance costs, the outputs from the model may include other related variables,
such as projected input use, emissions, and demand for new capacity in the regulated industry. In
some cases, linear programming models may also include supply and demand representations (e.g.,
elasticities) of multiple markets and therefore more closely resemble the partial equilibrium
models described in Section 8.3.2.
While the estimated change in expenditures incurred by the regulated sector may be of policy
interest, it is not equal to social cost when input or output prices change. If the linear programming
model captures changes in market prices in response to the policy, then it is possible to use the
model outputs to estimate a partial equilibrium estimate of social cost (e.g., changes in producer
and consumer surplus).
8,3,2 Par lis
In cases where the effects of a regulation are confined to a single or a few markets, partial
equilibrium single or multi-market models that incorporate anticipated demand and supply
responses can be used to estimate social cost.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Inputs into a partial equilibrium model may include regulatory costs estimated using a compliance
cost model and the supply and demand elasticities for the affected market (as well as cross-price
elasticities when there are multiple affected markets). The model then can be used to estimate the
change in market price and output. Changes in producer and consumer surplus reflect the social
cost of the regulation.
In a partial equilibrium model, the magnitude of the impacts of a regulation on the price and
quantity in the affected market depends on the shapes of the supply and demand curves in the
region at which expected changes are to occur. The shapes of these curves reflect the underlying
elasticities of supply and demand. These elasticities either can be estimated from industry and
consumer data or taken from previous studies. While in practice these models often assume perfect
competition, it is also possible to construct a partial equilibrium model that accounts for the role of
market power in production decisions.
If the elasticities used in an analysis are drawn from previous studies, they should reflect:
A similar market structure and level of aggregation;
The appropriate spatial resolution (i.e., local, regional, or national);65
Current economic conditions; and
The appropriate time horizon (i.e., short or long run).
In some cases, if the effects of a regulation are expected to spill over into adjoining markets (e.g.,
suppliers of major inputs or consumers of major outputs), partial equilibrium analysis can be
extended to these additional markets as well.
Advantages
Because they usually simulate only a single or small number of markets, partial equilibrium
models generally have fewer data requirements relative to a CGE approach and are more
straightforward to construct.
Partial equilibrium models are comparatively easy to use and interpret
Limitations
Partial equilibrium models are limited to cost estimation in a single or small number of
markets and do not capture broader effects in the overall economy.
Because partial equilibrium models are generally data driven and specific to a particular
application, they are usually not available "off-the-shelf' for use in a variety of analyses.
8.3.3 Computable General lodels
The most appropriate type of model to estimate the social cost of a regulation in a general
equilibrium framework is a computable general equilibrium model. This type of model is
comprehensive and internally consistent, accounting for budgetary and resource constraints
operating throughout the economy. A key advantage over the other types of models discussed in
this section is its ability to capture interactions between economic actors (often delineated with
multiple sectors and regions) and with pre-existing distortions (e.g., taxes, other regulations, or
externalities) across the entire economy. Relative to PE and compliance cost approaches, however,
65 For instance, Bernstein and Griffin (2006) estimated short-run price elasticities of demand for electricity in
the United States that varied from -0.04 to -0.31 by region, and long-run price elasticities of demand for
electricity that varied from about-0.05 to almost-0.6 by region.
8-30 Guidelines for Preparing Economic Analyses | 3rd edition
-------
CGE models are highly aggregate and often use simplified representations of production decisions
(e.g., perfect competition, characterization of abatement opportunities) within a sector. They may
also be more difficult to parameterize.
CGE models assume that an economy can be characterized by a set of conditions in which supply
equals demand in all markets. When the imposition of a regulation alters conditions in one market,
a general equilibrium model determines a new set of relative prices for all markets that return the
economy to a long-run equilibrium. These prices in turn determine changes in sector outputs and
household consumption of goods, services and leisure in the new equilibrium. In addition, the
model determines a new set of relative prices and demand for factors of production (e.g., labor,
capital and land) the returns to which compose business and household income. The social cost
of the regulation is estimated in CGE models as the change in economic welfare in the post-
regulation, simulated equilibrium compared to the pre-regulation, "baseline" equilibrium.66
CGE models are built using structural micro-theoretic foundations to capture behavioral
responses.67 In canonical CGE models,68 firms are generally assumed to be profit-maximizers with
constant returns to scale in production; households maximize utility from the consumption of
goods and services using a specific functional form; and markets are perfectly competitive. Multiple
household types can be included in the model (for instance, differentiated based on geography or
income) to calculate distributional impacts of policy changes. A common feature in many models is
an underlying model of international trade following Armington (1969) where preferences for
goods are differentiated by country of origin to allow for two-way trade for otherwise identical
goods. Labor and capital are typically fully mobile between sectors with labor fully employed and
no involuntary unemployment.
CGE models are generally more appropriate for analyzing medium- or long-term effects of
regulation, when most inputs are free to adjust and consumers can modify purchasing and labor-
leisure decisions in response to new prices. A longer time horizon also affords greater
opportunities for firms to change production processes (i.e., innovate). The time required to move
from one equilibrium to another after a new policy is introduced is not defined in a meaningful way
(and is usually assumed to be instantaneous). As such, CGE models are generally not well-suited for
66 Regulatory compliance creates a need for additional inputs to produce goods in the regulated sector along
with pollution abatement. While the total cost of these additional inputs can be derived from detailed
compliance cost estimates, it is not always clear how to allocate the total cost among the inputs specified in the
CGE model because CGE models are by their nature an aggregated, parsimonious representation of the economy.
67 Structural models explicitly specify underlying preferences, production and resource allocation in ways that
are consistent with economic theory. The calibration of structural or behavioral model parameters with actual
data ensures that the model represents important economic features while remaining in agreement with the
underlying theory (Woodford 2009).
68 Here, the term "canonical" is indicative of off-the-shelf models, or models with features that are most common
in the literature. In reality, a CGE model may contain several hundred sectors or only a few and may include a
single "representative" consumer or multiple household types. It may focus on a single economy with a simple
representation of foreign trade, or contain multiple countries and regions linked through an elaborate
specification of global trade and investment. The behavioral equations that govern the model allow producers to
substitute among inputs and consumers to substitute among final goods as the prices of commodities and factors
shift. The behavioral parameters can be econometrically estimated, calibrated or drawn from the literature. In
some models, agents may make intertemporal trade-offs in consumption and investment.
8-31 Guidelines for Preparing Economic Analyses | 3rd edition
-------
analyzing transition costs as the economy moves to the new equilibrium unless a transition path
can be appropriately specified.69
The case for using CGE models to evaluate a regulation's effects is strongest when the regulated
sector has strong linkages to the rest of the economy and the regulation is expected to affect most
firms in a broadly defined sector. Narrowly targeted regulations are more difficult to capture
without explicitly linking a CGE model to a detailed PE sector model (U.S. EPA 2017). Linking
models is discussed in Text Box 8.1. The extent to which CGE models will add value to the analysis
also depends on data availability (see Text Box 8.2 on input-output data efforts).70 When
developing a plan for analysis, analysts should consult with NCEE if they anticipate using a CGE
model to evaluate the effects of a regulation.
Note that absent a credible way to represent environmental externalities in a CGE model or the
benefits that accrue to society from mitigating them a CGE model's economic welfare measure is
incomplete.71 However, the inability to account for interactions between costs and benefits in a CGE
model does not invalidate their use to estimate costs or make it impossible to design consistent
approaches to cost and benefit estimation (U.S. EPA 2017). The possibility of incorporating benefits
into a CGE framework is discussed in Text Box 8.3.
Advantages
CGE models are best suited for estimating the cost of policies that will have a broad set of
economy-wide impacts, especially when indirect and feedback effects are expected to be
significant.
CGE models are most appropriate for analyzing medium- or long-term effects of policies or
regulations.
Limitations
Because of their equilibrium assumptions, CGE models are generally not appropriate for
analyzing short-run transition costs.
CGE models are highly aggregate and do not provide detailed cost estimates for narrowly
defined sectors or small geographic areas.
CGE models may be more difficult to parameterize and use highly simplified
representations of sector production and abatement decisions.
8.3.4 Oth i [n 'sfri"|ti1 r I- ised Models
Several other economy-wide approaches are referenced in the literature, including input-output (I-
0) models and 1-0 econometric models. These methods should not be used to estimate the social
cost of environmental regulation (U.S. EPA 2017).
69 For instance, Williams and Hafstead (2018) embed short run transitional unemployment costs in a general
equilibrium model.
70 Data limitations are a significant obstacle for all of the modeling approaches discussed in Section 8.3, both in
terms of achieving the granularity needed to adequately represent a regulation and to estimate its effects.
71 An expanding body of work has begun to include non-market goods in CGE models (Smith et al. 2004;
Carbone and Smith 2008).
8-32 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 8.2 - Input-Output Data and Open-Source Initiatives
Input-output (1-0} data are a basic input into any CGE model. An 1-0 table assembles data in a
tabular format that describes the interrelated fl ows of m arket goods and factors of production
over the course of a year. It may consist of hundreds of sectors or just a few sectors. In the
United States, the Bureau of Economic Analysis (BEA) provides a time series of national level 1-0
accounts with multiple levels of sectoral aggregation (between 15 and 402 sectors) based on
North American Industry Classification System codes. For more information on constructing 1-0
tables, see Miller and Blair (2009), Horowitz and Planting (2009), and
https://www.bea.gov/industry/input-output-accounts-data.
Below is an aggregated 1-0 table for the U.S. for 2022 based on BEA data. The columns for the
individual sectors denote how much of each commodity is used to produce that sector's output
(cost of annual production). A sector's cost schedule (upstream sectoral linkages) is composed
of intermediate inputs, factors of production (labor and capital) and tax payments. Payments to
factors (wages and profits) and tax payments comprise sectoral value added. Take the
agricultural sector - intermediate input costs consisted of $195 billion of agricultural inputs,
$129 billion of manufactured inputs, $338 billion of other intermediate inputs and $286 billion
of value added, for a total of $948 billion in input costs. The row for each commodity shows how
that commodity is consumed (also known as downstream linkages). For the agricultural sector,
$711 billion is consumed as intermediate inputs for sectoral production ($195 + $6 + $443 +
$66), while $237 billion is consumed as final demand (i.e., C + G + I + (X-M), or $246 + $0 - $25 +
$15). In this framework, the total output receipts must equal total input costs.
1-0 Table for the United States (2022)
O)
3
Utilities
c
o
B
c
3
ts
c
o
n:
t
Services
Account
3
U
bo
<
Mining
-------
In a CGE framework, columns for the individual sectors determine input shares used to
calibrate production functions. Columns for final demand determine expenditure shares
used to calibrate household expenditure functions. These data, along with "transactions and
transfers between institutions related to the distribution of income in the economy," form
the basis of the social accounting matrix (Miller and Blair 2009). Constructing a social
accounting matrix requires reorganizing the data shown above to link sources of household
income to expenditures. 1-0 accounts with sub-national or international detail are not
provided by the BE A but have been established by others. For instance, the Global T rade and
Analysis Project (GTAP) compiles and reconciles data from many sources to have consistent
sectoral and agent aggregations across countries (https://www.gtap.agecon.purdue.edu/).
Further, the Wisconsin National Data Consortium (WiNDC), develops consistent subnational
1-0 tables based on publicly available data fhttp: //windc.wisc.edu A
8.3.4.1 Input-Output Models
1-0 models are highly disaggregated empirical descriptions of the interrelated flows of good and
factors of production.72 They are generally static and assume a fixed, strictly proportional
relationship between inputs and outputs via multipliers.73 Although their specifications can
sometimes be partially relaxed, input-output models embody the assumptions of fixed prices and
technology, which do not allow for the substitution that normally occurs when goods become more
or less scarce. Similarly, most input-output models are demand driven and not constrained by
limits on supply, which would normally be transmitted through increases in prices. While some of
the rigidities in the models may be reasonable assumptions in the very short run or for regional
analysis with limited ties to the broader national economy, they limit the applicability of 1-0 models
for evaluating medium- to long-run effects or national issues. For instance, the lack of resource
constraints and substitution effects that occur over the longer run means that 1-0 models tend to
overestimate the effects of a policy.74 Importantly, the 1-0 approach does not necessarily account
for shifts in economic activity toward the pollution abatement sector (e.g., when the directly
regulated sector purchases pollution abatement equipment or services to comply with the
regulation). Because input-output models do not include flexible supply-demand relationships or
the ability to estimate changes in producer and consumer surpluses, they are not appropriate for
estimating social cost75
72 Miller and Blair (2009) is a standard reference on input-output analysis.
73 The assumption that output changes translate directly to proportional changes in inputs is not empirically
founded and therefore should not be used, even in the short run, because it ignores the potential for factor
substitution. Such shifts may change the labor-, capital-, energy- or materials-intensity of production.
74 Studies that rely on 1-0 models often calculate some combination of direct, indirect and induced effects. Direct
effects are the changes in output that result from an increase in the cost of inputs (e.g., fuel) in the directly
regulated sectors, using the fixed, proportional relationship mentioned above. Indirect effects of a regulation are
calculated by using the 1-0 relationship between outputs in the directly affected sectors and required inputs in
related sectors (e.g., suppliers). Induced effects are general re-spending effects that result from changes in
household income.
75 See U.S. Chamber of Commerce and NERA Consulting 2013; OECD 2004, and Dwyer et al. 2006.
8-34 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 8.3 - Separability between Benefits and Costs
When estimating the benefits and costs of environmental regulation, it is almost always
assumed that the two are separable, such that the beneficial impacts of the polices do not
meaningfully affect the factors that determine the cost of the policy, and vice versa. This is
due, in part, to a lack of empirical evidence regarding the sign or importance of the
relationship between environmental quality, which is typically not priced in the
marketplace, and market goods (Carbone and Smith 2008).
Benefits and costs are non-separable when either the compliance costs borne by firms or
households interact with and alter the valuation of environmental benefits (other than
through changes in environmental contaminants), and/or the beneficial impacts of the
regulation alter the costs. Non-separability may occur for several reasons. The costs of an
environmental policy may alter the budget constraint for households for instance, when
compliance costs are passed on to consumers as higher prices for goods such as electricity,
and this, in turn, affects their willingness to pay for the beneficial impacts of the policy. It
may also be the case that changes in environmental quality and health status lead to changes
in household behavior, which for large policies could affect relative prices in equilibrium and
the cost of complying with the policies. For example, greenhouse gas mitigation polices in the
electricity sector may reduce future demand for space cooling and therefore electricity, in
turn reducing the costs of complying with the policy.
Ongoing work also suggests that reductions in mortality risk may affect how households
smooth consumption over time (i.e., through savings), which may interact with pre-existing
capital taxes or affect the price of investment in pollution abatement capital (Marten and
Newbold, 2017). The fact that changes in environmental quality and health status can affect
behavior in markets underpins the revealed preference approaches for estimating
willingness to pay discussed in Chapter 7.
As noted by the SAB (U.S. EPA 2017), when either the costs or benefits of a regulation are
estimated while holding the other constant, any potential non-separability between costs
and benefits is missed, which complicates comparing them and calculating social net
benefits. The specific magnitude and ultimate impact of non-separability on the net benefits
of environmental regulations is an empirical question that requires additional study and is
the subject of an emerging literature (Sue Wing 2011). The SAB noted that potential non-
separability for large policies does not invalidate estimates of costs and benefits using
existing methods. However, caution should be applied when obvious interactions exist.
8.3.4.2 input-Output Econometric Models
1-0 based econometric models integrate the high level of detail from an input-output model with
the forecasting properties of a macro-econometric forecasting model. Unlike standard 1-0 models,
this approach accounts for supply-demand conditions in the economy, including resource
constraints, through a series of accounting (e.g., savings equal investment) and econometrically
estimated relationships (Hahn and Hird 1991). Feedbacks between supply and demand occur via
econometric equations (CGE models accomplish this via a price mechanism and market clearing
assumptions (West 1995)). The predictions generated by this type of model "are integrated and
simultaneously determined ... price increases in one sector are translated into cost and price
8-35 Guidelines for Preparing Economic Analyses | 3rd edition
-------
increases in other sectors" (Portney 1981). This is a key advantage over standard 1-0 models that
assume away these effects.
While CGE models assume full market clearing, 1-0 econometric models assume imperfect
knowledge of product and factor markets, with an emphasis on tracking short run disequilibrium
(West 1995). This makes them particularly attractive for analyzing transition costs. However, a
major drawback of 1-0 econometric models is that because they are reduced-form models
predicated on historical relationships, they cannot take into account the possibility that a firm or
consumer may modify their long-run behavioral response to changes in policy (referred to as the
Lucas critique). The inability to account for this dependence invalidates these models for purposes
of policy evaluation outside of short-term forecasting (Schmidt and Wieland 2013; Fischer and
Heutel 2013; U.S. EPA 2017).
. i Tlh Mm I 1 ¦ Y\ H "> [ i [ i ^ Cj i\\11 ngu
Even when the analyst has determined what types of models are most appropriate for the
estimation of social cost, several important modeling decisions remain, including deciding on the
level of sectoral and regional aggregation, whether to use a static or dynamic framework, and how
to parameterize the model. In addition, analysts should evaluate key uncertainties and take care not
to double count, particularly when using outputs from one type of model as an input into another.
jregation
The level of sectoral and regional aggregation assumed in a model will determine what aspects of
the sector or economy can and cannot be captured explicitly in a regulatory analysis. Matching the
level of aggregation in a model to the level needed to evaluate a policy's main effects is important to
ensure that the analysis does not miss important contributors to the cost For example, consider
the effects of a new regulation on refrigerant gases in the frozen bakery products sector. In a CGE
model, the frozen bakery products sector is not typically separated out as its own sector. Instead, it
is captured in a more aggregate category food products along with many other related
industries such as soft drinks, cereal and chewing gum. As a result, the frozen bakery products that
are affected by the policy "may be too small a part of the model's food products sector to give
meaningful results due to 'aggregation bias.'76 Put another way, there are too many products in the
model's sector to accurately isolate the frozen bakery products industry" (Rivera 2003).
The level of aggregation can affect sectoral and economy-wide results. For instance, sectoral
disaggregation allows for a differentiated representation of production technologies, behavioral
parameters (e.g., elasticities) and emission intensities that may matter for estimating costs and
other impacts (Alexeeva-Talebi et al. 20 1 2).77 However, models that are highly specialized for
76 Caron (2012) defines "aggregation bias" for a specific variable as the difference in its value from an
aggregated model relative to a disaggregated model that has been re-aggregated after the fact to a comparable
level
77 Alexeeva-Talebi etal. (2012) and Caron (2012) found that the range and standard deviation of sectoral
impacts increased with disaggregation. In some cases, even the direction of the estimated impacts was reversed
relative to more aggregate results. However, while a highly aggregated model may not be a reliable predictor of
sub-sectoral impacts, for many applications, they found that these models produce satisfactory estimates of the
overall impacts on the economy.
8-36 Guidelines for Preparing Economic Analyses | 3rd edition
-------
capturing impacts in a specific sector will usually miss impacts on a broader set of sectors. It is also
important to consider how costs are allocated spatially (and temporally) to avoid a mismatch
between affected facilities' locations and the scale of the model. While proficient at capturing major
impacts and interactions between sectors, CGE models generally are not well-suited for focusing on
a single or small number of specialized sectors because of their level of aggregation.
8.4,2 Choos I- we . tic an [.' nar Ki -ntt- rk
It is possible to construct static or dynamic versions of all three types of economic models discussed
in this chapter (i.e., compliance cost, PE and CGE). In a compliance cost framework, the analyst may
assume that economic conditions are static or dynamic. If future economic conditions are expected
to change meaningfully, a dynamic framework should be used because compliance decisions may
be influenced by future economic conditions even when the regulation is not expected to
meaningfully influence production or prices. For example, if an affected source anticipates
operating for a long time it may choose a more capital-intensive compliance option over a less-
capital intensive option because there is a longer period over which it can recover the cost of that
investment. Similarly, if the number of affected sources is anticipated to change over time, then the
cost of complying may change over time. A dynamic compliance cost framework may also be
preferred if, for example, regulated sources may make anticipatory investments prior to a
regulation's compliance dates or to account for the potential for technological change (see Section
8.2.3.4).
When the analyst expects intertemporal effects of a regulation to be confined to the regulated
sector or a few related sectors, it may be appropriate to simply apply partial equilibrium analysis to
multiple periods. As with compliance cost models, relevant conditions, like expected changes in
market demand and supply over time, should be taken into account in the analysis. The costs in
individual years can then be discounted back to the initial year for consistency.
If the intertemporal effects of a regulation on non-regulated sectors are expected to be significant,
analysts also can estimate social cost using a dynamic CGE model. Dynamic CGE models can capture
the effects of a regulation on affected sectors throughout the economy. They can also address the
long-term impacts of changes in labor supply, savings, factor accumulation and factor productivity
on the process of economic growth. In a dynamic CGE model, social cost is estimated by comparing
values in the simulated baseline (i.e., in the simulated trajectory of the economy without the
regulation) with values from a simulation with the regulation in place.
Analysts should keep in mind that the evolution of variables in a dynamic model sometimes
depends on exogenously imposed assumptions that are not always easy to validate. For instance,
modelers sometimes need to constrain the pace at which some variables in the model change (e.g.,
how quickly technology changes) based on an external assessment of what is technically feasible.
Key exogenous assumptions should be clearly documented and explained. In some cases, it also
may be useful to explore the robustness of cost estimates to alternative assumptions.
8.4.J tions
Dynamic models must specify the ways in which households and firms formulate and update
expectations about future prices, returns, growth or other key economic variables. There are a
variety of ways to formulate expectations about the future, but they generally fall into two general
categories: backward-looking and forward-looking. With advances in computer power, forward-
looking expectations are the more common assumption in CGE models.
8-37 Guidelines for Preparing Economic Analyses | 3rd edition
-------
The two main backward-looking formulations are myopic and adaptive expectations. Myopic
households and firms do not anticipate future changes to the economy or regulatory setting, and do
not make investments or change consumption and savings behavior until the period when the
change takes effect (Paltsev and Capros 2013). Households and firms with adaptive expectations
base their expectations about the future primarily on past experiences and are, therefore, relatively
slow to modify behavior in response to new information.
In forward-looking models, households and firms have either perfect foresight or rational
expectations. A household or firm with perfect foresight knows what the future values of key
economic variables will be with certainty and incorporates this information immediately into
current decisions (Paltsev and Capros 2013). Modeling rational expectations allow for households
and firms to account for uncertainty in future conditions; in this case, they incorporate all relevant
information, both past and future, into decision-making and are assumed to get future values
correct on average. In other words, they do not systematically make forecasting errors.
There are several analytic implications tied to the degree of model foresight assumed in a dynamic
model. For instance, a backward-looking model may lead to higher estimates of compliance costs
and welfare impacts compared to a forward-looking model since it restricts the response flexibility
of households and firms relative to reality. However, in cases where they are assumed to have
perfect foresight but the future path of key variables is uncertain in reality, a deterministic forward-
looking model may underestimate the compliance costs and welfare impacts of regulation. Note
that many EPA regulations phase in standards or allow for intertemporal smoothing of compliance
(e.g., banking of emissions allowances) that could at least partially alleviate this concern. Section
5.5.1 provides additional information on the role of uncertainty on household and firm behavior
when estimating the impact of regulation.
Another consideration is the large number of variables and constraints that must be simultaneously
determined in a forward-looking model. This, in turn, restricts the level of detail that can be
included in the model, which may be critical to adequately assessing the social cost of a regulation.
As such, a more aggregate forward-looking CGE model should be viewed as a complement to
analysis supported by detailed compliance cost or PE sector model. All else equal, these
considerations are even more restrictive for forward-looking models of decision making under
uncertainty that also necessitate integrating over all temporal sources of uncertainty to form
household and firm expectations.
-Vk : Tim - *ps
Static models provide cost estimates for one period, typically a year. They either assume that
conditions are invariant over time or that the cost estimate is indicative of a typical or
representative period. Static models exist for all three frameworks discussed in this chapter (i.e.,
compliance cost, PE and CGE). As discussed above, if economic conditions are expected to change
over time, or if changes in behavior to come into compliance and/or to new market equilibria take
time, static models may provide incomplete estimates of costs (and benefits).
Most dynamic models operate using discrete time steps. Time steps between periods are chosen to
provide enough detail regarding the adjustment to policy over time, while using a manageable
number of time periods for computational reasons. For instance, because dynamic CGE models are
often solved over periods of 50 years or more, it is not always practical to solve the model for each
individual year. However, when using a dynamic CGE model, the year in which a regulation comes
into effect may not be explicitly modeled. Due to the expense and time required to adjust the model
and baseline, adding a new solution year may not be an option. In this instance, analysts may use
the model year closest to the year in which the regulation will come into effect as a proxy.
8-38 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Regulations that are introduced gradually or vary timing of compliance by region or state pose
additional challenges for model representation.
In addition, if the end-year chosen for a dynamic model stops short of capturing important
regulatory effects, the social cost estimate may be biased downward. When compliance costs
cannot be estimated for all future years, a forward-looking model may smooth them over time,
which can also lead to biased social cost estimates, though the direction of the bias will depend on
what is assumed about future compliance costs.
8.4,3 Model Parameterization
Regardless of the chosen modeling framework, there is a distinction between values determined
within the model (those that are endogenous) and values determined outside of the model (those
that are exogenous). Model parameterization is concerned with the latter.78 The values that are
imposed exogenously will depend on the type of model used to capture economic behavior.
In general, model parameterization takes place in two steps. First, the analyst attempts to
accurately represent the current structure of the sector(s) or markets of interest For compliance
cost models, this step typically relates to specifying relevant compliance options and constraints
(e.g., production capacities). In the case of CGE models, this step consists of characterizing a
baseline and calibrating model functions to a reference equilibrium. Second, parameter values are
chosen that best characterize economic relationships (i.e., the curvature of different functions) in
the model. In the case of compliance cost models, the analyst may need to specify constraints on
economic behavior (e.g., production levels, other regulatory requirements) and cost functions (e.g.
slopes, nonlinearities). For partial and general equilibrium models, this second step is more difficult
and requires the analyst to choose appropriate behavioral parameters (e.g., elasticities).
Parameter values can be estimated or based on existing values from the literature. While basing
parameter values on the existing literature is the more common approach, inconsistencies between
the underlying structure of the model and the empirical analyses from which values are drawn can
lead to modeled responses that are not supported by the underlying data of the empirical
analyses.79 To alleviate some of these concerns, some researchers have econometrically estimated
the model parameters in a framework that is consistent with the underlying model (e.g., Jorgenson
et al. 2013). If parameters are estimated by the analyst, preference should be given to using publicly
available data where possible. When borrowing estimates from the literature to parameterize a
model, analysts should use estimates that reflect the most recent scientific methods and data as
possible, discuss the reasons for choosing one value over another, and discuss any limitations. For
instance, available parameter values in the literature may not be regularly updated and produced
using data that are significantly older than the modeled year(s) of interest In cases where there is
78 Specifically, parameters are "terms in the model that are fixed during a model run or simulation but can be
changed in different runs, either to conduct sensitivity analysis or to perform an uncertainty analysis when
probabilistic distributions are selected" (U.S. EPA 2009).
79 This point also applies to instances where analysts estimate their own parameters for a model. Identifying
assumptions in the empirical framework need to be consistent with assumptions in the model of interest. For CGE
analysis illustrating this point, seeShoven and Whalley (1984) and Canova (1995).
8-33 Guidelines for Preparing Economic Analyses | 3rd edition
-------
no clear consensus in the literature on the most defensible estimates to use, sensitivity analysis to
understand the robustness of cost estimates to the parameters chosen is recommended.80
Often, a regulation covers many highly heterogeneous facilities where both the compliance options
available and abatement costs vary widely. When highly disaggregated source-level data are
unavailable, analysts may pursue a model plant approach to estimate compliance costs, where a
subset of individual facilities sharing certain characteristics (e.g., plant age, type of production
process, industrial sector) are represented by a single model plant Analysts also may use a model
plant approach to reduce the computational requirements of a compliance cost model.
The model plant is intended to represent the typical conditions of a group of facilities. While this
provides a way to overcome data limitations and simplify the model, parameterization can still
prove challenging. This is particularly true when conditions vary significantly across seemingly
similar facilities. For example, if an abatement technology exhibits positive economies of scale, the
compliance cost of an average-sized facility will not equal the average cost of all the facilities
represented by that model plant This is because, with economies of scale, the higher cost to smaller
facilities will outweigh the lower cost to larger facilities relative to the average-sized facility. In this
case, the compliance costs of the facilities represented by the model plant will be under-estimated.
It is therefore important for the analyst to carefully consider the number of model plants needed to
capture the heterogeneity among constituent facilities that could affect compliance cost estimates.
Assumptions on variables and parameters determined outside of the model can be important
drivers in applied CGE analysis. For instance, estimates of elasticities that help define production
processes and agent preferences are of particular interest because model results are often sensitive
to these parameters. CGE-derived social cost estimates are particularly sensitive to parameters that
affect behavior in labor markets due to their pre-existing distortions, such as the assumed
elasticities governing the labor-leisure choice of consumers and production elasticities between
factors of production (Marten et al. 2 019).81 Model results tend to be more sensitive to behavioral
assumptions, for instance, the values chosen for elasticities relative to other data inputs such as the
benchmark input-output data (see Text Box 8.2) (Elliot et al. 2012). Additional model closures often
used in CGE modeling, like a fiscal budget closure, can also impact social costs and/or incidence
depending on the modeling context. It may be important to conduct additional sensitivity analysis
around these assumptions to understand the robustness of results if there is uncertainty in which
closure mechanisms should be chosen.82
80 Moreover, while many models are parameterized by a point estimate, Hertel et al. (2007) suggest that the
confidence in modeled results may depend on the precision of the parameter estimate. The authors note that
standard errors derived from the estimation framework for the parameter can also be used to guide sensitivity
analysis. Also, see Section 8.4.4.
81 Previous research has also illustrated these sensitivities in other contexts. For instance, Shoven and Whalley
(1984) observe that results from CGE analyses of the U.S. tax system are sensitive to labor supply, saving and
commodity-demand elasticity assumptions. Fox and Fullerton (1991) find that estimates of welfare changes
associated with tax reform are more sensitive to assumptions about the elasticity of substitution between labor
and capital than the actual level of detail about the U.S. tax system in the model.
82 Closure rules are exogenous assumptions made in the CGE model that characterize aspects of the economy
that are not explicitly modeled. For instance, many CGE models hold government consumption fixed by assigning
a fiscal closure rule that assigns how budget surpluses or deficits induced by a policy are recycled back to
households (e.g., lump sum, through the tax system). Goulder (2013) summarizes the economics literature on the
implications of alternative revenue recycling closures for climate policy analysis.
8-40 Guidelines for Preparing Economic Analyses | 3rd edition
-------
T.-f.-f Unveil*"nin
Clear communication of uncertainties is critical for transparency of the analysis. Uncertainty in
social cost estimates can arise from uncertainty regarding the baseline, affected universe of
facilities, policy responses, the number of affected markets and the cost of compliance activities.
The degree to which these and other factors affect the confidence placed in the social cost estimates
should be carefully reported and quantified when appropriate and possible.
While some key uncertainties have implications for both benefits and costs (e.g., for the baseline or
affected facilities), several are unique to social cost estimation. For instance, estimates of
compliance costs are often "study-level" estimates, used by engineers to judge the economic
feasibility of projects prior to engaging in a costly planning process, and are associated with an
error (e.g., +/- 30%).83 In some cases, more precise cost estimates, described by engineers as
"scoping" or "detailed" estimates, may be available. When compliance costs are used to
approximate or generate social cost estimates, qualitative and quantitative information available on
the degree of precision in the underlying estimates should be prominently discussed to provide
appropriate context
Uncertainty regarding the costs of compliance will propagate through to the estimate of social costs
when used in a partial or general equilibrium model. Estimates of social costs may also be subject to
model and parameter uncertainty. Model uncertainty refers to uncertainty in a model's ability to
accurately represent underlying processes relevant to understanding how an intervention affects
the system of interest (for example, due to simplifications necessary to tractably model complex
systems) (NRC 2009). As noted in Section 8.4.3, challenges in parameterizing models, including the
choice of functional form, also may be a prominent source of uncertainty. Conducting sensitivity
analysis or more sophisticated probabilistic analysis across a tractable range of identified
uncertainties can provide information on the robustness of the central social cost estimates.84
As noted in Section 8.2.3.4, technical change and learning can have an important effect on future
compliance costs. Estimates about the effect of innovation will be inherently uncertain and, in some
cases, may not be available. Even so, the expectation is that technological change and learning
generally leads to lower social costs over time compared to a scenario that assumes no innovation
occurs; uncertainty in this case is asymmetric, as innovation is unlikely to increase future costs.
Uncertainty may also affect social cost estimates when projecting the costs of regulations that are
implemented by local or state jurisdictions in the future. For example, in illustrative attainment
analyses conducted for some National Ambient Air Quality Standards (NAAQS), once all identified
control technologies have been applied, some areas of the country may still be modeled as out of
compliance with the air quality standard. In these cases, it is uncertain how attainment will be
achieved and at what cost Similarly, in the case of deregulatery actions, how state and local
jurisdictions respond for example by potentially enacting protections in place of the forgone
federal standards can affect the ultimate cost (and benefits) of relaxing the federal standard. In
these cases, sensitivity analysis is useful for understanding the robustness of social cost estimates
to alternate assumptions.
83 For example, EPA's Air Pollution Control Cost Manual (U.S. EPA 2018) notes that "costs and estimating
methodology in this Manual are directed toward the "study-level" estimate with a probable error of+/-30
percent."
84 See Chapter 5 for further discussion of uncertainty and sensitivity analyses.
8-41 Guidelines for Preparing Economic Analyses | 3rd edition
-------
* -i * . ential for Doublt ' li ving
Because a regulation may have multiple effects through the economy, the analyst should take
particular care to avoid double-counting costs. For example, counting both the increased costs of
production to firms resulting from a regulation and the attendant increases in prices paid by
consumers for affected goods would mean counting the same costs twice, leading to an
overestimate of social cost Also, when reporting private costs for certain groups, the portion of
those costs that reflect social costs versus transfers to other groups should be clearly identified in
the analysis.
Even in a general equilibrium analysis, analysts must take care in selecting an appropriate measure
of social cost Calculating social cost by adding together estimates of the costs in individual sectors
can lead to double counting. Instead, focusing on measures of changes in final demand, so that
intermediate goods are not counted, can avoid the double-counting problem.
When analysts rely on multiple models that take fundamentally different approaches to cost
estimation, care should be taken to separately report and characterize each model's output to avoid
double-counting. For example, if a technology-rich PE model is linked to a CGE model, the estimate
of social costs comes from the CGE model. The social cost is not the sum of the costs from the CGE
and PE models. Furthermore, the cost estimate from a compliance cost model, for example the
increased expenditures on compliance activities in the sector, should not be reported as the social
cost of the regulation without further elaborating what this cost estimate represents, why it
provides a reasonable estimate of the social cost, and that it is not equivalent to the actual social
cost of the rule.
8-42 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 8 References
33 U.S.C. 1251-1387: The Clean Water Act of 1972 (P.L 92-500), October 18,1972. Available at:
https://www.law.cornell.edu/uscode/text/33/1251 (accessed: November 11, 2024).
42 U.S.C. 7401: The Clean Air Act of 1963 (P.L. 95-95). Available at:
https://www.law.cornell.edu/uscode/text/42/7401 (Accessed: November 11,2024).
Alexeeva-Talebi, V., C. Bohringer, A. Loschel, and S. Voigt. 2012. The Value-added of Sectoral Disaggregation:
Implications on Competitive Consequences of Climate Change Policies. Energy Economics 34(S2): S127-S142.
Annicchiarico, B., S. Carattini, C. Fischer, and G. Heutel. 2021. Business Cycles and Environmental Policy: A
Primer. Environmental and Energy Policy and the Economy, Vol. 3. (Eds.: M. Kotchen, T. Deryugina, and J.
Stock). University of Chicago Press.
Armington, P. 1969. A Theory of Demand for Products Distinguished by Place of Production. Staff Papers-
International Monetary Fund, 16(1): 159-178.
Arrow, K., M. Cropper, G. Eads, R. Hahn, L. Lave, R. Noll, P. Portney, M. Russell, R. Schmalensee, V.K. Smith, and
R. Stavins. 1996. Benefit-Cost Analysis in Environmental, Health, and Safety Regulation: A Statement of
Principles. AEI Press.
Ballard, C. and D. Fullerton. 1992. Distortionary Taxes and the Provision of Public Goods. Journal of Economic
Perspectives, 6(3): 117-131.
Baumol, W. and W. Oates. 1988. The Theory of Environmental Policy. 2nd Ed. New York: Cambridge University
Press.
Bellas, A. and I. Lange 2011. Evidence of Innovation and Diffusion Under Tradable Permit Programs.
International Review of Environmental and Resource Economics, 5(1): 1-22.
Bento, A. and M. Jacobsen. 2007. Ricardian Rents, Environmental Policy and the 'Double-dividend' Hypothesis.
Journal of Environmental Economics and Management, 53(1): 17-31.
Bernstein, M. and J. Griffin. 2006. Regional Differences in the Price-Elasticity of Demand for Energy. NREL
Report. February. Available at: https://www.nrel.gov/docs/fy06osti/39512.pdf (accessed December 23.
20201
Boardman, A., D. Greenberg, A. Vining, and D. Weimer. 2011. Cost-Benefit Analysis: Concepts and Practice. 4th
Ed. Upper Saddle River, NJ: Prentice Hall.
Bohringer, C. and T. Rutherford. 2008. Combining Bottom-up and Top-down. Energy Economics, 30(2): 574-
596.
Brown, J., J. Hastings, E. Mansur, and S. Villas-Boas. 2008. Reformulating Competition? Gasoline Content
Regulation and Wholesale Gasoline Prices .Journal of Environmental Economics and Management, 55(1): 1-19.
Bullock, D. 1993. Welfare Implications of Equilibrium Supply and Demand Curves in an Open Economy.
American Journal of Agricultural Economics, 75(1): 52-58.
Burtraw, D. and K. Palmer. 2008. Compensation Rules for Climate Policy in the Electricity Sector .Journal of
Policy Analysis and Management, 27(4): 819-847.
Burtraw, D. and D. A. Evans, 2009. Tradable Rights to Emit Air Pollution. Australian Journal of Agricultural and
Resource Economics, 53: 59-84.
Busse, M. and N. Keohane. 2007. Market Effects of Environmental Regulation: Coal, Railroads, and the 1990
Clean Air Act. RAND Journal of Economics, 38(4): 1159-1179.
Cai, Y. and V. Arora. 2015. Disaggregating Electricity Generation Technologies in CGE Models: A Revised
Technology Bundle Approach with an Application to the U.S. Clean Power Plan. Applied Energy, 154: 543-555.
Carbone, J. and V. K. Smith. 2008. Evaluating Policy Interventions with General Equilibrium Externalities.
Journal of Public Economics, 92(5-6): 1254-1274.
8-43 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Caron, J. 2 012. Estimating Carbon Leakage and the Efficiency of Border Adjustments in General Equilibrium
Does Sectoral Aggregation Matter? Energy Economics, 34(S2]: S111-S126.
Canova, F. 1995. Sensitivity Analysis and Model Evaluation in Simulated Dynamic General Equilibrium
Economies. International Economic Review, 36(2): 477-501.
Chakravorty, U., C. Nauges, and A. Thomas. 2008. Clean air regulation and heterogeneity in US gasoline
prices. Journal of Environmental Economics and Management, 55(1): 106-122.
Chan, H., B. Chupp, M. Cropper, and N. Muller. 2018. The impact of trading on the costs and benefits of the
Acid Rain Program. Journal of Environmental Economics and Management, 88:180-209.
Dean, T., R. Brown, and V. Stango. 2000. Environmental Regulation as a Barrier to the Formation of Small
Manufacturing Establishments: A Longitudinal Examination. Journal of Environmental Economics and
Management, 40: 56-75.
Dwyer, L., P. Forsyth, and R. Spurr. 2006. Assessing the Economic Impacts of Events: A Computable General
Equilibrium Approach." Journal of Travel Research, 45: 59-66.
Elliott, J., M. Franklin, I. Foster, T. Munson, and M. Loudermilk. 2012. Propagation of Data Error and
Parametric Sensitivity in Computable General Equilibrium Models. Computational Economics, 39: 219-241.
Executive Order 12866: Regulatory Planning and Review, Section 1(a), October 4,1993. Available at:
https://www.epa.gov/laws-regulations/summary-executive-order-12866-regulatory-planning-and-review
(accessed December 23,2020).
Farrow, S., and A. Rose. 2018. Welfare Analysis: Bridging the Partial and General Equilibrium Divide for Policy
Analysis. Journal of Benefit Cost Analysis, 9(1): 67-83.
Ferioli, F., K. Schoots, and B. van der Zwaan. 2009. Use and limitations of learning curves for energy
technology policy: A component-learning hypothesis. Energy Policy, 37(7): 2525-2535.
Ferris, A., R. Shadbegian, and A. Wolverton. 2014. The Effect of Environmental Regulation on Power Sector
Employment: Phase I of the Title IV S02 Trading Program. Journal of the Association of Resource and
Environmental Economists, 1(4): 521-553.
Ferris, A. and A. McGartland. 2014. A Research Agenda for Improving the Treatment of Employment Impacts
in Regulatory Impact Analysis. Does Regulation Kill Jobs?, ed. C. Conglianese, A. Finkel, and C. Carrigan.
University of Pennsylvania Press.
Fischer, C. and G. Heutel. 2013. Environmental Macroeconomics: Environmental Policy, Business Cycles, and
Directed Technical Change. Annual Review of Resource Economics, 5(1): 197-210.
Fischer, C., I. Parry, and W. Pizer. 2003. Instrument Choice for Environmental Protection When Technological
Innovation is Endogenous. Journal of Environmental Economics and Management, 45: 523-545.
Fowlie, M. 2010. Emissions Trading, Electricity Restructuring, and Investment in Pollution Abatement.
American Economic Review, 100(3): 837-69.
Fowlie, M., Reguant, M. and Ryan, S.P., 2016. Market-Based Emissions Regulation and Industry Dynamics.
Journal of Political Economy, 124(l):249-302.
Fox, T., and D. Fullerton. 1991. The Irrelevance of Detail in a Computable General Equilibrium Model.
Economics Letters, 36: 67-70.
Frey, E. 2013. Technology Diffusion and Environmental Regulation: The Adoption of Scrubbers by Coal-Fired
Power Plants. Energy Journal, 34(1): 177-205.
Gillingham, K., and D. Rapson. 2016. The Rebound Effect and Energy Efficiency Policy. Review of
Environmental Economics and Policy, 10(1): 68-88.
Goulder, L., I. Parry, and D. Burtraw. 1997. Revenue-Raising versus Other Approaches to Environmental
Protection: The Critical Significance of Preexisting Tax Distortions. RAND Journal of Economics, 28(4): 708-
731.
8-44 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Goulder, L. 2000. Environmental Policy Making in a Second-Best Setting. Economics of the Environment:
Selected Readings. 4th Ed, ed. R. Stavins. New York: W.W. Norton.
Goulder, L. 2013. Climate change policy's interactions with the tax system. Energy Economics, 40: S3-S11.
Gray, W. and R. Shadbegian. 2024. When Do Firms Shift Production across States to Avoid Environmental
Regulation? Land Economics, 100(3): 443-457.
Grubb, M., Drummond, P., Poncia, A., McDowall, W., Popp, D., Samadi, S., Penasco, C., Gillingham, K.T.,
Smulders, S., Glachant, M. and Hassall, G., 2021. Induced Innovation in Energy Technologies and Systems: A
Review of Evidence and Potential Implications for C02 Mitigation. Environmental Research Letters 16(4):
043007.
Gruenspecht, H. 1982. Differentiated Regulation: The Case of Auto Emission Standards. American Economic
Review, 72: 328-331.
Hahn, R. and J. Hird. 1991. The Costs and Benefits of Regulation: Review and Synthesis. Yale Journal on
Regulation, 8: 233-278.
Harberger, A. 1964. The Measurement of Waste. American Economic Review, 54(3): 58-76.
Hazilla, M. and R. Kopp. 1990. Social Costs of Environmental Quality Regulations: A General Equilibrium
Analysis .Journal of Political Economy, 98(4): 853-873.
Hertel, T., D. Hummels, M. Ivanic, and R. Keeney. 2007. How Confident Can We Be of CGE-based Assessments
of Free Trade Agreements? Economic Modeling, 24: 611-635.
Horowitz, K. and M. Planting. 2009. Concepts and Methods of the Input-Output Accounts. Bureau of Economic
Analysis, U.S. Department of Commerce.
Horowitz, J., and K. McConnell. 2002. A Review of WTA/WTP Studies. Journal of Environmental Economics and
Management, 44: 426-447.
IEc. 2005. An Overview of Major Economic Modeling Paradigms and Their Potential Application in OSWER
Analyses. Final Report prepared for the U.S. EPA Office of Solid Waste and Emergency Response, Economics,
Methods and Risk Analysis Division.
Jacobsen, M. and A. Bentham. 2015. Vehicle Scrappage and Gasoline Policy. American Economic Review,
105(3): 1312-1338.
Jaffe, A., R. Newell, and R. Stavins. 2002. Technological Change and the Environment. Environmental and
Resource Economics, 22: 41-69.
Jorgenson, D. and P. Wilcoxen. 1990. Environmental Regulation and U.S. Economic Growth. RAND Journal of
Economics, 21(2): 314-340.
Jorgenson, D., H. Jin, D. Slesnick, and P. Wilcoxen. 2013. "An Econometric Approach to General Equilibrium
Modeling," in Handbook of Computable General Equilibrium Modeling, ed. P. Dixon and D. Jorgenson. The
Netherlands: Elsevier.
Just, R., D. Hueth, and A. Schmitz. 2005. Welfare Economics of Public Policy: A Practical Approach to Project and
Policy Evaluation. Northampton, MA: Edward Elgar Publishing.
Kellogg, R. and M. Reguant. 2021. Energy and Environmental Markets, Industrial Organization, and
Regulation, in Handbook of Industrial Organization, 5(1): 615-742.
Kiuila, 0. and T. Rutherford. 2013. The Cost of Reducing C02 Emissions: Integrating Abatement Technologies
into Economic Modeling. Ecological Economics, 87: 62-71.
Klemick, H., E. Kopits, A. Wolverton, and K. Sargent 2015. Heavy-duty Trucking and the Energy Efficiency
Paradox: Evidence from Focus Groups and Interviews. Transportation Research Part A: Policy and Practice, 77:
154-166
Klier, T. and J. Linn. 2016. The Effect of Vehicle Fuel Economy Standards on Technology Adoption. Journal of
Public Economics, 133(C): 41-63.
8-45 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Kokoski, M. and V.K. Smith. 1987. A General Equilibrium Analysis of Partial Equilibrium Welfare Measures:
The Case of Climate Change. American Economic Review, 77(3): 331-341.
Lahiri, S. and Y. Ono, 2007. Relative Emission Standard versus Tax under Oligopoly: The Role of Free Entry.
Journal of Economics, 91:107-128.
Lanz, B. and S. Rausch. 2011. General Equilibrium, Electricity Generation Technologies and the Cost of Carbon
Abatement: A Structural Sensitivity Analysis. Energy Economics, 33(5): 1035-1047.
Mankiw, N.G. 1998. Principles of microeconomics (Vol. 1). Elsevier.
Marten, A., R. Garbaccio, and A. Wolverton (2019). Exploring the General Equilibrium Costs of Sector-Specific
Environmental Regulations. Journal of the Association of Environmental and Resource Economists, 6(6): 1065-
1104.
Marten, A. and S. Newbold. 2017. Economy-wide Effects of Mortality Risk Reductions from Environmental
Policies. NCEE Working Paper #2017-03. National Center for Environmental Economics. U.S. EPA.
McFarland, J., J. Reilly, and H. Herzog. 2004. Representing Energy Technologies in Top-down Economic
Models using Bottom-up Information. Energy Economics, 26(4): 685-707.
Miller, R. and P. Blair. 2009. Input-Output Analysis: Foundations and Extensions. 2nd Ed. Cambridge University
Press.
Millimet, D., R. Santanu, and A. Sengupta. 2009. Environmental Regulations and Economic Activity: Influence
on Market Structure. Annual Review of Resource Economics, 1: 99-118.
Murray, B., A. Keeler, and W. Thurman. 2005. Tax Interaction Effects, Environmental Regulation, and "Rule of
Thumb" Adjustments to Social Cost Environmental and Resource Economics, 30(1): 73-92.
National Research Council (NRC). 2009. Science and Decisions: Advancing Risk Assessment. Washington, DC:
National Academies Press.
Nelson, R., T. Tietenberg, and M. Donihue. 1993. Differential Environmental Regulation: Effects on Electric
Utility Capital Turnover and Emissions. Review of Economics and Statistics, 77(2): 368-373.
Organisation for Economic Cooperation and Development (OECD). 2004. Environment and Employment: An
Assessment. Working Party on National Environmental Policy. ENV/EPOC/WPNEP(2003)11/FINAL.
OMB. 2023. Circular A-4, Regulatory Analysis, November 9, 2023. Available
at:https://www.whitehouse.gov/wp-content/uploads/2023/ll/CircularA-4.pdf (accessed June 13, 2024).
Paltsev, S. and P. Capros. 2013. Cost Concepts for Climate Change Mitigation. Climate Change Economics, 4
(SI): 1340003.
Paltsev, S., J. Reilly, H. Jacoby, and J. Morris. 2009. Appendix B: Measuring the Cost of Climate Policy of The
Cost of Climate Policy in the United States. Report 173. MIT Joint Program on the Science and Policy of Global
Change.
Parry, I. 2003. Fiscal Interactions and the Case for Carbon Taxes over Grandfathered Carbon Permits. Oxford
Review of Economic Policy, 19(3): 385-399.
Parry, I. 2005. Fiscal Interactions and the Costs of Controlling Pollution from Electricity. RAND Journal of
Economics, 36(4): 849-869
Parry, I. and A. Bento. 2000. Tax Deductions, Environmental Policy, and the "Double Dividend" Hypothesis.
Journal of Environmental Economics and Management, 39(1): 67-96.
Phaneuf, D. and T. Requate. 2017. A Course in Environmental Economics: Theory, Policy, and Practice.
Cambridge University Press.
Pizer, W. and R. Kopp. 2005. "Calculating the Costs of Environmental Regulation," in Handbook of
Environmental Economics, ed. K.G. Maler and J.R. Vincent, Volume 3. Amsterdam: North-Holland.
8-46 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Portney, P. 1981. The Macroeconomic Impacts of Federal Environmental Regulation. Natural Resources
Journal 21(3): 459-488.
Rausch, S. and V. Karplus. 2014. Markets versus Regulation: The Efficiency and Distributional Impacts of U.S.
Climate Policy Proposals. Energy Journal, 35(S1): 199-227.
Requate, T. 1997. Green Taxes in Oligopoly if the Number of Firms is Endogenous. Finanzarchiv/Public
Finance Analysis: 261-280.
Requate, T. 2006. "Environmental Policy under Imperfect Competition," in The International Yearbook of
Environmental and Resource Economics 2006/2007. A Survey of Current Issues, ed. Tietenberg, T. and Folmer,
H. Cheltenham, UK: Edward Elgar Publishing. 120-207.
Rivera, S. 2003. Key Methods for Quantifying the Effects of Trade Liberalization. International Economic
Review January/February: 1-8.
Rogerson, R. 2015. A Macroeconomic Perspective on Evaluating Environmental Regulations. Review of
Environmental Economics and Policy, 9(2): 219-238.
Ryan, S. 2012. The Costs of Environmental Regulation in a Concentrated Industry. Econometrica, 80(3): 1019-
1061.
Sagar, A. and B. van der Zwaan. 2006. Technological Innovation in the Energy Sector: R&D, Deployment, and
Learning-By-Doing. Energy Policy, 34: 2601-2608.
Schafer, A. and H. Jacoby. 2005. Technology Detail in a Multi-sector CGE Model: Transport under Climate
Policy. Energy Economics, 27(1): 1-24.
Schmidt, S. and V. Wieland. 2013. "The New Keynesian Approach to Dynamic General Equilibrium Modeling:
Models, Methods and Macroeconomic Policy Evaluation," in The Handbook of CGE Modeling, 1.
Shoven, J., and J. Whalley. 1984. Applied General-Equilibrium Models of Taxation and International Trade: An
Introduction and Survey .Journal of Economic Literature, 22(3): 1007-1051.
Smith, V.K., H. Sieg, H.S. Banzhaf, and R. Walsh. 2004. General Equilibrium Benefits for Environmental
Improvements: Projected Ozone Reductions under EPA's Prospective Analysis for the Los Angeles Air Basin.
Journal of Environmental Economics and Management, 47(3): 559-584.
Sue Wing, I. 2006. Representing Induced Technological Change in Models for Climate Policy Analysis. Energy
Economics, 28(5-6): 539-562.
Sue Wing, I. 2011. "Computable General Equilibrium Models for the Analysis of Economy-Environment
Interactions," in Research Tools in Natural Resource and Environmental Economics, ed. A. Batabyal and P.
Nijkamp. World Scientific.
Tietenberg, T. 2002. Environmental and Natural Resource Economics. 6th d. New York, NY: Harper Collins
Publishers.
West, G. 1995. Comparison of Input-Output, Input-Output + Econometric and Computable General
Equilibrium Impact Models at the Regional Level. Economic Systems Research, 7(2): 209-227.
Williams, R, and M. Hafstead. 2018. Unemployment and Environmental Regulation in General Equilibrium.
Journal of Public Economics, 160: 50-65.
Wolverton, A., A. Ferris, and N. Simon. 2019 Retrospective Evaluation of the Costs of Complying with Light-
Duty Vehicle Surface Coating Requirements. Journal of Benefit-Cost Analysis, 10(1): 39-64.
Woodford, M. 2009. Convergence in Macroeconomics: Elements of the New Synthesis. American Economic
Journal: Macroeconomics, 1(1): 267-269.
U.S. Chamber of Commerce and NERA Economic Consulting. 2013. Impacts of Regulations on Employment:
Examining EPA's Oft-Repeated Claims that Regulations Create Jobs. Available at:
https://www.uschamber.eom/sites/default/files/documents/files/020360 ETRA Briefing NERA Study final
.pdf (accessed December 6, 2024)
8-47 Guidelines for Preparing Economic Analyses | 3rd edition
-------
U.S. EPA. 1997. The Benefits and Costs of the Clean Air Act: 1970-1990, Prepared by the Office of Air and
Radiation and the Office of Policy, Planning and Evaluation. EPA/410/R-97/002.
U.S. EPA. 2007. Benefits and Costs of Clean Air Act Direct Costs and Uncertainty Analysis. EPA-COUNCIL-
07-002.
U.S. EPA. 2009. Guidance on the Development, Evaluation, and Application of Environmental Models.
EPA/100/K-09/003.
U.S. EPA. 2011. Benefits and Costs of the Clean Air Act 1990-2020, the Second Prospective Study. Available at:
https://www.epa.gov/clean-air-act-overview/benefits-and-costs-clean-air-act-1990-2020-second-
prospective-study (accessed December 6, 2024)
U.S EPA. 2016b. Cost Reduction through Learning in Manufacturing Industries and in the Manufacture of Mobile
Sources. EPA-420-R-16-018.
U.S. EPA. 2017. SAB Advice on the Use of Economy-Wide Models in Evaluating the Social Costs, Benefits, and
Economic Impacts of Air Regulations. September 29. Available at:
https://yosemite.epa.gOv/sab/SABPRODUCT.NSF/0/4B3BAF6C9 EA6F503852581AA0057D565/$File/EPA-
SAB-17-012.pdf (accessed December 6,2024).
U.S. EPA. 2018. Air Pollution Cost Control Manual. Available at: https: //www.epa.gov/economic-and-cost-
analysis-air-pollution-regulations/cost-reports-and-guidance-air-pollution#cost%20manual (accessed
December 6,2024).
U.S. EPA. 2020. Handbook on Valuing Changes in Time Use Induced by Regulatory Requirements and Other
EPA Actions. EPA-236-B-15-001.
van der Zwaan, B. and A. Rabl. 2004. The Learning Potential of Photovoltaics: Implications for Energy Policy.
Energy Policy, 32:1545-1554.
Yeh, S. and E. Rubin. A Review of Uncertainties in Technology Experience Curves. Energy Economics, 34(3):
762-771.
8-48 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 9 - Economic Impacts
A wide variety of economic impacts can occur as a consequence of
environmental policy. Analysis of who will experience gains and who will be
burdened by a regulation, and analysis of the nature and magnitude of
regulatory impacts, provides important information for decision makers,
stakeholders and the broader public. An economic impact analysis (EIA]1
identifies and quantifies a wide range of regulatory impacts including market-
based impacts such as changes in employment, prices, profitability and plant
closures; as well as impacts outside the marketplace (e.g., impacts on state
and local governments]. An EIA identifies specific groups that may benefit or
be burdened by a policy and assesses the impacts they experience. Affected
groups may include consumers, industries, small businesses, workers,
communities, tribes and governments. Using this definition of an EIA, this
chapter discusses issues relevant to estimating the economic impacts of EPA
policies. An EIA can be tailored to improve understanding of specific
regulatory impacts. However, in some instances, EPA has been directed to
conduct an EIA, as explained in Section 9.2 of this chapter. Subsequent
sections begin with a review of frameworks that provide a general
understanding of economic impacts, followed by guidance for assessing each
impact category.
This chapter primarily focuses on market impacts due to compliance costs.
However, Section 9.5.6 is a discussion of the impacts of benefits (changes in
environmental quality and public health] and several other sections, such as
Section 9.5.1.5, briefly discuss specific beneficial impacts. Impacts on
governments and non-profits are discussed in Section 9.5.4; and a
consideration of economy-wide impacts from both costs and benefits is
discussed in Section 9.5.5. Chapter 10, "Environmental Justice and Life Stage
Considerations," complements the current chapter by discussing how
regulation might change the distribution of environmental quality and health
risks across minority and low-income populations, and by life stage.
1 At the EPA, an EIA differs from a Regulatory Impact Analysis (RIA). The latter Is frequently used
interchangeably with "economic analysis" and may contain analyses of benefits, costs and economic impacts; in
other words, an EIA is often contained within an RIA. For more information, see Chapter 1.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
9.1 Background
Analyzing economic impacts sheds light on the distribution across groups of costs, transfers,2
benefits and other economic outcomes induced by regulation. An EIA may include a broad range of
measures including monetized metrics such as profit or price changes, as well as non-monetized
metrics such as changes in employment or the likelihood of plant closures. The crux of an EIA is
understanding these changes experienced by specific groups. In contrast, a BCA focuses on
measuring aggregate social net benefits and is concerned with economic efficiency which requires
that benefits outweigh costs, irrespective of to whom net benefits accrue. Thus, the two types of
analyses use different measures. Unlike aggregate benefit and cost measures calculated for a BCA,
the impact measures included in an EIA need not be mutually exclusive. For example, an impact
that appears simultaneously in two related markets, such as costs in the regulated sector and
revenues in the pollution control sector, can be included and appear as two impacts in an EIA. In
BCA, where the focus is on aggregate efficiency, transfers which, by definition, shift money from one
group to another will not impact estimates of net benefits. However, because transfers affect who
experiences gains or burdens from a policy, they may be key within an EIA (OMB 2023).
Despite these important differences, analyses of economic impacts in an EIA and of social benefits
and social costs in a BCA are complementary, as both shed light on the consequences of regulation.
When conducted for the same policy, both types of analyses should use a consistent baseline and
set of assumptions. Generally, both analyses have similar scopes; that is, if it is appropriate for the
analysis of social costs to extend to markets beyond the regulated industry then it would likely be
appropriate for the EIA as well. Both analyses should explain underlying assumptions, explore the
sensitivity of results to assumptions and inputs, strive for transparency and include documentation
and references (OMB 2023).
Whether regulatory consequences are measured in terms of economic impacts, changes in social
welfare or both, ultimately the focus is on how people are affected. An EIA that analyzes
profitability, for example, is studying potential impacts on the income of firm owners or
shareholders. Analysis of employment impacts sheds light on impacts on workers. An EIA that
estimates changes in prices is concerned about impacts on consumers. To complicate matters, many
impacts estimated in an EIA give insight into changes that might affect multiple groups. For
example, an increased likelihood of plant closure affects both firm owners and workers.
9.2 Statutes and Policies
Multiple statutes and policies contain directives for an EIA that are applicable across media.3 The
following statutes and executive orders (EOs), described more fully in Chapter 2, directly address
economic impacts.
2 Transfers are shifts of money or resources from one part of the economy to another such as tax payments. See
Section 8.2.2.2 for a discussion of compliance costs and transfers. Circular A-4 defines a transfer as "... a shift in
money (or other item of value] from one party to another. More generally, when a regulation generates a gain
for one group and an equal-dollar-value loss for another group, the regulation is said to cause a transfer from
the latter group to the former." (OMB 2023)
3 The EPA's Action Development Process (ADP') Library is a resource for analysts who wish to access relevant
statutes, EOs or Agency policy and guidance documents. Besides the broadly applicable statutes and directives
3-2 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Regulatory Flexibility Act (RFA), as amended by the Small Business Regulatory Enforcement
Fairness Act (SBREFA) (1996);
Unfunded Mandates Reform Act (UMRA) (1995);
EO 12866, Regulatory Planning and Review (1993) as amended by Executive Order
14094, Modernizing Regulatory Review (2023);
EO 13132 (1999), "Federalism;"
EO 13175 (2000), "Consultation and Coordination with Indian Tribal Governments;"
EO 13211 (2001), "Actions Concerning Regulations That Significantly Affect Energy Supply,
Distribution or Use."
Together with OMB's Circular A-4, these directives highlight features of affected entities that may be
relevant for EIAs. Table 9.1 lists the features identified by these directives and offers examples of
potentially affected groups.
Table 9.1 - Features of Potential Relevance to Economic Impact Analyses
as Identified by Statutes, Executive Orders and Guidance
Feature
Statute, Order or
Directive
Examples of Potentially Affected Economic Groups
Sector
UMRA; EO 12866; EO
13132; EO 13175; OMB
Circular A-4
Producers; industries; state, county, local, territorial, or
tribal governments.
Entity size
RFA/SBREFA; UMRA; EO
12866, OMB Circular A-4
Businesses, governmental jurisdictions, not-for-profit
organizations. Analyze small entities separately.
Time; Dynamics
OMB Circular A-4
Groups (e.g., consumers, workers, producers, firms,
industries) experiencing transitional or long-run impacts.
Geography
UMRA; EO 12866; OMB
Circular A-4
Regions, states, counties,
non-attainment areas, local or regional markets.
Energy
EO 13211
Energy sector (i.e., developers, distributors, generators, or
users of energy resources).
9.3 Connections between Economic Impacts and
Frameworks of Distributional Effects
Virtually any economic measure of the consequences of a regulation may be included in an EIA.4 To
accommodate this degree of flexibility, an EIA is not constrained or governed by an operating
framework. However, there are several conceptual frameworks in the economics literature that
discussed in this section, there are also environmental statutes with specific applicability that require
consideration of impacts on certain populations (such as impacts on labor; see Section 9.5.1.4), or that may
require analysis of impacts for facilities potentially eligible for regulatory variances.
4 For textbook discussions of the meaning and usefulness of impact analysis, see Field and Field (2005) and
Tietenberg (2006).
9-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
provide insight into the meaning and interpretation of impact categories. Section 9.3.1 provides a
summary.
While a worthwhile analytic objective for environmental policy is to estimate the net welfare
changes experienced by each affected group in an economy, the EPA does not currently conduct
such analyses. The information needed to distribute shares of regulatory costs, benefits and
transfers among groups and estimate each group's net welfare change is not available. This is
explained In Section 9.3.2.
, , i : iceptual Frameworks
A deeper understanding of economic impacts can be achieved by drawing connections to
conceptual frameworks of distributional effects. These frameworks, often presented in terms of
welfare effects, are useful for understanding parts of an EIA because they illustrate the different
pathways through which regulatory costs are distributed across population groups.5 Expenditures
are incurred by regulated entities to comply with environmental mandates, standards, permit
requirements, taxes and so on. Compliance expenditures may be passed on partially or fully to
other groups.6 For example, costs may be experienced by firm owners or shareholders through
lower profits, or passed on to consumers through higher prices. Or, costs may be passed on to
workers through changes in labor compensation, and/or on to the owners of other factors of
production through reduced rates of return to land and capital.7 The portion of the cost
experienced by these different groups depends on a variety of factors including the time-frame
under consideration, the characteristics of the regulated market such as the elasticity of demand
relative to the elasticity of supply and whether there are barriers that prevent new firms or imports
from entering the market Some costs may trickle through to related markets. While in practice
economists cannot always measure the extent of cost pass-through, existing frameworks help shed
light on the variety of ways that costs percolate through the economy.8
A framework developed by Harberger (1962) to better understand the distributional effects
(incidence) of taxation provides insight into who bears the costs of environmental regulation.
Effects are separated into two broad categories: those falling on the sources of income including
owners of firms, labor, capital or land; and those falling on the uses of income, or consumption, due
to changing prices. Harberger's simple two sector, two good model representing a perfectly
competitive closed economy with perfectly mobile factors of production suggests that a tax on one
input could lead to either, or both, source-side and use-side effects. Adapting the model to
5 If the regulated entity is not a profit-maximizing firm, then the principles discussed in this section are likely not
relevant. We address impacts on governments and non-profits in Section 9.5.4.
6 For a more detailed discussion, see Tietenberg (2002, 2006), which is the basis for the discussion in this
paragraph. Useful textbook discussions are also provided by Kolstad (2000) and Field and Field (2005). For a
review of the empirical literature, see Bento (2013). For a discussion specifically of the effects of command and
control regulations, see Fullerton and Heutel (2010).
7 Throughout this chapter, all factors of production are represented by either land (natural resources), labor
(human resources) or capital (man-made resources).
8 The following sources provide frameworks for understanding distributional impacts of environmental
regulation: Christiansen and Tietenberg 1985; Baumol and Oates 1988; Field and Field 2005; Tietenberg 1992,
2002,2006; Serret and Johnstone 2006; Kristrom 2006; Fullerton 2009,2011; Robinson et al. 2016; Fullerton
and Heutel 2010; Fullerton and Muehlegger 2019.
3-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
represent an environmental tax shows a use-side burden on purchasers of the commodity in the
taxed sector; and a source-side impact on factors affected by the tax (Fullerton and Muehlegger
2019). Many other existing frameworks also categorize distributional effects according to the route
through which the effect is transmitted (product prices, profits, shifts in factor compensation)
which is then traced to the group on which the effect falls (consumers; owners of firms, land or
capital; workers).
Figure 9.1 illustrates how Robinson et al. (2016) conceptualize one set of pathways through which
regulatory compliance costs may eventually be distributed across population groups. These
pathways help contextualize metrics that often appear in an EIA. The groups experiencing
economic impacts as described in Section 9.5 (producers, workers, other factors of production,
consumers, communities and the overall economy) are related to one or more of the three routes
through which regulatory compliance costs flow.9 The groups themselves, however, do not always
align perfectly with the three groups identified in the figure (consumers, employees and owners).
For example, the figure does not directly represent "producers," yet impacts on producers are
commonly analyzed at the EPA, and directives to consider some producer impacts are given by
statute or EO.10 Impacts on producers will ultimately be felt by all the people who together make up
affected firms (owners and shareholders, workers and other owners of productive factors).11 Other
impact categories discussed in Section 9.5, such as impacts on labor or employees, are more directly
represented by Figure 9.1.12 The right-hand box conceptualizes how costs might be experienced
across different population groups; for example, among regions or among households with
different demographic characteristics. This is a common endpoint for an EIA as explained in the
sections below on specific impact categories for example, Section 9.5.2 explains how price
increases might be experienced differently by high- versus low-income consumer groups.
Fullerton (2016) offers a more nuanced framework for disaggregating regulatory consequences. He
identifies the following potential cost-related effects on the regulated market:13 (1) an increased
cost of production results in an increase in the price of the regulated good affecting people who
purchase the good; (2) decreased production reduces revenues and changes relative returns to
9 Government and non-profit organizations are also discussed In Section 9.5, but they are structured differently
than private firms and are not well represented by Figure 9.1.
10 For example, RFA/SBREFA and EO 13211 (2001) direct agencies to consider impacts on small firms, and on
energy producers, respectively.
11 Through impacts on producers, regulatory costs could also affect upstream suppliers of inputs (e.g., coal) by
leading them to lower their prices, thinking that if they do not, the regulated facilities (e.g., power plants) could
shut down.
12 Some "changes" in Figure 9.1 may be measured as economic impacts, welfare changes or possibly both. For
more context on Figure 9.1's "changes in employee income and employment,"see Text Box 9.1 on labor impacts
and benefit cost analysis.
13 Some of these effects may be negligible or may not occur at all. Fullerton (2016) also identifies channels
through which distributional effects can occur on the benefits side. For example, asset prices can be affected by
environmental quality improvements (e.g., improvements could be capitalized into land and housing prices (and
some households could be dislocated due to higher rents). See Sections 9.5.1.5 and 9.5.3; and Chapter 10 for more
discussion.
3-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
workers and firm owners and factors of production; (3) restrictions on pollution create scarcity
rents14 for owners of firms, capital, and/or land; (4) transitional impacts occur as the economy
adjusts to a new equilibrium, for instance, if workers must search for new jobs; and (5) gains and
losses are capitalized into asset prices such as corporate stock prices rising due to an expected
future flow of scarcity rents.15
Figure 9.1 - Example Framework to Map Distribution of Compliance Costs
(Robinson et al. 2016)16
A few key insights for EIA can be gleaned from these frameworks:
Differentiating between impacts that occur in the short- and long-run is important.
The short-run refers to the period in which only some factors of production are variable
(e.g., labor) while others are fixed (e.g., capital equipment), and consumers are constrained
by existing household assets, commitments, and information. In policy contexts, the short-
run is sometimes referred to as a transition period. The long-run refers to the period in
which all factors of production are variable, the aforementioned consumer constraints are
relaxed, and the economy returns to equilibrium (i.e., all prices and quantities have fully
14 Scarcity rents represent a measure of welfare: "This producer's surplus which persists in long-run competitive
equilibrium is called scarcity rent." (Tietenberg 2006). For a discussion of scarcity rents created by
environmental regulations through pollution restrictions and captured by firms in the form of higher profits, see
Fullerton and Metcalf(2001). See Buchanan and Tullock (1975) for a discussion of the potential for scarcity
rents under a quota or a cap-and-trade policy where permits are distributed for free.
15 For an interesting example, see Fullerton (2011) where this framework is applied to a specific environmental
policy (a carbon permit system) by linking measurable outcomes to welfare changes.
16 Reproduced with author permission.
9-6 Guidelines for Preparing Economic Analyses | 3rd edition
-------
adjusted to the new regulation). There are likely to be different implications for the
economic impacts of a policy in the short-run compared to the long-run. For example, in the
long-run, consumers are better equipped to switch to substitute goods, and firms are better
equipped to switch to producing different outputs and to make entry and exit decisions.
These time frames also have different implications for workers (see Section 9.5.1.4).
The distribution of impacts among market participants depends on the nature of the
affected market(s). Market characteristics including the extent of competition and the
elasticity of demand relative to the elasticity of supply determine the allocation of impacts
among consumers, labor and owners of firms, capital, and other resources. All things equal,
competitive markets pass regulatory costs through to consumers to a greater extent than
markets in which firms have monopoly power. Firms in very competitive markets do not
earn excess profit and have no choice but to pass on costs if they want to stay in business. Of
course, the reduced quantity demanded at higher prices may force them to close. Firms with
market power have incentive to absorb a portion of regulatory costs since raising the price
they charge reduces the quantity consumers demand of their products and reduces
profits.17 Relative elasticities are also important. In an imperfectly competitive market, the
portion of the cost borne by producers increases with a greater elasticity of demand relative
to elasticity of supply (and the portion borne by consumers increases with a greater
elasticity of supply relative to elasticity of demand).18
Impacts may differ within market participant categories. Substantial heterogeneity of a
regulation's impacts is often experienced within groups. In practice, firms and their
circumstances are not identical, so compliance may be more burdensome for some firms
than for others.19 For example, small firms may have fewer units of production over which
to spread compliance costs, or some firms may have technologies that are more expensive
to adapt to regulatory requirements. Similarly, consumers and their circumstances are not
identical. People purchase varying bundles of goods and therefore will not be uniformly
affected by price changes. Also, the same incremental change in consumption will affect
individuals differently depending on their baseline levels of consumption; those with higher
levels will value a small change in consumption less (referred to as the diminishing
marginal utility of consumption). Industries, factors of production and other market
participant categories can be affected differently as well. In Section 9.5, we discuss the
conditions associated with divergent impacts for each impact category.
This section has discussed frameworks that shed light on the potential distribution of compliance
costs. Several papers also consider the distribution of health benefits or environmental quality (e.g.,
Fullerton 2016; Robinson et al. 2016; Pearce 2006). For example, Robinson etal. trace the effects of
hazard reduction on changes in human risks and the valuation of those changes. See Section 10.2.1
in Chapter 10 for a discussion of this literature.
17 For a discussion of economic impacts on a representative firm and on the market, in a supply and demand
model with perfect competition and under monopoly, see Tietenberg (2006), pp. 510-516.
18 See Fullerton and Metcalf (2002).
19 Heterogeneity in impacts may also be the result of regulatory design (e.g., differentiation of standards by
facility vintage). This possibility is discussed in Section 9.5.1.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
ft.-..,2 Disaggregated W Ii"-i - 10"i"- > I*
An analysis that disaggregates welfare effects (social costs and benefits) and transfers across
relevant groups is a worthwhile goal. The analyst could estimate the net welfare changes
experienced by each affected group in an economy, which in principle, if all regulatory
consequences by group were fully described, might obviate the need for an EIA. In practice,
however, many obstacles prevent a complete distributional analysis of welfare effects. For instance,
at the EPA it is typical that the social costs and benefits of an environmental regulation are
estimated for different groups. The former is usually estimated for firms that must comply with
regulatory requirements, but the ultimate incidence of those compliance costs among owners,
workers, and consumers (as costs are passed through to profits and prices, for example) is not
typically estimated. Social benefits are estimated for individuals experiencing changes in
environmental risks or conditions. Sparse information regarding the overlap between the groups
bearing the costs and experiencing the benefits makes calculating disaggregated net welfare effects
particularly challenging.
A different possibility to achieve disaggregated welfare effects would be converting the economic
impacts included in an EIA experienced by different groups into welfare changes and summing
across effects. Unfortunately, current models and data prevent such a detailed exercise. Consider
business closures, for example. They might decrease profits to owners and upstream firms and
cause workers to become unemployed. One would need to have information about effects on
upstream firms (i.e., those that would be affected and by how much) as well as information on
affected workers (e.g., the forgone wages of unemployed workers, the length of time they remain
unemployed and their wages once they are successfully re-employed). Such detailed information is
typically not available. Text Box 9.1 discusses the inherent difficulties of estimating social welfare
effects associated with employment impacts.
Finally, to estimate group specific social benefits, analysts would need group-differentiated
estimates of willingness to pay for the variety of environmental quality changes caused by EPA
rules. While the existing literature contains evidence of variability in willingness to pay for public
environmental goods among income (and other) groups, it does not contain a full suite of such
estimates20; and the use of any specific estimate would be controversial without significant public
review.
20 For discussion and examples, see Banzhafet. al (2019) and Chapter 10 on willingness-to-pay in the
environmental justice literature; Banzhafand Walsh (2008) for empirical evidence of household sorting in
response to toxic air emissions; and Ito and Zhang (2020) for evidence of variable WTP for clean air in China.
9-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 9.1 - Labor Impacts and Benefit Cost Analysis
In a benefit-cost analysis, some portion of changes in employment may also affect social welfare,
but there are many theoretical and practical challenges to accounting for them. One challenge is
how to estimate transition costs to workers experiencing involuntary job loss and unemployment
Including all resulting earnings losses would overstate social costs if they are transfers of
economic rents - for example, if displaced workers were highly paid relative to their productivity
(Hall 2011).
In addition to earnings losses, workers may incur transition costs due to relocation across labor
markets, health impacts or other impacts on well-being that are not well-measured (Smith 2015;
Kuminoff, Schoellman, and Timmins 2015). Transition costs may be higher during a recessionary
period, when overall labor demand is already reduced due to nationwide declines in production,
which can lengthen the time needed to locate new employment (Bartik 2015). These costs may be
higher for certain categories of workers such as those whose skills are specially adapted for the
sector experiencing reduced labor demand. For example, effects may differ by workers' age. For
involuntary job loss, older workers with more human capital may face larger earnings losses for
fewer years of remaining labor force time in their careers than otherwise similar workers who
are young. Older workers experiencing involuntary job loss may have access to more resources
from lifetime earnings, private insurance or access to social programs. Otherwise, similar younger
workers may face larger costs because capital market imperfections prevent borrowing against
their future lifetime earnings.
Likewise, quantifying changes in health or welfare due to an environmental regulation that affects
workers, for example by improving their productivity or their ability to work, is challenging. An
emerging literature documents these benefits; for reviews see Aguilar-Gomez etal. (2022) and
Graff Zivin and Neidell (2013, 2018). These are just some of the issues to consider regarding
potential welfare effects of labor impacts. Economists do not yet have a unified theory that
incorporates employment impacts measured as social welfare effects into benefit cost analysis.
For discussions, see Hall (2011), Ferris and McGartland (2014), and Smith (2015) who conclude
that more work is needed in this area.
With caution, we also mention an analytic construct for further considering net welfare by detailed
groups. A Social Welfare Function (SWF) establishes criteria under which efficiency and equity
outcomes are transformed into a single metric, making them directly comparable. To do this, SWFs
make assumptions regarding how society places different values on incremental changes in
measures of well-being across individuals or groups (see Adler 2012, 2019, for a discussion). OMB
(2023) outlines an option to implement a SWF in which individual- or group-specific WTP
estimates are weighted differently. The weights assign lower values to incremental increases in
consumption accruing to individuals with higher baseline consumption relative to people with
lower baseline consumption (to account for diminishing marginal utility of consumption).21
Implementation of this approach requires estimates of costs and benefits for each individual or
each income group conditional on their baseline income and cannot rely on estimates of the
average WTP across the whole population. Such average estimates are common in analyses of
environmental regulations - EPA's estimate of the value of statistical life is an example. OMB's
optional approach reflects one possible SWF; however, given its subjective nature, there is no clear
21 Please see OMB (2023) Section lO.e.fora detailed explanation.
9-9 Guidelines for Preparing Economic Analyses | 3rd edition
-------
consensus in the literature regarding how to value different distributions of welfare improvements.
For these reasons, SWFs are not currently recommended when conducting regulatory analysis at
the EPA.
Despite an inability to estimate the net welfare effects experienced by different groups affected by
regulation, estimates of economic impacts improve understanding of the pathways through which
welfare changes can occur, e.g., through business closures, or by restructuring markets or by
increasing housing values in a community. Impact measures may also be useful for identifying
individuals who might be strongly affected for example the firms likely to close; whereas net
welfare changes among groups might average out such strong effects so that their severity is
overlooked. In addition, certain impact categories are examined to respond to statutory and
executive order directives. Instead of focusing directly on welfare effects, this chapter provides
information for qualitatively and quantitatively assessing changes in a wide variety of economic
impacts that are expected to have an effect on welfare.
a ! -n.-jytr .\[i fmp,v~r
An EI A should develop a profile of baseline conditions among groups expected to experience
important effects of the rule. These are the conditions occuring in the absence of the rule or policy
over the period of analysis. For example, the profile could include the number of regulated firms,
their average size, and their average profitability. These metrics would be estimated for the year
the rule takes effect and for the remaining timeframe of analysis. An EIA may also include two
additional components: a preliminary analysis to screen for the magnitude of incremental impacts
and an in-depth examination of expected important impacts. For each component of an EIA,
analysts should highlight key analytic limitations and uncertainties. This section discusses the
baseline profile, the preliminary analysis, and the in-depth examination, and identifies potentially
useful data sources.
eline Profile
An EIA should develop a baseline profile that describes the industries, consumers, workers, or
other groups that are expected to experience important incremental effects of a regulation.22 The
profiles will overlap with baseline profiles developed for other components of a regulatory
analyses, such as the cost analysis.
The effects of some regulations may extend beyond participants in directly regulated markets,
affecting, for instance, upstream or downstream markets, or complementary or substitute product
markets. Often the markets involved in pollution control activities are affected. We will refer to the
latter as the environmental protection sector and note that it may overlap with upstream markets.
The following information can contribute to an industry profile:
22 For more about how to define and describe baselines, see Chapter 5. For more about developing a baseline for
governments or non-profit organizations, see Section 9.5.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
The affected North American Industrial Classification System (NAICS) industry codes
(NAICS is the standard used by federal statistical agencies in classifying business
establishments);23
Industry summary statistics, including total employment, revenue, costs, number of
establishments, number of firms, size of firms, and race and gender profile of firm owners
and workers;
Baseline industry structure, including competitive structure, market concentration and
degree of vertical integration within the industry;
Characteristics of supply and demand (e.g., relative elasticities);
Industry trends including growth rates, expected changes in technology and financial
conditions;
Openness to and reliance on international trade;
Pre-existing environmental and other regulations and associated compliance behavior;
Barriers to entry; and
Diversity of production technologies among firms.
The baseline socioeconomic characteristics of groups expected to experience consequential
economic gains or burdens due to a regulation are also important and may include consumers,
workers, business owners, shareholders, renters, community members and others. Attributes to
consider include:
Income and poverty levels;
Age distribution;
Employment status;
Community characteristics such as unemployment rate;
Geographic location and mobility; and
Pollution burdens.
The potential relevance of these market conditions and socioeconomic characteristics within the
context of a specific impact category is discussed in Section 9.5.
P[-Hmihit lalysis
During the early stages of regulatory analysis, a preliminary analysis to explore the potential for
important impacts can be useful and may be as simple as systematically thinking through the
expected impacts of a regulation and qualitatively describing them. When data are sparse, it may
still be possible to roughly estimate some regulatory impacts. For example, to screen for significant
impacts on small businesses, analysts can compare a rule's estimated annualized costs per
regulated facility to estimated annual revenues of affected small facilities to determine whether the
ratio of regulatory costs to facility revenues violates established thresholds.24
While the EPA has established thresholds that suggest when impacts on small entities are
significant, in most cases the criteria for when an impact warrants additional analyses are not well
defined and may depend on the condition of the economy. For example, during an economic
23 For more information on classifying industries by NAICS codes, see
h tips:/'/www, census, aov/eos/www/n a ics/.
24 See Final Guidance for EPA Rulewriters: Regulatory Flexibility Act as amended by the Small Business
Regulatory Enforcement Fairness Act (U.S. EPA 2006a).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
recession, impacts on workers may be a concern. Or, the timing of regulatory impacts may be
relevant, including the period of anticipation of an upcoming compliance date, and whether effects
are expected to grow or diminish or affect different groups over time. The context or location
within which economic impacts are experienced is important For example, reduced demand for
labor in a small town with declining job opportunities might have bigger labor market impacts than
in a larger city with abundant work opportunities. Or, when a trade exposed industry is the subject
of regulation, there may be concerns regarding potential loss of domestic market share. Finally, if
analysts suspect important impacts beyond directly regulated industries, the scope of analysis can
be broadened, even if data and tools permit only qualitative assessments.
cIiitD -f;fh' E'v."[mn."iion
Analysts may conduct an in-depth examination of the impact categories identified as likely to be
important by the preliminary analysis. Substantial adverse impacts deserve special attention. If
possible, a partial equilibrium analysis of affected markets can yield greater insights into impacts
relative to an engineering cost analysis alone.25 For example, with information on demand and
supply elasticities in affected markets, analysts can move to a more refined analysis that examines
the pathways through which costs would travel (e.g., consumer prices versus producer profits and
input prices including wages). With regional- and firm-specific demand and supply information,
analysts might also be able to shed light on how impacts vary across regions and firms. It may also
be possible to link together several sector-specific partial equilibrium models with a multi-market
model to examine linked impacts on regulated and related markets. If appropriate, a general
equilibrium model can offer insights into impacts on a broad spectrum of markets and groups
across the economy (see Section 9.5.5).
9.4,4 Data
Analysts may have access to proprietary data or detailed plant-level data (which may be
confidential business information) collected through the rulemaking process that can be leveraged
in an economic impact analysis. However, often data must be sought elsewhere. Table 9.2 describes
available data sources that might be useful for analyzing economic impacts. The right-hand column
gives examples of groups or impact categories under analysis for which each data source might be
useful. Note that quantitative estimates of some economic impacts may not be possible because of
inadequate household-, firm- or community-specific data (including elasticity estimates). Data that
are available are often aggregated to the sector, or jurisdiction, level.
25 For a discussion of partial equilibrium and other market and engineering models, see Chapter 8 on Analyzing
Costs.
3-12 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Table 9.2 - Examples of Available Data Sources for Analyzing Economic
Impacts of EPA Regulations
Source
Examples of types of data
Examples of relevant
groups/impact categories
U.S. Bureau of Labor Statistics
Consumer Expenditure Survey:
https: // www.bls.sov/cex/
Expenditures, income and
demographic characteristics of U.S.
consumers.
Consumers, Communities.
U.S. Bureau of Labor Statistics -
Current Employment Statistics:
https://vvww.bls.sov/ces/
Establishment-level estimates of
nonfarm employment, hours and
earnings by industry.
Sectors or Industries,
Producers, Labor,
Communities.
U.S. Bureau of Labor Statistics -
Current Population Survey:
https://www.bls.eov/cps/
Household level data on employment,
unemployment, persons not in the
labor force, hours of work, earnings
and other characteristics.
Labor Communities.
U.S. Bureau of Labor Statistics -
Producer Price Index:
https://www.bls.eov/Dni/
Index of producer output prices, by
detailed industry.
Sectors or Industries,
Producers.
U.S. Census Bureau - Longitudinal
Employer-Household Dynamics:
https: / /lehd.ces.census.gov/
Statistics on employment, earnings
and job flows at detailed levels of
geography and industry for different
demographic groups.
Sectors or Industries, Labor,
Producers, Government
entities.
Published research specific to an
industry or sector.
Demand and supply elasticities,
regional supply and demand
information, and other specific
estimates of interest.
Sectors or Industries,
Consumers, Producers.
University of Wisconsin -
Wisconsin National Data
Consortium:
htto://windc. wisc.edu/
Open-source datasets for economic
analysis, for U.S. states and counties,
with state, sector and region
economic activity.
Sectors or Industries,
Consumers, Producers,
Government entities.
U.S. Census States & Local Areas:
https: //data, census. eov/all?s=01
0XX00US$04G0000
Demographic and socioeconomic
information.
Consumers, Government
entities.
U.S. Census State and County
Quickfacts:
https: / / www. ce n s u s. so v / q u i c k fa
cts / fact/table /US/PST0452 21
Demographic and socioeconomic
information.
Consumers, Government
entities.
U.S. Census Bureau - American
Housing Survey:
https: / /www.census.sov / prosra
ms-survevs/ahs.html
Data on the housing and construction
industry, homeownership, and
characteristics of homes.
Housing and Construction
Industry, Consumers,
Government entities,
Communities.
U.S. Department of Housing and
Urban Development Aggregated
USPS Administrative Data on
Address Vacancies:
https: / / www.huduser.gov/portal
/ datasets / usps.html
Occupancy status.
Communities, Government
entities.
9-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Source
Examples of types of data
Examples of relevant
groups/impact categories
U.S. Census Bureau - American
Community Survey:
https; //www.census.sov/proara
ms-survevs/acs
Detailed population and housing
information, by community
Sectors or Industries, Labor,
Producers, Consumers,
Government entities,
Communities.
Trade Publications and
Associations
Market and technological trends,
sales, location and ownership
changes.
Sectors or Industries.
U.S. Census Statistics of U.S.
Businesses:
https: //www.census.eov/proera
ms-survevs/susb.html
National and subnational economic
activity by enterprise size and
establishment industry.
Producers, Small businesses,
Non-profits, Government
entities.
U.S. Bureau of Economic Analysis:
httDs://www.bea.eov/data
Economic statistics on U.S. production
(e.g., GDP], consumption, investment,
exports and imports, and income and
saving. National, Regional, Industry
and International economic accounts
Sectors or Industries,
Producers, Labor, Consumers,
Government entities,
Communities, International
competitiveness.
U.S. Census Bureau - Annual
Survey of Manufacturers:
https: / /www.census.sov/proara
ms-survevs/asm.html
Statistics for manufacturing
establishments
Discontinued after 2021, transitioned
to the Annual Integrated Economic
Survey:
https://www.census.gov/programs-
surveys/aies.html
Manufacturing sector,
Producers.
U.S. Census Bureau - Economic
Census:
https: //www.census.gov/progra
ms-survevs /economic-
census.html
Sector-level sales, value of shipments,
number of employees and
establishments, value added, cost of
materials, capital expenditures,
household and community
characteristics
Sectors or Industries,
Producers, Consumers,
Communities.
U.S. Department of Commerce
Industry & Trade Outlook
Periodically published book -
most recently in 2000
Industry, trends, international
competitiveness and regulatory
events.
Sectors or Industries.
New York University. Margins by
Sector:
http: / /pages.stern.nvu.edu/~ada
modar/New Home Paee/datafile
/margin, html
Profit margins: gross income and net
income based.
Sectors or Industries,
Producers, Businesses.
Internal Revenue Service.
Statistics of Income Bulletin
httDs://www.irs.gov/pub/irs-
soi/16winbul.pdf
Tax receipts, deductions and profits.
Sectors or Industries,
Producers, Businesses.
Dun & Bradstreet Information
Services: www.dnb.com
NAICS code, address, facility and
parent firm revenues and
employment.
Sectors or Industries,
Producers, Businesses.
9-14 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Source
Examples of types of data
Examples of relevant
groups/impact categories
Standard & Poors:
www.standardandpoors.com
Quarterly financial information for
publicly held firms, line-of-business
and geographic segment information
and Standard and Poor's (S&P)
ratings.
Sectors or Industries,
Producers, Businesses.
Value Line Industry Reports:
http://www.valueline.com/Stock
s/Industries.aspx
Industry overviews, company
descriptions and outlook, and
performance measures.
Sectors or Industries,
Producers, Businesses.
Securities and Exchange
Commission Filings and Forms:
https: / /www.sec.eov/edear.shtm
i
Income statement and balance sheet,
working capital, cost of capital,
employment, regulatory history,
foreign competition, lines of business,
ownership and subsidiaries, and
mergers and acquisitions.
Sectors or Industries,
Producers, Businesses.
U.S. Energy Information
Administration - Electricity Data:
httDs://www.eia.gov/electricitv/da
ta.php
Statistics on electric power plants,
capacity, generation, fuel
consumption, sales, prices and
customers.
Energy sector and subsectors
(e.g., oil, natural gas, coal,
nuclear energy sources),
Customers.
United States Utility Rate
Database (URDB)26
httDs://oDenei.org/wiki/Utilitv Rat
e Database
Rate structure information for electric
utilities in the United States. The
URDB includes rates for utilities based
on the authoritative list of U.S. utility
companies maintained by the U.S.
Department of Energy's Energy
Information Administration.
Energy sector and subsectors
(e.g., oil, natural gas, coal,
nuclear energy sources),
Customers.
U.S. Department of Commerce
Pollution Abatement Costs and
Expenditures Survey:
https://www.census.g0v/ec0n/0
verview/mul 100.html
Pollution abatement costs for
manufacturing facilities by industry,
state, and region. Data is limited to
annually from 1973 to 1994, with the
exclusion of 1987; and 1999 and
2005.
Sectors or Industries,
Producers, Businesses.
S&P, Moody's and Fitch state and
city bond ratings.
Financial strength indicator.
Government entities.
U.S. Department of Commerce
Census of Governments:
https://www.census.g0v/ec0n/0
verview/goOlOO.html
Revenue, expenditures debt,
employment, payroll, assets for
counties, cities, townships and school
districts.
Government entities.
United Nations, International
Trade Statistics Yearbook.
Foreign trade volumes for selected
commodities and major trading
partners.
Sectors or Industries,
Producers, Businesses.
26 Rates are posted annually by the National Renewable Energy Laboratory (NREL), under funding from, the U.S.
Department of Energy's Solar Energy Technologies Program, in partnership with Illinois State University's
Institute for Regulatory Policy Studies.
9-15 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Source
Examples of types of data
Examples of relevant
groups/impact categories
U.S. International Trade
Commission:
https: / / www.usitc.sov/research
and analvsis.htm
Investigative Reports.
Sectors or Industries,
Producers, International Trade
Global Trade Analysis Project:
https: //www.gtap.agecon.purdue.
edu/databases / defaultasp
Global data base describing bilateral
trade patterns, production,
consumption and intermediate use of
commodities and services.
Sectors or Industries,
Producers, International trade.
9.5 Impact Categories
This section provides guidance for assessing specific impact categories. Categories discussed are
not mutually exclusive; rather, they have a high likelihood of overlap. For example, impacts on
producers (employees and owners) likely overlap with impacts on the communities where they are
located. Impact categories discussed in this section are:
Producers and factors of production.
Consumers.
Communities.
Governments and non-profits.
Economy-wide.
Benefits of improved environmental quality or health.
The discussion that follows usually considers the impacts of new compliance activities. However, it
is also relevant to reductions in compliance activities which generally would produce impacts going
in the opposite direction.
9.5.1 Impacts on Producers and Factors of Production
Compliance activities typically increase production costs to regulated industries. This may affect
many different impact categories which are listed below and discussed in this section:
Production.
Profitability and plant closures.
Small businesses.
Labor.
Land and capital.
Related markets.
Energy sector.
Competitiveness.
Effects may vary by industry or firm characteristics, production technologies, pollution intensities,
policy design and more. There may be different effects in the long-run versus the short-run, and
according to whether one-time, ongoing, or transitional costs are being considered. Ongoing costs
are to maintain the newly achieved state of environmental quality. Transitional costs stem from
adjusting from one state of environmental quality to another (Baumol and Oates 1988).
9-16 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Consideration of the effect on small businesses is mandated by statute; consideration of impacts on
the energy sector is directed by executive order.27
If regulatory costs are small and/or distributed widely, there may be negligible impacts on
producers. However, even if the average impact across firms is small, some producers, such as
those facing the highest abatement costs, may be substantially affected. The following subsections
discuss how to assess impacts on producers and factors of production.
icts on Production
In response to substantial regulatory costs, the supply curve in the directly regulated market may
shift upward in the area near the market price which typically leads to higher prices and lower
output28 Reductions in industry output are usually driven by a mix of increased and lowered
operating rates at existing plants, closure of some plants and/or reduced future growth in
production relative to the baseline. This section discusses circumstances that influence changes in
output at the firm or facility (for firms that own more than one plant) level. Such changes can be
combined with industry characteristics such as the number and size or regional distribution of
firms to assess total changes in production.
At least two conditions can cause environmental regulation to have different impacts across firms,
and lead to changes in both the number and size of the average firm (Tietenberg 2006). The first is
significant heterogeneity in firm or facility cost structures; the second is regulatory requirements
that differ depending on firm characteristics.
Variability in cost structures can cause variation in the magnitude of regulatory costs and, while not
always the case, can lead to differences in the magnitude and direction of changes in output across
producers. For example, total industry output may decline or shift from the highest cost plants to
more efficient competitors. To better understand the extent of heterogeneity in how firms might
adjust production in response to regulatory requirements, a profile of baseline conditions is useful.
If available, detailed industry, firm or plant-level information may provide insights into how
production processes and baseline costs might vary across facilities and how this variation might
lead to different incremental costs of a regulation. For example, the ease with which facilities can
accommodate pollution control equipment may vary, or there could be variability in the ability to
substitute less hazardous chemicals for more toxic ones. Some firms may have to finance abatement
equipment and activities. For such firms, the cost and availability of financing can affect production
decisions.29 Ultimately, what analysts will need are the differences across firms in post-regulatory
costs. Firms may be able to maintain or even increase production levels if after absorbing
compliance costs, their production costs fall below the highest cost firms. Or they may decrease
27 See Chapter 2 and Section 9.2 which refers to theRFA as amended by the SBREFA, and to EO13211.
28 In the post-policy equilibrium, if the production costs of the marginal firm are not notably affected by the
regulation, then it is possible that the production and price effects can be de minimis even ifinframarginal firms
face notable compliance costs.
29 Analysts should carefully consider private market interest rates and other financing costs that firms might
face. A detailed consideration is presented in chapter 10 of the documentation for EPA's Integrated Planning
Model (IPM)for the power sector. Financing costs are represented as the weighted average cost of capital in
which firms finance projects with a combination of debt and equity. Merchant power providers are assumed to
face higher financing costs than utilities (U.S. EPA 2024a). See also Section 6.4 of these Guidelines on selecting
private discount rates.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
production if, after absorbing compliance costs, their production costs are among the highest in the
market.
The second cause of variable impacts across firms are variable regulatory requirements. Vintage-
based regulations that vary with the age of facilities may differentiate between existing and future
pollution sources, with future sources regulated more stringently. In other cases, firms in regions
with high baseline pollution may face stricter emission controls.30 In general, regulatory
requirements that vary by firm characteristics will shift economies of scale and can affect the
distribution of output among firms as well as firms' average level of output For example, firms may
respond to policies that differ across plant locations by relocating production to a less-regulated
area within the U.S. The greater the degree to which firms take advantage of this ability to shift
production across space to reduce compliance costs, the more likely it is that overall domestic
production does not change substantially. The outcome could be plant closure(s) and
accompanying plant opening(s) due to relocations, with distributional effects on affected areas.31
Shifts in production from domestic to foreign sources can also occur and are discussed in more
detail in Section 9.5.1.8.
. iii i cts on Profi »l iln »it I ant Closures
Regulatory costs can reduce profits and increase the possibility of plant closures. The industry
profile (see Section 9.4) describes baseline industry growth and financial conditions at regulated
firms. To assess changes in profits due to a regulation, analysts should compare the expected
change in market price to the change in production costs after accounting for compliance activities.
This increment should be multiplied by expected changes in output to estimate how profits change.
Industries and firms that are relatively profitable in the baseline will be better able to absorb any
new compliance costs that are not passed on to consumers. In cases where facilities have different
baseline pollution controls or different production technologies, those with lower costs after
meeting a new environmental standard will be better able to maintain profitability relative to other
firms and may increase their market share. These firms may even be able to increase profitability if
their costs of compliance increase by less than the increase in market price.
Discussing the likelihood of baseline closures improves understanding about the likelihood of
closures attributable to the regulation.32 Note that vertically or horizontally integrated facilities
might not be viable as stand-alone operations but may continue to operate based on their
contribution to the business line.
If pollution restrictions limit production of industry output, profitability may be affected. There
may be different profitability impacts for new versus existing firms. This may be the case, for
example, with vintage-differentiated regulation that imposes less rigorous pollution controls on
30 The firms experiencing less-stringent regulation might be more likely to see expanding market shares relative
to their counterparts, though some empirical evidence suggests this is not the case (Tietenberg 2006 citing
Pashigian 1984 and Pittman 1981; Greenstone 2002).
31 Shadbegian and Wolverton (2010) survey the plant location literature which suggests that firms reallocate
production (Gray and Shadbegian 2010), plant entry (List et al. 2003), or plant exit (Kahn 1997) in response to
environmental regulations.
32 For example, the EPA's documentation for its power sector model, IPM, includes detailed information on
power plants that have made public announcements of future closures, and this information can inform a
baseline analysis (U.S. EPA 2024a).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
existing relative to new firms.33 If market demand is increasing, new firms can enter but face higher
costs which negatively impact profitability. Existing firms can benefit through newly created
scarcity rents, with positive impacts on profitability. Over the long run, the likelihood of plant
closures may change if older plants with higher emissions are kept in operation for longer than was
expected in the baseline scenario.
Analysis of impacts on regulated firms' financial conditions involves the use of available financial
data. Impacts can be assessed by examining direct compliance costs as a percent of a firm's average
revenues, profits, or sales. An upper-bound assumption is that compliance costs are borne entirely
by the regulated industry (i.e., none are passed through to consumers). When data allow, assessing
the ratio of regulatory costs to profits is useful.34 Due to data limitations, analysts may only have
access to industry average revenues or sales. Calculating the ratio of full compliance costs to
average firm revenues gives some sense of the magnitude of compliance activities relative to
production activities without directly addressing the effect on profitability. When data on firm
profits are available, the ability of firms to pass costs through to prices should be considered.
Additional challenging issues affect ex-ante analysis of the effect of compliance spending on
profitability. First, economic models are simplified representations of complex economic systems.
They can be useful for estimating effects on groups but often are not reliable predictors of firm or
facility-level decisions.35 Second, common simplifying assumptions about firm decision-making
include perfect foresight, where agents know precise values for all economic variables in all future
years, and perfect information, where precise values drive decision-making so that a one-cent
difference between costs and revenues can be the difference between continued operation versus
closure.36 Such assumptions may perform well when describing aggregate behavior, but they often
run counter to the everyday complex and uncertain decision-making by managers, which is
remarkably difficult to model.37 There is typically little information regarding the economic
decision maker's expectations about the future (e.g., the firm's profitability, costs, revenues and
market conditions) and how those expectations respond to new conditions, such as a new
regulation. Indeed, many decisions are multi-faceted.
For example, management decisions about plant closures often result from the cumulative effect of
multiple factors, such as financial distress, unfavorable market conditions and aging equipment,
rather than any single factor such as a new environmental regulation. Finally, facility-specific,
rather than firm-specific, financial information is preferred for assessing profitability and
particularly for assessing the likelihood of plant closures. However, it is often difficult to find. For
instance, while financial data for publicly held companies is available, it is often too aggregated to
shed light on specific business practices or management decisions. For these reasons it is important
33 See Tietenberg (2006) Chapter 21 for more discussion and for references to literature finding evidence of a
new-source bias in environmental regulations.
34 Several sources in Table 9.2 provide information on industry profitability. See the table entries labeled, "New
York University, Margins by Sector," and "Internal Revenue Service, Statistics of Income."
35 Some models use "model plants" to represent specific plant or unit types and solve a linear programming
problem by choosing compliance strategies to minimize costs across the model plants (see Section 8.4.3).
36 This is referred to in the literature as the "penny-switching effect."See Krey and Riahi (2009).
37 The financial literature points to managers' individual characteristics and biases that can affect corporate
decision-making, e.g., risk aversion, confident or pessimistic approaches, misestimation of financial market data,
or loss aversion. For a brief survey of the literature on behavioral corporate finance, see Malmendier and Tate
(2015).
3-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
for analysts to describe the main limitations of the analysis when evaluating the incremental impact
of a regulation on firm profitability or the likelihood of plant closures.
ffmfvu' -n iii - sses
The RFA requires agencies to define small business according to the Small Business
Administration's (SBA) small business size standard regulations.38 As another option, the RFA
authorizes any agency to adopt an alternative definition of small business, "where appropriate to
the activities of the Agency," after consulting with the Chief Counsel for Advocacy of the SBA and
after opportunity for public comment If adopted, the agency must publish the alternative definition
in the Federal Register. The analytical tasks associated with complying with the RFA include a
screening analysis for "significant economic impacts on a substantial number of small entities"
(SISNOSE). The small businesses to be included in the analysis are those that are directly regulated;
that is, those that are subject to the rule's requirements. If a small business does not have an
obligation imposed directly by the regulation, then EPA guidance is that it should be excluded from
the analysis.
Care should be exercised when distributing regulatory costs experienced by small businesses over
multiple years. The annualization of compliance costs should rely on an estimate of the private
discount rate that reflects the cost of capital. In general, the private discount rate will reflect the
risk associated with the regulated entity in question. The cost of capital will also be affected by the
ability of affected firms to deduct debt from their tax liability.
Some small businesses may be liquidity constrained and find it challenging to spread costs over
multiple periods as they may face difficulty in raising external capital, including external debt. This
issue may differentially affect women-owned, minority-owned, rural small businesses and very
small businesses (firms with revenues less than $100,000 annually) (Federal Reserve Banks 2023,
2024, Morazzoni and Sy 2022, Fairlie et al. 2020, Cole 2020). For example, the Federal Reserve
Banks (2023) analysis finds that even though startups by people-of-color are just as likely to apply
for financing through financial institutions/lenders as are startups by White individuals, they are
less likely to receive the requested funds. Analysts should consider whether the costs faced by
liquidity-constrained small businesses are best modeled as being fully incurred during the year in
which they are borne.
In order to determine SISNOSE, the EPA conducts a screening analysis for both proposed and final
rules based on a percentage of sales as an economic impact for small businesses (a "sales test") (U.S.
EPA 2006a).39 While the analytic objective includes better understanding the effect of regulatory
costs on profitability, on the likelihood of plant closure or plant cutbacks, and so on, in practice
sparse data on profitability often limits an analysis to examining compliance costs as a percent of
average firm revenues or sales. As discussed in Section 9.5.1.2, ex-ante analysis of the effect of
compliance spending on profitability presents a difficult challenge.
"Small Entities" are defined by the RFA but "substantial number" is not specified. The EPA has
broad guidelines including example thresholds for determining SISNOSE certification, but generally
recommends three factors in determining "significant impact" and "substantial number":
38 See U.S. Small Business Administration (2022) for SBA's size standards.
39 See also Chapter 2. For a discussion of the screening analysis for small governments and small non-profits, see
Section 9.5.4.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
1. Magnitude of economic impact that may be experienced by regulated small entities;
2. Total number of regulated small entities that may experience the economic impact; and
3. Percentage of regulated small entities that may experience the economic impact
If the screening analysis reveals that a rule cannot be certified as having no SISNOSE, then the RFA
requires a regulatory flexibility analysis be conducted for the rule, which includes a description of
the economic impacts on small entities. Further analysis examining other types of impacts, as
discussed elsewhere in this chapter, in relation to small businesses, may provide additional
information for decision makers.40
:ts on Labor
Evaluation of employment impacts is required by many of the major environmental statutes.41
Impacts can vary according to baseline labor market conditions; employer and worker
characteristics such as industry, occupation, skill-level and region; and the type of workforce
adjustment or job transition. Employment impacts may occur in the regulated and environmental
protection sectors, in upstream or downstream sectors, or in sectors producing substitutes or
complements. As economic activity shifts in response to a regulation, typically there will be a mix of
declines and gains in employment in different parts of the economy over time. This section focuses
on labor demand42 and on employment impacts measured as changes in employment levels. An
employment impact analysis will describe both positive and negative changes in employment to
present a complete picture. For most situations, employment impacts are assessed as part of an EIA,
and should not be included in the formal BCA.43 See Text Box 9.1, above, for a discussion of social
costs and employment effects within BCA.
When the economy is at full employment as in long-run equilibrium, a regulation may reallocate
employment among economic activities rather than affect the general employment level, and in the
short-run may lead to transitional employment effects, such as workers involuntarily separated
from their jobs (Arrow et. al. 1996, Hafstead and Williams 2020).
Economic theory of labor demand indicates that employers affected by environmental regulation
may increase their demand for some types of labor, decrease demand for other types, or for still
other types, not change it at all. A variety of papers have provided frameworks for understanding
the employment impacts of regulation. Morgenstern et al. (2002) decompose the labor
consequences in a regulated industry facing increased abatement costs. They identify three
separate components. First, there is a demand effect caused by higher production costs raising
market prices. Higher prices reduce consumption (and production) reducing demand for labor
within the regulated industry. Second, there is a cost effect: as production costs increase, plants use
more of all inputs including labor to produce the same level of output. For example, pollution
abatement activities that require additional labor services to produce the same level of output.
40 See EPA Final Guidance for EPA Rulewriters: Regulatory Flexibility Act as amended by the Small Business
Regulatory Enforcement Fairness Act (U.S. EPA 2006a) for details on complying with the RFA.
41 Relevant statutes include the Clean Air Act, section 321(a); the Clean Water Act; section 507(e); the Toxic
Substances Control Act, section 24; the Solid Waste Disposal Act, section 7001(e); and the Comprehensive
Environmental Response, Compensation and Liability Act, (section 110(e).
42 See Section 9.5.6 and Chapter 7 Section 7.2. for examples of how environmental regulation may also affect
labor supply through changes in worker health and productivity (e.g., Graff Zivin and Neidell 2012,2013, 2018).
43 Except to the extent that labor costs are part of total costs in a BCA.
3-21 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Third, there is a factor-shift effect: post-regulation production technologies may be more or less
labor intensive (i.e., more/less labor is required per dollar of output). A different paper, Deschenes
(2014), describes environmental regulations as requiring additional capital equipment for pollution
abatement that does not increase productivity. This can be included in a labor demand model as an
increase in the rental rate of productive capital. These higher production costs induce regulated
firms to lower output and decrease labor demand (an output effect) as well as shift away from the
use of more expensive capital toward increased labor demand (a substitution effect).44 Berman and
Bui (2001) discuss how affected firms' overall labor demand could increase, decrease or remain
unaffected, depending, in part, on the labor-intensity of environmental protection activities needed
for regulatory compliance compared to the labor-intensity of producing output. To study labor
demand impacts empirically, a growing literature has compared employment levels at facilities
subject to an environmental regulation to employment levels at similar facilities not subject to that
environmental regulation; some studies find no employment effects, and others find significant
differences. For a review of recent empirical evidence, see Gray et al. (2023).
In practice, an EIA evaluates potential changes and shifts, positive and negative, in employment
levels by industry or other affected groups, and describes transitional employment effects for
affected groups of workers. While employment impacts are measured as changes in employment
levels by industry or affected group, workers affected by changes in labor demand due to regulation
may experience a variety of transitional effects including job gains or involuntary job loss and
unemployment (Smith 2015; Schmalensee and Stavins 2011; Congressional Budget Office 2011;
and OMB 2015). Transitional, or adjustment, costs may occur as workers shift out of current
employment and into other, potentially less desirable jobs (for example, jobs that are lower paying
or in a less desirable location); or into unemployment; or exit the labor force sooner than otherwise
(Walker 2013). Workers involuntarily displaced from declining industries or occupations, with long
job tenure, or living in areas where labor mobility is low or unemployment is high, may be
especially likely to face challenges in finding comparable re-employment (Baumol and Oates 1988).
If displaced workers' job search challenges are significant and keep them from employment, then
from a resource perspective, their labor is underutilized, similar to a stranded asset45 Involuntary
job loss can lead to significant earnings losses for displaced workers, and may involve periods of
unemployment as well as other impacts, such as negative health effects (Jacobson, LaLonde and
Sullivan 1993; Sullivan and von Wachter 2009a, 2009b).46 Text Box 9.2 discusses involuntary job
loss, unemployment impacts and health and wealth effects.
44 For an overview of the neoclassical theory of production and factor demand, see Chapter 9 ofLayard and
Walters (1978). For a discussion specific to labor demand, see chapter 4 of Borjas (1996). When using this
theoretic framework, authors have conceptualized regulation as an increase in the price of pollution (Greenstone
2002; Holland 2012), an increase in the price of capital (Deschenes 2014), an increase in energy prices
(Deschenes 2012), an increase in pollution abatement costs (Morgenstern et al. 2002), or with pollution
abatement requirements modeled as quasi-fixed factors of production (Berman and Bui 2001).
45 Worker displacement can ultimately affect communities' provision of public services. See Morris et al. (2019)
and Black et al. (2005) for examples of coal mining-reliant counties In Appalachia. See Section 9.5.3 for
discussion of impacts on communities.
46 Involuntary job loss refers to job displacement that results from employer decisions and that is unrelated to
worker performance, e.g. plant closings, mass layoff events and other firm-level employment reductions (Farber
2017; Sullivan and von Wachter 2009b; Chan and Stevens 2001).
3-22 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 9.2 - Unemployment Impacts, Health, and Wealth
Empirical studies indicate unemployment is associated with increased mortality risk for those in
their early and middle careers, but "whether unemployment is causally related to mortality
remains an open question ... and recent research has begun to focus on possible confounding,
mediating and moderating factors" (Roelfs et al. 2011, p. 2). The figure below shows the complex
relationships most related to environmental regulation between workforce status, regulation,
wealth and health. As line (3} indicates, a bi-directional relationship exists between
unemployment and health. Causality is difficult to identify for the unemployed population:
increased mortality risk may be caused by unemployment itself, independent of pre-existing
health status, or it may be caused by a decline in health resulting from a workforce status change
(e.g., job loss, unemployment). The first causal pathway is potentially informative for regulatory
analysis, but many studies lack detail to isolate it
Workforce Status
Employed
*> Unemployed
Wealth
1. Environmental regulation
protects human health.
2. Employment impacts (e.g.,
workforce adjustment],
3. Workforce status affects
health; health affects
workforce status.
4. Workforce status affects
wealth (e.g., unemployment
reduces wealth].
5. Wealth affects health.
Regulation
A nascent economic literature uses detailed worker data to explore the effect of plant closures or
mass layoff events on health. Sullivan and von Wachter (2009b) find increased mortality rates
among displaced male workers with long job tenure in Pennsylvania and, in a study of displaced
Austrian male workers, Kuhn et al. (2009) find that involuntary job loss negatively affected
mental health. A study of plant downsizing in Norway found that displaced workers were more
likely to utilize disability pensions than comparable workers in non-downsized plants (Rege et
al. 2009). In a meta-analysis of studies on unemployment and health, Picchio and Ubaldi (2023)
find on average a small negative effect of unemployment on health and, when the identification
strategy relies on exogenous unemployment shocks like plant closure, the effect becomes
smaller. Positive health impacts of moving from unemployment to a job may also exist (e.g.,
decreased depression) (van der Noordtetal. 2014).
The economics literature has found connections between wealth and health (indicated by line
(5)). Sullivan and von Wachter (2009a) find that higher variability of earnings is associated with
increased mortality. Dobkin et al. (2018) find that adverse health events measured by hospital
admissions can lead to reduced earnings and increased risk of bankruptcy for those without
health insurance.
The utility of these findings for regulatory analysis depends on whether involuntary job loss and
unemployment are expected impacts. The prospect of such an impact is shown by line (2). If
expected, the analysis may describe the likelihood of plant closures and employment impacts for
affected workers. But analysts should use caution transferring published empirical estimates on
adverse health impacts. Some studies use samples that may not correspond well to affected
9-23 Guidelines for Preparing Economic Analyses | 3rd edition
-------
workers in the policy scenario and some lack detailed data on key worker characteristics (e.g.,
"involuntariness" of job separation (Sullivan and von Wachter 2009b); if job loss was health-
related (Burgard etal. 2007).
While regulatory analyses may estimate employment impacts of regulations, it is challenging to
identify associated job displacement at the firm- or plant-level. Both Curtis (2018) and Hafstead
and Williams (2018) find workforce adjustments occur through reduced hiring rates rather
than increased job separations. Reduced hiring rates could still imply that workers spend more
time unemployed, though this may have a smaller impact than increased job separations. In a
survey of firms experiencing mass layoffs, government regulation is rarely a stated reason (U.S.
BLS 2011). More research is needed.
Workforce adjustments can be costly to firms as well as workers, so employers may choose to
adjust their workforce gradually over time through natural attrition (retirements, voluntary
separations) or reduced hiring, rather than incur costs associated with job separations (layoffs or
other firm-level employment reductions). Curtis (2018) estimates changes in industry employment
levels over time due to an environmental regulation and finds that changes occurred slowly
through reduced hiring rates, and not through increased job separations. Hafstead and Williams
(2018) find a similar result for the regulated sector, of employment levels decreasing through slow
hiring and natural attrition rather than increased separations, when modeling a carbon tax.
As a result of shifts in the demand for labor, environmental regulation might also induce wage
effects. However, firms generally avoid adjusting existing employees' wages downward (Walker
2013; Curtis 2018). Nominal wage rigidity has been attributed to many causes, not least is the
potential impact of lowering wages on employee morale (Howitt 2002). Another factor suggesting
very limited wage impacts in the specific context of environmental regulation, is that regulated
firms are often a fraction of employers in affected labor markets and thus are not influential enough
to affect industry wage rates (Berman and Bui, 2001).
The remainder of this section describes practical approaches to employment impacts analysis.
Estimating Labor Impacts: An employment impact analysis provides a baseline profile of
potentially affected employers and workers, labor market conditions and possibly potentially
affected communities. The analysis discusses or estimates potential changes or shifts in
employment due to a regulation. Both positive and negative employment changes should be
examined, including for example, possible employment impacts in the regulated sector as well as
the environmental protection sector. When feasible, analysts can describe direct changes expected
in the use of labor by the regulated sector for compliance requirements.47 In cases where impacts
are anticipated, and if data and modeling allow, analysts can describe employment impacts due to
changes in production, revenues or expenditures by the regulated sector and potentially also by
related sectors.
A baseline employment profile may include the size of the affected labor force, the degree to which
affected labor markets are concentrated among few employers, the amount of labor mobility, job
turnover, job search rates and the affected workers' regional or occupational unemployment
47 These labor costs (in dollars) are already included in the cost analysis of an RIA as they are costs to regulated
firms (see Chapter 8 for more information). They can also be described within an employment impact analysis,
and may be converted from dollar value labor costs to numbers of employees, or annual full-time equivalent
(FTE'), etc.
9-24 Guidelines for Preparing Economic Analyses | 3rd edition
-------
rates.48 Recent employment trends may be relevant Characteristics of affected workers, such as
sector, industry, occupation, earnings, experience and job skills, may be described. If employment
impacts are expected to be concentrated in certain communities, those communities could be
characterized. Table 9.2 lists examples of possible data sources that may be helpful in developing a
baseline employment profile.
To examine the incremental impacts of a regulation on employment, analysts should keep in mind
that labor demand may be affected differently in the short-run compared to the long-run. For
example, the RIA for the 2024 New Source Performance Standards for Greenhouse Gas Emissions
from Certain Units includes employment impact estimates for the power sector both for short-run
effects (e.g., construction-related employment needs) as well as long-run or recurring non-
construction employment due to shifts in the use of fuels in electricity generation (U.S. EPA 2024b).
For many regulations, assessing employment impacts will be limited to a qualitative discussion. It
will include the baseline profile described above, and the likely direction of change of employment
levels in affected sectors and occupations. A discussion of any concentrated employment impacts,
regionally or otherwise, would be useful. Information on the ability or limitations of workers to
respond to shifts in labor demand should be considered.
A quantitative analysis may project changes in employment in affected sectors by occupation or
among other groups of workers (e.g., by region). The quantitative estimates can use information
from the compliance cost analysis if the labor requirements for expected compliance activities are
provided. Examples of compliance activities include installation, operation and maintenance of
pollution control equipment; as well as monitoring, inspecting, reporting and recordkeeping. For
example, the RIA for the EPA's Safer Communities by Chemical Accident Prevention Final Rule
included estimates of changes in the number of labor hours required for compliance activities
among different occupations and for different sized facilities.49 Its analysis of labor impacts
examined how many total labor hours on average per year would be required for certain
provisions, and whether new workers would likely be hired. The analysis discussed which rule
provisions would likely require additional labor hours, the occupations of workers needed, and
whether the work was short- or longer-term.
In quantitative analyses, aggregated labor hours should be converted to estimates of annual
average job-years or full-time equivalents (FTEs).50 When these estimates are small relative to
average employment at a representative facility or firm, a reasonable assumption may be that
existing employees or contractors would take on the tasks for regulatory compliance rather than a
facility or firm adjusting the size of its workforce.
48 See, for example, Smith (2015) on local labor market conditions and unemployment, and Baumol and Oates
(1988) Chapter 15, on reemployment prospects and consideration of workers in communities characterized by
one or two large employers.
49 See U.S. EPA (2023a).
50 A job-year is not an individual job and is not necessarily a permanent or full-time job. Instead it is the work
performed by one FTE employee in one year. For example, 20 job-years may represent 20 full-time jobs or 40
half-time jobs in a given year, or any combination of full- and part-time workers such that the total is equivalent
to 20 FTE employees. In practice, for example, if the cost analysis for a regulation estimates a need for 1 million
labor hours per year in the regulated sector to conduct compliance activities, this could be converted to
approximately 480job-years by dividing 1 million by the annual work hours for a full-time employee, which
equals 2,080.
3-25 Guidelines for Preparing Economic Analyses | 3rd edition
-------
While transparent, the quantitative approach just outlined only addresses a subset of employment
impacts as it does not address shifts in labor demand associated with potential changes in output in
the regulated, or related, sectors. When a regulatory cost analysis project shifts in output due to
compliance costs or shifts in the composition of production within the regulated sector (e.g., shifts
in the electricity generation fuel-mix) a more detailed analysis may be possible. In these specific
cases, analysts can estimate employment impacts by multiplying the change in output by the
average amount of labor per unit of output (or per value of shipments) in the sector. This gives an
approximation of the output effect, a potentially important type of employment impact51 The U.S.
Census and U.S. Bureau of Labor Statistics (BLS) provide estimates of the units of labor associated
with expenditures (or value of output/sales) at the industry-level. A limitation of this type of
analysis is that in practice producer-level employment impacts will likely differ from aggregate,
industry-level employment impacts. For example, relatively more efficient firms may expand output
(and employment) to pick up the slack as less efficient producers contract (Jaffe et al. 1995;
Tietenberg 2002; and Christiansen and Tietenberg 1985).
Detailed industry information is useful to develop disaggregated employment estimates for related
sectors. For example, as part of estimating labor impacts in regulatory analyses of air pollution
regulations affecting the electric power sector, the EPA examined coal mining by region.52 The EPA
combined estimates of changes in coal demand with detailed estimates of coal supply and regional
coal mining productivity data available from the U.S. Energy Information Administration (U.S. EIA).
Labor productivity differed significantly across geographic regions, e.g., in 2018 labor productivity
in Virginia was 2.07 short tons of coal per labor hour, in Texas it was 6.73, and in Wyoming, it was
26.63 (U.S. EPA 2023b). This level of detail informed the analysis of employment impacts.
Approaches for estimating the employment impacts of environmental regulation are evolving.
Analysts are encouraged to engage the EPA's National Center for Environmental Economics early in
the process when developing a strategy for evaluating the employment impacts of a regulation.
Analysts should describe the methods used in a quantitative employment impacts analysis -
whether it analyzes changes in pollution abatement activities alone or combined with changes in
production - and explain analytical limitations, which might include:
Use of an estimation approach that produces partial employment impacts and does not fully
measure all potential changes in regulated and related sectors.
Application of average labor-to-cost or labor-to-output ratios instead of the change in labor
expected in response to incremental increases or decreases in costs or production.
Estimation of labor-to-cost, or labor to-output, ratios at the industry-level that reflect the
labor component of pre-regulation costs or production rather than post-regulation costs or
production. This is a limitation because such ratios can be influenced by the regulation.
Use of available labor ratio data that may be for industrial sectors not well-aligned with the
affected sectors.
Heterogeneity of firm- or facility-level responses to regulation, especially those of marginal
facilities operating at the tail end of productive efficiency, may be glossed over by labor
ratio data typically available at the sector level only.
51 Data on labor per unit of output would be a proxy for the overall effect on labor demand in the regulated
sector. These data are based on past production processes and therefore are not directly useful for measuring a
substitution effect between labor and other productive inputs when compliance activities are required in the
regulated industry.
52 U.S. EPA (2023b), "U.S. EPA Methodology for Power Sector-Specific Employment Analysis."
9-26 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Cautionary Notes: Analysts should proceed with caution regarding the following approaches
sometimes used to estimate quantitative employment impacts of regulation.
Transferring Certain Empirical Estimates: Morgenstern etal. (2002) estimated the effect of
pollution abatement expenditures on the quantity of labor in four highly polluting and regulated
industries. However, a later attempt to replicate and extend this research failed. Analysts should
not rely on the empirical estimates from Morgenstern et al. (2002). Likewise, analysts should not
rely on the estimates from Belova etal. (2013, 2015) as the authors "recommend that the EPA
refrain from using these results until the underlying cause(s) for the implausibly large estimates in
the employment effects found in Belova et al. (2013a) are uncovered and resolved."53 We highlight
Morgenstern et al. (2002) because of its prominence in the prior edition of the EPA's Guidelines for
Preparing Economic Analysis (2010). The theoretical model in Morgenstern etal. (2002) remains
valid.
Input-Output Analysis: As described in Section 8.3.4.1, input-output analysis can provide
employment impact estimates. This type of analysis is most suitable for analyzing detailed sectoral
impacts of regional, state, or local policies in the short term. In general, input-output models should
not be used for estimating impacts of national regulations because they do not allow prices,
production processes or technologies to adjust over time. As a result, they represent a very short-
term response to regulation and are better equipped to represent the response of a single region to
a small regulatory change which is not expected to affect prices.54 They are of limited use for
analyzing large regulatory changes or regulations that are national in scope.55
Plant Closures and Employment; Section 9.5.1.2 discusses difficulties in assessing the likelihood
of plant closures given a dearth of data and a limited ability to model key factors, such as
expectations of future profitability. Even in cases when estimates of the likelihood of plant closures
are available, estimating employment impacts from them can be difficult Employment impacts
associated with plant closures may differ from the projected decline in plant output. Firms face
labor adjustment costs, and, for example, multi-plant firms may choose to transfer workers,
potentially those more skilled and experienced, to other locations (Ferris, Shadbegian and
Wolverton 2014). Or, as noted above, production and employment may shift between firms, away
from higher cost plants towards more efficient competitors. Such heterogeneity implies that
employment impacts at the firm or plant-level can differ in direction from industry-level
employment impacts. Analysts should consider these possibilities.
its on Other Productive Factors: Land and Capital
In addition to labor impacts, environmental regulation can lead to changes in the demand for, and
value of, other factors of production employed by regulated firms. Economists label these other
factors of production as land (any natural resource), and capital (any man-made resource). In
general, environmental regulation is expected to have varying effects across factors, and tracing
53 Quote is from Belova et al. (2015). Note that Belova et al. (2013a) inside the quote is identical with Belova et
al. (2013), cited above.
54 Even for regional analyses, input-output models tend to overestimate impacts. "They typically include
exogenous multipliers that magnify direct effects on output and employment based on the assumption that all
new economic activity will recirculate within the regional economy. Input-output models tend to ignore
displacement of workers or resources that might occur outside the region under analysis" (U.S. EPA 2011).
55 The underlying data can be useful for identifying related sectors, e.g., upstream and downstream.
3-27 Guidelines for Preparing Economic Analyses | 3rd edition
-------
impacts back to specific factors is difficult (Fullerton 2009). Estimating changes in the quantities
demanded of broad categories of land and capital is more practical.56 There are two separate and
valid ways to represent the value of factors of production: earnings per period (also called rates of
return) or asset values. The latter is the discounted present value of the future stream of earnings
generated by the productive factor.
The relationship between changes in regulated firms' price and quantity of output, and changes in
their factor demands or factor returns, can be complicated. In response to stricter environmental
regulation, factors used intensively by the regulated industry might experience reduced demand
and/or returns. If a unit of capital is not perfectly mobile, or a type of natural resource is taken out
of production, it may lose value and impose a burden on the owners (Fullerton and Muehlegger
2019). For example, if a regulation induces firms to switch from high-carbon coal to lower emitting
natural gas, then the value of coal will decline, and the stock value of coal-intensive businesses
could decline as well. How fast an asset may return to production will affect the extent of burden. A
coal mine that closes may become valueless as the land may be quite difficult to switch to a different
use. It could even become a liability. Factors that are complements to pollution abatement might
experience an increase in demand or returns; while those that are highly mobile with similarly
valued alternative uses should hold their value. There are two general expectations for the long-run
response to environmental regulation. One is for land and capital to shift away from high-emission
activities toward lower emitting ones, including the environmental protection sector; another is for
land and capital to shift towards less regulated uses. Regionally differentiated impacts on capital
and land are possible when the stringency of pollution control varies by region.
To estimate how the costs of compliance are passed through to and distributed across productive
factors, analysts need the cross-price elasticities between these factors. When this is not available,
analysts can examine current production practices and the input biases of anticipated abatement
activities to inform a qualitative discussion of likely impacts on productive factors.
In general, income earned from ownership of land and capital (or of firms) tends to make up a
greater proportion of earnings for higher-income households. Thus, an increase in regulatory costs
passed through to households via lower returns to capital tend to be progressive, placing a greater
share of the burden on wealthier households.57 The magnitude of the impact on owners and
investors depends on the proportion of their portfolio affected by the change.
A different impact on factors of production stems from improved environmental quality which can
be capitalized into the price of nearby land, and buildings (including housing). The increase in
property or asset values accrues to the owners at the time of the improvement.58 The degree to
which the land and buildings are owner-occupied versus rented, and the degree to which the
increased value is passed on in the form of higher rents, will influence who experiences positive
versus negative impacts of the environmental improvement. If landlords increase rents to the point
of forcing out renters, then the renters may experience transitional impacts from relocation
activities. Identifying how owners and renters respond to improved environmental quality is a
complicated exercise and quantitative analysis is challenging. A qualitative discussion can be useful.
Related literature and modeling challenges are discussed in the final paragraphs of Section 10.2.1.
56 Land and capital may also be rented or supplied under contract. When not owned by the regulated firm, the
impacts are considered upstream, as discussed in Section 9.5.1.6.
57 For more details, see Rausch et al. (201 l)or Fullerton and Metcalf (2002).
58 If land improvements are concentrated and substantial, there could be community-wide effects. See Section
9.5.3 for a discussion of impacts on communities.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
f / fmfvu; i' it related Markets
An environmental regulation may affect markets other than those that are directly regulated.
Related markets may be positively affected, such as those in the environmental protection industry
or those producing substitutes; or negatively affected, such as those producing complements, or
those who are up- or downstream from the regulated industry (note that the environmental
protection sector may overlap with upstream markets). If the regulation causes a firm to use
different inputs or new technologies, then the producers of the new inputs will gain, while the
producers of the old ones will be burdened. Consumers in the related markets may experience
impacts as well (see Section 9.5.2). Downstream impacts may accrue to firms who purchase the
regulated firms' outputs. In general, when analyzing related markets, analysts should consider the
same potential impacts as for directly regulated markets.
If substantial impacts on related industries are expected, it will be useful to include firm sizes, profit
margins, growth rates and more, in a baseline profile of the related industries. For instance, when
the regulated sector sells an intermediate good or service (e.g., electricity), questions that might be
relevant include: What proportion of the purchasing firms are small or face narrow profit margins?
Are substitute inputs readily available? What proportion of the purchasing firms' spending goes to
the regulated firms?
Partial equilibrium models that represent significantly affected, related markets may be useful,
although sparse data and resources may limit their use. For regulations that are expected to
substantially affect many related markets, an economy-wide model as described in Section 9.5.5
might be considered, though the additional conditions described there should also be satisfied.
i cts on Ener/; I; i' i itribution or Use
EO 13211 (2001) directs agencies to prepare a Statement of Energy Effects for "significant energy
actions," which are defined as significant regulatory actions (under EO 12866) that also are "likely
to have a significant adverse effect on the supply, distribution or use of energy."59 OMB guidance
suggests that adverse effects could include any of the following:
Reductions in crude oil supply in excess of 10,000 barrels per day;
Reductions in fuel production in excess of 4,000 barrels per day;
Reductions in coal production in excess of 5 million tons per year;
Reductions in natural gas production in excess of 25 million mcf per year;
Reductions in electricity production in excess of 1 billion KWH per year or in excess of 500
MW of installed capacity;
Increases in energy use required by the regulatory action that exceed any of the thresholds
above;
Increases in the cost of energy production in excess of 1%;
Increases in the cost of energy distribution in excess of 1%; or
Other similarly adverse outcomes.
A regulatory action also may have adverse effects if it is likely to:
Adversely affect, in a material way, productivity, competition or prices in the energy sector;
59 See Section 2.1.6 and especially see OMB (2001).
3-23 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Adversely affect, in a material way, energy productivity, competition or prices within a
region;
Create a serious inconsistency or otherwise interfere with an action taken or planned by
another agency regarding energy; or
Raise novel legal or policy issues adversely affecting the supply, distribution or use of
energy arising out of legal mandates, the President's priorities or the principles set forth in
EOs 12866 (1993) and 13211 (2001).
For actions that may be significant under EO 12866 (1993), particularly for those that impose
requirements on the energy sector, analysts must be prepared to examine the energy effects listed
above.
its on Domestic and International Competitiveness
Competitiveness impacts are regulatory impacts that change the distribution of market power
among firms or sectors, either domestically or internationally. Unfair advantage may accrue to
producers that are free from regulatory constraints, or that face less expensive regulation. Or, high
fixed costs that are incurred to comply with environmental regulation may cause production to
become concentrated among fewer firms, enhancing their monopoly profits over the long run.
Regulatory constraints may differ among specific subsets of sectors or firms: existing versus new,
or small versus large.60 If some firms find it less costly to comply with a regulation, they may benefit
competitively at the expense of other regulated firms. Analysts may wish to consider the extent to
which production is shifted toward plants with higher-than-average productivity (Jaffe et al. 1995).
As with other impact categories, the extent to which a regulation leads to effects on competitiveness
depends on the interaction between the regulated firms' absorption of compliance costs and their
market structure.61 In general, greater compliance flexibility is expected to reduce competitiveness
effects.62
A first step to gauge the potential for competitiveness effects is the baseline profile of affected
industries. The profile should identify which domestic and international firms compete with
regulated entities, and their basic market structures. Do competitors face expensive environmental
regulation? Is the output produced by regulated firms differentiated from that of competitors,
potentially reducing impacts on competition? The literature suggests an increased likelihood of
competitiveness effects for industries in which compliance costs are high relative to total
production costs.63
Consideration of the impact of new environmental regulation in three key areas is particularly
germane to competitiveness effects. First, lack of access to debt or equity markets to finance market
entry, including regulatory costs, can represent significant barriers to entry.64 Over the long run,
this can change market structures and reduce competitiveness. Second, a regulation may have an
60 Section 9.5.1.2 discusses the impacts of differentiated regulation that occurs when existing firms are
regulated with greater leniency than new firms.
61 The importance of this interaction is discussed by Iraldo et al. (2011).
62 Evidence for this is presented by Iraldo et al. (2011) and Jaffe et al. (1995).
63 See Iraldo et al. (2011).
64 See the discussion about small business access to credit In Section 9.5.1.3.
3-30 Guidelines for Preparing Economic Analyses | 3rd edition
-------
impact on market concentration. A potentially useful measure of concentration is the Herfindahl-
Hirschman index (HHI), which is the sum of the squares of the market shares of each firm in a given
market. The U.S. Department of Justice uses the HHI to estimate changes in market concentration
due to mergers and acquisitions. Post-merger HHI values that are below 1,000 are considered
"unconcentrated," between 1,000 and 1,800 are regarded as moderately concentrated and above
1,800 are considered highly concentrated.65
Finally, the impact of regulation on the market position of domestic firms relative to their foreign
counterparts is important Domestic environmental regulations may have global economic
implications because the costs of domestic producers increases relative to foreign producers.66
Analyses of impacts on international competitiveness have been concentrated on the most
pollution- or energy-intensive and most trade-exposed industries because they are most likely to
face regulatory requirements and least able to pass compliance costs to consumers.67 For example,
in the context of unilateral climate policy, proposed legislation has focused on potential
competitiveness impacts on trade-exposed domestic energy firms.68 Quantifying these effects can
be complex and may require a multi-country computable general equilibrium model. There are
three classes of indicators of impacts on international competitiveness: the degree to which net
exports change, the degree to which production shifts overseas (i.e., pollution haven effect), and the
relative change in investment from domestic (regulated) producers to producers in other countries
(Jaffe etal. 1995).
: [[iTjwr v11 ;umers
Measuring impacts on consumers is straightforward when environmental policy regulates
consumer behavior. Requirements for automobile emissions tests or product bans such as the Final
Rule on Methylene Chloride in Paint and Coating Removal for Consumer Use (U.S. EPA 2019) have
impacts on consumers through time costs and fees. More frequently, environmental regulatory
requirements are imposed on producers. In these cases, there is a less obvious potential impact on
consumers as a result of producers passing through or transferring regulatory costs to purchasers
of their products through increased prices. To understand costpass-through to consumers, analysts
typically examine the expected impacts of a regulation on prices of final goods. Also relevant are the
characteristics of consumers purchasing the goods. Of course, firms may also be consumers of
regulated products and as such are covered in Section 9.5.1.6 "Impacts on Related Markets."
New environmental requirements typically raise the cost of production in directly regulated
industries, causing an upward shift in the market supply curve (that is, an increase in the price
65 For more information, see https://www.iustice.aov/atr/herfindahl-hirschman-index.
66 A related literature examines how differences in environmental regulation across countries, states, or sectors
may result in increased emissions in less regulated countries, also called emissions leakage. For instance, see
Bohringer et al. (2012) and Fischer and Fox (2012).
67 Carbone and Rivers (2017) discuss the impacts of environmental regulation on international competitiveness.
In general, the literature has found relatively small effects (Jaffe etal. 1995; Aldy and Pizer2015; Carbone and
Rivers 2017). Jaffe et al. (1995) point out that concerns about industry competitiveness may also ultimately
affect consumers as net exports decline and in the long-run imported goods become more expensive as the
economy returns to balanced trade.
68 For a survey of the literature on competitiveness impacts of unilateral climate change policy, see Carbone and
Rivers (2017). For a policy relevant discussion, see U.S. EPA (2016).
3-31 Guidelines for Preparing Economic Analyses | 3rd edition
-------
producers require for each quantity supplied). In response, consumers will do without or with less
of the product, and/or pay a higher price, thus bearing some of the burden of regulatory costs.
A good starting point to analyze potential impacts on consumers purchasing output from the
regulated sector is to gather information on the determinants of the elasticity of demand relative to
the elasticity of supply for the affected goods. To gauge elasticity of demand, a useful consideration
is whether the product is considered necessary by the purchaser, has many substitutes or its
purchase makes up a substantial portion of the consumer budget.69 Consumer impacts may be
smaller if there are good substitutes that are comparably priced causing a high demand elasticity
and smaller price change. There also may be small changes in output prices if compliance
expenditures are low relative to total production costs.
To gauge elasticity of supply, analysts should assess how easily firms can increase or decrease
production quantities. Information on the flexibility of capital equipment and buildings for shifting
into different types of production would be useful; for example, understanding whether excess
capacity can be used to produce comparably valued output
The characteristics of the regulated industry also influence the share of costs passed on to
consumers.70 Under noncompetitive conditions, when firms in the regulated industry have market
power, less cost-pass-through via prices is likely.71 All else equal, if the same compliance
requirements are placed on two markets that differ in terms of the degree of competition among
firms, the one with less competition (e.g., due to barriers to entry such as restricted access to a
scarce natural resource) will generally bear a higher share of those costs than the more competitive
market. For firms with market power, raising price will lower sales; therefore, these firms will
generally absorb some portion of regulatory costs (Tietenberg 2006). A market consisting of
producers that have different cost structures, perhaps because they use different technologies or
are of different sizes or ages, will lead to heterogeneity in the degree of pass-through of compliance
expenditures. Finally, the structure of related markets may affect cost pass-through. For example,
Preonas (2023) finds that distortions in the rail industry (an upstream market to the regulated one)
led railroads to reduce coal markups when downstream power plant demand for coal declined. This
suggests that regulatory costs faced by an industry may sometimes be partially absorbed by related
markets, shielding consumers from price increases.
A qualitative discussion of the factors that can affect impacts on consumers may be useful.
However, analysts may be able to locate empirical estimates of demand and supply elasticities. If
possible, analysts should select elasticity estimates that reflect the focus of analysis. For example, to
understand potential differences in the pass-through of regulatory costs into prices over time,
analysts should examine estimates of elasticity in the short-run compared to the long-run; to
understand differences in cost-pass-through across communities, analysts should examine regional
demand elasticity measures.
Combining an estimated price increase with information on the share of the consumer's budget
spent on the product will improve understanding of the impacts on households. There is a
possibility that budget shares may vary substantially across consumers. Even if price increases are
small, specific groups of consumers may still be affected if the product is a necessity for which low
69 For more information about the determinants of elasticities, see Appendix A: Economic Theory, Section A.4.1
Elasticities.
70 For cases when government or non-profit organizations are the producers, see Section 9.5.4.
71 For example, see Ganapti et. al. (2020) on incomplete pass-through of energy input costs and imperfect
competition in the manufacturing sector.
3-32 Guidelines for Preparing Economic Analyses | 3rd edition
-------
income households spend a substantial portion of their budget For example, the share of income
spent on energy or water by low income households is larger than for others, so energy or water
price increases may affect them more.72 This effect may be strengthened by the flexibility among
higher income households to purchase substitutes with substantial upfront costs such as efficient
appliances. However, it is also important to consider whether existing government programs may
help mitigate the impact of price increases on consumers.
If consumer impacts are expected to be nonnegligible, information on affected consumers such as
their age distribution, income level or residential location should be gathered to contribute to a
baseline profile. Nationwide averages of these variables may be appropriate if consumers are
broadly distributed across the country.
In some cases, assessing the impact of a regulation on consumers can be complex.73 Analyzing
policies with limited use patterns such as pesticides or paint removers may be challenging due to
inaccessible or sparse data. Other complicating factors are associated with goods for which price-
or rate-setting is complex. For example, to explore the extent to which proposed air pollution
control costs will be experienced by different electricity consumers, the analysis would need to
include information on how the policy affects consumers served by cost-of-service utilities,
compared to deregulated electricity providers. Any assistance available for low-income or other
consumers to offset rate increases is also relevant; as is variability in consumption patterns among
categories of customers. If regulatory costs are large, economy-wide models may lend additional
insight into how impacts affect consumers across the economy (see Section 9.5.5). Such models may
also examine the interaction with existing government transfer programs.74
|[?)'p,-vi\? "-.[i \ "-'[mmmh'i-
Environmental regulation may have significant impacts on some specific communities or
neighborhoods. Facility closures or production curtailments provide an example of locally
concentrated economic impacts that could be acute in areas with limited economic opportunities.
Displaced workers who live in such communities may be especially challenged as they search for
comparable re-employment Out-migration by displaced workers and families may cause
reductions in the demand for products in the local goods sector. Tax revenues may decline with
negative impacts on the provision and quality of community public goods. As the local economy
shrinks, property values may decline. For example, regulation on coal-fired power plants could
have negative impacts on coal-dependent communities. Mine closures and employment cuts can
affect others in the community as the economic base and local tax revenues decline (Baumol and
72 The share of income spent on energy falls as income increases. Some studies have found that policies that
increase energy prices are regressive, placing a greater burden on lower income households (e.g., Burtraw et al.
2009; Hassett et al. 2009; Williams et al. 2015). Other studies account for the indexing of transfer payments to
inflation and find that the burden of a carbon tax is roughly proportional to permanent income, and so is neither
regressive, nor progressive (Cronin etal. 2019). See Deryugina etal. (2019) for a discussion of some of these
energy policy studies.
73 Cory and Taylor (2017) conduct a detailed analysis of spending by low-income households and explore the
potential impacts on health spending caused by price changes induced by safe drinking water standards.
74 Some government transfer payments like Social Security are indexed to inflation and may provide some
protection of purchasing power for lower income households.
3-33 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Oates 1988; Black etal. 2005; Morris etal. 20 1 9).75 Impacts of changing business conditions that
spread across industries in the same community are often approximated by "local multipliers"
(Moretti 2010; Osman and Kemeny 2022). Such multipliers measure the broader changes in
employment and wage income across communities. When appropriate, analysis of baseline
economic conditions at the community level can help identify where the regulated industry is a key
driver of the local economy, signaling the potential for multiplier effects.
Community-level health impacts can be exacerbated by a combination of localized concentrations of
emissions from one or more sources, and community-wide exposure to other stressors. Locations
with such combinations of risks are often referred to as "hot spots." and may reflect baseline
conditions or be caused or aggravated by environmental regulation. Relevant issues to consider
may include proximity to multiple pollution sources, specific exposure pathways, and drivers of
differential susceptibility. For a full discussion see EPA's Technical Guidance for Assessing
Environmental Justice in Regulatory Analysis (2024d).
Localized improvements in environmental quality, such as hazardous site cleanup, can reduce
health risks and improve local property values thereby increasing the local tax base, and potentially
in the long run, increasing investments in local public and private goods. If low-income residents
are largely renters, then they could be burdened by increases in land values and subsequent
increases in rent due to improved environmental quality, while at the same time property owners
could enjoy higher rent payments. Property owners who reside in their own homes may be
burdened through property tax increases. The higher property taxes and rental payments may
cause some residents to move. The turnover may cause cost-of-living increases that further burden
remaining low-income residents even beyond increased rents and property taxes. Low-income
residents who relocate face transactions costs and do not experience the benefits of improved
environmental quality.76
When localized impacts of environmental policies are expected, a baseline profile of affected
communities will be informative.77 Data on the unemployment rate, average income level, the
poverty rate, whether the community is rural or urban, and its growth rate can help inform policy
makers as to the relative disadvantages faced by affected communities.78
75 Historic funding levels are being directed to coal mining and power plant communities to help rebuild and
diversify their economies. The Interagency Working Group on Coal & Power Plant Communities & Economic
Revitalization presents information on funding eligibility and more at https://eneraycommunities.aov/prioritv-
enerav-communiti es/#.
76 When a community that has experienced improved environmental quality undergoes a widescale turnover to
higher income households, this Is described as environmental gentrification. For further discussion of
gentrification in housing markets, see Section 10.2.1 of this guidance document, Section 8.2.5.1 of EPA's
Handbook on the Benefits, Costs, and Impacts of Land Cleanup and Reuse (U.S. EPA 2011), and/or Banzhafand
McCormick (2012).
77 For details regarding examining environmental justice communities, see Chapter 10.
78 For a discussion on contributors to higher susceptibility, see EPA's Technical Guidance for Assessing
Environmental Justice in Regulatory Analysis (U.S. EPA 2024c), Section 4.2, which addresses susceptibility or
vulnerability within groups such as communities.
3-34 Guidelines for Preparing Economic Analyses | 3rd edition
-------
c * a hii'i >'i" » i, )V6iiiiii >nts and Non-Profits
State and local governments and their residents, and non-profit organizations may incur costs or
bear the burden of costs from EPA regulations. The frameworks and impacts discussed above apply
to private markets. Governments and non-profits are distinctive because they are not motivated by
profits. Analysts should consider potential impacts to governments and non-profits, including
short- and long-run impacts.79 Useful measures for evaluating impacts on these types of entities
include assessments of the difficulty of paying regulatory costs and of continuing to provide
services.
Examples of important impacts on government include water treatment costs paid by municipally-
owned water authorities to comply with water quality standards. Air pollution controls required of
power plants may affect municipally-owned electric companies. Implementation and enforcement
costs associated with a variety of environmental regulations may impose costs on state or local
government. If regulation affects the local tax base, then there may be impacts on government
revenues or expenditures that may affect the provision of local public or private goods and services.
For example, some coal-mining counties in the United States derive a significant portion of their
budgets from coal-related revenues. Policies to restrict carbon pollution that reduce coal
production could significantly affect such communities causing the loss of local public goods and
lowered property values.80
To understand economic impacts on state, local and tribal governments, analysts should develop a
baseline profile potentially including the following relevant factors:
Size of the population in the community;
Property values;
Household income levels (e.g., median and/or income range);
Age distribution;
Unemployment rate;
Foreclosure rate; and
Revenue amounts by source.
If property taxes are the major revenue source, then the assessed value of property in the
community and the percentage of this assessed value represented by residential versus commercial
and industrial property may be important If a government entity serves multiple communities,
such as a regional water or sewer authority, then information for all the communities in the service
area may be relevant.
To gain insight into the ability of governments to finance new regulatory costs, U.S. EPA's Clean
Water Act Financial Capability Assessment Guidance (U.S. EPA 2024d) suggests examining baseline
financial capability by exploring indicators of debt, socioeconomic conditions and success regarding
financial management.81 Analysts can obtain the community's bond or credit rating, which is itself
determined by an assessment of financial health. For governments that rely on property taxes for
79 In some cases, EPA has been directed to consider impacts on government and non-profits. For example, UMRA
requires assessment of impacts to state, local and tribal governments. The RFA as amended by SBREFA requires
assessment of impacts to small entities including governments and non-profits (see Section 9.2 and Chapter 2).
80 Morris et al. (2019) study three counties with high labor shares engaged in coal mining and conclude that a
third or more of their budgets may be funded with coal-related revenue.
81 The EPA uses U.S. EPA (2024d) to assess implementation ofCWA requirements. The assessments affect
negotiations for Clean Water Act compliance schedules.
3-35 Guidelines for Preparing Economic Analyses | 3rd edition
-------
income, analysts might consider the amount of debt that must be repaid through property taxes
(known as net debt) per capita; or the net debt relative to the value of taxable properties. Property
tax revenues relative to full market value of properties may be a useful indication of the property
tax burden (U.S. EPA 2024d). Table 9.3 provides thresholds used by the Office of Enforcement and
Compliance Assurance (OECA) and the Office of Water (OW) to indicate weak, mid-range or strong
financial wellbeing of government entities.82
To screen for significant impacts on governments, analysts may wish to consider new regulatory
costs per capita, the ratio of per capita costs to median household income and lowest quintile
income, the latter especially in communities with households that have difficulty paying for their
water services. Depending on these values, further analysis might be desirable. 83 Further analysis
should consider a government entity's options for funding new costs or how new process
requirements could change operating procedures. For example, what is the availability of new loans
or grants and user fees? Are there other viable routes for increasing funds available to finance new
regulatory costs? Do new processes alter the quality or quantity of goods and services provided to
residents? Other factors that are potentially relevant are the historic trend in government revenues;
the capability of the revenue sources to shoulder additional financial burdens; and the magnitude of
the benefits from the rule enjoyed by citizens.
Finally, indirect impacts on state, local and tribal government may be important if a policy changes
local property values or employment rates or has other community-wide impacts. For example,
brownfield grants to assess or clean up land may cause small increases in local property values
which could raise property tax revenues (Sullivan 2017). On the other hand, a policy that
exacerbates unemployment, for example, could cause more spending on assistance programs.
EPA regulations may also affect non-profit organizations. For example, non-profit hospitals face
costs from hazardous waste disposal requirements. A baseline profile for non-profits should
consider:
Entity size and size of community served;
Goods or services provided;
Operating costs; and
Amount and sources of revenue.
If the entity is raising its revenues through user fees or charging a price for its goods or services
(such as university tuition), then the income levels of its clientele are relevant. If the entity relies on
contributions, then it would be helpful to know the financial and demographic characteristics of its
contributors and beneficiaries. If it relies on government funding (such as Medicaid) then possible
future changes in these programs would be informative.
82 For another source that explores approaches for assessing the health of a local government, see McDonald
(2018).
83 For instance, when assessing regulatory costs, the EPA considers financial impact as low if costs per
household are less than 1% of median household income, mid-range if it is 1-2% of median household income
and high if it is greater than 2% (U.S. EPA 2024d). Also, see the discussion of financial and rate model analyses in
Alternative 2 in U.S. EPA (2024d). Spreadsheet tools to help users evaluate the economic impacts of water quality
decisions can be found at https://www.epa.aov/wqs-tech/economic-auidance-water-quality-
standards#spreadsheet.
3-36 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Table 9.3 - Indicators of Economic and Financial Weil-Being of Government
Entities
Indicator
Strong
Mid-Range
Weak
Bond Rating
AAA - A (S&P) or
Aaa - A (Moody's) or
AAA - A (Fitch Ratings)
BBB (S&P) or
BAA (Moody's) or
BBB (Fitch Ratings)
BB - D (S&P) or
Ba - C (Moody's) or
BB - D (Fitch Ratings)
Overall Net Debt as a
Percent of Full Market
Property Value
Below 2%
2% - 5%
Above 5%
Unemployment Rate
More than 1 Percentage
Point Below the National
Average
± 1 Percentage Point of
National Average
More than 1 Percentage
Point Above the
National Average
Median Household
Income
More than 25% Above
Adjusted National MHI
± 25% of Adjusted
National MHI
More than 25% Below
Adjusted National MHI
Property Tax Revenues
as a Percent of Full
Market Property Value
Below 2%
2% - 4%
Above 4%
Property Tax Collection
Rate
Above 98%
94% - 98%
Below 94%
Source: Table B-l of U.S. EPA 2024d.
To screen for impacts on non-profits, analysts can compare regulatory costs to baseline revenues or
operating expenses. Regulatory costs can also be compared to baseline asset values or, after
accounting for debts, net asset values. If these ratios are large, insights would be gained from
information on the relative importance, size and growth rate of the non-profit, the nature of the
population being served and the vulnerability of revenues and donors.
Impacts on Small Governments and Small Non-Profits
Consideration of impacts on small governments and small non-profits is required by the RFA as
amended by SBREFA.84 The RFA defines a small governmental jurisdiction as the government of a
city, county, town, school district or special district with a population of less than 50,000. As with
the definition of small business, the RFA authorizes agencies to establish alternative definitions of
small government after opportunity for public comment and publication in the Federal Register.
Any alternative definition must be "appropriate to the activities of the agency" and "based on such
factors as location in rural or sparsely populated areas or limited revenues due to the population of
such jurisdiction" (U.S. EPA 2006a). Under the RFA, economic impacts on small governments are
included in the screening analysis for significant economic impacts on a substantial number of
small entities (SISNOSE), and any required regulatory flexibility analysis. In order to determine
84 See Chapter 2 and Section 9.2 for more information.
9-37 Guidelines for Preparing Economic Analyses | 3rd edition
-------
SISNOSE for small governments, the EPA conducts a screening analysis for both proposed and final
rules based on annualized compliance costs as a percentage of revenue (U.S. EPA 2006a).
The Unfunded Mandates Reform Act (UMRA) uses the same definition of small government as the
RFA, with the addition of tribal governments. Section 203 of UMRA requires the Agency to develop
a "Small Government Agency Plan" for any regulatory requirement that might "significantly" or
"uniquely" affect small governments. In general, "impacts that may significantly affect small
governments include but are not limited to those that may result in the expenditure by them
of $100 million [adjusted annually for inflation] or more in any one year." Other indicators that
small governments are uniquely affected may include whether they would incur higher per-capita
costs due to economies of scale, a need to hire professional staff or consultants for implementation,
or requirements to purchase and operate expensive or sophisticated equipment85
The RFA requires separate consideration of regulatory impacts on small non-profits and defines
one as a non-profit "enterprise which is independently owned and operated and is not dominant in
its field." Agencies are authorized to establish alternative definitions "appropriate to the activities
of the agency" after providing an opportunity for public comment and publication in the Federal
Register. Under the RFA, direct economic impacts on small non-profit organizations are included in
the SISNOSE screening analysis, and if required, the regulatory flexibility analysis for a rule. In
order to determine SISNOSE for small non-profits, the EPA conducts a screening analysis for both
proposed and final rules based on annualized compliance costs as a percentage of operating
expenditures.86
c * * \z ' tnomy-Wid Impacts
The more interconnected a regulated sector is with the rest of the economy, the greater the
likelihood that a regulation will affect related markets. If a regulation is expected to affect markets
with (i) significant cross-price effects between markets, and (ii) significant pre-existing distortions,
it may be appropriate to examine economy-wide impacts in a supplemental analysis (U.S. EPA
2017). Pre-existing market distortions that could be exacerbated by environmental regulations
include taxes or subsidies on labor, energy or capital; monopoly or monopsony power; price
controls; or other government regulations that change the way markets operate.
Computable general equilibrium (CGE) models are particularly effective at assessing long-run
economy-wide impacts.87 These include the allocation of employment or other factors of
production across sectors, the distribution of output by sector and the distribution of income
among households. For example, regulations in the power sector may cause electricity prices to
increase. The price increase will affect all industries that use electricity as an input to production, as
85 Guidance on complying with Section 203 of UMRA, "Interim Small Government Agency Plan," is available on
the EPA's intranet site, ADP Library.
86 See Table 1, "Recommended Quantitative Metrics for Economic Impact Screening Analyses" of U.S. EPA 2006a.
87 CGE models assume that for some discrete time period an economy can be characterized by a set of conditions
in which supply equals demand in all markets. When the imposition of a regulation alters conditions in one
market, the model determines a new set of relative prices that return the economy to its long-run equilibrium.
While highly aggregate in nature, CGE models capture substitution possibilities between production,
consumption and trade; interactions between economic sectors; and interactions with pre-existing distortions.
Thus, they provide information on changes outside the directly regulated sector. See Chapter 8 for more
discussion.
3-38 Guidelines for Preparing Economic Analyses | 3rd edition
-------
well as households. A CGE model can assess the distribution of consequent changes in production
and consumption. By design, the basic capacity to describe and evaluate these sorts of impacts
exists to some extent within every CGE model. More detailed impacts (e.g., effects on a certain type
of facility or on an environmental endpoint such as drinking water) are difficult to capture in a CGE
model due to model dimensionality and/or data constraints.
The simplest CGE models typically include a single representative consumer, a set of relevant
production sectors, and a government sector within a single-country, static framework. Additional
complexities can be specified. A CGE model can be solved dynamically over a longer time horizon,
incorporating intertemporal decision-making on the part of consumers or producers. These
decisions have implications for the treatment of savings, investment and the long-term profile of
consumption and capital accumulation. Consumers can be divided into income quintiles or deciles,
and producers disaggregated into a variety of regions and sectors, each producing a set of unique
commodities. The government, in addition to implementing a variety of taxes and other policy
instruments, may provide a public good or run a deficit. CGE models can be international in scope,
consisting of many countries or regions linked by international flows of goods and capital. The
behavioral equations that characterize economic decisions may take on simple or intricate
functional forms.
While CGE modeling is complex, the effort may be worthwhile when impacts are likely to be
substantial and widespread and when appropriate data (e.g., input-output tables, elasticities) are
available. Text Box 5.3 and Chapter 8 discuss detailed criteria for judging model quality. Feedback
from the Science Advisory Board (SAB) identified several guiding principles as to when economy-
wide modeling is appropriate for assessing economic impacts of regulation (U.S. EPA 2017).
Aspects of a CGE model that could affect suitability include degree of temporal, sectoral and
geographic disaggregation; time horizon; the way in which firm and household expectations about
the future are modeled; the types of impacts that can be forecast; and the approach for representing
the policy instrument CGE models may be useful as a supplement to other analytic approaches to
evaluate sectoral effects (including shifts in labor or capital between sectors), impacts on energy
supply and energy prices and effects on consumers. In some instances, linking a CGE model to
sector models may be a useful way to leverage the relative advantages of both approaches in a
single comprehensive framework (U.S. EPA 2017).88
CGE models have limitations. Many are not designed to illuminate certain types of impacts, such as
short-run or transitional impacts. For example, a standard forward-looking CGE model that
assumes full employment and instantaneous market adjustments is ill-suited to evaluate overall
employment impacts or the potential for short-run disequilibria in labor and capital markets.
Analysts interested in evaluating the short-run impacts of a policy should select a different
framework for analysis. Finally, relatively few CGE models incorporate feedback from changes in
pollution; instead, they mainly focus on private markets.
A partial equilibrium model of multiple markets that considers the interactions between a
regulated market and other closely related markets may be a practical alternative to a CGE model.
Such models require estimates of demand and supply elasticities and cross-price elasticities for
included markets. Partial equilibrium models may be appropriate for regionally-based or sector-
specific regulations that are too narrowly defined to be adequately captured in more aggregate CGE
models.
88 See Text Box 8.1 for more discussion of model linking.
3-33 Guidelines for Preparing Economic Analyses | 3rd edition
-------
The SAB recommends that analysts apply the simplest model that is adequate to address the policy
question at hand and consider a suite of models when possible (U.S. EPA 2017). A balance should be
struck between capturing detail and complexity in the model versus transparency and tractability
of the analysis.
As with all economic models, economy-wide and partial equilibrium models are simplified
representations of complex economic systems built to assess relationships between economic
factors. They are useful for estimating effects on groups but are not reliable predictors of firm or
facility-level decisions. See Section 9.5.1.2 for further explanation of the common simplifying
assumptions about firm decision-making.
[[iT[v*>t i I- - nefits
Environmental benefits are generally nonmarket effects and as such pose special analytic
challenges. As with costs, the benefits from improved environmental quality or health can accrue to,
and may differ among, a wide variety of individuals. A key determinant of differential impacts is
whether environmental improvements differ among affected groups (due to different exposure
pathways, for example), or are uniform but have variable impacts due to differences in pre-existing
factors such as baseline exposures or health (for more discussion, see U.S. EPA 2024c, especially
Chapter 4).
The literature provides several potential frameworks for explicitly considering variability in the
impacts of benefits across groups. Typically, these frameworks start with defining environmental
damages as a function of exposure and individual susceptibility to environmental stressors, then
they identify sources of susceptibility and finally they assess the impacts from environmental
regulation (see, for instance, Hsiang etal. 2019; Gee and Payne-Sturges 2004; and Morello-Frosch
and Jesdale 2006).
Useful information to improve understanding of the distribution of regulatory benefits includes:
The types of health effects or other benefits;
Population groups to whom the benefits are expected to accrue;
How exposure varies across the affected groups; and
How beneficial outcomes vary across population groups.
In addition to accruing to those who directly experience a reduced health risk, health and
environmental quality benefits may also accrue to people who own homes near improved
environmental quality, or to employers whose workers enjoy improved health and increased labor
productivity, as well as to others.
Chapter 10 discusses how to analyze health effects and benefits for specific populations of concern
(i.e., by income, race/ethnicity and age). The data and methods discussed there may be relevant for
analyzing the distribution of benefits on other categories of people, on communities or on the
general population. Sometimes analysts may wish to account not only for the ways in which
changes in the regulated sector affect the distribution of benefits, but also how price and quantity
responses across the economy affect the distribution of benefits, or how changes in environmental
quality affect prices and quantities. Absent a partial-equilibrium or economy-wide model that
explicitly incorporates benefits, relatively rare in the literature, these indirect impacts are difficult
to evaluate.
3-40 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 9 References
2 U.S.C. 48: Unfunded Mandates Reform Act of 1995 (P.L. 104-4), March 22,1995. Available at:
https://www.congress.gOv/104/plaws/publ4/PLAW-104publ4.pdf (accessed December 3, 2024).
5 U.S.C. 601-612: The Regulatory Flexibility Act of 1980 (P.L. 96-354), as amended by the Small Business
Enforcement Fairness Act (P.L. 104-121), 1996. Available at:
https://www.congress.gov/104/bills/s942/BILLS-104s942rfh.pdf (accessed December 3, 2024).
Adler, M. 2012. Well-Being and Fair Distribution: Beyond Cost-Benefit Analysis. New York: Oxford University
Press.
Adler, M. 2019. Measuring Social Welfare: An Introduction. New York: Oxford University Press.
Aguilar-Gomez, S., Dwyer, H., Graff Zivin, J. and Neidell, M., 2022. This is air: The "nonhealth" effects of air
pollution. Annual Review of Resource Economics, 14(1), pp.403-425.
Aldy, J.E. and W.A. Pizer.2015. The Competitiveness Impacts of Climate Change Mitigation Policies. Journal of
the Association of Environmental and Resource Economists, 2(4): 565-595.
Arrow, K. J.; M. L. Cropper; G. C. Eads; R. W. Hahn; L. B. Lave; R. G. Noll; Paul R. Portney; M. Russell; R.
Schmalensee; V. K. Smith and R. N. Stavins. 1996. Benefit-Cost Analysis in Environmental, Health, and Safety
Regulation: A Statement of Principles. American Enterprise Institute, the Annapolis Center, and Resources for
the Future; AEI Press. Available at: https://www.aei.org/research-products/book/benefit-cost-analysis-in-
environmental-health-and-safety-regulation/ (accessed December 23, 2020).
Banzhaf, H.S., and R.P. Walsh. 2008. "Do People Vote with Their Feet? An Empirical Test of Tiebout." American
Economic Review, 98 (3): 843-63.
Banzhaf, H.S. and E. McCormick. 2012. Moving Beyond Cleanup: Identifying the Crucibles of Environmental
Gentrification. The Political Economy of Environmental Justice. Stanford University Press, Stanford, CA.: 23-51.
Banzhaf, H.S., L. Ma, and C. Timmins. 2019. Environmental Justice: The Economics of Race, Place, and
Pollution .Journal of Economic Perspectives 33(l):158-208. Banzhaf, H.S., and Walsh, R.P. 2008. Do People Vote
with Their Feet? An empirical test of Tiebout's mechanism. American Economic Review 98(3) 843-63.
Bartik, T.J., 2015. The Social Value of lob Loss and Its Effect on the Costs of U.S. Environmental
Regulations. Review of Environmental Economics and Policy. 9(2): 179-197.
Baumol, W.J. and W.E. Oates. 1988. The Theory of Environmental Policy. 2nd Ed. New York: Cambridge
University Press.
Belova, A., W.B. Gray, J. Linn, and R.D. Morgenstern. 2013. "Environmental Regulation and Industry
Employment: A Reassessment," Discussion Papers, U.S. Census Bureau, Center for Economic Studies 2K132B,
Washington, DC. Available at: https: //papers.ssrn.com/sol3 /papers.cfm7abstract id=2306753 (Accessed
December 3,2024).
Belova, A., W.B. Gray, J. Linn, R.D. Morgenstern, and W. Pizer. 2015. Estimating the Job Impacts of
Environmental Regulation. Journal ofBenefit-Cost Analysis, 6(2): 325-340.
Bento, A.M. 2013. Equity Impacts of Environmental Policy. Annual Review of Resource Economics, 5: 181-96.
Berman, E. and L. T. M. Bui. 2001. Environmental Regulation and Labor Demand: Evidence from the South
Coast Air Basin. Journal of Public Economics, 79(2): 265-295.
Black, D., T. McKinnish, and S. Sanders. 2005. The Economic Impact of the Coal Boom and Bust. The Economic
Journal. 115 (April): 449-476.
Bohringer, C., E.J. Balistreri, T.F. Rutherford. 2012. The role of border carbon adjustment in unilateral climate
policy: Overview of an Energy Modeling Forum Study (EMF 29). Energy Economics, 34(2): S97-S110.
Borjas, George J. 1996. Labor Economics. McGraw-Hill.
3-41 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Buchanan, J.M. and G. Tullock. 1975. Polluters' Profits and Political Response: Direct Controls Versus Taxes.
The American Economic Review, 65(1): 139-147.
Burgard, S., J. Brand, and J. House. 2007. Toward a Better Estimation of the Effect of Job Loss on Health.
Journal of Health and Social Behavior 48: 369-384.
Burtraw, D., R. Sweeney, and M. Walls. 2009. The Incidence of U.S. Climate Policy: Alternative Uses of
Revenues from a Cap-and-Trade Auction. National Tax Journal, 62(3): 497-518.
Carbone, J.C., and N. Rivers. 2017. The Impacts of Unilateral Climate Policy on Competitiveness: Evidence
from Computable General Equilibrium Models. Review of Environmental Economics and Policy, 11(1): 24-42.
Chan, S. and A. H. Stevens. 2001. Job Loss and Employment Patterns of Older Workers .Journal of Labor
Economics 19: 4484-521.
Christiansen, G.B. andT. Tietenberg. 1985. Distributional and Macroeconomic Aspects of Environmental
Policy. Handbook of Natural Resource and Energy Economics, 1, North Holland, Amsterdam; Oxford. 345-393.
Cole, Rebel A. 2020. Bank Lending to Rural vs. Urban Firms in the United States Before, During, and After the
Great Financial Crisis. U.S. Small Business Administration Office of Advocacy, https://advocacy.sba.gov/wp-
content/uploads/2020/05/REPORT-Cole-Rural-and-Urban-Lending pd.pdf (accessed December 3, 2024).
Congressional Budget Office. 2011. Statement of Douglas W. Elmendorf, Director, before the Senate Budget
Committee, "Policies for Increasing Economic Growth and Employment in 2012 and 2013" (November 15)
Available at: http: //www.cbo.gov/sites/default/files/ll-15-0utlook Stimulus Testimonv.pdf (accessed
December 3,2024).
Cory, D. C. and L. D. Taylor. 2017. On the Distributional Implications of Safe Drinking Water Standards. Journal
of Benefit Cost Analysis, 8(1): 49-90.
Cronin, J.A., D. Fullerton, and S. Sexton. 2019. Vertical and Horizontal Redistributions from a Carbon Tax and
Rebate. Journal of the Association of Environmental and Resource Economists, 6(S1): S169-S208.
Curtis, E. M. 2018. Who Loses under Cap-and-trade programs? The Labor Market Effects of the NOx Budget
Trading Program," Review of Economics and Statistics, 100(1): 151-66.
Deryugina, T., D. Fullerton, and W.A. Pizer. 2019. An Introduction to Energy Policy Trade-Offs between
Economic Efficiency and Distributional Equity. Journal of the Association of Environmental and Resource
Economists, 6(S1): S1-S6.
Deschenes, 0. 2012. "Climate Policy and Labor Markets," Ch. 2 in The Design and Implementation of U.S.
Climate Policy, ed. D. Fullerton and C. Wolfram. University of Chicago Press.
Deschenes, 0. 2014. Environmental regulations and labor markets. IZA World of Labor, 22:1-10.
Dobkin, C., A. Finkelstein, R. Kluender, and M. Notowidigdo. 2018. The Economic Consequences of Hospital
Admissions. American Economic Review, 108(2): 308-352.
Executive Order 12866: Regulatory Planning and Review, Section 1(a), October 4,1993. Available at:
https: //www.archives.gov/files/federal-register/executive-orders/pdf/12866.pdf (accessed December 3,
2024).
Executive Order 13132: Federalism, August 10,1999. Available at:
https: //www.federalregister.gOv/documents/1999 /08/10/99-20729 /federalism (accessed December 3,
2024).
Executive Order 13175: Consultation and Coordination with Indian Tribal Governments, November 9,2000.
Available at: https: //www.federalregister.gov/documents/2000/ll /09 /00-29003 /consultation-and-
coordination-with-indian-tribal-governments (accessed December 3,2024).
Executive Order 13211: Actions Concerning Regulations that Significant Affect Energy Supply, Distribution, or
Use, May 18,2001. Available at: https://www.govinfo.gov/content/pkg/WCPD-2001-05-21/pdf/WCPD-
2001-05-21-Pg769.pdf (accessed December 3, 2024).
9-42 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Fairlie, R.W., A. Robb and D.T. Robinson. 2022. "Black and White: Access to Capital Among Minority-Owned
Startups," Management Science, 68(4).
Farber, H. S. 2017. Employment, Hours, and Earnings Consequences of Job Loss: US Evidence from the
Displaced Workers Survey. Journal of Labor Economics, 35(S1): S235-S272.
Federal Reserve Banks. 2023. Report on Startup Firms Owned by People of Color: Findings from the 2022
Small Business Credit Survey. Small Business Credit Survey.
https://www.fedsmallbusiness.org/reports/survey/2023/2023-report-on-startup-firms-owned-by-people-
of-color (accessed December 3, 2024).
Federal Reserve Banks. 2024. Report on Employer Firms: Findings from the 2023 Small Business Credit
Survey. Small Business Credit Survey. https: / /www.fedsmallbusiness.org/reports /survey /2024/2Q24-report-
on-emplover-firms (accessed December 3, 2024).
Ferris, A. E. and A. McGartland. 2014. "A Research Agenda for Improving the Treatment of Employment
Impacts in Regulatory Impact Analysis," in Does Regulation Kill Jobs? Edited by Cary Coglianese, Adam M.
Finkel, and Christopher Carrigan, 2014, Philadelphia, Pennsylvania: University of Pennsylvania Press.
Ferris, A. E., R. Shadbegian, A. Wolverton. 2014. The Effect of Environmental Regulation on Power Sector
Employment: Phase I of the Title IV S02 Trading Program. Journal of the Association of Environmental and
Resource Economics, 1(4): 521-553.
Field, B.C. and M.K. Field. 2005. Environmental Economics: An Introduction. 4th Ed. New York: McGraw-
Hill/Irwin.
Fischer, C. and A. Fox. 2012. Comparing Policies to Combat Emissions Leakage: Border Carbon Adjustments
Versus Rebates .Journal of Environmental Economics and Management, 64(2): 199-216.
Fullerton, D. 2009. Does environmental protection hurt low-income families? CESifo Forum, 10(1): 45-49.
Fullerton, D. 2011. Six Distributional Effects of Environmental Policy. Risk Analysis. 31(6): 923-929.
Fullerton, D. 2016. "Introduction." In D. Fullerton, ed., Distributional Effects of Environmental and Energy
Policy, Aldershot, UK: Ashgate.
Fullerton, D. and G. Heutel. 2010. The General Equilibrium Incidence of Environmental Mandates. American
Economic Journal: Economic Policy, 2(3): 64-89
Fullerton, D. and G. Metcalf. 2001. Environmental Controls, Scarcity Rents, and Pre-existing Distortions.
Journal of Public Economics, 80(2): 249-267.
Fullerton, D and G. Metcalf. 2002. "Tax Incidence," in Handbook of public economics, ed. A. Auerbach, and M.
Feldstein. Amsterdam: Elsevier.
Fullerton, D. and E. Muehlegger. 2019. Who Bears the Economic Burdens of Environmental Regulations?
Review of Environmental Economics and Policy, 13(1): 62-82.
Ganapti, S., J. Shapiro, and R. Walker. 2020. Energy Cost Pass-Through In U.S. Manufacturing: Estimates and
Implications for Carbon Taxes. American Economic Journal: Applied Economics 12(2): 303-42.
Gee, G.C. and D.C. Payne-Sturges. 2004. Environmental Health Disparities: A Framework Integrating
Psychosocial and Environmental Concepts. Environmental Health Perspectives, 112(17): 1645-1653.
Graff Zivin J. and M. Neidell. 2012. The Impact of Pollution on Worker Productivity. American Economic
Review, 102(7): 3652-73.
Graff Zivin, J. and M. Neidell. 2013. Environment, Health, and Human Capital. Journal of Economic Literature,
51(3): 689-730.
Graff Zivin, J. and M. Neidell. 2018. Air Pollution's Hidden Impacts. Science, 359(6371): 39-40.
Gray, W. and R. Shadbegian. 2010. When Do Firms Shift Production Across States to Avoid Environmental
Regulation? Working Paper. U.S EPA National Center for Environmental Economics. Available at:
3-43 Guidelines for Preparing Economic Analyses | 3rd edition
-------
https://www.epa.gov/environmental-economics/working-paper-when-do-firms-shift-production-across-
states-avoid (accessed December 3,2024).
Gray, W.B., R. Shadbegian, and A. Wolverton. 2023. Environmental Regulation and Labor Demand: What Does
the Evidence Tell Us? Annual Review of Resource Economics, 15(1), pp.177-197.
Greenstone, M. 2002. The Impacts of Environmental Regulations on Industrial Activity: Evidence from the
1970 and 1977 Clean Air Act Amendments and the Census of Manufactures. Journal of Political Economy,
110(6): 1175-1219.
Hafstead, M.A., and R.C. Williams III. 2020. Jobs and environmental regulation. Environmental and energy
policy and the economy, 1(1), pp.192-240.
Hafstead, M.A., and R.C. Williams III. 2018. Unemployment and Environmental Regulation in General
Equilibrium .Journal of Public Economics, 160: 50-65.
Hall, R. 2011. Discussion comment on Davis and von Wachter "Recessions and the Cost of Job Loss," Brookings
Papers on Economic Activity, 2: 57-72.
Harberger, A.C. 1962. The Incidence of the Corporation Income Tax. The Journal of Political Economy, 70(3):
215-240.
Hassett, K.A., A. Mathur, and G.E. Metcalf. 2009. The Incidence of a U.S. Carbon Tax: A Lifetime and Regional
Analysis. Energy Journal, 30(2): 157-79.
Holland, S. 2012. "Spillovers from Climate Policy to Other Pollutants," Ch. 5 in The Design and Implementation
of U.S. Climate Policy, D. Fullerton and C. Wolfram, ed. University of Chicago Press.
Howitt, Peter. 2002. Looking inside the Labor Market: A Review Articl e. Journal of Economic Literature, 40(1):
125-138.
Hsiang, S., P. Oliva, and R. Walker. 2019. The Distribution of Environmental Damages. Review of
Environmental Economics and Policy, 13(1): 83-103.
Iraldo, F., F. Testa, M. Melis, and M. Frey. 2011. A Literature Review on the Links between Environmental
Regulation and Competitiveness. Environmental Policy and Governance, 21(3): 210-222.
Ito, K., and Zhang, S. 2020. Willingness to pay for clean Air: Evidence from air purifier markets in China.
Journal of Political Economy 128(5):1627-72.
Jacobson, L., R. LaLonde, and D. Sullivan. 1993. Earnings Losses of Displaced Workers. American Economic
Review, 83: 685-709.
Jaffe, A.B., S.R. Peterson, P.R. Portney, and R.N. Stavins. 1995. Environmental Regulation and the
Competitiveness of U.S. Manufacturing: What Does the Evidence Tell Us? Journal of Economic Literature 33:
132-163.
Kahn, M.E. 1997. Particulate Pollution Trends in the United States. Regional Science and Urban Economics,
27(1): 87-107.
Kolstad, C. 2000. Environmental Economics. Oxford: Oxford University Press.
Krey, V., and K. Riahi. 2009. Risk Hedging Strategies under Energy System and Climate Policy Uncertainties.
IIASA Interim Report. IIASA, Laxenburg, Austria, IR-09-028.
Kristrom, B. 2006. Framework for assessing the distribution of financial effects of environmental policy. The
Distributional Effects of Environmental Policy, ed. Y. Serret and N. Johnstone. Gloucestershire, UK: OECD
Environment Directorate and Edward Elgar Publishing.
Kuhn, A., R. Lalive, and J. Zweimuller. 2009. The Public Health Costs of Job Loss .Journal of Health Economics,
28: 1099-1115.
Kuminoff, Nicolai V., Todd Schoellman, and Christopher Timmins. 2015. Environmental Regulations and the
Welfare Effects of Job Layoffs in the United States: A Spatial Approach. Review of Environmental Economics
and Policy, 9(2): 198-218.
9-44 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Layard, P.R.G. and A. A. Walters. 1978. Microeconomic Theory. McGraw-Hill, Inc. Chapter 9.
List, J. A., D. L. Millimet, P. G. Fredriksson, and W. W. McHone. 2003. Effects of Environmental Regulations on
Manufacturing Plant Births: Evidence from a Propensity Score Matching Estimator. The Review of Economics
and Statistics, 85(4): 944-952.
Malmendier, U. and G. Tate. 2015. Behavioral CEOs: The Role of Managerial Overconfidence. Journal of
Economic Perspectives, 29(4): 37-60.
McDonald, B.D. 2018. Local Governance and the Issue of Fiscal Health. State and Local Government Review,
50(1): 46-55.
Morazzoni, M., and A. Sy. 2022. Female entrepreneurship, financial frictions and capital misallocation in the
US .Journal of Monetary Economics. 129: 93-118.
Morello-Frosch, R., and B.M. Jesdale. 2006. Separate and Unequal: Residential Segregation and Estimated
Cancer Risks Associated with Ambient Air Toxics in U.S. Metropolitan Areas. Environmental Health
Perspectives, 114(3): 386-393.
Moretti, E., 2010. Local multipliers. American Economic Review, 100(2), pp.373-377.
Morgenstern, R.D., W.A. Pizer, and J. Shih. 2002. Jobs Versus the Environment: An Industry- Level Perspective.
Journal of Environmental Economics and Management 43: 412-436.
Morris, A., N. Kaufman, and S. Doshi. 2019. The Risk of Fiscal Collapse in Coal Reliant Communities. Columbia
University, Center on Global Energy Policy. Available at:
https://www.energypolicy.columhia.edu/puhlications/risk-fiscal-collapse-coal-reliant-communities/
(accessed November 12,2024).
0MB. 2001. Memoranda 01-27 (Guidance for Implementing E.0.13211). From Mitchell E. Daniels, Jr. July 13.
https://www.whitehouse.gOv/wp-content/uploads/2017/ll/2001-M-01-27-Guidance-for-Implementing-
E.0.-13211.pdf
0MB. 2015. 2015 Report to Congress on the Benefits and Costs of Federal Regulations and Agency
Compliance with the Unfunded Mandates Reform Act. Available
at: https: //obamawhitehouse.archives.gov/sites/default/files/omb/inforeg/2015 cb/2015-cost-benefit-
reportpdf (accessed December 3, 2024).
0MB. 2023. Circular A-4, Regulatory Analysis, November 9, 2023. Available at:
https://www.whitehouse.gov/wp-content/uploads/2023/ll ZCircularA-4.pdf (accessed June 13, 2024).
Pashigian, P. B. 1984. The Effect of Environmental Regulation on Optimal Plant Size and Factor Shares. The
Journal of Law and Economics, 27(1): 1-28.
Osman, T., and T. Kemeny. 2022. Local job multipliers revisited. Journal of Regional Science, 62(1), pp.150-
170.
Pearce, D. W. 2006. Framework for assessing the distribution of environmental quality. The Distributional
Effects of Environmental Policy, ed. Y. Serret and N. Johnstone. Gloucestershire, UK: OECD Environment
Directorate and Edward Elgar Publishing.
Picchio, M and M. Ubaldi. 2023. Unemployment and health: A meta-analysis. Journal of Economic Surveys.
Pittman, R.W. 1981. Issue in Pollution Control: Interplant Cost Differences and Economies of Scale. Land
Economics, 57(1): 1-17.
Preonas, Louis. 2023. Market Power in Coal Shipping and Implications for U.S. Climate Policy. The Review of
Economics and Statistics, 91(4): 2508-2537.
Rausch, S., G. E. Metcalf, and J. M. Reilly. 2011. Distributional Impacts of Carbon Pricing: A General Equilibrium
Approach with Micro-Data for Households. Energy Economics, 33(1): S20-S33.
Rege, M., M. Votruba, and K. Telle. 2009. The Effect of Plant Downsizing on Disability Pension Utilization.
Journal of European Economic Association, 7(4): 754-785.
3-45 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Robinson, L. A., J. K. Hammitt, and R. J. Zeckhauser. 2016. Attention to Distributions in U.S. Regulatory
Analyses. Review of Environmental Economics and Policy, 10 (2): 308-28.
Roelfs, D. J., E. Shor, K. W. Davidson, and J. E. Schwartz. 2011. Losing Life and Livelihood: A Systematic Review
and Meta-Analysis of Unemployment and All-Cause Mortality. Social Science & Medicine, 72(6): 840-54.
Schmalensee, R. and R. Stavins. 2011. A Guide to Economic and Policy Analysis for the Transport Rule.
Whitepaper. Boston, MA. Exelon Corp.
Serret, Y. and N. Johnstone. 2 006. The Distributional Effects of Environmental Policy, ed. Y. Serret and N.
Johnstone. Gloucestershire, UK: OECD Environment Directorate and Edward Elgar Publishing.
Shadbegian, R.J. and A. Wolverton. 2010. Location Decisions of U.S. Polluting Plants: Theory, Empirical
Evidence, and Consequences. International Review of Environmental and Resource Economics, 4:1-49.
Smith, V. K. 2015. Should Benefit-Cost Methods Take Account of High Unemployment? Symposium
Introduction. Review of Environmental Economics and Policy, 9(2): 165-178.
Sullivan, K. A. 2017. Brownfields Remediation: Impact on Local Residential Property Tax Revenue. Journal of
Environmental Assessment Policy and Management. 29(3): 1-20.
Sullivan, D. and T. von Wachter. 2009a. Average Earnings and Long-Term Mortality: Evidence from
Administrative Data. American Economic Review, Papers and Proceedings, 99(2): 133-138.
Sullivan, D. and T. von Wachter. 2009b. Job Displacement and Mortality: An Analysis Using Administrative
Data. The Quarterly Journal of Economics, 124(1): 1265-1306.
Tietenberg, T. 1992. Environmental and Natural Resource Economics. 2nd Ed. Chapter 20 Pollution Control
Policy: Distributional Effects. New York: HarperCollins Publishers Inc.
Tietenberg, T. 2002. Environmental and Natural Resource Economics. 6th Ed. Boston, MA: Pearson, Addison
Wesley, Inc.
Tietenberg, T. 2006. Environmental and Natural Resource Economics. 7th Ed. Boston, MA: Pearson, Addison
Wesley, Inc.
U.S. EPA. 2006a. Final Guidance for EPA Rulewriters: Regulatory Flexibility Act as amended by the Small
Business Regulatory Enforcement Fairness Act. Available at:
https://www.epa.gOv/sites/production/files/2015-06/documents/guidance-regflexact.pdf (accessed
November 13, 2024).
U.S. EPA. 2006b. A Framework for Assessing Health Risks of Environmental Exposures to Children (Final),
Washington, DC, 21 EPA/600/R-05/093F. Available at:
http://cfpub.epa.gov/ncea/cfin/recordisplay.cfm?deid=158363 (accessed April 1,22 2020).
U.S. EPA. 2010. Guidelines for Preparing Economic Analysis, https://www.epa.gov/sites/default/files/2017-
08/documents/ee-0568-50.pdf (Accessed August 29, 2024).https://www.epa.gov/sites/default/files/2017-
08/documents/ee-0568-50.pdf
U.S. EPA. 2011. Handbook on the Benefits, Costs, and Impacts of Land Cleanup and Reuse. EPA-240-R-11-001.
https://www.epa.gov/sites/production/files/2017-08/documents/ee-0569-02.pdf (accessed October 15,
2024).
U.S. EPA. 2016. Economy-Wide Modeling: Use of CGE Models to Evaluate the Competitiveness Impacts of Air
Regulations," Jun 17, 2016, Prepared for the U.S. EPA Science Advisory Board Panel (SAB) on Economy-Wide
Modeling of the Benefits and Costs of Environmental Regulation.
https://www.epa.gov/system/files/documents/2023-02/competitiveness%20memo 061716.pdf (accessed
November 13, 2024).
U.S. EPA. 2017. Science Advisory Board (SAB). "SAB Advice on the Use of Economy-Wide Models in Evaluating
Social Costs, Benefits, and Economic Impacts of Air Regulations." EPA-SAB-17-012. Available at:
https://www.epa.gOv/system/files/documents/2023-02/EPA-SAB-17-012 l.pdf (accessed November 13,
2024).
3-46 Guidelines for Preparing Economic Analyses | 3rd edition
-------
U.S. EPA. 2019. Economic Analysis of Final Rule - Regulation of Methylene Chloride, Paint and Coating
Remover. TSCA Section 6(a) (EPA Docket EPA-HQ-OPPT-2016-0231; RIN 2070-AK07). Available at:
https://www.regulations.gOv/document/EPA-HO-OPPT-2016-0231-0990 (accessed December 3. 20241
U.S. EPA. 2023a. Regulatory Impact Analysis: Safer Communities by Chemical Accident Prevention Final Rule.
Office of Emergency Management. August 30.
U.S. EPA. 2023b. U.S. EPA Methodology for Power Sector-Specific Employment Analysis. Available at:
https://www.epa.gov/system/files/documents/2023-
04/U.S.%20EPA%20Methodologv%20for%20Power%20Sector-Specific%20Emplovment%20Analvsis .pdf
(accessed July 24, 2024).
U.S. EPA. 2024a. Documentation for 2023 Reference Case: EPA's Power Sector Modeling Platform 2023 Using
IPM. Available at: https: //www.epa.gov/power-sector-modeling/documentation-2023-reference-case.
U.S. EPA. 2024b. Regulatory Impact Analysis for New Source Performance Standards for Greenhouse Gas
Emissions from New, Modified, and Reconstructed Fossil Fuel-Fired Electric Generating Units; Emission
Guidelines for Greenhouse Gas Emissions from Existing Fossil Fuel-Fired Electric Generating Units; and
Repeal of the Affordable Clean Energy Rule. Available at: https: IIwww.regulations.gov /document / E PA- H 0 -
OAR-2023-0072-8913 (accessed July 24, 2024).
U.S. EPA. 2024c. Technical Guidance for Assessing Environmental Justice in Regulatory Actions. Available at:
https://www.epa.gov/environmentaliustice/technical-guidance-assessing-environmental-iustice-regulatory-
analysis.
U.S. EPA. 2024d. Clean Water Act Financial Capability Assessment Guidance. March. EPA 800-B-24-001.
Available at: https: //www.epa.gov/system/files/documents/2023-01 /cwa-financial-capahility-assessment-
guidance.pdf.
U.S. Small Business Administration. 2022. Table of Small Business Size Standards: Matched to North American
Industry Classification System Codes. Available at: https: //www.sba.gov/sites/default/files/2023-
06/Table%20of%20Size%20Standards Effective%20March%2017%2C%202023%20%282%29.pdf
(accessed October 10,2024).
van der Noordt M, H. IJzelenberg, M. Droomers, and K. Proper. 2014. Health effects of employment: a
systematic review of prospective studies." Occupational and Environmental Medicine, 71:730-736.
Walker, W.R. 2013. The Transitional Costs of Sectoral Reallocation: Evidence from the Clean Air Act and the
Workforce. Quarterly Journal of Economics, 128:1787-1835.
Williams, R. C, H. Gordon, D. Burtraw, J. C Carbone, and R. D. Morgenstern. 2015. The Initial Incidence of a
Carbon Tax across Income Groups. National Tax Journal, 68(1): 195-214.
3-47 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 10 - Environmental Justice
and Life Stage Considerations
Instead of focusing on quantifying and monetizing total benefits and costs, an
evaluation of the impacts of a regulation examines how a regulation allocates
benefits, costs, transfers and other outcomes across specific groups of
interest. Chapter 9 describes approaches to quantify economic impacts across
a wide array of groups that may be of interest to decisionmakers. This chapter
overlaps with Chapter 9 in some respects many of the economic impact
categories it discusses are also potentially relevant here but it is distinct in
several ways. First, this chapter specifically considers the possible impacts of
a regulatory action on people of color, low-income or Indigenous populations
(i.e., the focus of environmental justice] and on children and older adults (i.e.,
life stages] due to their increased vulnerability to health effects from
pollution. Second, while variation in the benefits and costs of regulation across
these population groups is a significant consideration, this chapter also
discusses the importance of characterizing changes in human health
endpoints and environmental risk.
1 i P ¦ Mir'-" Pir-i'itr'""o P-Pkro
Consideration of how economic and human health effects vary across specific population groups
and life stages arises from several executive orders (EOs), directives and other documents.12 The
Agency also has developed separate guidance to provide direction to analysts on conducting
environmental justice analyses. Together, these orders, directives and policies provide a solid
foundation for considering effects on population groups from an environmental justice (EJ) and life
stage standpoint in the rulemaking process.
In addition to the general guidance in the Office of Management and Budget's (OMB's) Circular A-4
(0MB 2023) regarding distributional analysis, several EOs, described more fully in Chapter 2,
directly address different types of effects for population groups of concern:3
1 EPA's Regulatory Management Division's Action Development Process Library
(http://intranet.epa.aov/adplibrarv/adp) is a resource for accessing relevant statutes, executive orders and EPA
policy and guidance documents in their entirety.
2 Some environmental statutes also identify population groups that may merit additional consideration. See EPA
Legal Tools to Advance Environmental Justice (U.S. EPA 2022)for a review of legal authorities under the
environmental and administrative statutes administered by the EPA.
3 This chapter addresses analytical components of EOs 12898 and 14096 and does not cover other components
such as ensuring proper outreach and meaningful involvement.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
EO 12898, "Federal Actions to Address Environmental Justice in Minority Populations and
Low-Income Populations" (1994), calls on each federal agency to make achieving EJ part of
its mission to the greatest extent practicable and permitted by law "by identifying and
addressing, as appropriate, disproportionately high and adverse human health or
environmental effects of its programs, policies, and activities on minority populations and
low-income populations in the United States."
EO 14096, "Revitalizing Our Nation's Commitment to Environmental Justice for All" (2023),
supplements EO 12898 and calls on agencies to, as appropriate and consistent with
applicable law, identify, analyze, and address:
o Disproportionate and adverse human health and environmental effects..., including
those related to climate change and cumulative impacts of environmental and other
Burdens on communities with environmental justice concerns;
o Historical inequities, systemic barriers, or actions related to any Federal regulation,
policy, or practice that impair the ability of communities with environmental justice
concerns to achieve or maintain a healthy and sustainable environment; and
o Barriers related to Federal activities that impair the ability of communities with
environmental justice concerns to receive equitable access to human health or
environmental benefits, including benefits related to natural disaster recovery and
climate mitigation, adaptation, and resilience.
EO 13045, "Protection of Children from Environmental Health Risks and Safety Risks"
(1997), states that each federal agency shall ensure that its policies, programs, activities,
and standards address disproportionate risks to children that result from environmental
health or safety risks.
EO 13175, "Consultation and Coordination with Indian Tribal Governments" (2000), calls on
federal agencies to have "an accountable process to ensure meaningful and timely input by
tribal officials in the development of regulatory policies that have tribal implications."
EO 12866, "Regulatory Planning and Review" (1993), explicitly allows for consideration of
"distributive impacts" and "equity" when choosing among alternative regulatory
approaches, unless prohibited by statute.4
EO 14094, "Modernizing Regulatory Review" (2023), supplements, reaffirms, and amends
12866, and confirms that "[rjegulatory analysis, as practicable and appropriate, shall
recognize distributive impacts and equity, to the extent permitted by law."
10.2 Environmental Justice
This section offers a high-level summary of analytic expectations and recommendations for
evaluating environmental justice concerns for EPA regulatory actions, consistent with Technical
Guidance for Assessing Environmental Justice in Regulatory Analysis [EJ Technical Guidance) (U.S.
EPA 2024a). Analysts should consult the EJ Technical Guidance for additional detail on analytic
approaches and considerations.
4EO13563, issued in January 2011, supplements and reaffirms the provisions ofEO 12866.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
An analysis of EJ concerns for regulatory actions should address three questions:5
Baseline: Are there existing EJ concerns associated with environmental stressors affected
by the regulatory action for population groups of concern?6
Regulatory options: For the regulatory option(s) under consideration, are there potential
EJ concerns associated with environmental stressors that are affected by the regulatory
action for population groups of concern?
Mitigation or exacerbation of effects: For the regulatory option(s) under consideration,
are EJ concerns exacerbated, mitigated, or unchanged compared to the baseline?
These questions provide the framework for analyzing the effects of a regulatory action on
population groups of concern. The extent to which an analysis can address all three questions will
vary due to data limitations, time and resource constraints, and other technical challenges. These
challenges will vary by media and regulatory context, including the availability of information
generated from human health risk and exposure assessments, or other components of the
regulatory analysis.
The EPA encourages analysts to document key reasons why a particular question cannot be
addressed to help identify future priorities for filling key data and research gaps. A lack of existing
data and methods does not mean there is no EJ concern, so identifying such gaps is important
Regardless of the approach taken, the highest quality and most relevant data should be applied in a
manner consistent with the OMB and EPA data quality guidelines (0MB 2019; U.S. EPA 2012; U.S.
EPA 2002) and the Peer Review Handbook (U.S. EPA 2015).
The term "disproportionate" is used here to refer to differences in effects or risks that are extensive
enough that they may merit Agency action and should include consideration of cumulative impacts
or risks where appropriate and consistent with applicable law (U.S. EPA 2022). In general, the
determination of whether a difference in effects or risks is disproportionate is ultimately a policy
judgment which, while informed by analysis, is the responsibility of the decision-maker.7 The terms
"difference" or "differential" indicate an analytically discernible (or measurable) distinction in
effects or risks across population groups. It is the role of analysts to assess and present differences
in anticipated effects across population groups in the baseline and for the regulatory options, using
the best available information (both quantitative and qualitative) to inform the decision-maker and
the public.
5 An EJ concern is the actual or potential lack of just treatment or meaningful involvement of any population
group, community, or geographic area (e.g., associated with differences in income, race, color, national origin,
Tribal affiliation, or disability status) in the development, implementation, and enforcement of environmental
laws, regulations, and policies. For analytic purposes, this concept refers specifically to disproportionate and
adverse health and environmental effects that may exist prior to or be created by the regulatory action.
6 The term environmental stressor encompasses the range of chemical, physical, or biological agents,
contaminants, or pollutants that may be subject to a regulatory action.
7 A finding of disproportionate and adverse effects is neither necessary nor sufficient for the EPA to address
them. The Agency's statutory and regulatory authorities provide a broad basis for protecting human health and
the environment and do not require a demonstration of disproportionate effects to protect the health or
environment of any population.
10-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
i"m. c. i" P«ipulatr»n "Vxff;; \«"Micerri
At an early stage of the analysis, analysts need to identify the population groups of concern relevant
to a specific regulatory context8 The concept of vulnerability can be used to help identify
population groups of concern.9 For example, analysts can combine available data on baseline
health, demographic, socioeconomic, or other relevant indicators (including those related to
cumulative impacts, historic inequities and systemic barriers, and lack of access) to identify
characteristics in affected communities that correlate with increased vulnerability to
environmental exposure or lack of opportunity for public participation (Fann et al. 2011).
While the EPA does not have rigid criteria for identifying population groups of concern, E.O.s 12898
and 14096 reference race, ethnicity, national origin, low-income, disability status, Tribal affiliated
and Indigenous populations, and those engaged in cultural or subsistence practices. Note that
population groups of concern may be clustered within specific communities or geographically
dispersed (e.g., unhoused populations, migrant workers). Underserved communities or populations
also may warrant consideration.10 See the EJ Technical Guidance (U.S. EPA 2024a) for more in-depth
discussion.
It may be useful in some contexts to analyze population groups in combination or to evaluate
additional aspects of diversity within a specific population group of concern (e.g., by life stage,
gender), particularly when some individuals within a population group may be at greater risk for
experiencing disproportionate and adverse effects (e.g., due to unique exposure pathways).
Analysts should rely on the Office of Management and Budget (OMB) or other federal statistical
agencies (e.g., U.S. Census Bureau), when available, to define relevant population groups (or
combinations thereof) for a specific regulatory action. Note that analysis of additional population
groups is not a substitute for examining the population groups explicitly mentioned in the E.O.s.
i". 1.2 Mc m reps of an EJ Analysis
Conducting a preliminary analysis may be a useful first step to identifying what level of assessment
is feasible and appropriate to support the regulatory action. In addition, it can help identify the
extent to which a regulatory action may raise EJ concerns that need further evaluation. Feasibility is
informed by a technical evaluation of available data and methods, including:
Scientific literature that discusses the effects of the stressor(s) being regulated on
population groups of concern;
Information received via public comments, technical reports, press releases, or other
documentation discussing the environmental and health effects of the stressor(s) being
regulated for population groups of concern, including information on other relevant
environmental or non-environmental stressors;
8 The term population groups of concern is used instead of the term subpopulations to include "population
groups that form a relatively fixed portion of the population (e.g., based on ethnicity)." See the EPA's Early Life
Stages website: http://www,epa.aov/children/early-life-staaes.
9 Note that specific terminology and definitions related to vulnerability may be provided by statute.
10 Examples of other characteristics that may be relevant in some regulatory contexts include linguistic
isolation, occupation, rurality, and employment status, among others.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Availability of spatially disaggregated data for population groups that may live, work, or
play in close proximity to the stressor(s) being regulated, or may otherwise be affected by
the stressor(s); or
Availability of methods for conducting in-depth analysis (e.g., proximity-based approach,
risk- or exposure-assessment, and mixed methods approach).
If the preliminary analysis reveals that the scientific literature and data are unavailable or of
insufficient quality to pursue an in-depth analysis that characterizes how exposure, risk, or health
effects are distributed across population groups, analysts are expected to explain why additional
analysis is not possible. In particular, analysts are encouraged to discuss relevant evidence, key
limitations, and sources of uncertainty highlighted in the published literature. Some impacts that
cannot be quantified may still represent important effects that should be considered in the analysis.
When conducting further evaluation is determined to be both feasible and appropriate, an EJ
analysis typically includes five main steps (Figure 10.1). We briefly describe each step below.
Figure 10.1 - Main Steps of an EJ Analysis
Identify
Regulated
Sources
Where are
regulated sources
located?
Do health and
environmental
risks vary with
source
characteristics?
Do the regulatory
options vary with
source
characteristics?
Describe
Environmental
Stressors
¦ Is the pollutant
spatially
distributed?
What is known
about fate and
transport?
What are the
relevant health
outcomes?
Characterize
Affected
Populations
¦ What factors
might drive
higher exposure?
What factors
might drive
differential
exposure?
How do health
outcomes vary by
population group
or community?
Compare Affected
and Compariosn
Populations
Given data and
methods, are you
only
characterizing the
baseline or also
regulatory
options?
How are you
presenting and
characterizing
results?
Conduct
Sensitivity
Analysis
¦ What are key
uncertainties and
limitations?
Are there specific
pockets of
concern?
What are they key
assumptions?
Questions to Ask at Each Step
How can meaningful engagement inform the EJ analysis?
What data are available and at what spatial scale?
What tools are available to model exposure and/or risk?
How are the effects distributed across sources and communities?
(1) Identify the sources being regulated: Before analysts can identify the populations and
communities being affected by a regulatory action, it is important to first characterize the regulated
sources: where are they located? Are there particular characteristics of the regulated sources that
contribute to higher exposure and/or risk of health effects? Do the regulatory options vary with
these characteristics? For instance, are some sources subject to greater stringency or other
regulatory requirements that would be important to account for in the EJ analysis?
(2) Describe the environmental stressor: The spatial distribution of health and welfare
outcomes is a relevant consideration for some regulatory actions. In these cases, evidence on the
fate and transport of the environmental stressor can help determine the populations and
10-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
communities potentially exposed. In other cases, the regulatory action's effects may be more
widespread. It is also important to understand which specific health effects are of greatest
relevance for a given regulatory context The benefits analysis and, when conducted, the human
health risk and exposure assessments can be important sources for this information.
(3) Characterize affected populations: It is important to understand what factors may contribute
to EJ concerns. How are individuals being exposed? Are there unique pathways or other factors that
drive higher exposures for some population groups? Recognizing underlying contributors within a
specific regulatory context is important for properly assessing EJ concerns and can aid in the design
of regulatory options. This may include evidence of already overburdened communities, including
the cumulative effects of exposure to multiple environmental or non-environmental stressors on
human health and well-being.
(4) Compare the affected and comparison groups: To answer each of the three analytic
questions, analysts need to characterize the exposure and risk of health effects for population
groups of concern in the baseline and for the regulated action relative to a comparison population
group (see Section 10.2.6). This allows analysts to gauge the extent to which effects for the affected
population are similar or different than they are for the comparison group and how they vary
across population groups.
(5) Conduct sensitivity analysis: Due to the inherent limitations and uncertainties associated with
analyses of EJ concerns, conducting sensitivity analysis around key assumptions is particularly
important for clearly communicating results to the public.
Figure 10.1 also identifies four overarching questions that are relevant throughout the EJ analytic
process:
1. How can meaningful engagement inform the EJ analysis?
Meaningful engagement can help analysts to identify and sometimes help fill information
and data needs. It can also help analysts to identify factors such as unique pathways or pre-
existing vulnerabilities that may contribute to exposure and/or risk for affected
populations. See U.S. EPA (2024a) and U.S. EPA (2024b) for more information.
2. What data are available and at what spatial scale?
The quality and availability of data are key determinants in the scope and complexity of the
EJ analysis. In some cases, analysts will have data at the individual level for the
environmental stressor being regulated, allowing for a detailed, rigorous analysis. In other
cases, analysts may need to rely on proxies for individual-level effects. Data relevant to the
EJ analysis may include, but are not limited to, the demographic and socioeconomic
characteristics of populations that may be exposed to environmental stressors from
regulated sources, what each regulated source is emitting or discharging, and pre-existing
health conditions or other environmental and non-environmental stressors that increase
the vulnerability and therefore the risk of experiencing a health effect for some population
groups.
3. What tools are available to model exposure and/or risk?
Analysts have a choice among several scientifically defensible methods to assess EJ
concerns associated with a regulatory action, including proximity-based analysis, exposure
and risk modeling, and combining qualitative and quantitative approaches. The choice of a
specific analytic method for the EJ analysis is often driven by data availability. Together, the
data and methods utilized directly influence what conclusions can be drawn regarding EJ
concerns for specific population groups or communities. Chapter 6 of the EJ Technical
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Guidance (U.S. EPA 2024a) discusses the main methods available and the potential
advantages and disadvantages of each in more detail.11
4. How are the effects distributed across sources and communities?
In some cases, extensive differences in effects among population groups of concern may
occur in only a few geographic locations. Referred to as hot spots, these locations are
typically exposed to localized concentrations of emissions from one or more sources along
with other stressors. In these cases, it may be appropriate to tailor the analysis to evaluate
effects in a few specific areas. Identifying the potential for hot spots early helps analysts
develop appropriate sources of data and analytic approaches, which may differ from those
used for a broader analysis (see Section 10.2.7.5).
i m. , hf- M?)i?)'- ndations for Analy - ? >~i EJ Ticerns
The EJ Technical Guidance (U.S. EPA 2024a) makes five overarching recommendations to ensure a
high-quality EJ analysis, while also recognizing the need for flexibility to reflect policy
considerations and technical challenges within a particular regulatory context. The
recommendations are intended to bring greater consistency across EJ analyses as they strive to
answer the three analytic questions but are not prescriptive and do not mandate the use of a
specific approach. Analysts should use their best professional judgement to decide on the type of
analysis that is feasible and appropriate within a specific regulatory context.
While these recommendations and best practices are intended as a starting point, they should not
be interpreted as limiting the scope of the EJ analysis. It is recommended that analysts thoughtfully
tailor their analysis to the rule context and incorporate new data and methods as they become
available. Ultimately, the EPA strives to innovate and improve upon EJ analyses as the state of
science continues to evolve. The five overarching recommendations are:
1. When risks, exposures, outcomes, or benefits of the regulatory action are quantified,
some level of quantitative EJ analysis is recommended.
a. Analysts should present information on estimated health and environmental risks,
exposures, outcomes, benefits, or other relevant effects disaggregated by race,
ethnicity, income, and other relevant demographic and socioeconomic categories
when feasible and appropriate.
b. When such data are not available, it may still be possible to evaluate potential risk
or exposure using other metrics (e.g., proximity to affected facilities, cancer or
asthma prevalence, or evidence of unique exposure pathways for specific population
groups) in a scientifically defensible way.
c. When health and environmental outcomes or benefits are not quantified or
disaggregated by race, ethnicity, income, or other relevant demographic and
socioeconomic categories, analysts should present available quantitative and/or
qualitative information that sheds light on EJ concerns that may arise.
11 For an overview of proximity-based analysis, including a discussion of various spatial analysis techniques
used in the literature, see also Chakraborty et al. (2011), Chakraborty and Maantay (2011), and Mohai and Saha
(2007).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
2. Analysts should integrate EJ into the planning of a risk assessment conducted for the
regulatory action.12
3. Analysts should strive to characterize the distribution of risks, exposures, or
outcomes within each population group, not just average effects.
a. In particular, analysts should pay attention to whether populations in the upper tail
of the distribution face the highest adverse risks, exposures or health effects.
4. Analysts should follow best practices appropriate to the analytic questions at hand
(see Text Box 10.1).
5. As relevant, analysts should consider any economic costs or challenges that may be
exacerbated by the regulatory action for relevant population groups of concern.
a. For instance, it may be appropriate to consider how low-income populations are
affected by price changes or to consider the distribution of economic costs (i.e.,
private and social costs) more broadly from an EJ perspective.
iM. haracterizing the Basf "ir 3to )tions for In-
Dep nalysis
The five main steps of an in-depth EJ analysis can be applied to characterize baseline conditions,
evaluate the effects of the regulatory options, and make comparisons between the two to evaluate
the three analytic questions, when data and methods allow.
The 0MB (2023) defines the baseline as "an analytically reasonable forecast of the way the world
would look absent the regulatory action being assessed, including any expected changes to current
conditions over time." It includes the characteristics of current populations and how they are
affected by pollutant(s) prior to the regulatory action under consideration. As the 0MB definition
implies, however, the baseline is not a static concept. In particular, the 0MB notes that analysts
may need to consider the evolution of the market, compliance with other regulations, and the
future effect of current government programs and policies, as well as other relevant external
factors to project future baseline conditions. As discussed in chapter 5, how future regulations or
policies affect the baseline specification is complex and requires consideration of many factors.
Ideally all potential influences on baseline conditions would be estimated, but it is generally not
practicable to do so. Anticipated changes in baseline demographic composition may also be
relevant in an EJ context. Per the recommendations in Section 10.2.2, the baseline for the EJ
analysis, including the geographic scope, year of analysis, and health and other effects, should be
consistent with how it is specified in other parts of the regulatory analysis.
12 For more information on this recommendation, see Chapter 5 of U.S. EPA (2024a).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 10.1 - Current Best Practices for Evaluating EJ Concerns
Use the best available science while relying on current, generally accepted Agency
procedures for conducting risk assessment and economic analysis,
Use existing frameworks and data from other parts of the regulatory analysis,
supplemented as appropriate.
Be consistent with the basic assumptions underlying other parts of the regulatory analysis,
such as using the same baseline and regulatory option scenarios.
Use the highest quality and most relevant data available. Discuss the overall quality and
main limitations of the data.
Identify relevant population groups of concern and discuss available evidence of factors
that make them vulnerable to adverse effects (e.g., unique pathways; cumulative exposure
to multiple stressors; behavioral, biological, or environmental factors).
Consider unique pathways for individuals that rely on cultural or subsistence practices and
relevance for Tribal or Indigenous populations, when practicable.
Carefully select and justify the choice of comparison population group.
Carefully select and justify the choice of the geographic unit of analysis and discuss any
challenges or aggregation issues related to the choice of spatial scale.
Analyze and compare effects in baseline and across policy scenarios to show differences in
effects.
Present summary metrics for each population group and the comparison population group
and characterize differences between them.
When data allow, characterize the distribution of risks, exposures, or outcomes within each
population group, not just average effects.
Disaggregate data to reveal important spatial differences (e.g., demographic information
for each source/place] when feasible and appropriate.
Clearly describe data sources, assumptions, analytic approaches, and results.
Summarize the main conclusions of differences in exposure or health risk between
analyzed population groups based on the available evidence.
Discuss key sources of uncertainty or potential data biases (e.g., sample size, proximity as a
surrogate for exposure) and how they may influence the results.
When possible, conduct sensitivity analysis for key assumptions or parameters that may
affect findings.
Qualitatively describe behavioral responses not accounted for in the analysis that could
affect the level or distribution of exposure or health risks (e.g., dynamic spatial or temporal
effects, averting or adaptive behavior).
Make elements of the Ej analysis as straightforward and easy for the public to understand
as possible.
When data and methods allow, an EJ analysis can also examine the distribution of effects for each
regulatory option - different configurations of the regulatory action being considered. This analysis
is based on a prediction of the state of the world under the regulatory options. For the analysis of E]
concerns, analysts are encouraged to examine how the exposure, risk of health or environmental
effects, or other outcomes of the regulatory action are distributed across population groups for the
regulatory options being considered, where practicable. The EJ analysis can then evaluate the
change in the exposure or risk of relevant environmental and health effects for each regulatory
option compared to the baseline. In addition to identifying whether the regulatory action is
10-9 Guidelines for Preparing Economic Analyses | 3rd edition
-------
expected to exacerbate, mitigate, or leave baseline EJ concerns unchanged, the analysis should shed
light on the extent and distribution of these changes.
With these three sets of information - effects in the baseline, effects under the regulatory options,
and a comparison of the two -analysts can characterize the distribution of environmental and
health effects associated with a regulatory action, thus answering all three EJ analytic questions.
Note that a constant reduction in risk or exposure across population groups will likely not mitigate
EJ concerns if there are differences in baseline environmental quality or health risk across
population groups or communities (Maguire and Sheriff 2011). Conceptually, an EJ concern is only
completely mitigated when there is no difference in the distribution of effects across population
groups for the regulatory options being considered - i.e., everyone is experiencing the same
environmental quality or health risk post-regulation.
i ^ >. viontoAssf I i . icerns
In general, the type of analysis that can be conducted depends on the availability and quality of
data. In some cases, spatially resolved, individual-level data may be most appropriate and relevant
for an analysis of EJ concerns. In other cases, distance from a regulated source may be the best
available metric. At times, the best available information may be qualitative, including local
knowledge from affected communities and Tribes (e.g., Indigenous Knowledge, also referred to as
Traditional Ecological Knowledge). In all cases, analysts should use the highest quality and most
relevant data and information.
When data are missing or incomplete, it is recommended that analysts document what specific
types of data are unavailable or of insufficient quality, including but not limited to cases where the
data are available but not of the desired granularity (spatially or temporally) and/or available for
only subsets of the population. Text Box 10.2 illustrates how data quality may affect the level of
analysis.
Recognizing the importance of data quality, data needed to conduct an EJ analysis may include:
o Demographic and socioeconomic characteristics (e.g., race, ethnicity, income);
o Location of pollution sources (e.g., latitude/longitude coordinates, zip code, county);
o Historical, current, and projected emissions or concentrations of stressor(s) relevant to the
regulatory action;
o Prevalence of specific exposure pathways that may increase risk for some population
groups;
o Health effects (e.g., hospital and emergency admissions, race and ethnicity-stratified
mortality rates, race and ethnicity-stratified asthma or other morbidity rates);
o Other environmental or non-environmental stressors that may be risk- or effect modifiers
(e.g., indoor air concentrations, vulnerability to effects of climate change);
o Risk coefficients stratified by population groups of concern (e.g., race, ethnicity, income);
and
o Distribution of economic costs, when relevant (see Section 10.x).
)
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Text Box 10.2 - Data Quality and Spatial Resolution in the Context of Air
Quality Regulations
Analysts' ability to address how a regulatory action changes the distribution of risk across
population groups depends on the quality and spatial resolution of the data available. Finer-scale
air quality, health, and demographic data allow one to assess the distribution of effects across
population groups and to have greater confidence in the conclusions drawn from these data.
When air quality data are lacking or only available at a coarse level, the ability to assess change in
risk across populations and other conclusions is more limited.
An example in limited data environments: Using race-stratified county-level mortality and
morbidity data, analysts can calculate population-weighted mortality rates by county. Analysts
can then use a highly aggregated baseline air quality modeling projection (e.g., 12 or 36 km) to
identify population groups most exposed to air pollution. Using geospatial tools, it is possible to
combine the two sources of data. The coarse geographic scale of air quality information may
inhibit the analyst's ability to detect meaningful differences in effects among and between
groups. When risk coefficients are unavailable, it is not possible to estimate health effects
separately for each population group.
An example in data-rich environments: Using finely resolved air quality data, analysts can identify
at a highly disaggregated level (e.g., 1 km) population groups that experience the highest
exposure to air pollution. Analysts can also identify population groups that exhibit the highest
baseline incidence or prevalence rates for air pollution health effects. Using geospatial tools,
analysts can spatially combine the two data sources. Using race-specific or standard risk
coefficients analysts can then estimate health effects for each population group.
Three types of information are frequently used as inputs into an EJ analysis: demographic and
socioeconomic data, emission data for regulated sources, and data on pre-existing health conditions
or other factors that may increase an individual's vulnerability when exposed to releases from
regulated sources. The U.S. Census Bureau is the recommended source for demographic and
socioeconomic data in an EJ analysis. It produces several national-level data products that report
demographic and socioeconomic characteristics at relatively fine spatial scales (e.g., census tract,
block group), including the decennial Census, the American Community Survey (ACS), and the
American Housing Survey (AHS). See Section 6.3 of the EJ Technical Guidance (U.S. EPA 2024a) for
a detailed discussion of these and other sources of data and information to assess EJ concerns.
10.2.6 Analytic Methods
A variety of scientifically defensible methods can be used to assess EJ concerns associated with
regulatory actions. The choice of analytic method is most often driven by data availability. Analysts
may also rely on a combination of methods when analyzing a regulatory action. The conclusions
that can be drawn from the analysis will vary depending on the method used. See the EJ Technical
Guidance (U.S. EPA, 2024a) for a discussion of specific analytic methods and their relative
advantages and limitations when evaluating EJ concerns.
Considerable uncertainty may exist about key relationships and health outcomes, such as how a
reduction in emissions or other types of releases from a given source translates into ambient
environmental quality and how it, in turn, translates into the human health effects of interest This
is particularly problematic if uncertainties differ across population groups. For instance, if an
overexposed population group is more responsive to exposure (i.e., individuals in the group
10-11 Guidelines for Preparing Economic Analyses | 3rd edition
-------
experience greater adverse health effects per unit of exposure), then using exposure alone as a
proxy will underestimate the health risk posed by a stressor to that group. On the other hand, if
proximity to a pollutant source does not correlate with exposure, it could overstate potential
differences in health risk. Analysts should select the method that is most appropriate for the
available data, recognizing time and resource constraints.
Regardless of the analytic approach used, the EJ analysis should be presented in a transparent way
and include the following:
o Information about the specific population groups and individuals affected by the regulatory
action;
o Main exposure pathways and expected health and environmental outcomes;
o Evidence for why risk, exposure, or outcomes may vary by population group, including the
role of other relevant environmental and non-environmental stressors;
o Relevant geographic scale;
o Descriptions of the main methods of analysis used;
o Descriptions of key data or modeling assumptions;
o Summary statistics for the baseline and each regulatory option (both the mean and
distribution) by population group;
o An easy-to-understand description of what the summary statistics show;
o Conclusions based on the information available;
o Sensitivity analysis to examine the robustness of results across options presented; and
o Data quality, key sources of uncertainty, and limitations that affect conclusions regarding
potential differential effects.
Analysts should follow best practices appropriate to the questions under consideration (see Text
Box 10.1). If it is not feasible to follow a particular best practice, analysts should explain why this is
the case.
iM. * iialytic Considerations
Regardless of the analytic approach taken, analysts make a number of key decisions that can have a
substantial effect on the results of the analysis, including: the geographic and temporal scope of the
analysis; how to specify the comparison population group; how to spatially identify and aggregate
effects across affected and unaffected populations; whether to conduct analysis from a community
and/or facility perspective; and how to evaluate underlying variability, including the potential for
hotspots.
An important general strategy in analyzing EJ concerns is the use of sensitivity analysis. Due to the
uncertainties associated with the analytic decisions discussed below, sensitivity analysis around
key assumptions is often critical for clearly communicating results to the public.
ivt . " i - ic and I - nip i »l ope
The geographic scope of analysis for an EPA regulatory action is often the entire United States since
requirements typically apply nationwide. However, in some cases the effects of a regulatory action
are expected to be concentrated in specific regions or states. In such cases, it may make sense to
analyze and present differences in health and environmental outcomes across population groups at
both a national and a sub-national level. Because the geographic scope can affect the results of the
analysis (Baden et al., 2007), analysts should make certain that the scope is relevant for the
Guidelines for Preparing Economic Analyses | 3rd edition
-------
regulatory action under consideration. In addition, it is important to keep in mind that differences
in health and environmental outcomes in one region or state may not necessarily hold in other
regions or states.
It may be important to evaluate regulatory action effects on both shorter and longer time horizons.
For instance, while a regulatory action may result in near-term reductions in emissions, changes in
health and other risks may occur on a longer timeframe. In some cases, effects may even be felt
intergenerationally (e.g., climate change) and the analysis may accordingly extend beyond the
current generation to include a robust discussion of far-future health effects and costs. In general,
the period of time over which the analysis is conducted should also be consistent with other parts
of the regulatory analysis.
The scope of the analysis should generally match the scope used in other parts of the regulatory
analysis (e.g., benefit-cost analysis). However, in some situations, using a different time horizon or
spatial scale may be appropriate when considering EJ. For example, phasing in of regulatory
requirements or relocation of polluting activities in response to the regulatory action may result in
EJ concerns due to effects that occur on a time horizon or spatial scale that differs from other effects
considered in the regulatory analysis. If such situations arise, analysts should clearly articulate the
reasons for considering an alternative time horizon.
Another aspect of characterizing temporal scope is adequately anticipating the long run dynamic
effects of a regulatory action (Cain et al., 2024). The literature uses spatial sorting models to
examine how regulations may affect residential location choice but typically focuses on a specific
city or region (e.g., Kuminoff etal., 2015; Redding and Rossi-Hansberg, 2017).1314 Spatial sorting
can occur when improved environmental quality is capitalized into housing values, attracting
higher-income households and shifting renters and lower-income households to less expensive
neighborhoods with lower environmental quality (Melstrom and Mohammadi, 2022). On the other
hand, some residents may be more likely to move into high-risk zones due to differences in housing
prices (Bakkensen and Ma, 2020). Given the challenges of modeling these types of effects on a
national scale, it is recommended that analysts qualitatively discuss possible household responses
based on the available literature, while acknowledging the limitations of the analysis.
tvt I " I1 ,n>|MriSi-,[i Pi-«pulati [] v [ 'fjp
To evaluate differences in effects for population groups of concern, results need to be presented
relative to another group, typically referred to as a comparison population group. How the
comparison population group is selected has important implications for evaluating differences in
health, risk, or exposure effects across population groups of concern. It is possible to define the
comparison population group as individuals with similar socioeconomic characteristics in areas of
the state, region, or nation unaffected by the regulatory action (i.e., within-group comparison) or as
individuals with different socioeconomic characteristics within the affected areas (i.e., across-group
comparison).
13 One exception is Fan, et al (2018). They link spatial sorting and economy-wide models of the United States to
explore where people migrate in response to increased risk of extreme temperatures, while accounting for wage
and housing price feedbacks.
14 Likewise, while hedonic price methods may be useful for demonstrating how changes in environmental
quality factor into housing prices, predicting the effect of such price changes on household migration by race or
income may be infeasible.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Analysts should aim to define the comparison population group for an across-group comparison as
similar as possible to the population group of concern, but without the socioeconomic
characteristics defining the group of concern. For example, analysts could compare the proportion
of low-income households within areas affected by the regulatory action to the proportion of non-
low-income households within the same affected areas. If analysts have fate-and-transport
information on emissions, they can compare the average concentrations faced by low-income
households within the affected areas to those faced by non-low-income households living in the
same areas. Thus, the results from an across-group comparison indicate how the likelihood of risk
or exposure within the affected areas varies with demographic and socioeconomic characteristics.
A within-group comparison compares the likelihood of risk or exposure for a specific demographic
or socioeconomic group in affected areas to the likelihood of risk or exposure for that same
demographic or socioeconomic group elsewhere. Again, analysts should aim for the comparison
group to be as similar as possible to the population group of concern but without the risk or
exposure of interest For example, analysts can compare the proportion of low-income households
within areas affected by the regulatory action to the proportion of low-income households in
unaffected areas. Similarly, if analysts have information on the fate and transport of emissions, they
can compare the average concentrations faced by low-income households within the affected areas
to those faced by low-income households living in areas unaffected by the regulatory action.
If a regulatory action is expected to differentially affect populations within a given area (e.g.,
communities living near regulated facilities or in a specific region), then a combination of within-
and across-group comparisons can demonstrate whether there are differences between specific
population groups of concern and the general population. In some contexts, it may make sense to
define the comparison population group at a sub-national level to reflect differences in
socioeconomic composition across geographic regions. See Section 6.5 of the EJ Technical Guidance
(U.S. EPA, 2024a) for a more discussion of selecting the appropriate comparison group.
pt y tial Identificati n -nd Aggregating Effects
The spatial distribution of health and welfare outcomes is a relevant consideration for some
regulatory actions, such as those that reduce emissions from point sources that have fairly localized
effects or when there is a differential distribution of associated health or environmental effects. In
other cases, the regulatory action's effects may be more widespread, and spatial distribution is less
relevant (e.g., when exposure to a chemical substance depends on its purchase, use, transport, or
disposal).
When exposures, risks, or human health effects are spatially distributed, analysts need to determine
how to spatially identify affected and unaffected populations. The nature of the stressor(s) should
guide analysts' choices of the geographic area of analysis. Some air pollutants, for example, may be
emitted out of tall stacks and travel long distances, affecting individuals hundreds of miles away
from the sources and thereby making it appropriate to choose a relatively large geographic area. In
contrast, water pollutants or waste facilities may have more localized effects, making it appropriate
to select relatively small areas for analysis. Likewise, an assessment of local effects from point
sources - including possible traffic, odors, and noise implications from changes in production - may
call for more spatially resolved data than those that affect regional air quality.
Complications can arise when the spatial resolution of the analysis is either too refined or too
coarse. See the EJ Technical Guidance (U.S. EPA, 2024a) for a discussion of these challenges.
Analysts are encouraged to discuss the approach used to create buffers and aggregate geospatial
data, as what is most appropriate will vary with the stressor(s) affected and data used in the
Guidelines for Preparing Economic Analyses | 3rd edition
-------
analysis, and to provide a transparent justification of their choice. In some cases, it may be helpful
to consider multiple buffers to evaluate the effects of a regulatory action, for instance, because of
uncertainty regarding fate and transport of a specific environmental stressor or because the
regulatory action affects environmental stressors that travel different distances.
pt if u iin Cf'iiuiM.M K'.- ;ed Perspectives
Exposure to other environmental and non-environmental stressors can increase the vulnerability of
individuals or population groups to negative health effects from exposure to a specific
environmental hazard. While explicit modeling of these interactions is often not feasible, analysts
can shed light on this issue by evaluating and presenting results using not only a facility, but also a
community-based perspective.
An analysis with a facility-based perspective primarily considers who may be exposed to sources
regulated by the specific action under consideration. For example, such an analysis would examine
proximity, emissions, concentrations, or risk associated with each regulated source in conjunction
with the demographic and socioeconomic characteristics of those most likely to be exposed.
However, communities may be affected by multiple sources of pollution relevant to characterizing
risk for a specific regulatory action. An analysis that takes a community-based perspective
considers proximity, emissions, concentrations, or risk to a given community from multiple nearby
sources of pollution to which individuals are exposed, accounting for the possibility that certain
communities face increased vulnerability due to a greater number of nearby pollution sources.
pt - f »b' itin^ f m erlying Variability and Identifying t tential Hot
Spots
In addition to presenting aggregate results for population groups of concern affected by the
regulatory action, it is important to understand the extent to which there are heterogeneous effects,
both within specific population groups as well as across communities, given that communities often
vary widely in the risks they face from the affected sources as well as from other environmental and
non-environmental stressors. When data allow, analysts should characterize the distribution of
risks, exposures, or outcomes within each population group of concern, not just average effects,
with particular attention paid to the characteristics of populations at higher risk of exposure. When
relying on proximity-based analysis, differentiating results by key facility characteristics that may
be correlated with risk (e.g., plant age, capacity, production levels, accident history, types of
chemicals stored on site) can be useful.
It is also important to evaluate the potential for hot spots, with particular attention paid to the
communities in the upper end of the distribution of exposure or risk. Hot spots refer to geographic
areas with higher levels of localized concentrations of emissions from one or more sources within a
larger geographic area with more "normal" environmental quality. Hot spots may result from
baseline conditions, such as exposure to other pre-existing stressors within the community. It is
also possible that hot spots may be created, exacerbated, or mitigated following a regulatory action.
Relevant issues to consider may include proximity to multiple sources of pollution, specific
exposure pathways, and other drivers of increased vulnerability. Qualitative or other sources of
data may also help to identify specific population groups or communities where a more detailed
analysis is warranted.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
iM. vr ' laracteri :ii ;" - nalytic Results
Once an EJ analysis has been conducted, analysts face choices about how to characterize and
communicate the results. Analysts need to present summary metrics for relevant population groups
of concern and the comparison population group and characterize the differences between them.
This section discusses the way in which information from the analysis can be summarized and
presented, including the choice of summary metrics, ways of displaying the results in tables, maps
or other visual displays, and the distinction between statistical and policy significance when
interpreting results.
I. f ['Tories
Simple summary measures can be used to characterize the distribution of health and
environmental effects in the baseline and for regulatory options relative to appropriate comparison
population groups. Analysts should consider characterizing results of the EJ analysis using more
than one type of summary metric to provide a richer picture of potential effects. For instance,
relative ratios can facilitate comparisons across groups or locations because all ratios are in
common units. However, without presenting information on the absolute levels of risk or exposure,
it is not possible to determine if either group is at risk of experiencing a potential health effect
Analysts should also present information that communicates underlying heterogeneity in the data,
such as the degree of spread in the data relative to the mean (i.e., standard deviation).
Counts of the number of sources or geographic areas where the percent of a specific population
group living nearby exceeds a particular threshold (e.g., the state/national average or a specific
percentile) are not recommended. Counts are hard to interpret because they do not account for
differences in population size or density across geographic areas. It is more informative to display
metrics that characterize the full population or risk distribution to understand the extent to which
affected communities differ from the comparison group. See Section 6.6 of the EJ Technical Guidance
(U.S. EPA, 2024a) for more discussion of summary metrics.
.8.2 Displaying Results Visually
Tables, maps, and other visual displays help communicate a large amount of information in an
organized way to facilitate comparisons, convey results, and support discussion. Careful thought
should go into how information is presented, particularly when there are:
o Multiple comparison groups (e.g., state, U.S., rural areas);
o Different types of effects (e.g., pollutants, health effects, or other environmental metrics);
o Multiple categories of regulated facilities or types of sources;
o Many individual sources;
o Clustering of sources in specific geographic areas;
o Multiple scenarios (e.g., baseline, multiple regulatory options); or
o Sensitivity analysis around key analytic assumptions (e.g., buffer distance).
Analysts need to clearly explain how to interpret the information presented in tables, maps, or
figures to properly contextualize results and guard against erroneous conclusions (e.g., a large
percentage change from a small baseline value may not be a large change in absolute terms). Often
more than one table is needed to present results. In addition, holding or shading specific cells can
ease navigation of a dense table of results. Table 10.1 illustrates how results for multiple types of
Guidelines for Preparing Economic Analyses | 3rd edition
-------
sources and several distance buffers can be presented within a single table. This example also uses
shading to indicate values above the national average.
Visually displaying information in maps or figures can also help demonstrate how sources, risks,
and exposures are geographically distributed across population groups, including baseline
conditions and spatial clustering of sources. Note that it can be difficult to visually discern
differences between baseline and regulatory options in maps or figures unless differences are
large.15 However, differences not discernible on a map may still be important.
Additionally, it is important to consider how visual indicators can be used to characterize data
uncertainty, how geographic boundaries relevant to the analytic context (e.g., watershed, state, or
Tribal lands) are identified on a map, and how to select appropriate intervals for visually
representing the distribution of the data. For this reason, visual displays are only suggestive of
potential effects and should be accompanied by tables or other graphics that allow the reader to
access the underlying statistical information.
.8.3 Statistical Significance and Other Considerations
Tests of statistical significance can be used to examine whether the difference between the mean
values of two groups is due to factors other than chance. This can be done for a pairwise
comparison, which does not control for other factors, or via a regression approach, which allows
analysts to assess the relationship between two variables while controlling for other factors.
It may also be useful to examine parts of the distribution further from the mean (e.g., quantile
approaches) or to use approaches that can account for outliers, skewness or heteroscedasticity
(varying levels of spread) in the data. Note that the ability to test statistical significance is
predicated on having a sufficiently large sample size and, for parametric approaches, an assumed
distribution (e.g., normality).
It is important to understand that a statistical difference does not necessarily indicate that the
difference is meaningful from a policy perspective. For instance, analysts may find that low-income
households are more likely to be located near a pollution source than wealthier households, and
that this effect is statistically significant (i.e., the effect is statistically distinguishable from zero and
not due to sampling error). However, the difference in likelihood between these types of
households could still be quite small in magnitude. Analysts need to examine what the difference
implies (e.g., how different poverty is across geographic areas), and summarize those differences in
a manner appropriate for policy relevance.
15 For an overview of general mapping best practices to communicate EJ concerns, such as selecting a
projection, avoiding unintentional misrepresentation, and choosing a color scale to represent values, seeStieb et
al. (2019).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Table 10.1 - Example Summary Table for Proximity-Based Analysis Results
Race
Population
within 1 Mile
of Sites with
Legacy CCR
Sis
Population
within 3 Miles
of Sites with
Legacy CCR Sis
Population
within 1 Mile
of Sites with
CCRMUs
Population
within 3 Miles
of Sites with
CCRMUs
U.S.
Population
Asian
10.36%
4.66%
2.37%
2.82%
5.64%
Black or African
American
13.47%
17.03%
8.37%
13.73%
12.53%
Native Hawaiian/Pacific
Islander
.03%
.08%
.06%
.07%
.18%
Native American or
Alaskan Native
.79%
.93%
.86%
.78%
.82%
Other
16.65%
15.82%
20.63%
18.94%
12.82%
White
58.47%
61.47%
67.71%
63.67%
68.01%
Ethnicity
Population
within 1 Mile
of Sites with
Legacy CCR
Sis
Population
within 3 Miles
of Sites with
Legacy CCR Sis
Population
within 1 Mile
of Sites with
CCRMUs
Population
within 3 Miles
of Sites with
CCRMUs
U.S.
Population
Hispanic (any race")
26.27%
22.0%
32.61%
27.02%
19.24%
People of Color
Population
within 1 Mile
of Sites with
Legacy CCR
Sis
Population
within 3 Miles
of Sites with
Legacy CCR Sis
Population
within 1 Mile
of Sites with
CCRMUs
Population
within 3 Miles
of Sites with
CCRMUs
U.S.
Population
People of Color
52.42%
46.31%
46.70%
46.59%
41.14%
Poverty Level
Population
within 1 Mile
of Sites with
Legacy CCR
Sis
Population
within 3 Miles
of Sites with
Legacy CCR Sis
Population
within 1 Mile
of Sites with
CCRMUs
Population
within 3 Miles
of Sites with
CCRMUs
U.S.
Population
Households below the
poverty level
16.21%
16.11%
14.94%
14.9%
12.71%
Other
Sociodemographic
Factors
Population
within 1 Mile
of Sites with
Legacy CCR
Sis
Population
within 3 Miles
of Sites with
Legacy CCR Sis
Population
within 1 Mile
of Sites with
CCRMUs
Population
within 3 Miles
of Sites with
CCRMUs
U.S.
Population
Linguistically isolated
households
9.23%
5.42%
9.38%
6.05%
4.84%
Less than a high school
diploma
17.60%
14.02%
17.11%
15.57%
11.24%
Person with disability
15.53%
15.61%
14.66%
15.23%
12.70%
Source: Table 6-9. Estimated Percent of Key Sociodemographic Indicators Near Legacy CCR Surface
Impoundment (SI) and CCR Management Unit (MU) Sites (U.S. EPA, 2024c).
Finally, it is important to address and characterize uncertainty. Point estimates alone do not
provide information about whether estimates are robust to alternate assumptions, nor can they
convey the full range of potential outcomes. When statistical analysis is used, information such as
confidence intervals and variance should be presented. Sensitivity analysis can also play a role in
10-18 Guidelines for Preparing Economic Analyses | 3rd edition
-------
understanding the robustness of outcomes to key assumptions. Where the analysis is sensitive to
the choice of model or method used, this uncertainty should also be described. Uncertainty can also
be discussed by highlighting limitations in the literature, identifying caveats associated with results,
or highlighting gaps in the data. See Section 6.6 of the EJ Technical Guidance (U.S. EPA, 2024a) for
additional discussion.
iM.i , v ssing the Distribution vi" * n\? .-nv-:' vi.er Effects
This section addresses when it may be appropriate to evaluate how economic costs or challenges
are distributed across population groups, how compliance and enforcement may vary across
regulatory options under consideration, and the evaluation of non-health effects. We refer to costs
as defined Chapter 8.
I. . f d istribution of Economic Costs
Certain directives (e.g., E.0.13175, E.0.14008, and OMB Circular A-4) identify the distribution of
economic costs or challenges as an important consideration in regulatory analysis. The economics
literature also typically considers both costs and benefits when evaluating distributional
consequences of an environmental policy to understand its net effects. In the context of EJ, the
distribution of health or environment effects alone might convey an incomplete - and potentially
biased - picture of the overall burden faced by population groups of concern. For instance, if costs
are unevenly distributed such that low-income households bear a larger relative share, it is possible
that they may experience net costs even after accounting for environmental improvements.
Fullerton (2011) discusses six possible types of distributional effects that may result from an
environmental policy: costs to consumers via change in relative product prices; cost to producers or
factors of production via changes in the relative returns to capital and labor; the distribution of
scarcity rents (i.e., excess benefits due to restricted nature of a good, such as pollution permits); the
distribution of environmental quality improvements; temporary costs of adjustment and transition
(e.g., for capital and labor); and the capitalization of environmental improvements into asset prices
(e.g., land or housing values). That said, the consideration of economic costs in an EJ context may be
challenging, given a lack of data and methods in many instances.
Whether to undertake an analysis of economic costs as it pertains to EJ is a case-by-case
determination. It will depend on the relevance of the information for the regulatory decision at
hand, the likelihood that economic costs of the regulatory action will be concentrated among
particular types of households, and the availability of data and methods to conduct the analysis.16
Analysts should coordinate with economists from the Office of Policy when evaluating the potential
relevance of economic costs for EJ and the degree to which they can be discussed or analyzed.
In many cases, analysis of economic costs from an EJ perspective will not substantially alter the
assessment of distributional effects for population groups of concern. For instance, often the costs
of regulatory action are passed onto consumers as changes in prices or wages that are spread fairly
evenly across many households. When these price changes are small, the effect on an individual
16 Note that there may be other effects of a regulatory action (e.g., employment) beyond direct compliance and
social costs but understanding how all effects vary across population groups may not be feasible. For example,
data on the distribution of changes in employment across low-income households may be difficult to assess. See
Chapter 9.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
household also will likely be relatively small. In this case, further analysis is unlikely to yield
additional insights.
However, in some circumstances further exploration of the distribution of economic costs may offer
substantial insight because costs are expected to differentially affect specific population groups. For
example, further analysis may be warranted when costs to comply with the regulatory action
represent a noticeably higher proportion of income for some population groups; when some
population groups are less able to adapt to or substitute away from goods or services with now
higher prices; when changes in environmental quality or health and costs are likely to accrue to the
same set of individuals; when costs are concentrated on some types of households (e.g., renters) or
in specific geographic areas; when there are identifiable plant closures in or relocation of facilities
away from or into communities in which population groups of concern reside and work; or when
behavioral changes in response to the costs of the regulatory action leave population groups of
concern less protected than other groups.
While the Agency continues to investigate ways to improve incorporation of economic costs into an
analysis of EJ concerns, it recognizes that, even in cases where the information is relevant, data or
methods may not exist for full examination of the distributional implications of costs. In these
instances, the issue can be qualitatively discussed, and the limitations and assumptions associated
with characterizing costs explained. See the EJ Technical Guidance (U.S. EPA, 2024a) for further
discussion.
.9.2 Consid iiir C mplUnce an iCm 'n nt
Evidence suggests that compliance with environmental regulations can vary widely across sources
in ways that exacerbate pre-existing disparities (e.g., Allaire et al. 2018; Balazs et al. 2012; Fedinick
et al. 2019; McDonald and Jones, 2018). Analysts may want to consider whether regulated sources
have a history of significant non-compliance or enforcement actions taken against them under
various statutes or how capacity for monitoring and enforcement may differ across communities,
including those on Tribal lands. Past compliance issues may indicate pre-existing EJ concerns that
warrant further investigation.17
Analysts are encouraged to consider differences in compliance and ease of enforcement across
regulatory options in the EJ analysis. When there are pre-existing differences in risk or exposure,
options consistent with applicable law that improve monitoring coverage or encourage compliance
can reduce exposure in communities with EJ concerns (e.g., enhanced reporting requirements for
higher risk sources). Collecting, processing, and making publicly available real-time monitoring or
remotely sensed data may also be effective for enhancing public awareness and participation (U.S.
EPA 2021).
1,9,3 Gtlici C nsiderations
The distribution of non-health effects associated with environmental stressors affected by the
regulatory action may also be important to consider. For instance, certain population groups may
place a higher value on a cultural resource (e.g., spiritual or sacred sites). If a regulatory option
affects those resources, then the groups with a higher value will experience a different effect than
17 There is also a literature that explores whether the intensity of enforcement activities for environmental
regulations varies with demographics such as race and income (Konisky et al. 2021; Shadbegian and Gray 2012).
) Guidelines for Preparing Economic Analyses | 3rd edition
-------
groups that do not place a value on the cultural resource. Likewise, some regulatory options may
differentially affect access to specific recreational activities for some population groups.
Quantifying changes in non-health effects may be challenging. Often, data on the distribution of
baseline conditions for non-health effects are not easily available or are difficult to quantify, and/or
are not suitable for analyzing the effects of a regulatory action. For instance, data on some
ecosystem services (e.g., cultural uses of specific ecosystems) in the United States are quite limited
in availability compared to baseline health data, such as mortality incidence. Likewise, data and
models to assess how various regulatory options affect non-health related endpoints may not be
available.
10.3 Environmental Health for Children and Older Adults
Analysis may shed light on differential effects of regulation on children and older adults, both of
which are life-stage defined groups characterized by a multitude of unique behavioral, physiological
and anatomical attributes. EO 13045 requires that each federal agency address disproportionate
health risks to children. In addition, EPA's Children's Health Policy (U.S. EPA 1995) requires the
Agency to "consider the risks to infants and children consistently and explicitly as a part of risk
assessments generated during its decision-making process, including the setting of standards to
protect public health and the environment"18
There are two sets of important differences between children and adults regarding health effects.
First, there are differences in exposure to pollutants and in the nature and magnitude of health
effects resulting from the exposure. Children may be more vulnerable to environmental exposures
than adults because their bodily systems are still developing; they eat, drink and breathe more in
proportion to their body size; their metabolism may be significantly different especially shortly
after birth; and their behavior can expose them more to chemicals and organisms (e.g., crawling
leads to greater contact with contaminated surfaces, while hand-to-mouth and object-to-mouth
contact is much greater for toddler age children). In addition, since children are younger, they have
more time to suffer adverse health effects from exposure to contaminants. Second, individuals may
systematically place a different economic value on reducing health risks to children than on
reducing such risks to adults. In part this is because children cannot provide marginal willingness
to pay values for their own risk reductions, unlike adults, so children's health risk valuation
necessarily requires some model, implicit or explicit, about household decision making. These
models differ in their implications for valuation. The perceived or actual effects of a given health
outcome, too, may differ across children and adults. Empirical evidence also suggests that parents
value a given risk reduction to themselves differently than to their children, with willingness to pay
(WTP) for own risks generally valued less than those for children.19
Older adults also may be more susceptible to adverse effects of environmental contaminants due to
differential exposures arising from physiological and behavioral changes with age, disease status
and drug interactions, as well as the body's decreased capacity to defend against toxic stressors.
18 See https://www.epa.aov/children/epas-policy-evaluatina-risk-children for the original 1995 policy and the
2013 and 2018 reaffirmation memos.
19 See Gerking and Dickie (2013) for a review of both household decision making models for children's health
risk valuation and the empirical literature. U.S. EPA (2003) provides an overview of children's health valuation
issues in applied analysis.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Generally, many of the approaches described earlier in this chapter to characterize the distribution
of impacts may be adapted to evaluate environmental health risks by life stage.20 For example,
when proximity-based analysis is appropriate for evaluating EJ impacts, it might also be used to
examine whether children or older adults are disproportionately located near facilities of concern.
In such a case, the considerations described earlier about geography, defining the baseline and
comparison groups, and use of summary statistics would all apply.
1". ' -1 ~ ge I if- rvie
Evaluating the impacts of regulatory actions on children or older adults differs in an important way
from evaluating the same impacts on population groups of concern for EJ. For instance, when the
EPA evaluates disproportionate health risk impacts from environmental contaminants, it views
childhood as a sequence of life stages from conception through fetal development, infancy and
adolescence, rather than a distinct "subpopulation."
Use of the term "subpopulation" is ingrained in both EPA's past practices as well as various laws
that the EPA administers such as the Safe Drinking Water Act Amendments. Prior to publication of
revised risk assessment guidelines in 2005, the EPA described all groups of individuals as
"subpopulations." In the 2005 guidelines, the Agency recognizes the importance of distinguishing
between groups that form a relatively fixed portion of the population, such as those described in
Section 10.2, and life stages or age groups that are dynamic groups drawing from the entire
population.
The term "life stage" refers to a distinguishable time frame in an individual's life characterized by
unique and relatively stable behavioral and/or physiological characteristics associated with
development and growth. Since 2005, the EPA has characterized childhood as a life stage.21
iM-,;..: Analytical Considerations
Assessing the consequences of policies that affect the health of children or older adults requires
considerations that span risk assessment, action development and economic analysis. In each case,
existing Agency documents can assist in the evaluation.
ivt ' . i IV I id L" i1 mi- w nrfiit
Effects of pollution may differ depending upon age of exposure. Analysis of potentially
disproportionate impacts begins with health risk assessment but also includes exposure
assessment Many risk guidance and related documents address how to consider children and older
adults in risk and exposure assessment
20 In principle there is a potential distinction between factors that are fixed, such as race and sex, and those
defined by lifestages. The latter raises the possibility, at least, of examining effects through the lens of differences
in lifetime utility or well-being rather than focusing on a single life stage. See Adler (2008) for one proposal
consistent with this approach.
21 The 2005 Risk Assessment Guidelines "view childhood as a sequence of lifestages rather than viewing children
as a subpopulation, the distinction being that a subpopulation refers to a portion of the population, whereas a
life stage is inclusive of the entire population." (U.S. EPA 2005a).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
A general approach to considering children and childhood life stages in risk assessment is found in
A Framework for Assessing Health Risks of Environmental Exposures to Children (U.S. EPA 2006a).
The framework identifies existing guidance, guidelines and policy papers that relate to children's
health risk assessment It emphasizes the importance of an iterative approach between hazard,
dose response and exposure analyses. In addition, it includes a discussion of principles for weight-
of-evidence consideration that is, the critical evaluation of available and relevant data across
life stages.
EPA's 2005 Guidelines for Carcinogenic Risk Assessment (Cancer Guidelines) (U.S. EPA 2005a)
explicitly call for consideration of possible sensitive subpopulations and/or lifestages such as
childhood. The Cancer Guidelines were augmented by Supplemental Guidance for Assessing
Susceptibility from Early-Life Exposure to Carcinogens (U.S. EPA 2005b). Recommendations from
this supplement include calculating risks utilizing life stage-specific potency adjustments in
addition to life stage-specific exposure values which should be considered for all risk assessments.
EPA's Child-Specific Exposures Handbook (U.S. EPA 2008) and Highlights of the Child-Specific
Exposure Factors Handbook (U.S. EPA 2009c) help risk assessors understand children's exposure to
pollution. The handbook provides important information for answering questions about life stage
specific exposure through drinking, breathing and eating. EPA's guidance to scientists on selecting
age groups to consider when assessing childhood exposure and potential dose to environmental
contaminants is identified in Guidance on Selecting Age Groups for Monitoring and Assessing
Childhood Exposures to Environmental Contaminants (U.S. EPA 2005c).
While there is no standard framework for including economic and human health effects on older
adults in an analysis of the impacts of regulation, the EPA stresses the importance of addressing
environmental issues that may adversely impact them.22 These considerations are highlighted in
EPA's Exposure Factors Handbook (U.S. EPA 2011) and have led EPA's Office of Research and
Development to consider an exposure factors handbook specifically for the aging (see U.S. EPA
2007). Additionally, the toxicokinetic and toxicodynamic impacts of environmental agents in older
adults have been considered in EPA's document entitled Aging and Toxic Response: Issues Relevant
to Risk Assessment (U.S. EPA 2005d).
\ ti'Mi P velopment
Disproportionate impacts during fetal development and childhood are considered in EPA guidance
on action development, particularly the Guide to Considering Children's Health When Developing EPA
Actions: Implementing Executive Order 13045 and EPA's Policy on Evaluating Health Risks to Children
(U.S. EPA 2006b). The guide helps determine whether EO 13045 and/or EPA's Children's Health
Policy applies to an EPA action and, if so, how to implement the Executive Order and/or EPA's
Policy. The guide clearly integrates EPA's Policy on Children's Health with the Action Development
Process and provides an updated listing of additional guidance documents.
22 There is a lack of broad agreement about when this life stage begins. The U.S. and other countries typically
define this life stage to begin at the traditional retirement age of 65, but, for example, the U.N. has It begin at age
60 [U.S. EPA 2005d).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
lomic Analysis
While these Economic Guidelines provide general information on BCA of policies and programs,
many issues concerning valuation of health benefits accruing to children are not covered.
Information provided in the Children's Health Valuation Handbook (U.S. EPA 2003), when used in
conjunction with the Guidelines, allows analysts to characterize benefits and impacts of Agency
policies and programs that affect children.
The Handbook is a reference tool for analysts conducting economic analyses of EPA policies when
those policies are expected to affect risks to children's health. The Handbook emphasizes that
regulations or policies fully consider the economic impacts on children, including incorporating
children's health considerations into BCA, as well as a separate analysis focused on children.
Economic factors may also play a role in other analyses that evaluate children's environmental
health impacts. For example, because a higher proportion of children than adults live in poverty, the
ability of households with children to undertake averting behaviors might be compromised.23 This
type of information could inform the exposure assessment.
Analysis of who bears the costs and benefits of a policy also is complicated by the fact that
individual life stages change over time. For instance, because children eventually grow into adults,
health and other benefits of a policy that initially accrue mainly to children will also likely affect
them as adults. Likewise, while the costs of a policy are initially borne by current adults, they will
eventually be borne by the current set of children as they themselves become adults.
[[iv-1 sectici i h - r sen Environmental Justice and- \-bik1ren"?
Health
The burden of health problems and environmental exposures is often borne disproportionately by
children from low-income communities and minority communities (e.g., Arcury etal. 2021; Israel
etal. 2005; Lanphear etal. 1996; Mielke etal. 1999; Pastor etal. 2006; Schwartz etal. 2015). The
challenge for the EPA is to integrate both EJ and life stage susceptibility considerations, particularly
for children but also for older adults, where appropriate when conducting analysis. This is
especially true when short-term exposure to environmental contaminants, such as lead or mercury,
early in life can lead to life-long health consequences.
23 U.S. Census Historical Poverty Tables: People and Families -1959 to 2018.
https://www.census.aov/data/tables/tirne-series/derno/incorne-povertv/historical-poverty-people.htrnl
(accessed on April 1,2020).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 10 References
Adler, M. 2008. Risk Equity: A New Proposal. Harvard Environmental Law Review, 32.
Allaire, M., H. Wu, and U. Lall. 2018. National Trends in Drinking Water Quality Violations. Proceedings of the
National Academy of Sciences, 115(9): 20782083.
Arcury, T., H. Chen, T. Arnold, S. Quandt, K. Anderson, R. Scott, J. Talton, and S. Daniel. 2021. Pesticide
Exposure among Latinx Child Farmworkers in North Carolina. American Journal of Industrial Medicine, 64:
602-619.
Baden, B.M., D.S. Noonan, and R.M. Turaga. 2007. Scales of Justice: Is There a Geographic Bias in
Environmental Equity Analysis? Journal of Environmental Planning and Management, 50(2): 163-85.
Bakkensen, L., and L. Ma. 2020. Sorting over Flood Risk and Implications for Policy Reform. Journal of
Environmental Economics and Management, 104: 102362.
Balazs, C. L., R. Morello-Frosch, A.E Hubbard and I. Ray. 2012. Environmental Justice Implications of Arsenic
Contamination in California's San Joaquin Valley: A Cross-Sectional, Cluster-Design Examining Exposure and
Compliance in Community Drinking Water Systems. Environmental Health, 11(1): 1-12.
Cain, L., D. Hernandez-Cortes, C. Timmins, and P. Weber. 2024. Recent Findings and Methodologies in
Economics Research in Environmental Justice. Review of Environmental Economics and Policy, 18(10): 116-
142.
Chakraborty, J., and J. Maantay. 2011. "Proximity Analysis for Exposure Assessment in Environmental Health
Justice Research," in J. Maantay and S. McLafferty (Eds.), Geospatial Analysis of Environmental Health.
Geotechnologies and the Environment, vol 4. Springer-Verlag, pp. 111-138.
Chakraborty, J., J. A. Maantay, and J. Brender. 2011. Disproportionate Proximity to Environmental Health
Hazards: Methods, Models, and Measurement. American Journal of Public Health, 101(S1): S27-36.
Executive Order 12866: Regulatory Planning and Review, Section 1(a), October 4,1993. Available at:
https://www.epa.gov/laws-regulations/summary-executive-order-12866-regulatory-planning-and-review
(accessed December 6,2024).
Executive Order 12898: Federal Actions to Address Environmental Justice in Minority Populations and Low-
Income Populations, February 11,1994. Available at: https://www.archives.gov/files/federal-
register/executive-orders/pdf/12898.pdf (accessed December 6, 2024).
Executive Order 13045: Protection of Children from Environmental Health Risks and Safety Risks, April 23,
1997. Available at: https://www.federalregister.gov/documents/1997/04/23/97-10695/protection-of-
children-from-environmental-health-risks-and-safety-risks (accessed December 6, 2024).
Executive Order 13175: Consultation and Coordination with Indian Tribal Governments, November 9, 2000.
Available at: https: //www.federalregister.gov/documents/2000/ll /09 /00-29003 /consultation-and-
coordination-with-indian-tribal-governments (accessed December 6,2024).
E.0.14008: Tackling the Climate Crisis at Home and Abroad. Federal Register 2021-02177 (27 January 2021).
Available at: https: //www.federalregister.gov/documents/2021/02 /01/2021-02177 /tackling-the-climate-
crisis-at-home-and-abroad (accessed December 6, 2024).
Executive Order 14096: Revitalizing Our Nation's Commitment to Environmental Justice for All, April 21,
2023. Available at: https: //www.federalregister.gov/documents/2023 /04/26/2023-08955 /revitalizing-our-
nations-commitment-to-environmental-iustice-for-all (accessed November 11, 2024).
Executive Order 13563: Improving Regulation and Regulatory Review. Available at:
https://obamawhitehouse.archives.gOv/the-press-office/2011/01/18/executive-order-13563-improving-
regulation-and-regulatorv-review (accessed December 6,2024).
Fan, Q., K. Fisher-Vanden, and H. Klaiber. 2018. Climate Change, Migration, and Regional Economic Impacts in
the U.S. Journal of the Association of Environmental and Resource Economists, 5(3): 643-671.
5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Fann, N., H. A. Roman, C.M. Fulcher, M.A. Gentile, B.J. Hubbell, K. Wesson, J.I. Levy. 2011. Maximizing Health
Benefits and Minimizing Inequality: Incorporating Local-Scale Data in the Design and Evaluation of Air
Quality Policies. Risk Analysis, 31(6): 908-922.
Fedinick, K.P., S. Taylor, and M. Roberts. 2019. Watered Down Justice: Communities of Color and theSDWA.
Natural Resources Defense Council Report R-19-09-A.
Fullerton, D. 2011. Six Distributional Effects of Environmental Policy. Risk Analysis, 31(6): 923-929.
Gerking, S. and M. Dickie. 2013. Valuing Reductions in Environmental Risks to Children's Health. Annual
Review of Resource Economics, 5: 245-260.
Israel, B., E. Parker, Z. Rowe, A. Salvatore, M. Minkler, J. Lopez, A. Butz, A. Mosley, L. Coates, G. Lambert, P.
Potito, B. Brenner, M. Rivera, H. Romero, B. Thompson, G. Coronoado, and S. Halstead. 2005. Community-
Based Participatory Research: Lessons Learned from the Centers for Children's Environmental Health and
Disease Prevention Research. Environmental Health Perspectives, 113(10): 1463-1471.
Konisky, D. M., C. Reenock, and S. Conley. 2021. Environmental Injustice in Clean Water Act Enforcement:
Racial and Income Disparities in Inspection Time. Environmental Research Letters, 16(8): 084020.
Kuminoff, N., T. Schoellman, and C. Timmins. 2 015. Environmental Regulations and the Welfare Effects of Job
Layoffs in the United States: A Spatial Approach. Review of Environmental Economics and Policy, 9(2): 198-
218.
Lanphear, B.D.; M. Weitzman; S. Eberly. 1996. Racial Differences in Urban Children's Environmental
Exposures to Lead. American Journal of Public Health, 86(1): 1460-1463.
Maguire, K., and G. Sheriff. 2011. Comparing Distributions of Environmental Outcomes for Regulatory
Environmental Justice Analysis. International Journal of Environmental Research and Public Health 8:1707-
1726.
McDonald, Y.J., and N.E. Jones. 2018. Drinking Water Violations and Environmental Justice in the United
States, 2011-2015. American Journal of Public Health, 108(10): 1401-1407.
Melstrom, R., and R. Mohammadi. 2022. Residential Mobility, Brownfield Remediation, and Environmental
Gentrification in Chicago. Land Economics, 98(1): 62-77.
Mielke, H.W., C.R. Gonzales, M.K. Smith, and P.W. Mielke. 1999. The Urban Environmental and Children's
Health: Soils as an Indicator of Lead, Zinc, and Cadmium in New Orleans, Louisiana, U.S.A. Environment
Research, 81:117-129.
Mohai, P., and R. Saha. 2007. Racial Inequality in the Distribution of Hazardous Waste: A National Level
Reassessment. Social Problems 54(3): 343-370.
OMB. 2019. Improving Implementation of the Information Quality Act. Available at:
https://www.whitehouse.gOv/wp-content/uploads/2019/04/M-19-15.pdf (accessed December 6, 2024).
OMB. 2023. Circular A-4, Regulatory Analysis, November 9, 2023. Available at:
https://www.whitehouse.gov/wp-content/uploads/2023/ll/CircularA-4.pdf (accessed June 13, 2024).
Pastor, Jr., M.; R. Morello-Frosch; and J.L. Sadd. 2006. Breathless: Schools, Air Toxics, and Environmental
Justice in California. Policy Studies Journal, 34(3): 337-362.
Redding, S., and E. Rossi-Hansberg. 2017. Quantitative Spatial Economics. Annual Review of Economics, 9: 21-
58.
Schwartz, N., C. von Glascoe, V. Torres, L. Ramos, and C. Soria-Delgado. 2015. Where They (Live, Work and)
Spray: Pesticide Exposure, Childhood Asthma and Environmental Justice among Mexican-American
Farmworkers. Health and Place, 32: 83-92.
Shadbegian, R. J., and W. B. Gray. 2012. "Spatial Patterns in Regulatory Enforcement," in Banzhaf, S. (Ed.). The
Political Economy of Environmental Justice, pp. 225-248.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Stieb, D., A. Huang, R. Hocking, D. Crouse, A. Osornio-Vargas, and P. Villeneuve. 2019. Using Maps to
Communicate Environmental Exposures and Health Risks: Review and Best-Practice Recommendations.
Environmental Research, 176:108518.
U.S. Environmental Protection Agency (U.S. EPA). 1995. Policy on Evaluating Health Risks to Children.
Available at: https: //www.epa.gov/environmental-economics/childrens-health-valuation-handbook-2003
(accessed December 6,2024).
U.S. EPA. 2002. Guidelines for Ensuring and Maximizing the Quality, Objectivity, Utility, and Integrity of
Information Disseminated by the Environmental Protection Agency (EPA/260R-02-008). Washington, D.C.: U.S.
EPA, Office of Environmental Information. Available at: http://www.epa.gov/sites/production/files/2015-
08 /documents /epa-info-quality-guidelines.pdf (accessed December 6, 2024)
U.S. EPA. 2003. Children's Health Valuation Handbook. Available at: https://www.epa.gov/environmental-
economics/childrens-health-valuation-handbook-2003 (accessed December 6,2024).
U.S. EPA. 2005a. Guidelines for Carcinogen Risk Assessment. EPA/630/P-03/001F. Available at:
https://www.epa.gov/risk/guidelines-carcinogen-risk-assessment (accessed December 6, 2024).
U.S. EPA. 2005b. Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens.
Available at: https: //www.epa.gov/risk/supplemental-guidance-assessing-susceptibility-early-life-exposure-
carcinogens (accessed December 6, 2024).
U.S. EPA. 2005c. Guidance on Selecting Age Groups for Monitoring and Assessing Childhood Exposures to
Environmental Contaminants. EPA/630/P-03/003F. Available at: https: //www.epa.gov/risk/guidance-
selecting-age-groups-monitoring-and-assessing-childhood-exposures-environmental (accessed December 6,
2024).
U.S. EPA. 2005d. Aging and Toxic Response: Issues Relevant to Risk Assessment. Available at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=156648 (accessed December 6, 2024).
U.S. EPA. 2006a. A Framework for Assessing Health Risks of Environmental Exposures to Children (Final),
Washington, DC, EPA/600/R-05/093F. Available at:
https://cfpuh.epa.gov/ncea/risk/recordisplay.cfm?deid=158363 (accessed December 6, 2024).
U.S. EPA. 2006b. EPA's Action Development Process - Guide to Considering Children's Health When
Developing EPA Actions: Implementing Executive Order 13045 and EPA's Policy on Evaluating Health Risks
to Children. Available at: https: //19ianuary2021snapshot.epa.gov/sites/static/files/2014-
05/documents/epa adp guide childrenhealth.pdf (accessed December 6, 2024).
U.S. EPA. 2007. Summary Report of a Peer Involvement Workshop on the Development of an Exposure
Factors Handbook for the Aging. EPA/600/R-07/061. Washington, DC. Available at:
https://assessments.epa.gov/risk/document/&deid%3D17192 3 (accessed December 6,2024).
U.S. EPA. 2008. Child-Specific Exposures Handbook. Final Report. EPA/600/R-06/096F. Available at:
https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=199243 (accessed December 6,2024).
U.S. EPA. 2009c. Highlights of the Child-Specific Exposure Factors Handbook. Final Report. EPA/600/R-08/135.
Available at: https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=200445 (accessed December 6, 2024).
U.S. EPA. 2011. Exposure Factors Handbook. Final Report. Available at:
https://cfpuh.epa.gov/ncea/risk/recordisplay.cfm?deid=236252 (accessed December 6, 2024).
U.S. EPA. 2012. Guidance for Evaluating and Documenting the Quality of Existing Scientific and Technical
Information. Washington, DC: U.S. EPA, Science and Technology Policy Council. Available at:
http://www.epa.gov/sites/production/files/2015-05/documents/assess3.pdf (accessed December 6,2024).
U.S. EPA. 2015. Peer Review Handbook. Science and Technology Policy Council. 4th Edition. EPA/100/B-
15/001. Available at: http: //www.epa.gov/osa/peer-review-handbook-4th-edition-2015 (accessed
December 6,2024).
U.S. EPA. 2021. Using All Appropriate Injunctive Relief Tools in Civil Enforcement Settlements. Memo. April
26. Available at: https://www.epa.gov/sites /default/files /2021-
Guidelines for Preparing Economic Analyses | 3rd edition
-------
04/documents/usingallappropriateiniunctiverelieftoolsinci vilenforcementsettlement0426.pdf (accessed
December 6,2024).
U.S. EPA. 2022. EPA Legal Tools to Advance Environmental Justice. Office of General Council. March. Available
at: https://www.epa.gov/ogc/epa-legal-tools-advance-environmental-iustice. (accessed October 28, 2024)
U.S. EPA. 2024a. Technical Guidance for Assessing Environmental Justice in Regulatory Analysis. Available at:
https://www.epa.gov/environmentaliustice/technical-guidance-assessing-environmental-iustice-regulatory-
analysis (accessed December 9, 2024).
U.S. EPA. 2024b. Achieving Health and Environmental Protection Through EPA's Meaningjitl Engagement Policy.
Available at: https: //www.epa.gov/environmentaliustice/epas-meaningful-engagement-policy (accessed
December 6,2024).
U.S. EPA. 2024c. Regulatory Impact Analysis for Hazardous and Solid Waste Management System: Disposal of
Coal Combustion Residuals from Electric Utilities; Legacy Surface Impoundments. Final rule. Available at:
https: //www.regulations.gov/document/EPA-HO-OI.EM-2020-0107-1067 (accessed December 6, 2024).
I Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 11 - Presentation of Analysis
a
This chapter provides some general guidance for presenting analytical results
to policy makers and others interested in environmental policy development.
Economic analyses play an important role throughout the policy development
process. From the initial, preliminary evaluation of potential options through
the preparation of a final economic analysis document, economic analysts
participate in an interactive process with policy makers. The fundamental goal
of this process is to collect, analyze and present information useful for policy
makers.
Economic analysis is often motivated by a desire to find an optimal outcome,
such as a degree of stringency in a regulation, or a level of provision of a
public good that yields the largest possible net benefits. Environmental
statutes sometimes mandate criteria other than economic efficiency, such as
best available control technology or lowest achievable emission rate. Policy
makers rely on quantitative analysis to promulgate these approaches. In
particular, they rely on analyses that delineate the costs, benefits or other
impacts of a wide range of control options.
This guidance for presenting inputs, analyses and results applies to all stages
of this process, not only to the final document embodying the completed
economic analysis. Conveying uncertainty effectively and reporting critical
assumptions and key unquantified effects to decision makers is critical at all
points in the policy-making process.
This chapter begins by providing general guidance on how to present the
results of economic analyses, with a particular emphasis on presenting
benefits and costs, including those that cannot be quantified and/or put into
dollar terms. The chapter then discusses the components, or inputs, of an
economic analysis, and how their effect on the economic analysis can best be
communicated.
11.1 Presenting Results of Economic Analyses
The presentation of the results of an economic analysis should be thorough and transparent. The
reader should be able to understand:
What the primary conclusions of the economic analysis are;
Which benefits arise from the statutory objective of the regulation and which do not;
11-1 Guidelines for Preparing Economic Analyses | 3rd edition
-------
How the benefits and costs were estimated;
What the important non-quantified and/or non-monetized effects are;
What key assumptions were made for the analysis;
What the primary sources of uncertainty are in the analysis; and
How those sources of uncertainty affect the results.
An economic analysis of regulatory or policy options should present all identifiable costs and
benefits that are incremental to the regulation or policy under consideration.
Benefits and costs should be reported in monetary terms whenever possible. In reality, there are
often effects that cannot be monetized, and the analysis needs to communicate the full richness of
benefit and cost information beyond what can be put in dollar terms. Benefits and costs that cannot
be monetized should, if possible, be quantified (e.g., expected number of adverse health effects
avoided or improved biodiversity). Benefits and costs that cannot be quantified should be presented
qualitatively (e.g., directional impacts on relevant variables). Section 11.1 contains more detailed
guidance on presenting this information in the U.S. Environmental Protection Agency's (EPA's)
economic analyses.
Agencies are also required to provide OMB with an accounting statement reporting benefit and cost
estimates when sending over each economically significant rule. Analysts should rely upon these
Guidelines and Circular A-4 for developing these estimates. Circular A-4 (2023) provides a suggested
format for this accounting statement.1
The results of economic analyses of environmental policies should generally be presented in three
sections.
Results from BCA. Estimates of the net social benefits should be presented based on the
benefits and costs expressed in monetary terms. Non-monetized and unquantifiable
benefits and costs should also be included and described in the presentation.
Results from cost-effectiveness analysis (CEA). Under OMB Circular A-4, CEA should
generally be performed for rules in which the primary effect is human health or safety.
Results of these analyses should also be presented when they are conducted.2
Results from economic impact analysis (EIA) and distributional assessments. Results
of the EIA should be reported, including predicted effects on prices, profits, plant closures,
employment and any other effects. Distributional impacts for particular groups of concern,
including small entities, governments and environmental justice populations should also be
presented.
The relative importance of these three sections will depend on the policy and statutory context of
the analysis.
1 The sample accounting statement is on p. 91 of Circular A-4 (2023).
2 The Institute of Medicine (IOM) (2006) issued recommendations to regulatory agencies on how to perform
health-based CEA. Examples of CEA can be found in appendices of several RIAs including those for particulate
matter (PM) National Ambient Air Quality Standards (NAAQS) [see Appendix G listed at
http://www.epa.aov/ttn/ecas/ria.html (accessed March 13, 2011)] and the Ground Water Rule [see Appendix H
listed at http://www.epa.gov/safewater/disinfection/gwr/regulation.html (accessed March 13, 2011)].
11-2 Guidelines for Preparing Economic Analyses | 3rd edition
-------
ii i i l*i- Thhiii;; nv- l - su! i I- -?neVh"- Analyses
When presenting the results of a BCA, the expected benefits and costs of all analyzed options should
be reported, including the proposed or finalized option and any alternatives. OMB's Circular A-4
(2023) recommends studying alternative levels of stringency in addition to the proposed or
finalized option, and the incremental costs and benefits would be reported for each increasingly
stringent option. Separate time streams of benefits and costs should be reported, in constant
(inflation-adjusted), undiscounted dollars. Per the discussion in Chapter 6, appropriately
discounted benefits and costs should be reported as well.
Ideally, all benefits and costs of a regulation would be expressed in monetary terms, but this is
almost never possible because of data gaps, unquantifiable uncertainties and other challenges. It is
important not to exclude an important benefit or cost category from BCA even if it cannot be placed
in dollar terms. Instead, such benefits and costs should be expressed quantitatively if possible (e.g.,
avoided adverse health impacts, number of species added). If important benefit or cost categories
cannot be expressed quantitatively, they should be discussed qualitatively. Of course, care should
be taken to avoid overlapping categories of benefits and costs and to avoid double-counting.
Quantifiable benefits and costs, properly discounted, should be compared to determine a
regulation's net benefits, even if important benefits or costs cannot be monetized. However, an
economic analysis should assess the likelihood that non-monetized benefits and costs would
materially alter the net benefit calculation for a given regulation.
Incremental benefits, costs and net benefits of moving from less to more stringent regulatory
alternatives should also be presented. If a regulation has particularly significant impacts on
population groups of concern, the various options' incremental impacts on these groups or source
categories should be reported. This should include a discussion of incremental changes in
quantified and qualitatively described benefits and costs.
Given the number of potential models presented in Chapters 7 and 8, the analyst should take care to
clearly indicate the correspondence between the benefit and cost estimates. For example, the cost
analysis may include results from a general equilibrium model, but the benefit analysis may only
include partial equilibrium effects. In this case, the cost side of the equation includes general
equilibrium feedback effects while the benefit side does not This difference should be clearly
presented and explained.
The tables at the end of this chapter contain templates for presenting information on regulatory
benefits and costs, including those that cannot be quantified or put into dollar terms. The analyst's
primary goal, using these tables, is to communicate the full richness of benefit and cost information
instead of focusing narrowly on what can be put in dollar terms. Some guiding principles for
constructing these tables follow.
All meaningful benefits and costs, including benefits arising from the statutory
objective of the regulation as well as other welfare effects, are included in all of the
tables even if they cannot be quantified or monetized. Not only does this provide
consistency for the reader, but it also maintains important information on the context of the
quantified and monetized benefits.
The types of benefits and costs are described briefly in plain terms to make them
clearer to the public and to decision makers, and they should be well-defined and
mutually exclusive, to the extent possible. Benefits should be grouped in a manner
consistent with the categories in Table 7.1 of Chapter 7, although the order and specific
characterization can be expected to vary by rule as needed.
11-3
Guidelines for Preparing Economic Analyses | 3rd edition
-------
The benefits are expressed first in natural or physical units (i.e., numbers) to provide
a more complete picture of what the rule accomplishes. These units are not discounted
as they would be in a CEA because the goal here is to describe what might be termed the
"physical scope" of the rule's benefits. It may be the case that physical or natural units are
not relevant for presenting costs.3
Explanatory notes accompany each benefit and cost entry and can be used to describe
whatever the most salient or important points are about scientific uncertainty, the type of
benefit or cost, how it is estimated or the presentation.
The benefit categories in these templates (e.g., improved human health, improved environment and
other benefits,) will need to be revised to reflect the benefits categories for the rule under
consideration. Likewise, cost categories may need to be revised to match the circumstances of the
individual rule. Simpler analyses may need only the overview (Table 11.1) and the final summary
(Table 11.4).
Table 11.1 is a quick-glance summary of regulatory benefits and costs, the extent to which they
could be quantified and monetized, and a reference to where they are more fully characterized or
estimated in the economic analysis. Some benefits may be described only qualitatively.
Table 11.2 reports benefits in non-monetary terms along with the units and additional explanatory
notes. The goal of this table is to communicate the physical scope of the regulation's benefits rather
than the dollar equivalent Benefits here do not need to be discounted to present value, but the time
associated with the quantities should be made clear (e.g., "annual" or "more than 10 years").
Table 11.3 reports benefits and costs in monetary terms along with totals for dollar-valued benefits
and costs. Here it is important to specify the reference year for the dollars (i.e., real terms), the
discount rate(s) used and the unit value and/or source.
Table 11.4 contains a template for bringing all this information together in summary that includes
the type of benefit or cost, how it is measured, its quantity and dollar benefits. When multiple
regulatory options are included in this table, it is appropriate for including in the regulatory
preamble as requested by OMB.
Consistent with recommendations in these Guidelines for communicating uncertainty, quantitative
entries should generally include a central or best estimate in addition to a range or confidence
interval. The ability to do this, of course, may be limited by data availability.
The templates provided in Tables 11.1-11.4 presume that the regulatory action is designed to
achieve health and environmental-protection benefits, albeit at some cost. In the case of a
deregulatory action, the structure of the templates may need to be reversed.
3 Note that, as described in Chapter 6, the undiscounted stream of the non-monetized effects should be presented
as they occur over time, and that these non-monetized effects generally should also still be discounted in benefit-
cost analysis and cost-effectiveness analysis if they are aggregated over time. See Section 6.1.
11-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Table 11.1 - Template for Regulatory Benefits and Costs Checklist
Benefits
Improved Human Health Benefits
Effect can be
Quantified?
(put in numeric
terms]
Effect can be
Monetized?
(putin dollar
terms]
More Information
(e.g., reference to section of
the economic analysis]
Reduced incidence of adult premature
mortality from exposure to PM2.5
~
e.g., see Section 5.2 of the
economic analysis
Reduced incidence of fetal loss from
reduced exposure to disinfection
byproducts
~
-
Notes and reference to
section of the economic
analysis
Unquantified human health benefit
with a brief description
-
-
Notes and reference
Improved Environment Benefits
Effect can be
Quantified?
(putin numeric
terms]
Effect can be
Monetized?
(putin dollar
terms]
More Information
(e.g., reference to section of
the economic analysis]
Fewer fish killed from reduced nutrient
loadings into waterways
Notes and reference
Improved timber harvest from lower
tropospheric ozone concentrations
~
~
Notes and reference
Other environmental benefit with a brief
description
-
-
Notes and reference
Other Benefits
Effect can be
Quantified?
(put in numeric
terms]
Effect can be
Monetized?
(put in dollar
terms]
More Information
(e.g., reference to section of
the economic analysis]
Reduced fuel expenditures from
improved efficiency in automobiles and
light trucks
~
Notes and reference
Other benefit with a brief description
-
-
Notes and reference
11-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Costs
Compliance Costs (Fixed)
Effect can be
Monetized?
(put in dollar terms]
More Information
(e.g., reference to section of the
economic analysis)
Research and Development
investments to meet new standard
~
Notes and reference
Capital Costs for new pollution control
equipment
Notes and reference
Compliance Costs (Variable)
Effect can be
Monetized?
(putin dollar terms)
More Information
(e.g., reference to section of the
economic analysis)
Operating Costs for pollution control
equipment
~
Notes and reference
Monitoring, reporting and
recordkeeping costs associated with
new requirements
s
Notes and reference
Transaction costs
-
Notes and reference
Other Opportunity Costs
Effect can be
Monetized?
(putin dollar terms)
More Information
(e.g., reference to section of the
economic analysis)
Transition costs
-
Notes and reference
Reduced output in the regulated
market
Notes and reference
Other costs with brief description
-
Notes and reference
11-6 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Table 11.2 - Template for Quantified Regulatory Benefits
Improved Human Health Benefits
Quantified
Benefits
(confidence
interval or range]
Units
More Information
(with possible reference to
section of the economic
analysis]
Reduced incidence of adult
premature mortality from
exposure to PM2.5
estimate
(range)
expected avoided
premature deaths
per year
e.g., range represents
confidence interval
Reduced incidence of fetal loss
from reduced exposure to
disinfection byproducts
estimate
(range)
expected avoided
fetal losses per
year
e.g., confidence interval
cannot be estimated. Range
based on alternative studies
Unqucmtifiedhuman health
benefit with a brief description
*
*
e.g., data do not allow for
quantification
Improved Environment Benefits
Quantified
Benefits
(confidence
interval or range]
Units
More Information
(with possible reference to
section of the economic
analysis]
Fewer fish killed from reduced
nutrient loadings into waterways
estimate
[range]
thousands of fish
per year
Notes
(reference]
Improved timber harvest from
lower tropospheric ozone
concentrations
estimate
[range]
thousands of
board feet per
year
Notes
(reference]
Other environmental benefit with
a brief description
*
*
Notes
(reference]
Other Benefits
Quantified
Benefits
(confidence
interval or range]
Units
More Information
(with possible reference to
section of the economic
analysis]
Fuel savings from improved
efficiency in automobiles and light
trucks
estimate
(range)
millions of gallons
of gasoline
reduced per year
Notes
(reference)
Other benefit with a brief
description
*
*
Notes
(reference)
Note: * indicates the benefit cannot be quantified with available information.
11-7 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Table 11.3 - Template for Dollar-Valued Regulatory Benefits and Costs
Dollar-Valued Benefits
Improved Human Health Benefits
Dollar Benefits
(millions per year)
Basis of
Value
More Information
Reduced incidence of adult premature
mortality from exposure to PM2.5
$ estimate
($ range]
e.g., $X based
on Agency
guidance
Notes
(reference)
Reduced incidence of fetal loss from
reduced exposure to disinfection
byproducts
*
Not available
Notes
(reference)
Unquantified human health benefit
with a brief description
*
*
e.g., data insufficient to
quantify
(reference)
Improved Environment Benefits
Dollar Benefits
(millions per year]
Basis of
Value
More Information
Fewer fish killed from reduced nutrient
loadings into waterways
$ estimate
($ range)
e.g., $X based
on WTP for
recreational
fishing
e.g., range reflects two
different valuation
approaches
(reference)
Improved timber harvest from lower
tropospheric ozone concentrations
$ estimate
($ range)
e.g., change in
consumer and
producer
surplus
e.g., estimated from market
model across several species
(reference)
Other environmental benefit with a
brief description
*
*
Notes
(reference)
Other Benefits
Dollar Benefits
(millions per year)
Basis of
Value
More Information
Fuel savings from improved efficiency
in automobiles and light trucks
$ estimate
($ range)
e.g., $X, based
on net-of-tax
average per
gallon price
e.g., there is debate on how
well fuel savings represent
consumer benefits
(reference)
Other benefit with a brief description
*
Not available
Notes
(reference)
TOTAL Benefits that can be monetized
($ millions per year)
$ estimate
($ range)
$ estimate
($ range)
$ estimate
($ range)
11-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Costs
Compliance Costs (Fixed)
Dollar Benefits
(millions per year]
Basis of
Value
More Information
(with possible reference)
R&D investments
$ estimate
($ range)
e.g., $X based
on industry
survey
Notes
(reference)
Capital Costs
$ estimate
($ range]
-
e.g., estimated from
engineering cost models
Compliance Costs (Variable)
Dollar Benefits
(millions per year)
Basis of
Value
More Information
(with possible reference)
Operating Costs
$ estimate
($ range)
e.g., estimated from
engineering cost models
Monitoring, Reporting and
Recordkeeping Costs
$ estimate
($ range)
e.g., $X based
on industry
estimates
e.g., industry survey with
55% response
Transaction Costs
$ estimate
($ range)
Notes
(reference)
Other Opportunity Costs
Dollar Benefits
(millions per year)
Basis of
Value
More Information
(with possible reference)
Transition Costs
$ estimate
($ range)
-
Notes
(reference)
Reduced output in the regulated
market
$ estimate
($ range)
-
Notes
(reference)
Other Costs
$ estimate
($ range)
-
Notes
(reference)
TOTAL Costs that can be monetized
($ millions per year)
$ estimate
($ range)
$ estimate
($ range)
$ estimate
($ range)
Note: * indicates the benefit cannot be quantified with available information.
11-9 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Table 11.4 - Template for Summary of Benefits and Costs
Notes: e.g., "annual average numbers; millions 2022 dollars annualized at 3% discount rate," best estimate [with
range).
Benefits
Improved Human
Health Benefits
Option 1
m
Option 1
[$]
Proposed or
Finalized
Option
m
Proposed or
Finalized
Option
($)
Option 3
Option 3
($)
Source,
limitations or
other key
notes
Reduced
incidence of adult
premature
mortality from
exposure to PM2.5
estimate
[range]
estimate
[range]
estimate
[range]
estimate
[range]
estimate
[range]
estimate
[range]
highlight most
important
points, as
needed
Reduced
incidence of fetal
loss from
exposure to
disinfection
byproducts
estimate
[range]
*
estimate
[range]
*
estimate
[range]
*
e.g., no
valuation data
exist. Effects
sensitive to
dose-response
model.
Unquantifled
human health
benefit with
description
*
*
*
*
*
*
e.g., risk data
insufficient for
quantification
Improved
Environment
Benefits
Option 1
m
Option 1
[$)
Proposed or
Finalized
Option
m
Proposed or
Finalized
Option
($]
Option 3
m
Option 3
[$]
Source,
limitations or
other key
notes
Fewer fish killed
from reduced
nutrient loadings
into waterways
estimate
[range]
estimate
[range]
estimate
[range]
estimate
[range]
estimate
[range]
estimate
[range]
Notes
Improved timber
harvest from
lower
tropospheric
ozone
concentrations
estimate
[rangej
estimate
[range]
estimate
[range]
estimate
[range]
estimate
[range]
estimate
[range]
Notes
Other
environmental
benefit with
description
*
*
*
*
*
*
Notes
11-10 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Other Benefits
Option 1 Option 1
m (S)
Proposed or
Finalized
Option
m
Proposed or
Finalized
Option
($)
Option 3
m
Option 3
($]
Source,
limitations
or other key
notes
Fuel savings from
improved efficiency
in light duty
vehicles
estimate
(range]
estimate
(range)
estimate
(range]
estimate
[range]
estimate
[range]
estimate
[range]
Notes
Other benefit with
description
*
*
*
*
*
*
Notes
TOTAL monetized
benefits
(annualized,
millions $2022)
$
estimate
(range]
$
estimate
[range]
$ estimate
[range]
$ estimate
[range]
$
estimate
[range]
$ estimate
[range]
e.g., range
maybe
overstated
due to
aggregation
(See Section
8.1]
Costs
Compliance Costs
(Fixed)
Option 1
($ millions]
Proposed or
Finalized
Option
m
Option 3
m
Source, limitations or other key
notes
R&D investments
$ estimate
[range]
$ estimate
[range]
$ estimate
[range]
Notes
(reference]
Capital Costs
$ estimate
[range]
$ estimate
[range]
$ estimate
[range]
e.g., estimated from engineering
cost models
Compliance Costs
(Variable)
Option 1
($ will ions]
Proposed or
Finalized
Option
m
Option 3
Source, limitations or other key
notes
Operating Costs
$ estimate
[range]
$ estimate
[range]
$ estimate
[range]
e.g., estimated from engineering
cost models
Monitoring and
Recordkeeping Costs
$ estimate
[range]
$ estimate
[range]
$ estimate
[range]
e.g., industry survey with 55%
response
Transaction Costs
$ estimate
[range]
$ estimate
[range]
$ estimate
[range]
Notes
(reference]
11-11 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Other Opportunity
Costs
Option 1
($ millions)
Proposed or
Finalized
Option
m
Option 3
m
Source, limitations or other key
notes
Transition Costs
$ estimate
(range)
$ estimate
(range)
$ estimate
(range)
Notes
(reference)
Other Costs
$ estimate
(range)
$ estimate
(range)
$ estimate
(range)
Notes
(reference)
Reduced output in the
regulated market
$ estimate
(range)
$ estimate
(range)
$ estimate
(range)
Notes
(reference)
TOTAL monetized
Costs (annualized,
millions $2022)
$ estimate
(range)
$ estimate
(range)
$ estimate
(range)
-
TO TAL Net Ben efits
that can be monetized
(annualized, millions
$2022)
$ estimate
(range)
$ estimate
(range)
$ estimate
(range)
-
Note: * indicates the benefit cannot be quantified with available information.
11.1.2 Presenting the Results of Cost-Effectiveness Analyses
When BCA is not possible, CEA may be the best available option. The cost-effectiveness of a policy
option is calculated by dividing the annualized cost of the option by non-monetary benefit
measures. Options for such measures range from quantities of pollutant emissions reduced,
measured in physical terms, to a specific improvement in human health or the environment,
measured in reductions in illnesses or changes in ecological services rendered.
The non-monetary measure of benefits used in a CEA must be chosen with great care to facilitate
valid comparisons across options. The closer the chosen measure is to the variable that directly
impacts social welfare, the better the measures will convey the weight of the consequences of the
alternatives and the more robust a CEA will be. Consider the following steps that a typical
environmental economic assessment follows:
Changes in emissions are estimated (e.g., tons of emissions); then
Changes in environmental quality (e.g., changes in ambient concentrations of a given air
pollutant) are estimated; then
Changes in human health or welfare (e.g., changes in illness or visibility) are estimated.
Each successive step in this sequence yields a better, and preferable, measure for CEA.
To illustrate, consider a typical air pollution scenario. Depending on where and when air pollutants
are released into the atmosphere, a given ton of a particular pollutant can have widely divergent
impacts on ambient air quality. Similarly, depending on when and where air quality changes, widely
different levels of human health impacts may result Particularly when different regulatory
approaches are under consideration (e.g., regulation of different source categories in different
locations), failing to standardize the analyses on the benefit measure that directly affects human
health or welfare will significantly reduce the value of the analysis to decision makers (and the
public).
11-12 Guidelines for Preparing Economic Analyses | 3rd edition
-------
When presenting the results of a CEA, the rationale for the selection of the non-monetary benefit
measure must be described in detail. The presentation of results should also include a discussion of
the limitations of the analysis, especially if an inferior measure, such as cost per ton of pollutant,
must be used.
CEA is most useful when the policy or regulation in question affects a single endpoint. When
multiple endpoints are affected (e.g., cancer and kidney failures), combining endpoints into a single
effectiveness measure is impossible unless appropriate weighting factors exist for the multiple
endpoints. The theoretically correct weights to apply are the dollar values associated with each
endpoint, but generally it is the absence of these values that necessitates CEA. Therefore, it is not
possible to compare a policy or regulation that reduces relatively more expected cancers, but fewer
expected cases of kidney failure, with one that has the opposite relative effects. When this occurs,
the effects of each option for each endpoint should be reported. A single endpoint may be selected
for calculating cost-effectiveness, while other endpoints can be listed as ancillary benefits (or, if
possible, their monetary value should be subtracted from the option's cost prior to calculating its
cost-effectiveness) (OMB 2023).
The most cost-effective option i.e., the option with the lowest cost per unit of benefit is not
necessarily the most economically efficient. Moreover, other criteria, such as statutory
requirements, enforcement problems, technological feasibility or quantity and location of total
emissions abated may preclude selecting the least-cost solution in a regulatory decision. However,
where not prohibited by statute, CEA can indicate which control measures or policies are inferior
options.
i i i * IV Thhiii;; in- l - su! i I P ind Distributional Analyses
EIA and distributional outcomes focus on disaggregating effects to show impacts separately for the
groups and sectors of interest If costs and/or benefits vary significantly among the sectors affected
by the policy, then both costs and benefits should be shown separately for the different sectors.
Presenting results in disaggregated form will provide important information to policy makers that
may help them tailor the rule to improve its efficiency and distributional outcomes.
The results of the EIA should also be reported for important sectors within the affected population
identifying specific segments of industries, regions of the country or types of firms that may
experience significant impacts or plant closures and losses in employment.
Reporting the results in distributional assessments may include the expected allocation of benefits,
costs or both for specific population groups of concern including those highlighted in the various
mandates. These include minorities, low-income populations, small businesses, governments, not-
for-profit organizations and vulnerable populations (including children). Where these mandates
specify requirements that depend on the outcomes of the distributional analyses, such as the
Regulatory Flexibility Act, the presentation of the results should conform to the criteria specified by
the mandate.
11.2 Communicating Sources of Uncertainty
While guidance on performing uncertainty analysis is in Chapter 5, it is also important to consider
how to communicate uncertainty in the analysis. Estimates of costs, benefits and other economic
impacts should be accompanied by indications of the most important sources of uncertainty
embodied in the estimates, and, if possible, a quantitative assessment of their importance.
11-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
In economic analysis, uncertainty encompasses two different concepts:
Statistical variability of key parameters; and
Incomplete understanding of important relationships.
Economic analyses of environmental policies and regulatory options will frequently have to
accommodate both concepts. The importance of statistical variability is commonly assessed using
Monte Carlo analyses. Expert elicitation techniques, can help close knowledge gaps surrounding key
relationships (see Chapter 5).
Ideally, an economic analysis would present results in the form of probability distributions that
reflect the cumulative impact of all underlying sources of uncertainty. When this is impossible, due
to time or resource constraints, results should be qualified with descriptions of major sources of
uncertainty. If at all possible, information about the underlying probability distribution should be
conveyed.
An economic analysis of an environmental regulation should carefully describe the data used in the
analysis, the models it relies on, major assumptions that were made in running the models and all
major areas of uncertainty in each of these elements. Presentations of economic analyses should
strive for clarity and transparency. An analysis that produces conclusions that can withstand close
scrutiny is more likely to provide policy makers with the information they need to develop robust
environmental policies.
11.2.1 Data
An economic analysis should clearly describe all important data sources and references used.
Unless the data are confidential business information or some other form of private data, they
should be available to policy makers, other researchers, policy analysts and the public. Providing
documentation and access to the data used in an analysis is crucial to the credibility and
reproducibility of the analysis.
EPA Order CIO 2105.0 (U.S. EPA 2023a and U.S. EPA 2023b) and the applicable federal regulations
established a mandatory quality system for the EPA. As required by the quality system, all EPA
offices have developed quality management plans to ensure the quality of their data and
information products.
In any economic analysis, there should be a clear presentation of how data are used and a concise
explanation of why the data are suitable for the selected purpose. The data's accuracy, precision,
representativeness, completeness and comparability should be discussed when applicable. When
data are available from more than one source, a rationale for choosing the source of the data should
be provided.
r r 2,2 Model Che - "n v fFjmptions
An economic analysis of an environmental regulation should carefully describe the models it relies
on, the major assumptions made in running the models (to be discussed more fully below) and any
areas of outstanding uncertainty. The analyst should take particular care to explain any results that
might be viewed as counterintuitive. In particular, analysts should be careful not to accept model
output blindly. Any model that is used without proper thought given to both its input and output
may become a "black box" insofar as nonsensical results may result from a misspecified scenario, a
coding error or any of a number of other causes.
11-14 Guidelines for Preparing Economic Analyses | 3rd edition
-------
In the process of conducting an economic analysis, it is sometimes necessary to bridge an
information gap by making an assumption. Analysts should not simply note the information gap but
should also justify the chosen assumption and provide a rationale for choosing one assumption
over other plausible options. The analyst should take care not to overlook information gaps that are
filled with a piece of information that is only slightly related to the desired information. Analysts
are advised to keep a running list of assumptions. This will make it easier to identify "key
assumptions" for the final report The likely impact of errors in assumptions should be
characterized both in terms of direction and magnitude of effect when feasible.
Maintaining a list of assumptions can benefit the analysis in several ways. In the short run, a list can
serve to focus analysts' attention on those assumptions with the greatest potential to affect net
benefits, possibly leading to new approaches to bridging an information gap. In the long run,
highlighting information gaps may encourage the EPA or others to devote attention and resources
to generating that information.
Whenever the likely errors in a particular assumption can be characterized numerically or
statistically, the factor is a good candidate for sensitivity analysis or uncertainty analysis,
respectively. In many cases, only a narrative description of the impact of errors in assumptions is
possible. The analyst should include a table that clearly lays out all of the key assumptions and the
potential magnitude and direction of likely errors in assumptions in the summary of results.
'f'f-.i.r. Mdressing Uncer , L'n --11 I v i mrr| i ons and Model
Choice
Every analysis should address uncertainties resulting from the choices the analyst has made. For
example, many economic analyses performed at the EPA include assessments of economic impacts
expected to occur decades into the future. Estimates of the future costs and benefits of a regulation
will be sensitive to assumptions about growth rates for populations, source categories, economic
activity and technological change, as well as many other factors. Sensitivity analyses on key
variables in the baseline scenario should be performed and reported when possible. This allows
the reader to assess the importance of the assumptions made for the central case. Some of these
variables may be affected by a regulation, particularly the assumed rate of technological innovation
(see Chapter 5 for additional guidance on specifying baselines).
The impact of using alternative assumptions or alternative models can be assessed quantitatively
in many cases through sensitivity analysis and presenting alternatives, as described in Chapter 5. In
addition to explaining the uncertainty in a model's parameters, analysts should discuss the
uncertainty generated by the choice of model. Multiple models are often available and choosing
among them is similar to making an assumption. Implicit in the choice of a model are many factors.
For example, one model may take long-run effects into account while another model does not
When possible, presenting results of an alternate model can inform the reader. When resource
limitations prevent the use of an alternative model, it is still often possible to predict the direction
and likely magnitude of the use of an alternate model, and the analyst should present this
information to the reader.
11.3 Use of Economic Analyses
The primary purpose of conducting economic analysis is to provide policy makers and others with
detailed information on a wide variety of consequences of environmental policies. One important
element these analyses have traditionally provided to the policy-making process is estimates of
social benefits and costs the economic efficiency of a policy. For this reason, these Guidelines
11-15 Guidelines for Preparing Economic Analyses | 3rd edition
-------
reflect updated information associated with procedures for calculating benefits and costs,
monetizing benefits estimates and selecting particular inputs and assumptions.
Determining which regulatory options are best even on the restrictive terms of economic efficiency
is often made difficult by uncertainties in data and by the presence of benefits and costs that can be
quantified but not monetized, or that can only be qualitatively assessed. Even if the criterion of
economic efficiency were the sole guide to policy decisions, social benefit and costs estimates alone
would not be sufficient to define the best policies.
A large number of social goals and statutory and judicial mandates motivate and shape
environmental policy. For this and other reasons, these Guidelines contain information concerning
procedures for conducting analyses of other consequences of environmental policies, such as
economic impacts and equity effects. This is consistent with the fact that economic efficiency is not
the sole criterion for developing good public policies.
Even the most comprehensive economic analyses are but part of a larger policy development
process, one in which no individual analytical feature or empirical finding dominates. The role of
economic analysis is to organize information and comprehensively assess the economic
consequences of alternative actions benefits, costs, economic impacts and equity effects and
the trade-offs among them. Ultimately statutory requirements dictate if and how the analytic
results are used in standard setting. In any case, these results, along with other analyses and
considerations, serve as important inputs for the broader policy-making process and serve as
important resources for the public.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Chapter 11 References
Institute of Medicine (IOM). 2006. Valuing Health for Regulatory Cost-Effectiveness Analysis, ed. Wilhelmine
Miller, Lisa A. Robinson, and Robert S. Lawrence. Washington, DC: The National Academies Press.
Jena, A., C. Mulligan, T.J. Philipson, and E. Sun. The Value of Life in General Equilibrium. NBER Working Paper
14157. Available at: http://www.nber.org/papers/wl41S7 (accessed December 23, 2020).
0MB. 2010a. 2010 Report to Congress on the Benefits and Costs of Federal Regulations and Unfunded
Mandates on State, Local, and Tribal Entities. Available at:
https: //www.whitehouse.gOv/sites/whitehouse.gov/files/omb/legislative/reports/2010 Benefit Cost Repo
rt.pdf (accessed December 23,2020).
0MB. 2023. Circular A-4, Regulatory Analysis, November 9, 2023. Available at:
https://www.whitehouse.gov/wp-content/uploads/2023/ll/CircularA-4.pdf (accessed June 13, 2024).
Sieg, H., V.K. Smith, H.S. Banzhaf, and R. Walsh. 2004. Estimating the General Equilibrium Benefits of Large
Changes in Spatially Delineated Public Goods. International Economic Review, 45(4): 1047-1077.
U.S. EPA. 1997d. Guiding Principles for Monte Carlo Analysis. Risk Assessment Forum, EPA/630/R-97/001.
March. Available at: https: //www.epa.gov/risk/guiding-principles-monte-carlo-analysis (accessed December
23,2020)
U.S. EPA. 2002f. Guidance for Quality Assurance Project Plans EPA QA/G-5. Office of Environmental
Information, EPA/240/R-02/009. December.
US EPA. 2023a. CIO 2105.4 Environmental Information Quality Policy. August21, 2023.
US EPA. 2023b. CIO 2105-P-01.4 Environmental Information Quality Procedure. August 21, 2023.
Yang, T., K. Maus, S. Paltsev, and J. Reilly. 2004. Economic Benefits of Air Pollution Regulation in the USA: An
Integrated Approach. MIT Joint Program on the Science and Policy of Global Change Report No. 113. Available
at: http: //web.mit.edu/globalchange/www/MITIPSPGC Rptll3.pdf (accessed December 23, 2020).
11-17 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Appendix A - Economic Theory
This appendix provides a brief overview of the fundamental theory underlying
the approaches to economic analysis discussed in Chapters 3 through 9. The
first section summarizes the basic concepts of the forces governing a market
economy in the absence of government intervention. Section A. 2 describes
why markets may behave inefficiently. If the preconditions for market
efficiency are not met, government intervention can be justified.1 The
usefulness of benefit-cost analysis [BCA] as a tool to help policy makers
determine the appropriate policy response is discussed in Section A.3.
Sections A.4 and A.5 explain how economists measure the economic impacts
of a policy and set the optimal level of regulation. Section A.6 concludes and
provides a list of additional references.
> i P IUj Mr
The economic concept of a market is used to describe any situation where exchange takes place
between consumers and producers. Economists assume that consumers purchase the combination
of goods that maximizes their well-being, or "utility," given market prices and subject to their
household budget constraint. Economists also assume that producers (firms) act to maximize their
profits. Economic theory posits that consumers and producers are rational agents who make
decisions taking into account all of the costs the full opportunity costs of their choices, given
their own resource constraints.2 The purpose of economic analysis is to understand how the agents
interact and how their interactions add up to determine the allocation of society's resources: what
is produced, how it is produced, for whom it is produced and how these decisions are made. The
simplest tool economists use to illustrate consumers' and producers' behavior is a market diagram
with supply and demand curves.
The demand curve for a single individual shows the quantity of a good or service that the individual
will purchase at any given price. This quantity demanded assumes the condition of holding all else
constant (i.e., assuming the budget constraint, information about the good, expected future prices,
prices of other goods, etc. remain constant). The height of the demand curve in Figure A.1 indicates
the maximum price, P, an individual with Qd units of a good or service would be willing to pay to
acquire an additional unit of a good or service. This amount reflects the satisfaction (or utility) the
individual receives from an additional unit, known as the marginal benefit of consuming the good.
Economists generally assume that the marginal benefit of an additional unit is slightly less than that
1 The EPA's mandates frequently rely on criteria other than economic efficiency, so policies that are not justified
due to a lack of efficiency are sometimes adopted.
2 Opportunity cost is the next best alternative use of a resource. The full opportunity cost of producing
(consuming) a good or service consists of the maximum value of other goods and services that could have been
produced (consumed) had one not used the limited resources to produce (purchase) the good or service in
question. For example, the full cost of driving to the store includes not only the price of gas but also the value of
the time required to make the trip.
A-1 Guidelines for Preparing Economic Analyses | 3rd edition
-------
realized by the previous unit The amount an individual is willing to pay for one more unit of a good
is less than the amount she paid for the last unit; hence, the individual demand curve slopes
downward. A market demand curve shows the total quantity that consumers are willing to
purchase at different price levels (i.e., their collective willingness to pay (WTP) for the good or
service). In other words, the market demand curve is the horizontal sum of all the individual
demand curves.
Figure A.1 - Market and Total WTP
Price
Figure A.2 - Marginal and Total Cost
The concept of an individual's WTP is one of the fundamental concepts used in economic analyses,
and it is important to distinguish between total and marginal WTP. Marginal WTP is the additional
amount the individual would pay for one additional unit of the good. The total WTP is the aggregate
amount the individual is willing to pay for the total quantity demanded (Qd). Figure A.1 illustrates
the difference between the marginal and total WTP. The height of the demand curve at a quantity
Qa-i gives the marginal WTP for the Qd-ith unit. The height of the demand curve at a quantity Q0 gives
the marginal WTP for the Qdth unit Note that the marginal WTP is greater for the Qd-ith unit. The
total WTP is equal to the sum of the marginal WTP for each unit up to Qd. The shaded area under the
demand curve from the origin up to Qd shows total WTP.
An individual producer's supply curve shows the quantity of a good or service that an individual or
firm is willing to sell (Qs) at a given price. As a profit-maximizing agent, a producer will only be
willing to sell another unit of the good if the market price is greater than or equal to the cost of
A-2 Guidelines for Preparing Economic Analyses | 3rd edition
-------
producing that unit The cost of producing the additional unit is known as the marginal cost.
Therefore, the individual supply curve traces out the marginal cost of production and is also the
marginal cost curve. Economists generally assume that the cost of producing one additional unit is
greater than the cost of producing the previous unit because resources are scarce. Therefore, the
supply curve is assumed to slope upward. In Figure A.2, the marginal cost of producing the Qsth unit
of the good is given by the height of the supply curve at Qs. The marginal cost of producing the Qs+ith
unit of the good is given by the height of the supply curve at Qs+i, which greater than the cost of
producing the Qsth unit, and greater than the price, P. The total cost of producing Qs units is equal to
the shaded area under the supply curve from the origin to the quantity Qs. The market supply curve
is simply the horizontal summation of the individual producers' marginal cost curves for the good
or service in question.
In a competitive market economy, the intersection of the market demand and market supply curves
determines the equilibrium price and quantity of a good or service sold. The demand curve reflects
the marginal benefit consumers receive from purchasing an extra unit of the good (i.e., it reflects
their marginal WTP for an extra unit). The supply curve reflects the marginal cost to the firm of
producing an extra unit. Therefore, at the competitive equilibrium, the price is where the marginal
benefit equals the marginal cost This is illustrated in Figure A.3, where the supply curve intersects
the demand curve at equilibrium price Pm and equilibrium quantity Qm.
A counterexample illustrates why the equilibrium price and quantity occur at the intersection of the
market demand and supply curves. In Figure A.3, consider some price greater than Pm where Qs is
greater than Qd (i.e., there is excess supply). As producers discover that they cannot sell off their
inventories, some will reduce prices slightly, hoping to attract more customers. At lower prices
consumers will purchase more of the good (Qd increases) although firms will be willing to sell less (Qs
decreases). This adjustment continues until Qd equals Qs. The reverse situation occurs if the price
becomes lower than Pm. In that case, Qd will exceed Qs (i.e., there is excess demand] and consumers who
cannot purchase as much as they would like are willing to pay higher prices. Therefore, firms will begin
to increase prices, causing some reduction in the Qd but also increasing Qs. Prices will continue to rise
until Qs equals Qd. At this point no purchaser or supplier will have an incentive to change the price or
quantity; hence, the market is said to be in equilibrium.
Economists measure a consumer's net benefit from consuming a good or service as the excess amount
that she is willing to spend on the good or service over and above the market price. The net benefit of
all consumers is the sum of individual consumer's net benefits i.e., what consumers are willing to
spend on a good or service over and above that required by the market This is called the consumer
surplus. In Figure A3, the market demands price Pm for the purchase of quantity Qm. However, the
demand curve shows that there are consumers willing to pay more than price Pm for all units prior to
Qm-
Therefore, the consumer surplus is the area under the market demand (marginal benefit) curve but
above the market price. Policies that affect market conditions in ways that decrease prices by
decreasing costs of production (i.e., that shift the marginal cost curve to the right) will generally
increase consumer surplus. This increase can be used to measure the benefits that consumers receive
from the policy.3
3 Section A.4.2 provides a more technical discussion of how consumer surplus serves as a measure of benefits.
A-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Figure A.3 - Market Equilibrium
Figure A.4 - Utility Possibility Frontier
V
c
\ H
A
\ UPF
\D
U,
On the supply side, a producer can be thought to receive a benefit if he can sell a good or service for
more than the cost of producing an additional unit i.e., its marginal cost. Figure A.3 shows that
there are producers willing to sell up to Qm units of the good for less than the market price Pm.
Hence, the net benefit to producers in this market, known as producer surplus, can be measured as
the area above the market supply (marginal cost) curve but below the market price. Policies that
increase prices by increasing market demand for a good (i.e., that shift the marginal benefit curve to
the right) will generally increase producer surplus. This increase can be used to measure the
benefits that producers receive from the policy.
Economic efficiency is defined as the maximization of social welfare. In other words, the efficient
level of production is one that allows society to derive the largest possible net benefit from the
market. This condition occurs where the (positive) difference between the total WTP and total costs
is the largest In the absence of externalities and other market failures (explained below), this
occurs precisely at the intersection of the market demand and supply curves where the marginal
benefit equals the marginal cost. This is also the point where total surplus (consumer surplus plus
producer surplus) is maximized. There is no way to rearrange production or reallocate goods so
that someone is made better off without making someone else worse off a condition known as
Pareto optimality. Notice that economic efficiency requires only that net benefits be maximized,
irrespective of to whom those net benefits accrue. It does not guarantee an "equitable" or "fair"
distribution of these surpluses among consumers and producers, or between sub-groups of
consumers or producers.
A-4 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Economists maintain that if the economic conditions are such that there are no market imperfections
(as discussed in Section A.2), then this condition of Pareto-optimal economic efficiency occurs
automatically.4 That is, no government intervention is necessary to maximize the sum of consumer
surplus and producer surplus. This theory is summarized in the two Fundamental Theorems of
Welfare Economics, which originate with Pareto (1906) and Barone (1908):
1. First Fundamental Welfare Theorem. Every competitive equilibrium is Pareto-optimal.
2. Second Fundamental Welfare Theorem. Every Pareto-optimal allocation can be achieved
as a competitive equilibrium after a suitable redistribution of initial endowments.
One graphical representation of these results is given in Figure A.4, which shows utility (welfare)
levels in a two-person economy.5 The curve shown is the utility possibility frontier (UPF) curve; the
area within it represents the set of all possible welfare outcomes. Each point on the negatively
sloped UPF curve is Pareto optimal since it is not possible to increase the utility of one person
without decreasing the utility of the other. If the initial allocation is at point A, then the set of
Pareto-superior (welfare-enhancing) outcomes include all points in the shaded area, bordered by H,
V and the UPF curve.6 If trading is permitted, the First Welfare Theorem applies and the market will
move the economy to a superior, more efficient point such as B. Then the Second Welfare Theorem
simply says that for any chosen point along the UPF curve, given a set of lump sum taxes and
transfers, an initial allocation can be determined inside the UPF from which the market will achieve
the desired outcome.7
If the market supply and demand curves reflect society's true marginal social cost and WTP, then a
laissez-faire market (i.e., one governed by individual decisions and not government authority) will
produce a socially efficient result However, when markets do not fully represent social values, the
private market will not achieve the efficient outcome (see Mankiw 2004, or any basic economics
text); this is known as a market failure. Market failure is primarily the result of externalities, market
power and inadequate or asymmetric information. Externalities are the most likely cause of the
failure of private and public sector institutions to account for environmental damages.
4 Technically, there are two types of efficiency. Allocative efficiency means that resources are used for the
production of goods and services most wanted by society. Productive efficiency implies that the least costly
production techniques are used to produce any mix of goods and services. Allocative efficiency requires that
there be productive efficiency, but productive efficiency can occur without allocative efficiency. Goods can be
produced at the least-costly method without being most wanted by society. Perfectly competitive markets in the
long run will achieve both of these conditions, producing the "right" goods (allocative efficiency) in the "right"
way (productive efficiency). These two conditions imply Pareto-optimal economic efficiency (see Varian 1992 or
any basic economics text for a more detailed discussion),
5 Another, perhaps more commonly used, graphical tool to explain the First and Second Welfare Theorems is an
Edgeworth box. See Varian (1992) or other basic economic textbook for a detailed discussion.
6 Note that efficiency could be obtained by moving along the vertical line V, which keeps utility of person 1 (Ui)
constant while increasing utility of person 2 (U2), or by moving along the horizontal line H, which only shows
improvements in utility for person 1. Moving to point B improves the utility for both individuals.
7 Note that outcomes on the frontier such as C and D, although efficient, may not be desired on equity (or
fairness) grounds.
A-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Externalities occur when markets do not account for the effect of one individual's decisions on
another individual's well-being.8 In a free market producers make their decisions about what and
how much to produce, taking into account the cost of the required inputs labor, raw materials,
machinery and energy. Consumers purchase goods and services taking into account their income
and their own tastes and preferences. This means that decisions are based on the private costs and
private benefits to market participants. If the consumption or production of these goods and
services poses an external cost or benefit on those not participating in the market, however, then
the market demand and supply curves no longer reflect the true marginal social benefit and
marginal social cost Hence, the market equilibrium will no longer be the socially (Pareto) efficient
outcome.
Externalities can arise for many reasons. Transactions costs or poorly defined property rights can
make it difficult for injured parties to bargain or use legal means to ensure that costs of damages
caused by polluters are internalized into their decision making.9 Activities that pose environmental
risks may also be difficult to link to resulting damages and often occur over long time periods.
Externalities involve goods that people care about but are not sold in markets.10 Air pollution
causes ill health, ecological damage and visibility impacts over a long time, and the damage is often
far from the source(s) of the pollution. The additional social costs of air pollution are not included
in firms' profit maximization decisions and so are not considered when firms decide how much
pollution to emit The lack of a market for clean air causes problems and provides the impetus for
government intervention in markets involving polluting industries.
8 More formally, an externality occurs when the production or consumption decision of one party has an
unintended negative (positive) impact on the profit or utility of a third party. Even if one party compensates the
other party, an externality still exists (Perman et al. 2003). See Baumol and Oates (1988) or any basic economics
textbook for similar definitions and more detailed discussion.
9 A property right can be defined as a bundle of characteristics that confer certain powers to the owner of the
right: the exclusive right to the choice of use of a resource, the exclusive right to the services of a resource and the
right to exchange the resource at mutually agreeable terms. Externalities arise from the violation of one or more
of the characteristics of well-defined property rights. This implies that the distortions resulting from an externality
can be eliminated by appropriately establishing these rights. This insight is summarized by the famous "Coase
theorem" which states that if property rights over an environmental asset are clearly defined, and bargaining
among owners and prospective users of the asset is allowed, then externality problems can be corrected and the
efficient outcome will result regardless of who was initially given the property right. The seminal paper is Coase
[I960).
10 Often these are goods that exhibit public good characteristics. Pure public goods are those that are non-
rivalrous in consumption and non-excludable (see Perman et al. (2003) for a detailed discussion of these, as well
as congestible and open access resources i.e., goods that are neither pure public nor pure private goods.)
Because exclusive property rights cannot be defined for these types of goods, pure private markets cannot
provide for them efficiently.
A-6 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Figure A.5 - Negative Externality
Price
- D (=MB)
0
Q* Q,
Quantity
'm
Figure A,5 illustrates a negative externality associated with the production of a good. For example, a
firm producing some product might also be generating pollution as a by-product. The pollution may
impose significant costs in the form of adverse health effects, for example on households
living downwind or downstream of the firm. Because those costs are not borne by the firm, the firm
typically does not consider them in its production decisions. Society considers the pollution a cost
of production, but the firm typically will not In this figure:
D is the market demand (marginal benefit) curve for the product;
MPC is the firm's marginal private real-resource cost of production, excluding the cost of the
firm's pollution on households;
MSD is the marginal social damage of pollution (or the marginal external cost) that the firm
is not considering; and
MSC is society's marginal social cost associated with production, including the cost of
poll utio n (MSC = MPC + MSD).
In an incomplete market, producers pay no attention to external costs, and production occurs
where market demand (D) and the marginal private real-resource cost (MPC] curves intersect at
a price P, and a quantity Q,. In this case, net social welfare (total WTP minus total social costs) is
equal to the area of the triangle P0PiX less the area of triangle XVZ.11 If the full social cost of
production, including the cost of pollution, is taken into consideration, then the marginal cost curve
should be increased by the amount of the marginal social damage (MSD) of pollution.12 Production
will now occur where the demand and marginal social cost (MSC) curves intersect at a price P*
and a quantity Q*. At this point net social welfare (now equal to the area of the triangle, P0P1X,
alone) is maximized, and therefore the market is at the socially efficient point of production. This
example shows that when there is a negative externality such as pollution, and the social damage
11 Recall from Section A.1 that total WTP is equal to the area under the demand curve from the origin to the
point of production (OPiZQm). Total costs (to society] are equal to the area under the MSC curve from the origin
to the point of production (OPoYQm).
12 When conducting BCA related to resource stocks, the MSD or marginal external cost is the present value of
future net benefits that are lost to due to the use of the resource at present. That is, exhaustible resources used
today will not be available for future use. These foregone future benefits are called user costs in natural resource
economics (see Scott 1953,1955). The marginal user cost is the user cost of one additional unit consumed in the
present and is added together with the marginal extraction cost to determine the MSC of resource use.
A-7 Guidelines for Preparing Economic Analyses | 3rd edition
-------
(external cost) of that pollution is not taken into consideration, the producer will oversupply the
polluting good.13 The shaded triangle (XYZ), referred to as the deadweight loss (DWL), represents
the amount that society loses by producing too much of the good.
A.3 Benefit-Cost Analysis
If a negative externality such as pollution exists, an unregulated market will not account for its cost
to society, and the result will be an inefficient outcome. In this case, there may be a need for
government intervention to correct the market failure. A correction may take the form of dictating
the allowable level of pollution or introducing a market mechanism to induce the optimal level of
pollution.14 Figure A.5 neatly summarizes this in a single market diagram. To estimate the total
costs and benefits to society of an activity or program, the costs and benefits in each affected
market, as well as any non-market costs or benefits, are added up. This is done through BCA.
BCA can be thought of as an accounting framework of the overall social welfare of a program, which
illuminates the trade-offs involved in making different social investments (Arrow et al. 1996). It is
used to evaluate the favorable effects of a policy action and the associated opportunity costs. The
favorable effects of a regulation are the benefits, and the foregone opportunities or losses in utility
are the costs. Subtracting the total costs from the total monetized benefits provides an estimate of
the regulation's net benefits to society. An efficient regulation is one that yields the maximum net
benefit, assuming that the benefits can be measured in monetary terms.
BCA can also be seen as a type of market test for environmental protection. In the private market, a
commodity is supplied if the benefits that society gains from its provision, measured by what
consumers are willing to pay, outweigh the private costs of producing the commodity. Economic
efficiency is measured in a private market as the difference between what consumers are willing to
pay for a good and what it costs to produce it Since clean air and clean water are public goods,
private suppliers cannot capture their value and sell it The government determines their provision
through environmental protection regulation. BCA quantifies the benefits and costs of producing
this environmental protection in the same way as the private market, by quantifying the WTP for
the environmental commodity. As with private markets, the efficient outcome is the option that
maximizes net benefits.
The key to performing BCA lies in the ability to measure both benefits and costs in monetary terms
so that they are comparable. Consumers and producers in regulated industries and the
governmental agencies responsible for implementing and enforcing the regulation (and by
extension, taxpayers in general) typically pay the costs. The total cost of the regulation is found by
summing the costs to these individual sectors (see Section A.4.3 for an example, excluding the costs
to the government). Since environmental regulation usually addresses some externality, the
benefits of a regulation often occur outside of markets. For example, the primary benefits of
drinking water regulations are improvements in human health. Once the expected reduction in
illness and premature mortality associated with the regulation is calculated, economists use a
13 Similarly, the private market will undersupply goods for which there are positive externalities, such as parks
and open space.
14 Chapter 4 discusses the various regulatory techniques and some non-regulatory means of achieving pollution
control.
A-8 Guidelines for Preparing Economic Analyses | 3rd edition
-------
number of techniques to estimate the value that society places on these health improvements.15
These monetized benefits can then be summed to obtain the total benefits from the regulation.
Note that in BCA gains and losses are weighted equally regardless of to whom they accrue.
Evaluation of the fairness, or the equity, of the net gains cannot be made without specifying a social
welfare function. However, there is no generally agreed-upon social welfare function and assigning
relative weights to the utility of different individuals is an ethical matter that economists strive to
avoid. Given this dilemma, economists have tried to develop criteria for comparing alternative
allocations where there are winners and losers without involving explicit reference to a social
welfare function. According to the Kaldor-Hicks compensation test, named after its originators
Nicholas Kaldor and J.R. Hicks, a reallocation is a welfare-enhancing improvement to society if:
1. The winners could theoretically compensate the losers and still be better off; and
2. The losers could not, in turn, pay the winners to not have this reallocation and still be as
well off as they would have been if it did occur (Perman et al. 2003).
While these conditions sound complex, they are met in practice by assessing the net benefits of a
regulation through BCA. The policy that yields the highest positive net benefit is considered welfare
enhancing according to the Kaldor-Hicks criterion. Note that the compensation test is stated in
terms of potential compensation and does not solve the problem of evaluating the fairness of the
distribution of well-being in society. Whether and how the beneficiaries of a regulation should
compensate the losers involves a value judgment and is a separate decision for government to
make.
Finally, BCA may not provide the only criterion used to decide if a regulation is in society's best
interest There are often other, overriding considerations for promulgating regulation. Statutory
instructions, political concerns, institutional and technical feasibility, enforceability and
sustainability are all important considerations in environmental regulation. In some cases, a policy
may be considered desirable even if the benefits to society do not outweigh its costs, particularly if
there are ethical or equity concerns.16 There are also practical limitations to BCA. Most importantly,
this type of analysis requires assigning monetized values to non-market benefits and costs. In
practice it can be very difficult or even impossible to quantify gains and losses in monetary terms
(e.g., the loss of a species, intangible effects).17 In general, however, economists believe that BCA
provides a systematic framework for comparing the social costs and benefits of proposed
regulations, and that it contributes useful information to the decision-making process about how
scarce resources can be put to the best social use.
! Mlea^snn:;.
sticities
The net change in social welfare brought about by a new environmental regulation is the sum of the
negative effects (i.e., loss of producer and consumer surplus) and the positive effects (or social
benefits) of the improved environmental quality. This is shown graphically for a single market in
15 Chapter 7 discusses a variety of methods economists use to value environmental improvements.
16 Chapter 9 addresses equity assessment and describes the methods available for examining the distributional
effects of a regulation.
17 Kelman (1981) argues that it is even unethical to try to assign quantitative values to non-marketed benefits.
A-9 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Figure A,5 above. The use of demand and supply curves highlights the importance of assessing how
individuals will respond to changes in market conditions. The net benefits of a policy will depend
on how responsively producers and consumers react to a change in price. Economists measure this
responsiveness by the supply and demand elasticities.
Figure A.6 - Demand Curve for Tuna
$/lb
The term "elasticity" refers to the sensitivity of one variable to changes in another variable. The
price elasticity of demand (or supply] for a good or service is equal to the percentage change in the
quantity demanded (or supplied) that would result from a 1% increase in the price of that good or
service. For example, a price elasticity of demand for tuna equal to -1 means that a 1% increase in
the price of tuna results in a 1% decrease in the quantity demanded. Changes are measured
assuming all other things, such as incomes and tastes, remain constant Demand and supply
elasticities are rarely constant and often change depending on the quantity of the good consumed
or produced. For example, according to the demand curve for tuna shown in Figure A.6, at a price of
$1 per pound, a 10% increase in price would reduce quantity demanded by 2.5% (from 8 lbs to 7.8
lbs). At a price of $4 per pound, a 10% increase in price would result in a 40% decrease in quantity
demanded (from 2 to 1.2 lbs). This implies that the price elasticity of demand is -0.25 when tuna
costs Sl/Ib but -4 when the price is $4/lb. When calculating elasticities it is important realize where
one is on the supply or demand curve, and the price or quantity should be stated when reporting an
elasticity estimate.
Elasticities are important in measuring economic impacts because they determine how much of a
price increase will be passed on to the consumer. For example, if a pollution control policy leads to
an increase in the price of a good, multiplying the price increase by current quantity sold generally
will not provide an accurate measure of impact of the policy. Some of the impact will take the form
of higher prices for the consumer, but some of the impact will be a decrease in the quantity sold.
The amount of the price increase that is passed on to consumers is determined by the elasticity of
demand relative to supply (as well as existing price controls). "Elastic" demand (or supply)
indicates that a small percentage increase in price results in a larger percentage decrease (increase)
A-10 Guidelines for Preparing Economic Analyses | 3rd edition
-------
in quantity demanded (supplied).18 All else equal, an industry facing a relatively elastic demand is
less likely to pass on costs to the consumer because increasing prices will result in reduced
revenues. In determining the economic impacts of a rule, supply characteristics in the industries
affected by a regulation can be as important as demand characteristics. For highly elastic supply
curves relative to the demand curves, it is likely that cost increases or decreases will be passed on
to consumers.
The many variables that affect the elasticity of demand include:
The cost and availability of close substitutes;
The percentage of income a consumer spends on the good;
How necessary the good is for the consumer;
The amount of time available to the consumer to locate substitutes;
The expected future price of the good; and
The level of aggregation used in the study to estimate the elasticity.
The availability of close substitutes is one of the most important factors that determine demand
elasticity. A product with close substitutes at similar prices tends to have an elastic demand,
because consumers can readily switch to substitutes rather than paying a higher price. Therefore, a
company is less likely to be able to pass through costs if there are many close substitutes for its
product. Narrowly defined markets (e.g., salmon) will have more elastic demands than broadly
defined markets (e.g., food) since there are more substitutes for narrow goods.
Another factor that affects demand elasticities is whether the affected product represents a
substantial or necessary portion of customers' costs or budgets. Goods that account for a
substantial portion of consumers' budgets or disposable income tend to be relatively price elastic.
This is because consumers are more aware of small changes in the price of expensive goods
compared to small changes in the price of inexpensive goods, and therefore may be more likely to
seek alternatives. A similar issue concerns the type of final good involved. Reductions in demand
may be more likely to occur when prices increase for "luxuries" or optional purchases. If the good is
a necessity item, the quantity demanded is unlikely to change drastically for a given change in price.
Demand will be relatively inelastic.
Elasticities tend to increase over time, as firms and customers have more time to respond to
changes in prices. Although a company may face an inelastic demand curve in the short run, it could
experience greater losses in sales from a price increase in the long run. Over time customers begin
to find substitutes or new substitutes are developed. However, temporary price changes may affect
consumers' decisions differently than permanent ones. The response of quantity demanded during
a one-day sale, for example, will be much greater than the response of quantity demanded when
prices are expected to decrease permanently. Finally, it is important to keep in mind that elasticities
differ at the firm versus the industry level. It is not appropriate to use an industry-level elasticity to
estimate the ability of only one firm to pass on compliance costs when its competitors are not
subject to the same cost.
Characteristics of supply in the industries affected by a regulation can be as important as demand
characteristics in determining the economic impacts of a rule. For relatively elastic supply curves, it
is likely that cost increases or decreases will be passed on to consumers. The elasticity of supply
18 Demand (or supply) is said to be "elastic" if the absolute value of the price elasticity of demand (supply) is
greater than one and "inelastic" if the absolute value of the elasticity is less than one. If a percentage change in
price leads to an equal percentage change in quantity demanded (supplied) (i.e., if the absolute value of elasticity
equals one), demand (supply) is "unit elastic."
A-11 Guidelines for Preparing Economic Analyses | 3rd edition
-------
depends, in part, on how quickly per unit costs rise as firms increase their output Among the many
variables that influence this rise in cost are:
The cost and availability of close input substitutes;
The amount of time available to adjust production to changing conditions;
The degree of market concentration among producers;
The expected future price of the product;
The price of related inputs and related outputs; and
The speed of technological advances in production that can lower costs.
Similar to the determinants of demand elasticity, the factors influencing the price elasticity of
supply all relate to a firm's degree of flexibility in adjusting production decisions in response to
changing market conditions. The more easily a firm can adjust production levels, find input
substitutes or adopt new production technologies, the more elastic is supply. Supply elasticities
tend to increase over time as firms have more opportunities to renegotiate contracts and change
production technologies. When production takes time, the quantity supplied may be more
responsive to expected future price changes than to current price changes.
Demand and supply elasticities are available for the aggregate output of final goods in most
industries. They are usually published in journal articles on research pertaining to a particular
industry.19 When such information is unavailable, as is often the case for intermediate goods,
elasticities may be quantitatively or qualitatively assessed.20 Econometric tools are frequently used
to estimate supply and demand equations (thereby the elasticities) and the factors that influence
them.
A.4,2 Measuring the Welfare 11 - r i 1 ,11111c - in Ei rit nini'-mtal
Goods
As introduced in Section A.1, changes in consumer surplus are measured by the trapezoidal region
below the ordinary, or Marshallian, demand curve as price changes. This region reflects the benefit
a consumer receives by being able to consume more of a good at a lower price. If the price of a good
decreases, some of the consumer's satisfaction comes from being able to consume more of a
commodity when its price falls, but some of it comes from the fact that the lower price means that
the consumer has more income to spend. However, the change in (Marshallian) consumer surplus
only serves as a monetary measure of the welfare gain or loss experienced by the consumer under
the strict assumption that the marginal utility of income is constant21 This assumption is almost
never true in reality. Luckily, there are alternative, less demanding monetary measures of
19 Another useful source of elasticity estimates is the recently developed EPA Elasticity Databank (U.S. EPA
2007d). In the absence of an encyclopedic, "Book of Elasticities," the Elasticity Databank, serves as a searchable
database of elasticity parameters across a variety of types (i.e., demand and supply elasticities, substitution
elasticities, income elasticities and trade elasticities) and economic sectors/product markets. The database is
populated with EPA-generated estimates used in Environmental Impact Assessment studies conducted by the
Agency since 1990, as well as estimates found in the economics literature. It can be accessed from the Technology
Transfer Network Economics and Cost Analysis Support website: http://www.epa.gov/ttnecasl/Elasticity.htm.
20 Final goods are those that are available for direct use by consumers and are not utilized as inputs by firms in
the process of production. Goods that contribute to the production of a final good are called intermediate goods.
It is of course possible for a good to be final from one perspective and intermediate from another (Pearce 1992).
21 See Perman et al. (2003), fust et al. (2005) or any graduate level text for a more thorough exposition of this
issue.
A-12 Guidelines for Preparing Economic Analyses | 3rd edition
-------
consumer welfare that prove useful in treatments of BCA. Intuitively, these measures determine
the size of payment that would be necessary to compensate the consumer for the price change. In
other words, they estimate the consumer's WTP for a price change.
As mentioned above, a price decline results in two effects on consumption. The change in relative
prices will increase consumption of the cheaper good (the substitution effect), and consumption
will be affected by the change in overall purchasing power (the income effect). A Marshallian
demand curve reflects both substitution and income effects. Movements along it show how the
quantity demanded changes as price changes (holding all other prices and income constant), so it
reflects both the substitution and the income effects. The Hicksian (or "compensated") demand
curve, on the other hand, shows the relationship between quantity demanded of a commodity and
its price, holding all other prices and utility (rather than income) constant. This is the correct
measure of a consumer's WTP for a price change. The Hicksian demand curve is constructed by
adjusting income as the price changes so as to keep the consumer's utility the same at each point on
the curve. In this way, the income effect of a price change is eliminated and the substitution effect
can be considered alone. Movements along the Hicksian demand function can be used to determine
the monetary change that would compensate the consumer for the price change.
Hicks (1941) developed two correct monetary measures of utility change associated with a price
change: compensating variation and equivalent variation. Compensating variation (CV) assesses how
much money must be taken away from consumers after a price decrease occurred to return them to
the original utility level. It is equal to the amount of money that would "compensate" the consumer
for the price decrease. Equivalent variation (EV) measures how much money would need to be given
to the consumer to bring the consumer to the higher utility level instead of introducing the price
change. In other words, it is the monetary change that would be 'equivalent' to the proposed price
change.
Before examining the implications of these measures for valuing environmental changes, it is useful
to understand CV and EV in the case of a reduction in the price of some normal, private good, Ci.22
This is shown with indifference curves and a budget line, as seen in Figure A.7.
Figi nee Curve
-.... I
22 The notation and discussion in this section follow Chapter 12 ofPerman et al. (2003).
A-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Assume that the consumer is considering the trade-off between Ci and all other goods, denoted by a
composite good, C2. The indifference curve, U0, depicts the different combinations of the two goods
that yield the same level of utility. Because of diminishing marginal utility, the curve is concave,
where increasing amounts of C? must be offered for each unit of C2 given up to keep the consumer
indifferent The budget line on the graph reflects what the consumer is able to purchase given her
income, V0, and the prices of the two goods P/ and P/, respectively.23 A utility-maximizing
consumer will choose quantities Ci? and C2, the point where the indifference curve is tangent to the
budget constraint24
Figure A.8 shows the change in the optimal consumption bundle resulting from a reduction in the
price of Ci. If the price of Ci falls, the budget line shifts out on the Ci axis because more Ci can be
purchased for a given amount of money. The consumer now chooses Cf and C2" at point b and
moves to a new, higher utility curve, t/j. CV then measures how much money must be taken away at
the new prices to return the consumer to the old utility level. That is, starting at point b and keeping
the slope of the budget line fixed at the new level, by how much must it be shifted downward to
make it tangent to the initial indifference curve, U01 It is, therefore, the maximum amount the
consumer would be willing to pay to have the price fall occur i.e., the precise monetary measure
of the welfare change.25 In Figure A.8, CV is simply given by the amount Y0 - Yh EV, on the other
hand, measures how much income must be given to the individual at the old price set to maintain
the same level of well-being as if the price change did occur. That is, keeping the slope of the budget
line fixed at the old level, by how much must it be shifted upwards to make it tangent to Ui? EV is,
then, the minimum amount of money the consumer would accept in lieu of the price fall. This too is
a proper monetary measure of the utility change resulting from the price decrease. In Figure A.8
then EV is the amount Y2 - Y0, leaving the individual at point/.
Figure A.8 - Change in Optimal Consumption Bundle
23 In Figure A 7, C2 is considered the numeraire good (i.e., prices are adjusted so that P2' is equal to 1),
24 For a review of the utility maximizing behavior of consumers, see any general microeconomics textbook,
25 In Figure A.8, this would result in a shift from Ci"to Ci*. This is known as the income effect of the price
change. The shift from Ci' to Ci* is considered the substitution effect.
A-14 Guidelines for Preparing Economic Analyses | 3rd edition
-------
CV and EV are simply measures of the distance between the two indifference curves. However, the
amount of money associated with CV, EV and Marshallian consumer surplus (MCS) is generally not
the same. For a price fall, it can be shown that CV < MCS < EV, and for a price increase, CV > MCS >
EV.26 Notice that in the case of a price decrease, the CV measures the consumer's willingness to pay
(WTP) to receive the price reduction and EV measures the consumer's willingness to accept (WTA)
to forgo the lower price. If the price of Ci were to increase, then the relationships between
WTP/WTA and CV /EV would be reversed. CV would measure the consumer's WTA to suffer the
price increase and EV would be the individual's WTP to avoid the increase in price.
In order to examine the implications of these measures for valuing changes in environmental
conditions, one can think of Ci in the above discussion as an environmental commodity, henceforth
denoted by E. Then an improvement in environmental quality (or an increase in an environmental
public good) resulting from some policy is reflected by an increase in the amount of E. Holding all
else constant, such an increase is equivalent to a decrease in the price of E and can be depicted as a
shifting outward of the budget line along the E axis.
Welfare changes due to an increase in E follow along the lines of the previous discussion.
However, because E is generally non-exclusive and non-divisible, the consumer consumption
level cannot be adjusted. Therefore, the associated monetary measures of the welfare change are
not technically CV and EV but are referred to as compensating surplus (CS) and equivalent surplus
(ES). In practice, however, the process is the same; a Hicksian demand curve is estimated for the
unpriced environmental good. Analogous to the preceding discussion, if there is an
environmental improvement, then CS measures the amount of money the consumer would be
willing to pay for the improvement that would result in the pre-improvement level of utility. For
the purposes of environmental valuation, this is the primary measure of concern when
considering environmental improvements. ES measures how much society would have to pay
the consumer to give the consumer the same utility as if the improvement had occurred. In other
words, this is how much the consumer would be willing to accept to not experience the gain in
environmental quality. If valuing an environmental degradation, then CS measures the WTA and
ES measures WTP.
Whereas statements can be made about the relative size of CV, EV and MCS for price changes of
normal goods, Bockstael and McConnell (1993) find that it is not possible to make similar
statements about CS, ES and MCS for a change in environmental quality.27 Given that environmental
quality is generally an unpriced public good, ordinary Marshallian demand functions cannot be
estimated, so it may seem irrelevant that one cannot say anything about how MCS approximates the
proper measure. However, Bockstael and McConnell's results are important in relation to indirect
methods for environmental valuation. However, most indirect valuation studies are based on
Marshallian demand functions in practice, in the hope of keeping the associated error small.
26 This can be seen by redrawing Figure A.8 using a graph of Marshallian and Hicksian demand curves. See
Perman etal. (2003) for a detailed explanation.
27 Willig (1976) shows that ordinary, or Marshallian, demand curves can provide an approximate measure of
welfare changes resulting from a price change. In most cases the error associated with using MCS, with respect
to CV or EV, will be less than 5% (see Perman et al. 2003).
A-15 Guidelines for Preparing Economic Analyses | 3rd edition
-------
A.4.3 Single Market, Multi-Market and General Equilibrium Analysis
Both supply and demand elasticities are affected by the availability of close complements and
substitutes. This highlights the fact that regulating one industry can have an impact on other, non-
regulated markets. However, this does not necessarily imply that all of these other markets must be
modeled. Changes due to government regulation can be captured using only the equilibrium supply
and demand curves for the affected market, assuming: (1} there are small, competitive adjustments
in all other markets; and (2) there are no distortions in other markets. This is referred to as partial
equilibrium analysis.
For example, suppose a new environmental regulation increases per unit production costs. The
benefits and costs of abatement in a partial equilibrium setting are illustrated in Figure A.9 where
the market produces the quantity Qm in equilibrium without intervention. The external costs of
production are shown by the marginal external costs (MFC] curve without any abatement. Total
external costs are given by the area under the MEC curve up to the market output, Qm. or the area of
triangle QmEO.
Figure A.9 - Benefits and Costs of Abatement
Supply -i- MAC
With required abatement production, costs are the total of supply plus marginal abatement costs
(MAC), shown as the new, higher supply curve in the figure. These higher costs result in a new
market equilibrium quantity shown as Q*. The social cost of the requirement is the resulting change
in consumer and supplier surplus, shown here as the total observed abatement costs
(parallelogram P0P1AC) plus the area of triangle ABC, which can be described as deadweight loss.
Abatement also produces benefits by shifting the MEC curve downward, reflecting the fact that each
unit of production now results in less pollution and social costs. Additionally, the reduced quantity
of the output good results in reduced external costs. The reduced external costs, i.e., the benefits,
are given by the difference between triangle QmEO and triangle Q*DO, represented by the shaded
area in the figure.
The net benefits of abatement are the benefits (the reduced external costs] minus the costs (the loss
in consumer and producer surplus). In the figure, this would equal the shaded area (the benefits)
minus total abatement costs and deadweight loss as described above.
While the single market analysis is theoretically possible, it is generally impractical for rulemaking.
As mentioned in Section A.3, this is often because the gains occur outside of markets and cannot be
linked directly to the output of the regulated market Therefore, BCA is frequently done as two
separate analyses: a benefits analysis and a cost analysis.
A-16 Guidelines for Preparing Economic Analyses | 3rd edition
-------
When a regulation is expected to have a large impact outside of the regulated market, then the
analysis should be extended beyond that market. If the effects are significant but not anticipated to
be widespread, one potential improvement is to use multi-market modeling in which vertically or
horizontally integrated markets are incorporated into the analysis. The analysis begins with the
relationship of input markets to output markets. A multi-market analysis extends the partial
equilibrium analysis to measuring the losses in other related markets.28
In some cases, a regulation can have such a significant impact on the economy that a general
equilibrium modeling framework is required.29 This may be because regulation in one industry has
broad indirect effects on other sectors, households may alter their consumption patterns when they
encounter increases in the price of a regulated good, or there may be interaction effects between
the new regulation and pre-existing distortions, such as taxes on labor. In these cases, partial
equilibrium analyses are likely to result in an inaccurate estimation of total social costs. Using a
general equilibrium framework accounts for linkages between all sectors of the economy and all
feedback effects and can measure total costs comprehensively.30
* 7 *, >1 sjt i,¦ J L'-v I "f P.-" f \Utj ''>J i
Following from the definition in Section A. 1, the most economically efficient policy is the one that
allows for society to derive the largest possible social benefit at the lowest social cost This occurs
when the net benefits to society (i.e., total benefits minus total costs) are maximized. In Figure A.10,
this is at the point where the distance between the benefits curve and the costs curve is the largest
and positive.
28 An example of the use of multi-market model for environmental policy analysis is contained in a report
prepared for EPA on the regulatory impact of control on asbestos and asbestos products (U.S. EPA 1989).
29 General equilibrium analysis is built around the assumption that, for some discrete period of time, an
economy can be characterized by a set of equilibrium conditions in which supply equals demand in all markets.
When this equilibrium is "shocked" through a change in policy or a change in some exogenous variable, prices
and quantities adjust until a new equilibrium is reached. The prices and quantities from the post-shock
equilibrium can then be compared with their pre-shock values to determine the expected impacts of the policy or
change in exogenous variables.
30 Chapter 8 provides a more detailed discussion of partial equilibrium, multi-market and general equilibrium
analysis.
A-17 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Figure A.10 - Maximized Net Benefits
$
Costs
Pollution
Abatement
Benefits
Note that this is not necessarily the point at which:
Benefits are maximized;
Costs are minimized;
Total benefits = total costs (i.e., benefit-cost ratio = 1);
Benefit-cost ratio is the largest; or
The policy is most cost-effective.
If the regulation were designed to maximize benefits, then any policy, no matter how expensive,
would be justified if it produced any benefit, no matter how small. Similarly, minimizing costs
would, in most cases, simply justify no action at all. A benefit-cost ratio equal to one is equivalent to
saying that the benefits to society would be exactly offset by the cost of implementing the policy.
This implies that society is indifferent between no regulation and being regulated; hence, there
would be no net benefit from adopting the policy. Maximizing the benefit-cost ratio is not optimal
either. Two policy options could yield equivalent benefit-cost ratios but have vastly different net
benefits. For example, a policy that cost $100 million per year but produced $200 million in benefits
has the same benefit-cost ratio as a policy that cost $100,000 but produced $200,000 in benefits,
even though the first policy produces substantially more net benefit for society.31 Finally, finding
the most cost-effective policy has similar problems because the cost-effectiveness ratio can be seen
as the inverse of the benefit-cost ratio. A policy is cost effective if it meets a given goal at least cost
31 Benefit-cost ratios are useful when choosing one or more policy options subject to a budget constraint. For
example, consider a case where five options are available and the budget is SI, 000. The first option will cost
$1,000 and will deliver benefits of$2,000. Each of the other four will cost $250 and deliver benefits of$750. If
options are selected according to the net benefits criterion, the first option will be selected, because its net benefits
are $1,000 while the net benefits of each of the other options are $500. However, if options are selected by the
benefit-cost ratio criterion, the other four options will be selected, as each of their benefit-cost ratios equal 3,
versus a benefit-cost ratio of 2 for the first option. In this case, choosing options by the net benefits criterion will
yield $1,000 in total net benefits, while choosing options by the benefit-cost ratio criterion will yield $500 in total
net benefits. In most cases, choosing options in decreasing order of benefit-cost ratios will yield the largest possible
net benefits given a fixed budget. This method will guarantee the optimal solution if the benefits and costs of each
option are independent, and if each option can be infinitely subdivided, simply select the options in decreasing
order of their benefit-cost ratios and once the budget is exceeded subdivide the last option selected such that the
budget constraint is met exactly (see Dantzig 1957). Also note that this strategy does not require measuring
benefits and costs in the same units, which means that it is directly useful for CEA (Hyman and Leibowitz 2000],
while the net-benefit criterion is not.
A-18 Guidelines for Preparing Economic Analyses | 3rd edition
-------
i.e., minimizes the cost per unit of benefit achieved. Cost-effectiveness analysis (CEA) can provide
useful information to supplement existing BCA and may be appropriate to rank policy options when
the benefits are fixed and cannot be monetized, but it provides no guidance in setting an
environmental standard or goal.
Conceptually, net social benefits will be maximized if regulation is set such that emissions are reduced
up to the point where the benefit of abating one more unit of pollution (i.e., marginal social benefit)32 is
equal to the cost of abating an additional unit (i.e., marginal abatement cost).33 If the marginal benefits
are greater than the marginal costs, then additional reductions in pollution will offer greater benefits
than costs, and society will be better off. If the marginal benefits are less than marginal costs, then
additional reductions in pollution will cost society more than they provide in benefits and will make
society worse off. When the marginal cost of abatement is equal to society's marginal benefit, no gains
can be made from changing the level of pollution reduction, and an efficient aggregate level of
emissions is achieved. In other words, a pollution reduction policy is at its optimal, most economically
efficient point when the marginal benefits equal the marginal costs of the rule.34
The condition that marginal benefits must equal marginal costs assumes that the initial
pollution reduction produces the largest benefits for the lowest costs. As pollution reduction
(i.e., regulatory stringency) is increased, the additional benefits decline and the additional
costs rise. While not always true, a case can be made that the benefits of pollution reduction
follow this behavior. The behavior of total abatement costs, however, will depend on how the
pollution reduction is distributed among the polluters since firms may differ in their ability to
reduce emissions. The aggregate marginal abatement cost function shows the least costly way
of achieving reductions in emissions. It is equal to the horizontal sum of the marginal
abatement cost curves for the individual polluters. Although each firm faces increasing costs of
abatement, marginal cost functions still vary across sources. Some firms may abate pollution
relatively cheaply, while others require great expense. To achieve economic efficiency, the
lowest marginal cost of abatement must be achieved first, and then the next lowest. Pollution
reduction is achieved at lowest cost only if firms are required to make equiproportionate
32 The benefits of pollution reduction are the reduced damages from being exposed to pollution. Therefore, the
marginal social benefit of abatement is measured as the additional reduction in damages from abating one more
unit of pollution.
33 The idea that a given level of abatement is efficient as opposed to abating until pollution is equal to zero
is based on the economic concept of diminishing returns. For each additional unit of abatement, marginal social
benefits decrease while marginal social costs of that abatement increase. Thus, it only makes sense to continue to
increase abatement until the point where marginal abatement benefits and marginal costs are just equal. Any
abatement beyond that point will incur more additional costs than benefits. Alternatively, one can understand the
efficient level of abatement as the amount of regulation that achieves the efficient level of pollution. If one
considers a market for pollution, the socially-efficient outcome would be the point where the marginal WTPfor
pollution equals the marginal social cost of polluting.
34 It is important to reemphasize the word "marginal" in this statement. Marginal, in economic parlance, means
the extra or next unit of the item being measured. If regulatory options could be ranked in order of regulatory
stringency, then marginal benefits equal to marginal costs means that the additional benefits of increasing the
regulation to the next degree of stringency is equal to the additional cost of that change.
A-13 Guidelines for Preparing Economic Analyses | 3rd edition
-------
cutbacks in emissions. At the optimal level of regulation, the cost of abating one more unit of
pollution is equal across all polluters.35
Figure A. 11 illustrates why the level of pollution that sets the marginal benefits and marginal costs
of abatement equal to each other is efficient.36 Emissions are drawn on the horizontal axis and
increase from left to right, The damages from emissions are represented by the marginal damage
(MD) curve. Damages may include the costs of worsened human health, reduced visibility, lower
property values, and loss of crop yields or biodiversity. As emissions rise, the marginal damages
increase. Ei represents the amount of emissions in the absence of regulation on firms. The costs of
controlling emissions are represented by the marginal abatement cost curve (MAC). As emissions
are reduced below E- the marginal cost of abatement rises.
Figure A.11 - Efficient Level of Pollution
Costs. Damages
The total damages associated with emissions level E* are represented by the area of the triangle
AE0E*, while the total abatement costs are represented by area AEiE*. The total burden on society of
this level is equal to the total abatement costs of reducing emissions from Ei to E* plus the total
damages of the remaining emissions, E*. That is, the total burden is the darkly shaded triangle,
EoAEl
Now assume that emissions are something other than E*. For example, suppose emissions were Ex,
which is greater than E*. Total damages for this level of emissions are equal to the area of the
triangle BE0EX, while total costs of abatement to this level is equal to the area CExEi. The total burden
on society of this level is the sum of the areas of the darkly shaded and the lightly shaded triangles.
This means that the excess social cost of choosing emissions Ex rather than E* is equal to the area of
the lightly shaded triangle, ABC. A similar analysis could be done if emissions levels were below
35 Thus, a regulation that requires all firms to achieve the same level of reduction will probably result in
different marginal costs for each firm and not be efficient (see Field and Field 2005 or any other environmental
economics text for a detailed explanation and example).
36 Figure A.ll illustrates the simplest possible case, where the pollutant is a flow (i.e., it does not accumulate
over time) and marginal damages are independent of location. When pollution levels and damages vary by
location, then the efficient level of pollution is reached when marginal abatement costs adjusted by individual
transfer coefficients are equal across all polluters. Temporal variability also implies an adjustment to this
equilibrium condition. In the case of a stock pollutant, marginal abatement costs are equal across the discounted
sum of damages from today's emissions in all future time periods. In the case of a flow pollutant, this condition
should be adjusted to reflect seasonal or daily variations (see Sterner 2003).
A-20 Guidelines for Preparing Economic Analyses | 3rd edition
-------
level, E*. Here, the additional abatement costs would be greater than the decrease in damages,
resulting in excess social costs. The policy that sets the emissions level at E* at the point where
marginal benefits of pollution reduction (represented by the MD curve) and the MAC curve
intersect is economically efficient because it imposes the least net cost on, and yields the highest
net benefits for, society. That is, the triangle E0AEi is the smallest shaded region that can be
obtained.
This section has focused on first-best optimal regulation when there are no pre-existing market
distortions. However, it is important to note that realizable policy outcomes will often be "second
best" due to information constraints, political constraints, imperfect competition and market
distortions created by tax and other government interventions. For example, many of the
emissions-based policies emphasized in these Guidelines may be less feasible for addressing
nonpoint source pollution, such as agriculture, which is less observable and more stochastic than
emissions from point sources. Agriculture is also subject to multiple non-environmental policy
distortions that must be considered in the measurement of the social benefits and costs of
regulating agriculture.
usion
The purpose of this appendix is to present a brief explanation of some of the fundamental
economics relevant to Chapters 3 through 9. It is not intended to provide a comprehensive
discussion of all microeconomic theory and its application to environmental issues. The interested
reader can turn to undergraduate or graduate level textbooks for a more thorough exposition of the
topics covered here. At the undergraduate level, Field and Field (2005) provide an introduction to
the basic principles of environmental economics. Tietenberg's (2002) and Perman etal.'s (2003)
presentations are more technical but still used primarily for undergraduate courses. Freeman
(2003) is the standard text for graduate courses in environmental economics and deals with the
methodology of non-market valuation. Supplemental texts that provide a good handle on
environmental economics with less technical detail include Stavins (2000a), and Portney and
Stavins (2000). Finally, general microeconomics textbooks (Mankiw 2004, and Varian 2005 atthe
undergraduate level; and Mas-Colell et al. 1995, Kreps 1990, and Varian 2005 at the graduate level),
and applied welfare economics textbooks (Just et al. 2005) are useful references as well.
A-21 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Appendix A References
Barone, E. 1908. The Ministry of Production in the Collectivist State. Giornale degli Economists as translated in
Hayek, 1935 (ed). Collectivist Economic Planning. London: Routledge.
Baumol, W.J. and W.E. Oates. 1988. The Theory of Environmental Policy. 2nd Ed. New York: Cambridge
University Press.
Bockstael, N.E. and K.E. McConnell. 1993. Public Goods as Characteristics of Non-market Commodities.
Economic Journal 103(420): 1244-1257.
Coase, R. 1960. The Problem of Social Cost. The Journal of Law and Economics, 3:1-44.
Dantzig, G.B. 1957. Discrete-variable Extremum Problems. Operations Research, 5(2): 266-277.
Field, B.C. and M.K. Field. 2005. Environmental Economics: An Introduction. 4th Ed. New York: McGraw-
Hill/Irwin.
Freeman III, A.M. 2003. The Measurement of Environmental and Resource Values: Theory and Methods. 2nd Ed.
Washington, DC: Resources for the Future.
Hicks, J.R. 1941. The Rehabilitation of Consumers' Surplus. The Review of Economic Studies, 8(2): 108-116.
Hyman, J.B. and S.G. Leibowitz. 2000. A General Framework for Prioritizing Land Units for Ecological
Protection and Restoration. Environmental Management, 25(1): 23-35.
Just, R.E., D.L. Hueth, and A. Schmitz. 2005. Welfare Economics of Public Policy: A Practical Approach to Project
and Policy Evaluation. Northampton, MA: Edward Elgar Publishing.
Kelman, S. 1981. Cost-Benefit Analysis: An Ethical Critique. Regulation, 5(1): 33-40.
Kreps, D.M. 1990. A Course in Microeconomic Theory. Princeton, NJ: Princeton University Press.
Mankiw, N.G. 2004. Principles of Economics. 3rd Ed. Mason, OH: South-Western.
Mas-Colell, A., M.D. Whinston, and J.R. Green. 1995. Microeconomic Theory. New York: Oxford University Press.
Pareto, V. 1906. Manual of Political Economy. 1971 translation of 1927 edition, New York: Augustus M. Kelley.
Pearce, D.W., ed. 1992. The MIT Dictionary of Modern Economics. 4th Ed. Cambridge, MA: MIT Press.
Perman, R., M. Common, J. McGilvray, J., and Y. Ma. 2003. Natural Resource and Environmental Economics. 3rd
Ed. Essex: Pearson Education Limited.
Portney, P. and R. Stavins. 2000. Public Policies for Environmental Protection. Washington, DC: Resources for
the Future.
Scott, A.D. 1953. Notes on User Cost. The Economic Journal, 63(250): 368-384.
Scott, A.D. 1955. Natural Resources, The Economics of Conservation (2nd Edition 1983). Ottawa: Carleton
University Press, Toronto, Don Mills: Oxford University Press.
Stavins, R. 2000a. Economics of the Environment. 4th Ed. New York: W.W. Norton.
Sterner, T. 2003. Policy Instruments for Environmental and Natural Resource Management. Washington, DC:
Resources for the Future.
Tietenberg, T. 2002. Environmental and Natural Resource Economics. 6th Ed. New York, NY: Harper Collins
Publishers.
U.S. EPA. 1989. Regulatory Impact Analysis of Controls on Asbestos and Asbestos Products: Final Report.
Prepared by the Office of Pesticides and Toxic Substances.
Varian, H. 1992. Microeconomic Analysis. New York: W.W. Norton & Co., Inc.
Varian, H. 2005. Intermediate Microeconomics: A Modern Approach, 7th Ed. New York: W.W. Norton & Co., Inc.
Willig, R.D. 1976. Consumer's Surplus without Apology. The American Economic Review, 66(4): 589-97.
A-22 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Appendix B - Mortality Risk Valuation
Estimates
Some U.S. Environmental Protection Agency [EPA] policies are designed to
reduce the risk of contracting a potentially fatal health effect such as cancer.
Reducing these risks of premature death provides welfare increases to those
individuals affected by the policy. These policies generally provide marginal
changes in relatively small risks. That is, these policies do not provide
assurance that an individual will not die prematurely from environmental
exposures; rather, they marginally reduce the probability of such an event. For
BCA, analysts generally aggregate these small risks over the affected
population to derive the number of statistical lives saved (or the number of
statistical deaths avoided] and then use a "value of statistical life" (VSL] to
express these benefits in monetary terms.
The risk reductions themselves can generally be classified according to the
characteristics of the risk in question (e.g., voluntariness or controllability]
and the characteristics of the affected population (e.g., age and health status].
These dimensions may affect the value of reducing mortality risks. Ideally the
VSL would account for all possible risk and demographic characteristics that
matter. It would be derived from the preferences of the population affected by
the policy, based on the type of risk that the policy is expected to reduce. For
example, if a policy were designed to remove carcinogens at a suburban
hazardous waste site, the ideal measure would represent the preferences for
reduced cancer risks for the exposed population in the area and would reflect
the changes in life expectancy that would result. Unfortunately, time and
resource constraints make it difficult if not impossible to obtain such unique
valuation estimates for each EPA policy. Instead, analysts need to draw from
existing VSL estimates obtained using well-established methods (see Chapter
7]-
This appendix describes the default VSL estimate currently used by the
Agency and its derivation, as well as how analysts should characterize and
assess benefit transfer issues that may arise in its application. Benefit transfer
considerations that are common to all valuation applications, including the
effect of most demographic characteristics of the study and policy
populations, are described in Chapter 7 Section 7.3 and will not be repeated
here.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
B.1 Central Estimate of VSL
Table B.l contains the VSL estimates that currently form the basis of the Agency's recommended
central VSL estimate. Fitting a Weibull distribution to these estimates yields a central estimate
(mean) of $10.7 million ($2022) with a standard deviation of $7.2 million.1'2 The EPA recommends
that the central estimate, updated to the base year of the analysis, be used in all benefits analyses
that seek to quantify mortality risk reduction benefits.
This approach was vetted and endorsed by the Agency when the 2000 Guidelines for Preparing
Economic Analyses were drafted.3 It remains the EPA's default guidance for valuing mortality risk
changes although the Agency has considered and presented alternatives.4
B.2 Other VSL Information
For most of mortality risk reductions, the EPA uniformly applies the VSL estimate discussed above.
For a period of time (2004-2008), the Office of Air and Radiation (OAR) valued mortality risk
reductions using a VSL estimate derived from a limited analysis of some of the available studies.
OAR arrived at a VSL using a range of $1 million to $10 million (2000$) consistent with two meta-
analyses of the wage-risk literature. The $1 million value represented the lower end of the
interquartile range from the Mrozek and Taylor (2002) meta-analysis of 33 studies. The $10 million
value represented the upper end of the interquartile range from the Viscusi and Aldy (2003) meta-
analysis of 43 studies. The mean estimate of $5.5 million (2000$) was also consistent with the mean
VSL of $5.4 million estimated in the Kochi et al. (2006) meta-analysis. However, the Agency neither
changed its official guidance on the use of VSL in rulemakings nor subjected the interim estimate to
a scientific peer-review process through the Science Advisory Board (SAB) or other peer-review
group.
During this time, the Agency continued work to update its guidance on valuing mortality risk
reductions. The EPA commissioned a report from meta-analytic experts to evaluate methodological
questions raised by the EPA and the SAB on combining estimates from the various data sources. In
1 The VSL was updated from the $4.8 million ($1990) estimate referenced in the 2000 Guidelines by adjusting
the individual study estimates for inflation using CPI-U and then fitting a Weibull distribution to the estimates.
The updated Weibull parameters are: location = 0, scale = 11.91, shape = 1.51 (updated from location = 0; scale =
5.32; shape = 1.51). The Weibull distribution was determined to provide the best fit for this set of estimates. See
U.S. EPA 1997a for more details.
2 This VSL estimate was produced using the Consumer Price Index (CPI). Some economists prefer using the GDP
deflator inflation index in some applications. The key issue for EPA analysts is to ensure that the chosen index is
used consistently throughout the analysis.
3 The studies listed in Table B.l were published between 1974 and 1991, and most are hedonic wage estimates
that may be subject to considerable measurement error (Black etal. 2003; Black and Kniesner 2003). Although
these were the best available data at the time, they are sufficiently dated and may rely on obsolete preferences
for risk and income. The Agency is currently considering more recent studies as it evaluates approaches to revise
its guidance.
4 The EPA engaged the SAB-EEAC on several issues including the use of meta-analysis as a means of combining
estimates and approaches for assessing mortality benefits when changes in longevity may vary widely (U.S. EPA
2006d; U.S. EPA 2016). see U.S. EPA 2017 for recent SAB recommendations.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
addition, the Agency consulted several times with the SAB Environmental Economics Advisory
Committee (SAB-EEAC) on the issue (e.g., U.S. EPA 2017).
Table B.1 - Value of Statistical Life Estimates (mean values in millions
of 2022 dollars)
Study
Method
Value of Statistical Life
Kniesnerand Leeth (1991 - US)
Labor Market
$1.34
Smith and Gilbert (1984)
Labor Market
$1.57
Dillingham (1985)
Labor Market
$2.02
Butler (1983)
Labor Market
$2.46
Miller and Guria (1991)
Contingent Valuation
$2.69
Moore and Viscusi (1988)
Labor Market
$5.60
Viscusi, Magat, and Huber (1991)
Contingent Valuation
$6.05
Marin and Psacharopoulos (1982)
Labor Market
$7.39
Gegax et al. (1985)
Contingent Valuation
$6.27
Kniesnerand Leeth (1991 - Australia)
Labor Market
$7.39
Gerking, de Haan, and Schulze (1988)
Contingent Valuation
$7.61
Cousineau, Lecroix, and Girard (1988)
Labor Market
$8.06
Jones-Lee (1989)
Contingent Valuation
$8.51
Dillingham (1985)
Labor Market
$8.73
Viscusi (1978)
Labor Market
$9.18
R.S. Smith (1976)
Labor Market
$10.30
V.K. Smith (1983)
Labor Market
$10.52
Olson (1981)
Labor Market
$11.64
Viscusi (1981)
Labor Market
$14.55
R.S. Smith (1974)
Labor Market
$16.12
Moore and Viscusi (1988)
Labor Market
$16.35
Kniesnerand Leeth (1991 - Japan)
Labor Market
$17.02
Herzog and Schlottman (1987)
Labor Market
$20.38
Leigh and Folsom (1984)
Labor Market
$21.72
Leigh (1987)
Labor Market
$23.29
Garen (1988)
Labor Market
$30.23
Derived from U.S. EPA (1997a) and Viscusi (1992), Updated to 2022$ with CPI-U.
B-3 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Until updated guidance is available, the Agency determined that a single, peer-reviewed estimate
applied consistently best reflects the SAB-EEAC advice received to date. Therefore, the VSL
described above that was vetted and endorsed by the SAB should be applied in relevant analyses
while the Agency continues its efforts to update its guidance on this issue.
T[ ,\[
Policy analysts valuing mortality risk reductions should account for differences in risk and
population characteristics between the policy and study scenarios and their potential effect on the
overall results. The ultimate objective of the benefit transfer exercise is to account for all of the
factors that significantly affect the value of mortality risk reduction in the context of the policy.
Analysts should carefully consider the implications of correcting for some relevant factors, but not
for others, recognizing that it may not be feasible to account for all factors.
istics
Risk characteristics appear to affect the value that people place on risk reduction. A large body of
work identifies eight dimensions of risk that affect human risk perception.5
1. Voluntary/involuntary
2. Ordinary/catastrophic
3. Delayed/immediate
4. Natural/man-made
5. Old/new
6. Controllable/uncontrollable
7. Necessary/unnecessary
8. Occasional/continuous
Transferring VSL estimates among these categories may introduce bias. There have been some
recent efforts attempting to quantitatively assess these sources of bias.6 These studies generally
conclude that voluntariness, control and responsibility affect individual values for safety, although
there is no consensus on the direction and magnitude of these effects.
In addition, environmental risks may differ from those that form the basis of VSL estimates in many of
these dimensions. Occupational risks, for example, are generally considered to be more voluntary in
nature than are environmental risks and may be more controllable. As part of the Agency's review of
our mortality risk guidance we are evaluating the literature from which the studies are drawn.
Support for quantitative adjustments in the empirical literature is lacking for most of these factors.
The SAB reviewed an Agency summary of the available empirical literature on the effects of risk
and population characteristics on WTP for mortality risk reductions (U.S. EPA 2000d). The SAB
review concludes that among the demographic and risk factors that might affect VSL estimates, the
5 A review of issues in risk perception is found in Lichtenstein and Slovic (2006). Other informative sources
include Slovic (1987), Rowe (1977), Otway (1977), and Fischoffetal. (1978).
6 Examples include Hammitt and Liu (2004), Sunstein (1997), Mendeloffand Kaplan (1990), McDaniels etal.
(1992), Savage (1993), Jones-Lee and Loomes (1994,1995,1996), and Covey etal. (1995).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
current literature can only support empirical adjustments related to the timing of the risk. The
review supports making the following adjustments to primary benefits estimates: (1) adjusting
WTP estimates to account for higher future income levels, though not for cross-sectional
differences in income; and (2) discounting risk reductions that are brought about in the future by
current policy initiatives (that is, after a cessation lag), using the same rates used to discount other
future benefits and costs. All other adjustments, if made, should be relegated to sensitivity analyses.
Increases in income over time. The economics literature shows that the income elasticity of WTP to
reduce mortality risk is positive, based on cross-sectional data. As a result, benefits estimates of
reduced mortality risk accruing in future years may be adjusted to reflect anticipated income
growth, using the range of income elasticities (0.08, 0.40 and 1.0) employed in The Benefits and
Costs of the Clean Air Act, 1990-2010.7 Recent EPA analyses have assumed a triangular distribution
from these values and used the results in a probabilistic assessment of benefits.8 At the time of this
writing, the EPA is engaged in a consultation with the SAB-EEAC on the appropriate range of
income elasticities and will update this guidance as needed.
Timing of reduced exposure and reduced risk. Many environmental policies are targeted at
reducing the risk of effects such as cancer, where there may be an extended period of time
between the reduced exposure and the reduction in the risk of death from the disease.9 This delay
between the change in exposure and realization of the reduced risk may affect the value of that
risk reduction. Most existing VSL estimates are based on risks of relatively immediate fatalities
making them an imperfect fit for a benefits analysis of many environmental policies. Economic
theory suggests that reducing the risk of a delayed health effect will be valued less than reducing
the risk of a more immediate one, when controlling for other factors.
K 7 Pi i ¦ \ ^ n TV *.^ ^ 'rh1 r P-if11"«,\pf ri'~
istics
Two population characteristics are particularly noteworthy for their potential effect on mortality
risk valuation estimates: age and health status of the exposed population. In September 2006, the
Agency requested an additional advisory from the SAB-EEAC on issues related to valuing changes in
life expectancy for which age and baseline health status are close correlates.10 Because the outcome
of this review is not yet available, we focus here on previous advice received from the SAB on
related questions.
Age. It has sometimes been posited that older individuals should have a lower WTP for changes in
mortality risk given the fewer years of life expectancy remaining compared to younger individuals.
This hypothesis may be confounded, however, by the finding that older persons reveal a greater
7 For details see Kleckner and Neuman (2000).
8 See, for example, pp. 6-84 of the Final Economic Analysis for the Stage 2 Disinfection Byproducts Rule (DBPR)
(U.S. EPA 2005d).
9 Although latency is defined here as the time between exposure and fatality from illness, alternative definitions
may be used in other contexts. For example, "latency" may refer to the time between exposure and the onset of
symptoms. These symptoms may be experienced for an extended period of time before ultimately resulting in
fatality.
10 U.S. EPA (2006d) summarizes much of the literature related to the effects of age and health status on WTP for
changes in mortality risk and includes the charge questions put to the SAB-EEAC on these issues.
B-5 Guidelines for Preparing Economic Analyses | 3rd edition
-------
demand for reducing mortality risks and hence have a greater implicit value of a life year (Ehrlich and
Chuma 1990). Several authors have attempted to explore potential differences in mortality risk
valuation estimates associated with differences in the average age of the affected population using
theoretical models of life-cycle consumption.11 In general, this literature has shown that the
relationship between age and WTP for mortality risk changes is ambiguous, requiring strong
assumptions to even sign the relationship.12 Empirical evidence is also mixed. A number of empirical
studies (mostly hedonic wage studies) suggest that the VSL follows a consistent "inverted-U" life-cycle,
peaking in the region of mean age.13 Others find no such statistically significant relationship and still
others show WTP increasing with age.14 Stated preference results are also mixed, with some studies
showing declining WTP for older age groups and others finding no statistically significant relationship
between age and WTP.15
In spite of the ambiguous relationship between age and WTP, two alternative adjustment
techniques have been derived from this literature. The first technique, value of statistical life-years
(VSLY), is derived by dividing the estimated VSL by expected remaining life expectancy. This is by far
the most common approach and presumes that: (1) the VSL equals the sum of discounted values for
each life year; and (2) each life year has the same value. This method was applied as an alternative
case in an effort to evaluate the sensitivity of the benefits estimates prepared for the EPA's
retrospective and prospective studies of the costs and benefits of the Clean Air Act (U.S. EPA 1997a;
U.S. EPA 1999).
A second technique is to apply a distinct value or suite of values for mortality risk reduction
depending on the age of incidence. However, there is relatively little available literature upon which
to base such adjustments.16
Neither approach enjoys general acceptance in the literature as they both require large
assumptions to be made, some of which have been contradicted in empirical studies. Since
published support is lacking, neither approach is recommended at this time.
Analysts are advised to note the age distribution of the affected population when possible,
especially when children are found to be a significant portion of the affected population.17 Although
11 See, for example, Shepard and Zeckhauser (1982), Rosen (1988), Cropper and Sussman (1988,1990), and
Johansson (2002).
12 See Evans and Smith (2006) for a recent summary.
13 See Jones-Lee etal. (1985), Aldy and Viscusi (2008), Viscusi andAldy (2007a, 2007b), and Kniesner et al.
(2006).
14 Viscusi and Aldy (2003) review more than 60 studies of mortality risk estimates from 10 countries and discuss
eight hedonic wage studies that explicitly examine the age-WTP relationship. Only five of the eight studies found
a statistically significant, negative relationship between age and the return to risk. Smith etal. (2004) and
Kniesner et al. (2006) find that WTP increases with age.
15 Krupnick et al. (2002) report that WTP for mortality risk reductions changes significantly with age after age
70. Alberini et al. (2004) find no difference in the WTP for younger age groups and find a 20% reduction for
those aged 70 and older. However, this difference was not statistically significant.
16 This second approach was illustrated in one EPA study (U.S. EPA 2002) for valuation of air pollution mortality
risks, drawing upon adjustments measured in Jones-Lee et al. (1985).
17 See U.S. EPA (2003a) for more information on the valuation of children's health risks. OMB's Circular A-4
advises agencies to use estimates of mortality risk valuation for children that are at least as large as those used
for adult populations (OMB 2003).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
the literature on the valuation of children's health risks is growing, there is still not enough
information currently to derive age-specific valuation estimates.
Health status. Individual health status may also affect WTP for mortality risk reduction. This is an
especially relevant factor for valuation of environmental risks because individuals with impaired
health are often the most vulnerable to mortality risks from environmental causes. For example,
particulate air pollution appears to disproportionately affect individuals in an already impaired
state of health. Health status is distinct from age (a "quality versus quantity" distinction) but the
two factors are clearly correlated and therefore must be addressed jointly when considering the
need for an adjustment Again, both the theoretical and empirical literatures on this point are mixed
with some studies showing a declining WTP for increased longevity with a declining baseline health
state (Desvousges et al. 1996) and other studies showing no statistically significant effects
(Krupnick etal. 2002).18
Application of existing VSLY approaches implicitly assumes a linear relationship in which each
discounted life year is valued equally. As Office of Management and Budget (OMB) (1996) notes
"current research does not provide a definitive way of developing estimates of VSLY that are
sensitive to such factors as current age, latency of effect, life years remaining, and social valuation of
different risk reductions." The second alternative, applying a suite of values for these risks, lacks
broad empirical support in the economics literature. However, the potential importance of this
benefit transfer factor suggests that analysts consider sensitivity analysis when risk data
essentially risk estimates for specific age groups are available. An emerging literature on the
value of life expectancy extensions, based primarily on stated preference techniques, is beginning to
help establish a basis for valuation in cases where the mortality risk reduction involves relatively
short extensions of life.19
elusion
Due to current limitations in the existing economic literature, these Guidelines conclude that, for the
present time, the appropriate default approach for valuing these benefits is provided by the central
VSL estimate described earlier. However, analysts should carefully present the limitations of this
estimate. Economic analyses should also fully characterize the nature of the risk and populations
affected by the policy action and should confirm that these parameters are within the scope of the
situations considered in these Guidelines. While a qualitative discussion of these issues is generally
warranted in EPA economic analyses, analysts should also consider a variety of quantitative
sensitivity analyses on a case-by-case basis as data allow. The analytical goal is to characterize the
impact of key attributes that differ between the policy and study cases. These attributes, and the
18 The fields of health economics and public health often account for health status through the use of quality-
adjusted life years (QALYs) or disability adjusted life years (DALYs). These measures have their place in
evaluating the cost-effectiveness of medical interventions and other policy contexts but have not been fully
integrated into the welfare economic literature on risk valuation. More information on QALYs can be found in
Gold etal. (1996) and additional information on DALYs can be found in Murray (1994).
19 It should be noted that many observers have expressed reservations over adjusting the value of mortality risk
reduction on the basis of population characteristics such as age. One of the ethical bases for these reservations is
a concern that adjustments for population characteristics imply support for variation in protection from
environmental risks. Another consideration is that existing economic methods may not capture social WTP to
reduce health risks. Chapter 9 details how some these considerations may be informed by a separate assessment
of equity.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
degree to which they affect the value of risk reduction, may vary with each benefit transfer exercise,
but analysts should consider the characteristics described above (e.g., age, health status,
voluntariness of risk and latency) and values arising from altruism.
As the economic literature in this area evolves, WTP estimates for mortality risk reductions that
more closely resemble those from environmental hazards may support more precise benefit
transfers. Literature on the specific methods available to account for individual benefit-transfer
considerations will also continue to develop. In addition, The EPA will continue to conduct
periodic reviews of the risk valuation literature and will reconsider and revise the
recommendations in these Guidelines accordingly. The EPA will seek advice from the SAB as
guidance recommendations are revised.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Appendix B References
Alberini, A., M. Cropper, A. Krupnick, and N.B. Simon. 2004. Does the Value of a Statistical Life Vary with Age
and Health Status? Evidence from the United States and Canada. Journal of Environmental Economics and
Management, 48: 769-792.
Black, D., J. Galdo and L. Liu. 2003. How Robust Are Hedonic Wage Estimates of the Price of Risk: The Final
Report. EPA Document R 82943001.
Black, D.A. and T.J. Kniesner. 2003. On the Measurement of Job Risk in Hedonic Wage Models .Journal of Risk
and Uncertainty, 27(3): 205-220.
Butler, R.J. 1983. Wage and Injury Rate Response to Shifting Levels of Workers' Compensation, In Safety and
the Work Force: Incentives and Disincentives in Worker's Compensation, ed. J.D. Worral. Ithaca: Cornell
University, ILR Press.
Cousineau J.M., R. Lacroix, and A.M. Girard. 1988. Occupational Hazard and Wage Compensating Differentials.
University of Montreal Working Paper.
Covey, J., M. Jones-Lee, G. Loomes, and A. Robinson. 1995. The Exploratory Empirical Study. In Exploratory
Study of Consumers' Willingness to Pay for Food Risk Reductions, Report to the Ministry of Agriculture,
Fisheries and Food, ed. D. Ives, B. Soby, G. Goats, and D.J. Ball.
Cropper, M.L. and F.G. Sussman. 1988. Families and the Economics of Risks to Life. American Economic
Review, 78(1): 255-60.
Cropper, M.L. and F.G. Sussman. 1990. Valuing Future Risks to Life. Journal of Environmental Economics and
Management, 19:160-174.
Desvousges, W.H., F.R. Johnson, S.P. Hudson, S.R. Gable, and M.C. Ruby. 1996. Using Conjoint Analysis and
Health-State Classifications to Estimate the Value of Health Effects of Air Pollution, Triangle Economic
Research, Raleigh, NC.
Dillingham, A.E. 1985. The Influence of Risk Variable Definition on Value of Life Estimates. Economic Inquiry,
24: 277-294.
Ehrlich, I. and H. Chuma. 1990. A Model of the Demand for Longevity and the Value of Life Extension. Journal
of Political Economy, 98(4): 761-782.
Evans, M.F. and V.K. Smith. 2006. Do We Really Understand the Age-VSL Relationship? Resource and Energy
Economics, 28: 242-261.
Fischoff, B., P. Slovic, S. Lichenstein, S. Reed, and B. Combs. 1978. How Safe is Safe Enough? A Psychometric
Study of Attitudes Towards Technological Risks and Benefits. Policy Sciences, 9:127-152.
Gegax, D., S. Gerking, and W. Schulze. 1985. Perceived Risk and the Marginal Value of Safety. The Review of
Economics and Statistics, 73(4): 589-596.
Gerking, S., M. de Haan, and W. Schulze. 1988. The Marginal Value of Job Safety: A Contingent Valuation Study.
Journal of Risk and Uncertainty, 1(2): 185-200.
Gold, M.R., J.E. Siegel, L.B. Russell, and M.C. Weinstein. 1996. Cost Effectiveness in Health and Medicine, New
York: Oxford University Press.
Hammitt, J.K. and J.T. Liu. 2004. Effects of Disease Type and Latency on the Value of Mortality Risk. Journal of
Risk and Uncertainty, 28(1): 73-95.
Herzog, Jr., H.W. and A.M. Schlottman. 1987. Valuing Risk in the Workplace: Market Price, Willingness to Pay,
and the Optimal Provision of Safety. University of Tennessee Working Paper.
Johansson, P.O. 2002. On the Definition and Age-Dependency of the Value of Statistical Life. The Journal of Risk
and Uncertainty, 25(3): 251-263.
Jones-Lee, M.W. 1989. The Economics of Safety and Physical Risk Oxford: Basil Blackwell.
B-9 Guidelines for Preparing Economic Analyses | 3rd edition
-------
Jones-Lee, M.W., M. Hammerton, and P.R. Philips. 1985. The Value of Safety: Results of a National Sample
Survey. Economic Journal, 95: 49-72.
Jones-Lee, M.W. and G. Loomes. 1994. Towards a Willingness-to-Pay Based Value of Underground Safety.
Journal of Transport Economics and Policy, 28: 83-98.
Jones-Lee, M.W. and G. Loomes. 1995. Scale and Context Effects in the Valuation of Transport Safety .Journal of
Risk and Uncertainty, 11:183-203.
Jones-Lee, M.W. and G. Loomes. 1996. The Tolerability of Third Party Risk and the Value of Risk Reduction Near
Airports. Report to the Civil Aviation Authority, CASPAR, Newcastle.
Kleckner, N. and J. Neuman. 2000. Update to Recommended Approach to Adjusting WTP Estimates to Reflect
Changes in Real Income. Industrial Economics, Inc. Memorandum to Jim DeMocker. U.S. Environmental
Protection Agency. September 30.
Kniesner, T.J. and J.D. Leeth. 1991. Compensating wage differentials for fatal injury risk in Australia, Japan,
and the United States. Journal of Risk and Uncertainty, 4(1): 75-90.
Kniesner, T.J, W.K. Viscusi, and J.P. Ziliak. 2006. Life-Cycle Consumption and the Age-Adjusted Value of Life.
Contributions to Economic Analysis & Policy 5(1): 1-36. Available at:
https: //ideas.repec.Org/a/bpi/beieap/vcontributions.5y2006iln4.html (accessed December 23, 2020).
Kochi, I., B. Hubbell, and R. Kramer. 2006. An Empirical Bayes Approach to Combining and Comparing
Estimates of the Value of a Statistical Life for Environmental Policy Analysis. Environmental and Resource
Economics, 34(3): 385-406.
Krupnick, A., A. Alberini, M. Cropper, N. Simon, B. O'Brien, R. Goeree, and M. Heintzelman. 2002. Age, Health,
and the Willingness to Pay for Mortality Risk Reductions: A Contingent Valuation Survey of Ontario Residents.
The Journal of Risk and Uncertainty, 24(2): 161-186.
Leigh, J.P. 1987. Gender, Firm Size, Industry and Estimates of the Value-of- Life. Journal of Health Economics, 6:
255-273.
Leigh, J.P. and R.N. Folsom. 1984. Estimates of the Value of Accident Avoidance at the Job Depend on
Concavity of the Equalizing Differences Curve. The Quarterly Review of Economics and Business, 24(1): 55-56.
Lichtenstein, S. and P. Slovic. 2006. The Construction of Preference. New York: Cambridge University Press.
Marin, A. and G. Psacharopoulos. 1982. The Reward for Risk in the Labor Market: Evidence from the United
Kingdom and Reconciliation with other Studies .Journal of Political Economy, 90(4): 827-853.
McDaniels, T.L., M.S. Kamlet, and G.W. Fischer. 1992. Risk Perception and the Value of Safety. Risk Analysis, 12:
495-503.
Mendeloff, J. and R.M. Kaplan. 1990. Are Twenty-fold Differences in 'Lifesaving' Costs Justified? A
Psychometric Study of the Relative Value Placed on Preventing Deaths from Programs Addressing Different
Hazards. In New Risks, ed. A. Cox Jr., and D.F. Ricci. New York, NY: Plenum Press.
Miller, T. and J. Guria. 1991. The Value of Statistical Life in New Zealand. Report to the New Zealand Ministry
of Transport, Land Transportation Division, #0-477-052 55-X.
Moore, M.J. and K.W. Viscusi. 1988. Doubling the Estimated Value of Life: Results using New Occupational
Fatality Data .Journal of Policy Analysis and Management, 7(3):476-490.
Moore, M.J. and W.K. Viscusi. 1988. The Quantity-Adjusted Value of Life. Economic Inquiry, 26(3): 369-388.
Mrozek, J. and L. Taylor. 2002. What Determines the Value of Life? A Meta Analysis. Journal of Policy Analysis
and Management, 21(2): 253-70.
Murray, C.L. 1994. Quantifying the Burden of Disease: The Technical Basis for Disability-Adjusted Life Years.
Bulletin of the World Health Organization, 72: 429-445.
Olson, C.A. 1981. An Analysis of Wage Differentials Received by Workers on Dangerous Jobs. Journal of Human
Resources, 16(2): 167-185.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
OMB. 1996. Economic Analysis of Federal Regulations under Executive Order No. 12866. (or Best Practices
document). January 11,1996.
OMB. 2003. Circular A-4, Regulatory Analysis, September 17, 2003. Available at:
https: //obamawhitehouse.archives.gov/omb/circulars a004 a-4/ (accessed December 23, 2020).
Otway, H.J. 1977. Review of Research on Identification of Factors Influencing Social Response to
Technological Risks, International Atomic Energy Agency, IAEA Paper CN-36/4, Vienna, (in Vol. 7 of Nuclear
Power and its Fuel Cycle, IAEA Proceedings: 95-118).
Rosen, S. 1988. The Value of Changes in Life Expectancy. Journal of Risk and Uncertainty, 1(3): 285-304.
Rowe, W.D. 1977. An Anatomy of Risk. New York, NY: Wiley.
Savage, 1.1993. An Empirical Investigation into the Effects of Psychological Perceptions on the Willingness-to-
Pay to Reduce Risk .Journal of Risk and Uncertainty, 6: 75-90.
Shepard, D.S. and R.J. Zeckhauser. 1982. Life-Cycle Consumption and Willingness to Pay for Increased
Survival. In The Value of Life and Safety, ed. M.W. Jones-Lee. Amsterdam, the Netherlands: North-Holland.
Slovic, P. 1987. Perception of Risk. Science, 30(4): 423-439.
Smith, R.S. 1974. The Feasibility of an 'Injury Tax' Approach to Occupational Safety. Law and Contemporary
Problems, 38(4): 730-744.
Smith, R.S. 1976. The Occupational Safety and Health Act: Its Goals and Achievements. Washington: American
Enterprise Institute.
Smith, V.K. 1983. The Role of Site and Job Characteristics in Hedonic Wage Models. Journal of Urban
Economics, 13: 296-321.
Smith, V.K. and C.C.S. Gilbert. 1984. The Implicit Valuation of Risks to Life: A Comparative Analysis. Economics
Letters, 16: 393-399.
Smith, V.K., M.F. Evans, H. Kim, and D.H. Taylor. 2004. Do the Near Elderly Value Mortality Risks Differently?
Review of Economics and Statistics, 86(1): 423-429.
Sunstein, C.R. 1997. Bad Deaths. Journal of Risk and Uncertainty, 14(3): 259-282.
U.S. EPA. 1997a. The Benefits and Costs of the Clean Air Act: 1970-1990, Prepared by the Office of Air and
Radiation and the Office of Policy, Planning and Evaluation. EPA/410/R-97/002.
U.S. EPA. 1999. The Benefits and Costs of the Clean Air Act: 1990-2010. EPA 410-R-99-001.
U.S. EPA. 2000d. SAB Report on EPA's White Paper Valuing the Benefits of Fatal Cancer Risk Reduction. EPA-
SAB-EEAC-00-013.
U.S. EPA. 2002d. Final Regulatory Support Document: Control of Emissions from Unregulated Nonroad
Engines. EPA-420-R-02-022. Available at:
https://nepis.epa.gov/Exe/ZyPDF.cgi/PI 002 KX6.PDF?Dockev=Pl 002KX6.PDF (accessed December 23,
2020).
U.S. EPA. 2003a. America's Children and the Environment Measures of Contaminants, Body Burdens and
Illnesses. 2nd Ed. EPA/240-R-03-001. February.
U.S. EPA. 2005d. Economic Analysis for the Final Stage 2 Disinfectants and Disinfection Byproducts Rule.
December 2005. EPA 815-R-05-010. Available at:
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P100500X.TXT (accessed December 23, 2020).
U.S. EPA. 2006d. Small Engine Buyback. Office of Air and Radiation. December 11, 2006. Available at:
http: //www.epa.gov/air/recipes/smallen.html (accessed August 7, 2008).
U.S. EPA. 2007g. SAB Advisory on EPA's Issues in Valuing Mortality Risk Reduction. Available at:
http://yo Semite, epa.gov/sab /sabproduct.nsf/4128007E7876B8F0852 573760058A978/$File/sab-08-
001.pdf (accessed December 23, 2020).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
U.S. EPA. 2016. Valuing Mortality Risk for Policy: A MetaAnalytic Approach. Available at:
https://sab.epa.gov/ords /sab/r/sab apex/sab/advisoryactivitydetail?pl8 id=242 0&clear=18&session=344
3671097498 (accessed October 28, 2024).
U.S. EPA. 2017. SAB Review of EPA's Proposed Methodology for Updating Mortality Risk. Available at:
https: //sab.epa.gov/ords /sab/r/sab apex/sab/advisoryactivitydetail?pl8 id=242 0&clear=18&session=344
3671097498 (accessed October 28, 2024).
Viscusi, W.K. 1978. Labor Market Valuations of Life and Limb: Empirical Evidence and Policy Implications.
Public Policy, 26(3): 359-386.
Viscusi, W.K. 1981. Occupational Safety and Health Regulation: Its Impact and Policy Alternatives. Research in
Public Policy Analysis and Management, 2: 281-299.
Viscusi, W.K. 1992. Fatal Tradeoffs: Public and Private Responsibilities for Risk. New York: Oxford University
Press.
Viscusi, W.K. and J.E. Aldy. 2003. The Value of a Statistical Life: A Critical Review of Market Estimates
throughout the World. Journal of Risk and Uncertainty, 27(1): 5-76.
Viscusi, W.K., W.A. Magat, and J. Huber. 1991. Pricing Environmental Health Risks: Survey Assessments of
Risk-Risk and Risk-Dollar Trade-Offs for Chronic Bronchitis. Journal of Environmental Economics and
Management, 21(1): 32-51.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Glossary
Abatement costs - Abatement costs are costs borne by firms when they are required to remove
and/or reduce environmental byproducts created during production.
Annualized value - An annualized value is a constant stream of benefits or costs. The annualized
cost is the constant amount that a party would have to pay at the end of each period t to add up to
the same cost in present value terms as the varying stream of costs being annualized. Similarly, the
annualized benefit is the constant amount that a party would accrue at the end of each period t to
add up to the same benefit in present value terms as the varying stream of benefits being
annualized.
Baseline - A baseline describes an initial, status quo scenario that is used for comparison with one
or more alternative scenarios. In typical economic analyses the baseline is defined as the best
assessment of the way the world would evolve absent the regulation or policy action.
Benefit-cost analysis (BCA) - A BCA is an evaluation of the social benefits and social costs of a
policy action. The social benefits of a policy are measured by society's willingness-to-pay for the
policy outcome. The social costs are measured by the opportunity costs of adopting the policy. BCA
addresses the question of whether the benefits for those who gain from the action are sufficient to,
in principle, compensate those burdened by costs such that everyone would be at least as well off as
before the policy. The calculation of net benefits (benefits minus costs) answers this question and
helps ascertain the economic efficiency of the policy. Where all benefits and costs can be quantified
and expressed in monetary units, BCA provides decision makers with a clear indication of the most
economically efficient alternative, that is, the alternative that generates the largest net benefits to
society (ignoring distributional effects).
Benefit-cost ratio - A benefit-cost ratio is the ratio of the net present value (NPV) of benefits
associated with a project or proposal, relative to the NPV of the costs of the project or proposal. The
ratio indicates the benefits expected for each dollar of costs. Note that this ratio is not an indicator
of the magnitude of net benefits. Two projects with the same benefit-cost ratio can have vastly
different estimates of benefits and costs.
Benefit transfer - Benefit transfer is the use of estimated values of environmental quality changes
drawn from primary (usually published) studies for the evaluation of similar changes of interest to
the analyst
Cessation lag - Cessation lag is the time between a reduction in exposure and the reduction in risk.
See latency for a definition of a related but distinct concept
Command-and-control regulation - Command-and-control regulation is a prescriptive regulation
that stipulates how much pollution an individual source or plant is allowed to emit and/or what
types of control equipment it must use to reduce pollution.
Compliance cost - A compliance cost is the private cost that a regulated entity incurs to comply
with a regulation for instance, through the planning, design, installation, and operation of
pollution abatement equipment.
Consumption rate of interest - Consumption rate of interest is the rate at which individuals are
willing to exchange consumption in one period (usually year) for consumption in the next period.
This rate reflects the individual's rate of time preference.
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Cost-effectiveness analysis (CEA) - CEA examines the costs associated with obtaining an
additional unit of an environmental outcome. It is designed to identify the least expensive way of
achieving a given environmental quality target, or the way of achieving the greatest improvement in
some environmental target for a given expenditure of resources.
Distributional analysis - Distributional analysis assesses changes in social welfare by examining
the effects of a regulation across different subpopulations and entities.
Distorted market - A distorted market is one that does not experience free and open competition
due to government interventions and/or prevailing market conditions. Examples of distortions
include externalities, regulations, pre-existing taxes, or imperfectly competitive markets.
Dollar year - The year to which the purchasing power of a dollar is indexed. For example, if costs
and benefits are reported in 2016 dollars, it means that the purchasing power of those costs and
benefits reflect what could have been bought in 2016.
Ecological production function - An ecological production function is a description of how
ecosystems combine inputs to produce ecosystem services that consumers enjoy directly or are
used in the production of goods or services that are enjoyed by consumers.
Economic efficiency - Economic efficiency can be defined as the maximization of social welfare.
Under the efficient level of production, there is no way to rearrange production or reallocate goods
such that someone is better off without making someone else worse off in the process.
Economic impact analysis (EIA) - Economic impact analyses (EIAs) examine how compliance
costs, transfers, and other policy outcomes are distributed across groups. EIAs describe and often
quantify outcomes such as changes in employment, plant closures or local government tax revenues
that provide insight into the economic consequences of regulation.
Emissions tax - An emissions tax is a charge levied on each unit of pollution emitted.
Environmental justice - Environmental justice is the fair treatment and meaningful involvement
of all people regardless of race, color, national origin or income with respect to the development,
implementation and enforcement of environmental laws, regulations and policies. Fair treatment
means that no group of people, including racial, ethnic or socioeconomic groups, should bear a
disproportionate share of the negative environmental consequences resulting from industrial,
government and commercial operations or policies. Meaningful involvement occurs when (1)
potentially affected community members have an appropriate opportunity to participate in
decisions about a proposed activity that may affect their environment and/or health; (2) the
public's contribution can influence the regulatory agency's decision; (3) their concerns will be
considered in the decision-making process; and (4) the decision makers seek out and facilitate the
involvement of those potentially affected.1
Expected value - Expected value is the probabilistically weighted outcome that defines a statistical
mean and a measure of the central tendency of a set of data. For a variable with a discrete number
of outcomes, the expected value is calculated by multiplying each of the possible outcomes by the
likelihood that each outcome will occur and then summing all of those values.
1 Definition taken from http://www.epa.aov/compliance/environmentaliustice/index.html (accessed December
22,2020)
ii Guidelines for Preparing Economic Analyses | 3rd edition
-------
Expert elicitation - Expert elicitation is a formal, highly structured and well-documented process
for obtaining the judgments of multiple experts. Typically, an elicitation is conducted to evaluate
uncertainty. This uncertainty could be associated with the value of a parameter to be used in a
model, the likelihood and frequency of various future events or the relative merits of alternative
models.
Externality - An externality occurs when the actions of one individual (or firm) have a direct,
unintentional and uncompensated effect on the well-being of other individuals or the profits of
other firms.
Flow pollutant - A flow pollutant is a pollutant for which the environment has some absorptive
capacity. It does not accumulate in the environment as long as its emission rate does not exceed the
absorptive capacity of the environment. Animal and human wastes are examples of flow pollutants.
General equilibrium - A general equilibrium modeling approach concurrently considers the effect
of a regulation across all sectors in the economy.
Hedonic price - Hedonic price, recreational demand or locational choice models may be regarded
as "reduced form" representations of ecological production from which the analyst can infer the
values individuals ascribe to ecosystem services by observing the choices they make, provided that
the analyst can adequately control for potentially confounding factors.
Hotspot - A hotspot is a geographic area with a high level of pollution/contamination within a
larger geographic area of low or "normal" environmental quality.
Kaldor-Hicks criterion - The Kaldor-Hicks criterion is really a combination of two criteria: the
Kaldor criterion and the Hicks criterion. The Kaldor criterion states that an activity will contribute
to Pareto optimality if the maximum amount the gainers are hypothetically prepared to pay is
greater than the minimum amount that the losers are hypothetically prepared to accept. Under the
Hicks criterion, an activity will contribute to Pareto optimality if the maximum amount the losers
are hypothetically prepared to offer to the gainers in order to prevent the change is less than the
minimum amount the gainers are hypothetically prepared to accept as a bribe to forgo the change.
In other words, the Hicks compensation test is conducted from the losers' point of view, while the
Kaldor compensation test is conducted from the gainers' point of view. The Kaldor-Hicks criterion
is widely applied in welfare economics and managerial economics. It forms an underlying rationale
for BCA.
Latency - Latency is the time between the increase in exposure to a pollutant and the increase in
health risk. See cessation lag for a definition of a related but distinct concept.
Marginal benefit - The marginal benefit is the benefit received from an incremental increase in the
consumption of a good or service.
Marginal cost - The marginal cost is the change in total cost that results from a unit increase in
output.
Marginal social benefit - The marginal social benefit is the marginal benefit received by the
consumer of a good (marginal private benefit) plus the marginal benefit received by other members
of society (external benefit).
Marginal social cost - The marginal social cost is the marginal cost incurred by the producer of a
good (marginal private cost) plus the marginal cost imposed on other members of society (external
cost).
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Market failure - A market failure occurs when the allocation of goods and services by the free
market is not economically efficient The most common causes of market failure are externalities,
market power and inadequate or asymmetric information. Externalities are the most likely cause of
market failure in an environmental context
Market-based approaches - Market-based approaches to environmental policy include instruments
such as taxes, fees, charges and subsidies. These approaches create a price incentive to reduce pollution
and leave decisions about the level of emissions to each source. Another example is an allowance
trading system, which sets the total quantity of emissions and then allows trading of permits among
firms.
Meta-analysis - Meta-analysis is an umbrella term for a suite of techniques that synthesize the
results of empirical research. This could include a simple ranking o results, a meta-analytic average
or other central tendency estimate, or a multivariate regression.
Net benefits - Net benefits are calculated by subtracting total costs from total benefits.
Net future value - Net future value is similar to NPV, however, instead of discounting all future
values back to the present, values are accumulated forward to some future time period for
example, to the end of the last year of a policy's effects.
Net present value (NPV) - The NPV is calculated as the present value of a stream of current and
future benefits minus the present value of a stream of current and future costs.
Non-use value - Non-use value is the value that an individual may derive from a good or resource
without consuming it, as opposed to the value obtained from use of the resource. Non-use values
can include bequest value, where an individual places a value on the availability of a resource to
future generations; existence value, where an individual values the mere knowledge of the existence
of a good or resource; and paternalistic altruism, where an individual places a value on others'
enjoyment of the resource.
Nudge - A nudge is a structural or design feature of an individual choice that alters people's
behavior in a predictable way without precluding any options or significantly changing their
economic incentives but that can still be easily avoided.
Operating costs - Operating costs are recurring expenditures associated with the operation and
maintenance of equipment, including salaries and wages, energy inputs, materials and supplies,
purchased services and maintenance or repair of equipment associated with pollution
abatement or waste management.
Opportunity cost - Opportunity cost is the value of foregone allocation during some resource
economic decision; the value of foregone allocation is often described as the "value of the next best
alternative use" of the resource." Opportunity cost need not be assessed in monetary terms. It can be
assessed in terms of anything that is of value to the person or persons doing the assessing. For
example, a grove of trees used to produce paper may have a next-best-alternative use as habitat for
spotted owls. Assessing opportunity costs is fundamental to assessing the true cost of any course of
action. In the case where there is no explicit accounting or monetary cost (price) attached to a course
of action, ignoring opportunity costs could produce the illusion that the action's benefits cost nothing
at all. The unseen opportunity costs then become the implicit hidden costs of that course of action.
Pareto efficiency or Pareto optimality - Pareto efficiency is an economic state in which it is
impossible to reallocate resources to make one individual better off without making another worse
off.
iv
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Partial equilibrium - A partial equilibrium modeling approach accounts for market changes in the
regulated sector. Market responses to the regulation may include reduced industry output or higher
prices as firms pass on some costs directly to consumers. The goal of a partial equilibrium approach
is to measure the net change in consumer and producer surplus relative to the pre-regulatory
equilibrium.
Performance-based standard - A performance-based standard is a pollution control standard that
requires polluters to meet a source-level emission standard without mandating the specific method
by which they must comply with the standard. A performance-based standard is defined in terms
of an emission level or an emission rate (i.e., emissions per unit of output or input).
Prescriptive regulation - A prescriptive regulation is a policy that stipulates how much pollution
an individual source or plant is allowed to emit and/or what types of control equipment or
approaches it must use to reduce pollution. Prescriptive regulations are also known as "direct
regulatory instruments" or "command-and-control" regulations.
Price elasticity of demand - Elasticity of demand measures the relationship between changes in
quantity demanded of a good and changes in its price. It is calculated as the percentage change in
quantity demanded that occurs in response to a percentage change in price. As the price of a good
rises, consumers will usually demand a lower quantity of that good. The greater the extent to which
quantity demanded falls as price rises, the greater is the price elasticity of demand. Some goods for
which consumers cannot easily find substitutes, such as gasoline, are considered price inelastic.
Note that elasticity can differ between the short term and the long term. For example, if the price of
gasoline rises, consumers will eventually find ways to conserve their use of the resource. Some of
these ways, like finding a more fuel-efficient car, take time. Hence gasoline would be price inelastic
in the short term and more price elastic in the long term.
Price elasticity of supply - Elasticity of supply measures the relationship between changes in
quantity supplied of a good and changes in its price. It is measured as the percentage change in
quantity supplied that occurs in response to a percentage change in price. For many goods the
quantity supplied can be increased over time, for example, by locating alternative sources or
investing in an expansion of production capacity. One might therefore expect that the price
elasticity of supply will be greater in the long term than the short term for such a good, that is, that
supply can adjust to price changes to a greater degree over a longer period of time.
Quality-adjusted life year (QALY) - QALY is a composite measure used to convert different types
of health effects into a common, integrated unit, incorporating both the quality and quantity of life
lived in different health states. This metric is commonly used in medical arenas to make decisions
about medical interventions.
Rebound effect - A rebound effect is the reduction in expected gains from improvements in the
energy efficiency of a technology that results from changes in consumer behavior. For example,
tighter vehicle fuel economy standards lead to rebound effects because these regulations make it
cheaper to consume energy or fuel on a per-unit basis causing demand for these services and
therefore emissions from them to increase.
Shadow price of capital - The shadow price of capital accounts for the social value of displacing
private capital investments. For example, when a public project displaces private sector
investments, the correct method for measuring the social costs and benefits requires an adjustment
of the estimated costs (and perhaps benefits as well) prior to discounting using the consumption
rate of interest This adjustment factor is referred to as the "shadow price of capital."
₯
Guidelines for Preparing Economic Analyses | 3rd edition
-------
Social benefits - Social benefits are the sum of all positive changes in societal well-being
experienced as a result of the regulation or policy action. Economists define benefits by focusing on
changes in individual well-being, referred to as welfare or utility. Willingness to pay (WTP) is the
preferred measure of these changes as it theoretically provides a full accounting of individual
preferences across trade-offs between income and the favorable effects.
Social cost - Social cost means the sum of all opportunity costs, or reductions in societal well-being,
incurred as a result of the regulation or policy action. These opportunity costs consist of the value lost
to society of all the goods and services that will not be produced and consumed as regulated entities
reallocate resources in order to comply with the regulation. To be complete, an estimate of social cost
should include both the opportunity costs of current consumption that will be foregone because of the
regulation, and also the losses that may result if the regulation reduces capital investment and thus
future consumption.
Social opportunity cost of capital - Social opportunity cost of capital is the rate at which
consumption in the next period is reduced because private investment is displaced by required
investments from policy. This is the rate at which society can trade consumption over time due to
productive capital.
Social rate of time preference - Social rate of time preference is the discount rate at which society
is willing to trade consumption in one period (usually year) for consumption in the next period.
Social welfare function - A social welfare function establishes criteria under which efficiency and
equity outcomes are transformed into a single metric, making them directly comparable. A
potential output of such a function is a ranking of policy outcomes that have different aggregate
levels and distributions of net benefits. A social welfare function can provide empirical evidence
that a policy alternative yielding higher net benefits, but a less equitable distribution of wealth,
ranks better or worse than a less efficient alternative with more egalitarian distributional
consequences.
Stock pollutants - A stock pollutant is a pollutant for which the environment has little or no
absorptive capacity, such as non-biodegradable plastic, heavy metals such as mercury, and
radioactive waste. A stock pollutant accumulates through time.
Subsidy - A subsidy is a kind of financial assistance, such as a grant, tax break or trade barrier, that
is implemented to encourage certain behavior. For example, the government may directly pay
polluters to reduce their pollution emissions.
Technology standard - A pollution control standard that mandates the use of a specific control
technology or production process by individual polluters.
Transaction costs - Transactions costs are the costs incurred when buying or selling a good or
service. They may include the costs of searching out a buyer or seller, bargaining and enforcing
contracts.
Transfers - Transfers are shifts in money or resources from one part of the economy (e.g.,
a group of individuals, firms, or institutions) to another in a way that does not affect the total
resources that are available to society.
Use value - Use value is the value that an individual may derive from consumption or use of a good
or resource.
Value of statistical life (VSL) - The VSL is the marginal rate of substitution (MRS) between
mortality risk and money, i.e., the willingness-to-pay (WTP) for small reduction in the risk of
premature mortality.
vi Guidelines for Preparing Economic Analyses | 3rd edition
-------
Value of statistical life year (VSLY) - The VSLY is an estimated dollar value for a year of statistical
life. In practice this metric is typically derived by dividing a VSL estimate by remaining life
expectancy or discounted remaining life expectancy. This approach usually assumes that each year
of life over the life cycle has the same value.
Willingness to accept (WTA) - WTA is the amount of compensation an individual would be willing
to take in exchange for giving up some good or service. In the case of an environmental policy, WTA
is the least amount of money that an individual would accept to forego an environmental
improvement (or endure an environmental decrement).
Willingness to pay (WTP) - WTP is the largest amount of money that an individual would pay to
receive the benefits (or avoid the damages) resulting from a policy change, without being made
worse off. In the case of an environmental policy, WTP is the maximum amount of money an
individual would pay to obtain an improvement (or avoid a decrement) in an environmental effect
of concern.
vii Guidelines for Preparing Economic Analyses | 3rd edition
------- |