INTEGRATED ENVIRONMENTAL
1 DECISION-MAKING IN THE
21st CENTURY
EVI EW DRAFI
May 3,1999
REPORT FROM THE EPA SCIENCE ADVISORY
BOARD'S INTEGRATED RISK PROJECT
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
SCIENCE ADVISORY BOARD
INTEGRATED RISK PROJECT
Steering Committee
CHAIR
Or. Genevieve M. Matanoski, The Johns Hopkins University, Baltimore, MD
MEMBERS
Dr. Joan M. Oaisey, Lawrence Berkeley Laboratory, Berkeley, CA
Dr. Paul Deisler (Consultant), Austin, TX
Dr. Mark A. Harwell, University of Miami, Miami, PL
Dr. Wayne Kachel, MELE Associates, Brooks AFB, TX
Dr. Alan Maki, Exxon Company, USA, Houston, TX
Dr. Paul R. Portney, Resources for the Future, Washington, DC
Dr. Milton Russell (Consultant), Joint Institute for Energy and Environment and U. Tenn., Knoxville, TN
Dr. Ellen Silbergeld, University of Maryland, Baltimore, MD
Dr. Robert Stavms (Consultant), Harvard University, Cambridge, MA
Dr. Paul H. Templet (Consultant), Louisiana State University, Baton Rouge, LA
Dr. Valerie Thomas (Consultant), Princeton University, Princeton, NJ
Dr. Bernard Weiss (Consultant), University of Rochester Medical Center, Rochester, NY
Dr. Marcia Williams (Consultant), Putman, Hayes & Bartlett, Inc., Los Angeles, CA
Dr. Terry F. Yosie (Consultant), Ruder Finn, Inc., Washington, DC
Dr. Terry F. Young, Environmental Defense Fund, Oakland, CA
SCIENCE ADVISORY BOARD STAFF
Ms. Stephanie Sanzone, Designated Federal Official, US EPA, Science Advisory Board
Mr. Thomas O. Miller, Designated Federal Official, US EPA, Science Advisory Board
Ms. Wanda Fields, Management Assistant, US EPA, Science Advisory Board
Ecological Risks Subcommittee
CHAIR
Dr. Mark A. Harwell, Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL
MEMBERS
Dr. William Adams, Kennecott Utah Copper Corp, Magna, UT
Dr. Steven M. Bartell, SENES Oak Ridge, Inc., Oak Ridge, TN
Dr. Kenneth W. Cummins, South Florida Water Management District, Sanibel, FL
Dr. Virginia Dale, Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN
Dr. Carol Johnston, Natural Resources Research Institute, University of Minnesota, Duluth, MN
Dr. Frederick K. Pfaender, Carolina Fed. for Environmental Programs, University of North Carolina, Chapel Hill, NC
Dr. William H. Smith, School of Forestry and Environmental Studies, Yale University, New Haven, CT
Dr. Terry F. Young, Environmental Defense Fund, Oakland, California
SCIENCE ADVISORY BOARD STAFF
Ms. Stephanie Sanzone, Designated Federal Official, US EPA, Science Advisory Board
Ms. Wanda R. Fields, Management Assistant, US EPA, Science Advisory Board
-------
Human Exposure and Health Subcommittee
CO-CHAIRS
Dr. Joan Daisey, Lawrence Berkeley Laboratory. Berkeley, CA
Dr. Bernard Weiss, Department of Environmental Medicine, University of Rochester Medical Center, Rochester, NY
MEMBERS
Dr. Stephen Ayres, School of Medicine, International Health Programs, Virginia Commonwealth University,
Richmond, VA
Dr. Paul Bailey, Mobil Business Resources Corporation, Product Stewardship & Toxicology, Paulsboro, NJ
Dr. George Daston, Miami Valley Laboratories, The Procter and Gamble Co., Ross, OH
Dr. Curtis Klaussen, Department of Pharmacology, University of Kansas Medical Center,
Kansass City, KS
Dr. Paul Lioy, Environmental and Occupational Health Sciences Institute, Rutgers University, Piscataway, NJ
Dr. William Pease, Environmental Defense Fund, Oakland, CA
Dr. Henry Pitot, McArdle Laboratory for Cancer Research, University of Wisconsin, Madison, Wl
Dr. Jonthan Samet, Department of Epidemiology, Johns Hopkins University, Baltimore, MD
Dr. Valerie Thomas, Center for Energy and Environmental Studies, Princeton University, Princeton, NJ
Dr. Lauren Zeise, Office of Environmental Health Hazard Assessment, California Environmental Protection Agency,
Berkeley, CA
SCIENCE ADVISORY BOARD STAFF
Mr. Samuel Rondberg, Designated Federal Official, U. S. EPA, Science Advisory Board
Ms. Mary L. Winston, Management Assistant, U. S. EPA, Science Advisory Board
Economic Analysis Subcommittee
CHAIR
Dr. Paul R. Portney, Resources for the Future, Washington, D.C. 20036
MEMBERS
Dr. Nancy E. Bockstael, Department of Agricultural and Resource Economics, University of Maryland, College Park,
MD Dr. Trudy Ann Cameron, Department of Economics, University of California, Los Angeles, CA
Dr. Maureen L. Cropper, The World Bank, Washington, DC
Dr. A. Myrick Freeman, Department of Economics, Bowdoin College, Brunswick, ME
Dr. Charles D. Kolstad, Department of Economics, University of California, Santa Barbara, CA
Dr. Robert Repetto, Economic Research Program, World Resources Institute, Washington, D. C.
Dr. Robert N. Stavins, John F. Kennedy School of Government, Harvard University, Cambridge, MA
Dr. Thomas H. Tietenberg, Dept. of Economics, Colby College, Waterville, ME
Dr. W. Kip Viscusi, Harvard Law School, Cambridge, MA
SCIENCE ADVISORY BOARD STAFF
Mr. Thomas Miller, Designated Federal Official, U.S. EPA, Science Advisory Board
Ms. Diana L. Pozun, Management Assistant, U.S. EPA, Science Advisory Board
-------
Valuation Subcommittee
CO-CHAIRS
Dr. Alan W. Maki, Exxon Company, USA, Houston, TX
Dr. Milton Russell (Consultant), Joint Institute for Energy & Environment and U. Tennessee, Knoxville, TN
MEMBERS
Dr. Stephen M. Ayres, Virginia Commonwealth University, Medical College of Virginia, Richmond, VA
Dr. Nancy E. Bockstael, Dept. of Agricultural and Resource Economics, University of Maryland, College Park, MD
Dr. Caron Chess (Consultant), Center for Environmental Communications, Rutgers University, New Brunswick, NJ
Dr. Virginia Dale, Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN
Dr. William H. Desvousges (Consultant), Triangle Economic Research, Durham, NC
Dr Thomas Dietz (Consultant), Department of Sociology and Anthropology, George Mason University, Fairfax, VA
Dr. A. Mynck Freeman, Department of Economics, Bowdom College, Brunswick, ME
Dr. Mark A. Harwell, Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL
Professor Jerry A. Hausman (Consultant), Department of Economics, Massachusetts Institute of Technology,
Cambridge, MA
Dr. Douglas E. MacLean (Consultant), Department of Philosophy, University of Maryland, Baltimore, MD
Dr. John W. Payne (Consultant), Fuqua School of Business, Duke University, Durham, NC
Dr. Edella Schlager (Consultant), School of Public Administration and Policy, University of Arizone, Tucson, AZ
Dr. Margaret Shannon (Consultant), Center for Environmental Policy and Administration, Syracuse University,
Syracuse, NY
Dr. Paul Templet (Consultant), Institute for Environmental Studies, Louisiana State University, Baton Rouge, LA
Dr. Terry F Young, Environmental Defense Fund, Oakland, California 94611
Dr. James Wilson (Consultant), Department of Resource Economics and Policy, University of Maine, Orono, ME
Science Advisory Board Staff
Mr. Thomas Miller, Designated Federal Official, U.S. EPA, Science Advisory Board
Ms. Diana Pozun, Management Assistant, U.S. EPA, Science Advisory Board
Risk Reduction Options Subcommittee
CO-CHAIRS
Dr. Wayne M. Kachel, MELE Associates, Brooks AFB, TX
Ms. Marcia Williams, (Consultant), Putman, Hayes & Bartlett, Inc., Los Angeles, CA
MEMBERS
Ann Bostrom (Consultant), School of Public Policy, Georgia Institute of Technology, Atlanta, GA
Ms. Dorothy P. Bowers (Consultant), Environmental and Safety Policy, Merck & CO., Inc., Whitehouse Station, NJ
Mr. Robert Frantz (Consultant), Corporate Environmental Programs, General Electnc Company,
Dr. Nina Bergan French, SKY+, Oakland, CA
Ms. Mary A. Gade (Consultant), Illinois Environmental Protection Agency, Springfield, IL
Mr. Bradford S. Gentry (Consultant), Yale University, The Center for Environmental Law & Policy, New Haven, CT
Dr. Ricardo R. Gonzalez (Consultant), Department of Radiological Sciences, Univ. of Puerto Rico School of
Medicine, San Juan, PR
Dr. Michael Greenberg (Consultant), The State University of New Jersey, Rutgers, Department of Urban Studies &
Community Health, New Brunswick, NJ
Dr. Linda E. Greer (Consultant), Natural Resources Defense Council, Washington, DC
Dr. Hilary I. Inyang, Center for Environmental Engineering and Sciences Technologies (CEEST), University of
Massachusetts, Lowell, MA
Dr. Charles D. Kolstad, University of California, Department of Economics, Santa Barbara, CA
Terrence J. McManus (Consultant), Corporate Environmental Affairs, Intel Corporation, Chandler, AZ
Dr. Wm. Randall Seeker, Energy & Environmental Research Corp., Irvine, CA 92718
SCIENCE ADVISORY BOARD STAFF
Ms. Kathleen W. Conway, Designated Federal Official, U.S. EPA, Science Advisory Board,
Ms. Dorothy M. Clark, Management Assistant, U.S. EPA, Science Advisory Board
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
INTEGRATED ENVIRONMENTAL DECISION-MAKING IN THE 21st
CENTURY
TABLE OF CONTENTS
PART I — THE FRAMEWORK
1. Proposed Framework for Integrated Environmental Decision-Making
1.1 Integrated Environmental Decision-Making 1-1
1.2 Scope of the Project 1-8
1.3 A Proposed Framework for Integrated Environmental Decision-Making 1-12
1.4 Nature of the Framework 1 -23
1.5 Benefits and Challenges of the Framework 1-33
1.6 References Cited 1-36
PART II — INPUTS TO ENVIRONMENTAL DECISION-MAKING: RISK COMPARISONS
Preface
2. Ecological Risks
2.1 Background 2-1
2.2 Objectives and Approach 2-3
2.3 ERS Ecological Risk Ranking Methodology 2-4
2.4 National-Scale Ecological Risk Ranking 2-23
2.5 An Effects-Backwards Methodology for Risk Rankings 2-32
2.6 References Cited 2-36
Appendix 2A. Ecological Risk Profiles 2-38
3. Human Health Risks
3.1 Introduction 3-1
3.2. The Environmental Health Risk Rating Methodology 3-4
3.3 Analysis and Reporting of Relative Risk Rating Survey Data 3-16
3.4 Correspondence Between Ecological and Health Risk Formats 3-18
3.5. Implications of Ratings 3-19
3.6 A Fuzzy Logic Approach 3-20
3.7 Extensions and Refinements of the Methodology 3-25
3.8 Summary and Conclusions 3-26
3.9 References Cited 3-27
Appendix 3A. Health Risk Assessment Introduction 3-28
Appendix 3B. Instructions 3-30
Appendix 3C. Risk Characterization Data Sheets 3-33
PART III — INPUTS TO ENVIRONMENTAL DECISION-MAKING: ECONOMICS AND
VALUATION
Preface
4. Benefit-Cost Analysis for Integrated Risk Decisions
4.1 Introduction 4-1
4.2 Fundamental Questions in the Economic Analysis of Risk 4-3
4.3. The Benefits of Risk Reduction 4-8
4.4 Costs of Environmental Protection 4-17
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
4.5 Comparing Total Benefits and Total Costs 4-22
4.6 Distributional Considerations 4-27
4.7 Conclusions 4-28
References Cited 4-31
Endnotes 4-33
5. Assessing the Value of Natural Resources
OVERVIEW 5-1
5.1 Introduction 5-4
5.2 Valuation and the Decision Context 5-6
5.3 The Nature of Values 5-14
5.4 The Economic Valuation Framework 5-20
5.5 The Importance of Deliberative Processes to Valuation 5-26
5.6 Additional Approaches to Valuation of Environmental Systems 5-37
5.7 Summary and Conclusions 5-47
5.8 References Cited 5-52
PART IV — INPUTS TO ENVIRONMENTAL DECISION-MAKING: RISK REDUCTION
APPROACHES
Preface
6. Risk Reduction Options
6.1 Introduction and Approach 6-1
6.2 Define the Problem 6-6
6.3 Develop Background Information 6-12
6.4 Identify the Spectrum of Risk Reduction Options 6-16
6.5 Establish Screening and Prioritization Criteria 6-33
6.6 Screen and Prioritize Potential Risk Reduction Options 6-40
6.7 Evaluate the Remaining Risk Reduction Options 6-48
6.8 Optimize the Options 6-54
6.9 Select an Option 6-59
6.10 Document the Process 6-65
6.11 Quantify Option Effectiveness 6-65
6.12 References Cited 6-71
PART V — IMPLEMENTATION AND PERFORMANCE EVALUATION
Preface
7. Performance Evaluation — The Design and Use of Environmental Report Cards
7.1 Introduction 7-1
7.2 Types of Performance Measures 7-3
7.3 Improved Report Cards 7-8
7.4 Implications for Existing Monitoring Systems 7-17
7.5 Summary and Recommendations 7-19
7.6 References Cited 7-22
PART VI — CONCLUSIONS AND RECOMMENDATIONS
8. Conclusions and Recommendations 8-1
APPENDIX: Summary of a Hypothetical Case Example
-------
PARTI THE FRAMEWORK
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2 CHAPTER 1. PROPOSED FRAMEWORK FOR INTEGRATED
3 ENVIRONMENTAL DECISION-MAKING
4
5 TABLE OF CONTENTS
6
7
8 1.1 Integrated Environmental Decision-Making 1-1
9 1.1.1 The Call for Integrated Decision-Making 1-1
10 1.1.2 Signs of Progress 1-3
11
12 1.2 Scope of the Project 1-8
13
14 1.3 A Proposed Framework for Integrated Environmental Decision-Making 1-12
15 1.3.1 Overview 1-12
16 1.3.2 Phase I: Problem Formulation 1-14
1.3.3 Phase II: Analysis and Decision-Making 1-17
lu 1.3.4 Phase III: Implementation and Performance Evaluation 1-20
19
20 1.4 Nature of the Framework 1-23
21 1.4.1 Major Characteristics 1-23
22 1.4.2 Types of Integration 1-25
23 1.4.3 Building on Previous Frameworks 1-29
24
25 1.5 Benefits and Challenges of the Framework 1-33
26 1.5.1 Benefits 1-33
27 1.5.2 Challenges 1-34
28
29 1.6 References Cited 1-36
30
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 CHAPTER 1. PROPOSED FRAMEWORK FOR INTEGRATED
2 ENVIRONMENTAL DECISION-MAKING
3
4
5 1.1 Integrated Environmental Decision-Making
6
7 1.1.1 The Call for Integrated Decision-Making
8
9 Environmental decision-making has progressed over the years with the accretion
10 of experience and knowledge, becoming ever more subtle, inclusive, and powerful.
11 Despite criticisms leveled against them, and despite inherent limitations, previous
12 environmental decisions and policies have spurred significant environmental and health
13 progress. First, the decisions have addressed a multitude of complicated
14 environmental problems. Second, they have offered broad perspectives on how we
15 formulate and support environmental research. Decision-makers currently draw upon
16 an eclectic mixture of tools and information to inform their decisions: ecological and
human health risk assessment; benefit/cost and cost-effectiveness models; expanded
risk communication and public participation; and measures for monitoring the results of
19 the decisions themselves. Although these tools provide essential inputs for decision-
20 making, they have typically been applied unevenly and to relatively narrow issues. In
21 the area of human health risks, for example, assessments often have been framed
22 around single stressors or classes of stressors in relatively specific exposure situations.
23 The deficiencies of such highly focused assessments are increasingly apparent; they
24 sidestep the complexities, interrelationships, and subtleties of environmental problems
25 as they actually confront us. Thus, a number of recent studies have urged the
26 Environmental Protection Agency (EPA) to begin to address environmental issues in a
27 more integrated way (e.g., NAPA, 1995; Presidential/Congressional Commission on
28 Risk Assessment and Risk Management, 1997).
29
30 Much of the fragmentation in EPA's approach to the control of environmental
31 problems has its roots in the statutory framework that guides the work of the Agency.
32 From its formation in 1970, the EPA has been given responsibility for implementing a
33 number of environmental statutes that mandate targeted actions to control specific
34 pollutants in specific media (e.g., Clean Air Act language regarding particulates in air) or
1-1
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 specific routes of exposure (e.g., Safe Drinking Water Act language regarding priority
2 pollutants in drinking water). The focus on assessing and controlling chemical
3 contaminants pollutant by pollutant in single media has resulted in an evolving
4 collection of federal laws and regulatory requirements that is neither systematic nor
5 comprehensive. Nonetheless, these laws have been largely successful in controlling
6 many of the targeted pollutants and have provided a strong national underpinning for an
7 effective environmental protection program comprised of federal, state, and local
8 controls.
9
10 Yet, despite these successes, there is a growing consensus, both within and
11 outside the Agency, that a more integrated approach to environmental management is
12 needed. Prioritizing and managing risks pollutant by pollutant and medium by medium
13 can be both inefficient at reducing the major burdens of environmental impacts on
14 human health and ecosystems and costly in the face of today's shrinking budgets. Of
15 still greater concern is the possibility that such a fragmented approach may cause us to
16 overlook significant environmental problems while we busy ourselves with
17 comparatively minor issues that contribute little to the overall protection of human
18 health and ecosystems. Further, in some instances, current statutes and regulations
19 prevent the Agency from considering all relevant risk, benefit/cost, or other information.
20 A 1995 report from the National Academy of Public Administration (NAPA) pointed out
21 that there are "no established criteria that the Agency might use to set priorities that cut
22 across statutory lines" and called on Congress and the Agency to give serious thought
23 to an "integrated statute that would provide multi-media decision-making authority to the
24 Agency" (NAPA, 1995). The SAB views the issue of statutory integration as a policy
25' discussion and outside the bounds of the present study. Even within the current
26 statutory framework, however, there are numerous opportunities for a more holistic
27 assessment of risks and risk management options, and more inclusive decision-making
28 approaches.
29
30 The call for integrated assessment of risks has included the need to consider
31 multiple sources, multiple routes of exposure, and multiple human health endpoints
32 (e.g., cancer, genetic effects, and developmental and reproductive toxicity) (see, for
33 example, NRC, 1994), and aggregate risks posed by multiple agents or stressors (e.g.,
34 endocrine disruptors or mixtures of polycyclic aromatic hydrocarbons). Although many
35 Agency risk assessments still focus on a single agent (e.g., lead, 1,3-butadiene), there
1-2
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 clearly has been an evolution in risk assessment methods towards more realistic,
2 multiple source, multiple pathway, multiple agent assessments.
3
4 1.1.2 Signs of Progress
5
6 In the 1983 publication entitled Risk Assessment in the Federal Government:
7 Managing the Process (NRC, 1983), commonly referred to as the "Red Book", an NRC
8 panel laid out the elements of risk assessment and risk management using terminology
9 that came to be the standard. These concepts were adopted by EPA Administrator
10 Ruckleshaus in 1984 and have formed the basis for much of the Agency's action to this
11 day. In summary, the NRC committee described the four steps of risk assessment as
12 hazard identification, dose-response assessment, exposure assessment, and risk
13 characterization, which was defined as "the estimated incidence of the adverse effect in
14 a given population" (NRC, 1983) (Figure 1-1). In addition, the NRC committee stressed
15 the scientific basis for risk assessment and the need for both quantitative and
16 qualitative expressions of risk. Risk management was viewed as "a decision-making
17 process that entails consideration of political, social, economic, and engineering
information with risk-related information to develop, analyze, and compare regulatory
19 options and to select the appropriate regulatory response..." (NRC, 1983).
20
21 The Red Book was extremely useful in articulating the risk assessment process
22 and its relationship to risk management. The paradigm was expressed, however, in
23 terms of single agents and single health effects in humans. Since that time, the Agency
24 has developed risk assessment guidelines to address a number of endpoints (i.e.,
25 cancer, reproductive and developmental toxicity, and neurotoxicity), as well as exposure
26 assessment, which is a component of the risk assessment model. In addition, the
27 Agency has taken steps to consider more integrated exposure scenarios, e.g., multi-
28 route exposures to mixtures of chemical agents associated with Superfund sites (U.S.
29 EPA, 1989) or combustor emissions (EPA, 1990; 1993a), and, in response to the Food
30 Quality Protection Act, to consider multiple pesticides and multiple routes of exposure in
31 assessing children's risk.
32
33
1-3
-------
[FILEAREDBOOK.PDQ]
RESEARCH
RISK ASSESSMENT
RISK MANAGEMENT
LaBoratory ind lulu
ootervations ol advaise
Main enacts ind
eipsoiuies 10 oartcHar
iginii
Hizara Identrfciuoir
(Deal IM agent causa
IM adverse efteef)
Inloimaien en
exnaeoiauon tnatnots loi
(ugn 10 ton dosa and
inmal 10 luiraan
Ooia-Raipona
As!»sjm«nc
(Whal is IM ralannshp
baiwaan easa and
ncdanca « Humans1)
FiaH maasuramants.
astmaiao aiposuras.
charactanann ol
pogdaieia
Eiposure Aisasunanc
(Wisai eiposum are
cunanfly eipnancad at
anueipalad under tittaren
conditions')
\
Oatttopmem el
regulatory option!
Rbk Chaiactaroauir
(What s ma asmaitd
nodanca ol tne adnrsa
allact in a gwan
populauoni)
Evahaton ol pubic
rieitnt. aconomc. socal
poSlcal consaquaneaa ol
\
Agancy oecawns ina
actions
Elements of risk assessment and risk management.
\-\
From:
National Research Council, Commission on Life Sciences, Committee on the Institutional Means for
Assessment of Risks to Public Health. 1983. Risk Assessment in the Federal Government:
Managing
the Process. National Academy Press. Washington, D.C. 191pp.
I-H
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 The Agency has also made significant progress in adapting the Red Book
2 paradigm to ecological risk assessment. In 1992, the Agency released its Framework
3 for Ecological Risk Assessment (U.S. EPA, 1992) which used the term "characterization
4 of ecological effects" to include both hazard identification and exposure assessment.
5 The Framework also added an explicit Problem Formulation phase prior to the analysis
6 of exposure and effects to emphasize the importance of articulating the problem and a
7 plan for analyzing and characterizing risk prior to conducting specific risk analyses. The
8 resulting framework contained three phases: Problem Formulation, Analysis, and Risk
9 Characterization (Figure 1-2). An expanded discussion of ecological risk assessment
10 principles and approaches was subsequently provided by the Agency in final Guidelines
11 for Ecological Risk Assessment (U.S. EPA, 1998). The guidelines note that "although
12 ecological risk assessments provide critical information to risk managers, they are only
13 part of the environmental decision-making process" (U.S. EPA, 1998). In addition to
14 assessing the relationship between a particular stressor and a particular effect, the
15 ecorisk guidelines set the stage for considering multiple effects (including cascading
16 effects) associated with a single stressor or source, as well as multiple causes of an
17 observed effect or change in ecological condition. The Agency has already applied the
ecological risk assessment paradigm, including the development of a conceptual model
19 relating various stressors and effects, to five watershed cases (for discussion, see SAB,
20 1997).
21
22 Three additional Agency developments merit brief mention here: a) guidance
23 and support for comparative risk analysis; b) extra-statutory approaches to
24 environmental protection; and c) guidance on planning and scoping for cumulative risk
26 assessment.
26
27 First, Comparative Risk Analysis (CRA) has been defined by the Agency as "both
28 an analytical process and a set of methods used to systematically measure, compare,
29 and rank environmental problems" (U.S. EPA, 1993b). The Agency, in its Unfinished
1-5
-------
Ecological Risk Assessment
Planning
(Risk Assessor/
Risk Manager/
Interested Parties
Dialogue)
it*?£feiS?S&£::
Charactenzstio n
of
Ecological
Effects
Characterization
of
Exposure
RJSKLCHARACTERJZATIO
A
T
Communicating Results
to the Risk Manager
^:~]Risk Management and "/
•.Communicating Results to
i-..^Interested Parties : :"/:..
Figure]
1992a).
The framework for ecological risk assessment (modified from U.S. EPA,
f -
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Business report (U.S. EPA, 1987), and the SAB, in Reducing Risk (SAB, 1990),
2 engaged in comparative risk analyses. In its 1990 report, the Board concluded that it
3 was possible, on a scientific basis, to distinguish between large risks and small risks
4 using a set of technical criteria. In the years that followed, the Agency promoted the
5 wide use of CRA as a process for setting priorities by integrating multiple stressors and
6 multiple types of risks within whole regions such as cities, states or even the nation
7 itself. Comparative risk analysis is intended principally as a policy-development and
8 broad resource-allocation tool. In contrast to Unfinished Business and Reducing Risk,
9 however, state and local-level comparative risk analyses have highlighted the role of the
10 public and stakeholder groups, in addition to the scientific/technical community, in
11 defining risk priorities. Support for broader inclusion of public values in decision-making
12 is a theme that has been echoed by a number of recent reports (e.g., NRC, 1996;
13 Presidential/ Congressional Commission on Risk Assessment and Risk Management,
14 1997).
15
16 Second, during the 1990s, the Agency has experimented with a number of
17 approaches to re-inventing environmental protection, including greater use of
community-based decision-making, voluntary cross-media emissions reductions (the
19 33/50 project; reference), integrated environmental agreements with states (National
20 Environmental Performance Partnership agreements), and voluntary regulatory reform
21 efforts with an array of stakeholders (e.g., Common Sense Initiative, Project XL).
22
23 Third, the Agency has recently issued guidance directing program offices to
24 "consider a broader scope that integrates multiple sources, effects, pathways,
25 stressors, and populations for cumulative risk analyses in all cases for which relevant
26 data are available" (U.S. EPA, 1997). The cumulative risk guidance also notes that on-
27 going Agency efforts to involve stakeholders "will provide the solid basis for engaging
28 interested and affected parties in risk assessment and risk management issues"
29
30 The SAB has not reviewed the role and adequacy of the science being used in
31 the planning and evaluation of these activities, including CRA. These endeavors are an
32 indication of movement in the Agency toward more integrated and inclusive methods of
33 environmental decision-making. These experiments, however, do not yet represent the
34 mainstream of EPA's efforts.
35
1-7
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 1.2 Scope of the Project
2
3 It is in this atmosphere that the SAB undertook the task of revisiting its 1990
4 report, Reducing Risk, to update and extend the thinking about how science can best
5 inform the decision-making process. The Charge to the SAB from the Agency included
6 requests to:
7
8 a) update the risk rankings in Reducing Risk using explicit scientific criteria and
9 the judgments of SAB panel members;
10 b) identify risk reduction opportunities and strategies;
11 c) identify uncertainties and data quality issues associated with the risk rankings;
12 d) assess costs and benefits of risk reduction options; and
13 e) propose a new framework for assessing the value of ecosystems.
14
15 The initial charge also included a request that the SAB explore techniques and
16 criteria for identifying emerging risks. However, the SAB concluded that its recent
17 report, Beyond the Horizon: Using Foresight to Protect the Environmental Future (SAB,
18 1995) provided criteria and suggestions germane to this charge question and so did not
19 elaborate further on the future risks as part of the Integrated Risk Project.
20
21 After careful consideration of the Charge and discussion with Deputy
22 Administrator Fred Hansen, the SAB concluded that it could best assist the Agency by
23 investigating approaches for accomplishing these goals and considering the inter-
24 relationship of these tasks in the broader decision-making context.
26
26 The project, known as the Integrated Risk Project, was guided by a Steering
27 Committee and five specialized Subcommittees working over several years. The
28 Subcommittees and their respective charges were as follows:
29
30 a) The Steering Committee (SC), chaired by Dr. Genevieve Matanoski, set
31 the overall direction for the project by defining scope and timetables. The
32 SC met periodically over the course of the project to: (1) assess the
33 progress and direction of the subcommittees; (2) ensure that the results
34 could be integrated into a comprehensive decision process for identifying
35 current and future environmental risks; and (3) review options for reducing
1-8
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 risks in a holistic context. The SC's efforts were designed specifically to
2 illustrate the relationship among the various factors influencing risk
3 management decisions (e.g., technical assessment of the risks and risk
4 reduction options, economic considerations, equity considerations, and so
5 forth).
6
7 b) The Ecological Risks Subcommittee (ERS), chaired by Dr. Mark
8 Harwell, was charged to assess and rank risks to ecosystems at the
9 national scale, as well as to suggest ways in which the risk ranking
10 methodology could be applied at smaller geographical scales (e.g.,
11 regional, state, or local). The group was also asked to explore
12 commonalities and differences with the Human Exposure and Health
13 Subcommittee (HEHS) methodology with the aim of integrating the two
14 ranking schemes.
15
16 c) The Human Exposure and Health Subcommittee (HEHS), co-chaired
17 by Drs. Joan Daisey and Bernard Weiss, was charged to develop a
methodology for assessing and ranking risks to human health, to consider
19 ways in which an integrated risk ranking could be produced that includes
20 both cancer and non-cancer risks, and to test the methodology for a
21 limited set of environmentally mediated health issues. The Subcommittee
22 was also asked to explore commonalities and differences with the ERS
23 methodology.
24
25 d) The Risk Reduction Options Subcommittee (RROS), co-chaired by Dr.
26 Wayne Kachel and Ms. Marcia Williams, was charged with developing a
27 methodology for selecting an optimal set of risk reduction options with due
28 regard for the human health and ecological risks (defined in terms of risks
29 associated with environmental stressors, locations, or exposure/transport
30 media). Because of the time constraints on the project, the RROS was
31 asked by the SC only to illustrate the methodology for one or more
32 example problems in lieu of addressing the wider range of risks
33 considered by the ERS and the HEHS.
34
35
1-9
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 e) The Economic Analysis Subcommittee (EAS), chaired by Dr. Paul
2 Portney, was charged with assessing current methods for estimating costs
3 and benefits (either physical or monetary) associated both with the
4 implementation of risk reduction strategies and with allowing risks to go
5 unaddressed. The EAS was also asked to consider those aspects of the
6 "net benefits" equation that cannot easily be monetized.
7
8 f) The Valuation Subcommittee (VS), co-chaired by Drs. Alan Maki and
9 Milton Russell, was charged to consider a new framework for assessing
10 the value of ecosystems to humans, including ecological services and
11 environmentally mediated health and quality of life values. The work of
12 the VS was intended to provide a wider societal view of risk and risk
13 reduction options than that derived from science-based risk assessments
14 and current methods of economic analysis.
15
16 Over the course of the project, the IRP SC and subcommittees held over 25
17 public meetings and teleconference calls. Although most of these meetings were held
18 in Washington, D.C., public sessions were also held in Berkeley and San Francisco,
19 CA; Atlanta, GA; New Orleans, LA; and Baltimore, MD.
20
21 The conceptual model that emerged-^the framework for Integrated
22 Environmental Decision-making (IED)—is one that emphasizes the dialogue and
23 interaction between risk assessors, risk reduction options analysts (e.g., engineers,
24 economists, and environmental law/policy experts), decision-makers, and the public.
26 This report describes the IED framework for making integrated decisions, and
26 recommends needed improvements to the tools and techniques required to both
27 implement and evaluate IED.
28
29 Throughout the study, the Steering Committee used stressors as a common
30 focus to link the components of the framework because stressors are the actual
31 physical, chemical, or biological changes that affect ecological or human systems, and
32 thus, will be the focus of risk reduction actions. However, the IED framework provides a
33 multi-dimensional approach to problem definition and options analysis, one that
34 recognizes the linkages between stressors, sources, pathways of exposure, and
35 adverse effects endpoints. It is this multi-dimensional approach that can lead to the
1-10
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 most effective overall reduction of risk.
2
3 The SAB, as a body of scientists, engineers, and economists, is best suited to
4 describing the scientific and technical analyses that should inform decision-making.
5 The project focused, therefore, on methods for assessing and comparing multiple risks
6 to human health or the environment; an approach for designing a risk reduction
7 program; qualitative and quantitative benefit/cost analyses; and considerations in the
8 design of performance evaluation systems. With regard to those aspects of the IED
9 process that rely on the application of public values (e.g., the selection of specific risk
10 goals and tradeoffs, and decision criteria), the SAB cannot represent the public.
11 However, the report does discuss the importance of public values to decision-making
12 and suggests ways in which EPA, and other decision-making bodies, can utilize
13 deliberative processes to engage stakeholders and the public in aspects of
14 environmental decision-making.
15
16 This report represents the SAB's broadest view of environmental decision-
17 making to date. The proposed framework is designed as a flexible guide that can be
used to address environmental problems of different size, scope, and location. The
19 effort to develop the IED has required the SAB to venture into risk management areas
20 that do not usually arise in most of its work. However, while methods for reaching
21 decisions about regulatory matters or policies are part of the IED framework, the SAB
22 offers no recommendations directly impinging on specific decisions or policies.
23
24
1-11
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 1.3 A Proposed Framework for Integrated Environmental Decision-Making
2
3 1.3.1 Overview
4
5 In order to encourage a more integrated approach to environmental protection,
6 the SAB proposes a framework for decision-making based on the recognition that
7 requires in-depth analysis of projected risk reduction under possible management
8 scenarios and selection of a preferred scenario, based on criteria such as feasibility,
9 cost-effectiveness, seriousness of the risks addressed, and equity. The final phase of
10 IED—Implementation and Performance Evaluation—is one in which the implementation
11 of risk reduction measures occurs and environmental results are monitored and
12 evaluated. Performance Evaluation provides critical feedback so that management
13 approaches can be fine-tuned and the extent and nature of remaining risks, and the
14 means for reducing those risks if necessary, can be re-evaluated.
15
16 The complete IED framework is an integrative scheme for making decisions
17 where many different variables, often interacting across physical, regulatory, and
18 organizational boundaries, can be considered simultaneously rather than in isolation by
19 the many types of participants. It allows for: a) the consideration of related clusters of
20 risks; b) the development of multiple risk reduction options; c) the definition of markers
21 for evaluating progress toward specific environmental goals; and d) consideration of
22 public preferences and values throughout the process.
23
24 Although the IED process requires the involvement of a broad spectrum of
25' participants (e.g., scientists, engineers, economists, decision-makers, and the public),
26 the different groups have unique roles to play. In other words, the framework does not
27 imply that "everyone must be involved in everything all the time." For example, just as
28 scientists cannot provide the perspective of the general public, members of the general
29 public cannot do the job of scientists. Decision-makers, after considering the various
30 sorts of information (data, views, and judgments) generated by the problem formulation
31 and analysis, must make the decision. The important message of the IED framework is
32 that the various groups involved must maintain effective communication with each other
33 throughout the process so that each can most effectively do its job within the overall
34 context.
35
1-12
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
/Information
Expert
Judgment
Values
/Information
Expert
Judgment
Values
Legal and
Institutional
Milieu
Figure 1-3. Integrated Environmental
Decision-Making Framework
PROBLEM FORMULATION
What are the most important environmental risks?
What are our environmental goals?
/•—
Risk 1} Goal Setting
vCompansons\
^
Preliminary
Options Analysis
t
ANALYSIS AND DECISION-MAKING
What are the best risk reduction opportunities?
How can we achieve our goals and objectives?
IMPLEMENTATION and
PERFORMANCE EVALUATION
How are we doing?
/frnplementatiorN /Monitoring \/Information A
t }\and Reporting JI Evaluation J
REPC
CAR
Is the nature
of the problem
changing?
RE
CA
Are
our
»ORT
RD
we meeting
ofyecf/Ves?
RT
1-13
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 As illustrated in Figure 1-3, the IED framework is intended to answer a series of
2 straightforward questions. What are the most important environmental risks? What are
3 our environmental goals? What are the best risk reduction opportunities? How can we
4 achieve our goals and objectives? How will we know whether or not we are meeting our
5 goals? What modifications in our approach are needed to improve environmental
6 results? Finding answers to these fundamental questions requires application of
7 scientific and technical assessment and analysis techniques, as well as political, policy,
8 and values-driven choices. The following sections describe the IED process, including
9 the use of the various analytical methods.
10
11 1.3.2 Phase I: Problem Formulation
12
13 The initial phase of the IED framework is Problem Formulation, in which
14 agreement should be reached among all participants—risk assessors, risk managers,
15 and interested and affected parties—about what needs to done by whom and why, if
16 not how. As shown in Figure 1-3, Phase I includes primarily three related tasks: Risk
17 Comparisons, Goal Setting, and Preliminary Options Analysis. In this initial Phase, the
18 discussions are at the level of planning, scoping, and screening, rather than the
19 detailed analyses conducted in Phase II.
1-14
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
Figure 1-4. Risk Comparisons
The term Risk Comparison is used in this report to denote the characterization and ranking of risks
posed by environmental stressors, where an environmental stressor is any physical, chemical, or biological
change or agent that could affect ecological or human health systems. In the case of a single stressor, this
analysis consists essentially of risk assessment approaches that have already been established for human
health risks (NRC, 1983) and for ecological risks (EPA 1992,1998). The information resulting from this
component is a characterization of the nature and magnitude of the risks posed by the stressor and
characterization of the systems at risk and elements of systems that may be exposed to the stressor.
In the multi-stressor case, it is important to consider a wide range of environmental risks
simultaneously so that the seriousness of nsks can be characterized relative to one another. This may include
comparison of nsks across different stressors affecting human health, ecosystems, and/or quality of life. Risk
comparisons may be done by comparing quantitative estimates of nsks where that is possible, or by qualitative
evaluation of nsks against some set of criteria. While quantitative risk comparisons are ideal, qualitative
compansons may be best suited when comparing non-commensurate risks (e.g., nsks to humans compared to
risks to ecosystems). Chapters 2 and 3 descnbe approaches to risk comparisons designed to utilize the best
available scientific information and expert judgment to categonze nsks (e.g., high, medium, and low) and make
transparent the influence of any value judgments or tradeoffs. The two approaches can be used to rank risks
within, but not across, the categories of human health or ecological nsks. This is because the ranking factors,
while analogous, differ in specifics for the two types of nsks.
The I ED framework also allows for the explicit assessment and companson of nsks to quality of life
(QOL), often defined as potential non-health impacts on humans from environmental change. Examples of
nsks often included in this category are aesthetic, economic, and equity impacts, as well as effects on peace of
mind, cultural or community identity, and recreational opportunities. Although some of these QOL risks can be
assessed via analysis of ecological risks or benefit/cost, the SAB did not specifically propose a method for
ranking QOL risks because the selection of QOL ranking criteria is largely a value-dnven, rather than scientific.
process and as such is more appropriately conducted by a broad group of public or stakeholder
representatives rather than by a technical panel. However, the EPA (1993) has developed a guidebook for
assessing quality of life risks that provides a starting point for such assessments. In most of the state
comparative risk projects, risks to human welfare or quality of life have been considered by a separate, non-
technical subcommittee that developed criteria and produced a ranked list of QOL risks.
The SAB also does not propose a formal process for merging the ecological and human health risk
issues into a single ranked list because such an activity appropriately would require consideration of many
non-technical issues, including political acceptability and societal values. However, the development of a
single pnoritized list of risks of concern, including nsks to ecosystems, human health, and quality of life, can be
developed with public/values deliberation and is an appropriate activity during Problem Formulation.
1 Risk comparison methods (see Figure 1 -4) are used to identify the sets of risks
2 to ecological, human health, and/or quality of life systems that will be the subject of
3 detailed consideration in Phase II. Preliminary analysis of risk reduction options is
4 important in Phase I in order to formulate the problem in terms that will be amenable to
5 the greatest overall risk reduction. Problem Formulation requires considerable dialogue
6 between the participants comparing risks and those involved with preliminary
7 identification of available options. This dialogue is designed to focus the problem and
1-15
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 to ensure consistency between the stressors and risks being considered and the risk
2 reduction opportunities that may be available.
3
4 Planning, scoping, and screening - including selection of endpoints - also
5 requires explicit input of societal values and stakeholder participation. For example,
6 while some of the ecological endpoints that are selected in the problem formulation
7 phase may be chosen strictly because of the value attached to their ecological role,
8 there are also ecological endpoints that will be chosen because of their particular
9 significance to society. Examples of ecological endpoints in this latter category include
10 both economically important species and endangered species. Similarly, human health
11 endpoints would likely include both risks to the general population and those relevant to
12 specific subsets of the population, such as children or the elderly, because of societal
13 concerns about their protection.
14
15 There is also an important role for decision-makers in this phase; for example,
16 decision-makers will be involved in helping to identify the important environmental
17 problems to be considered, identifying the sets of at-risk systems to address, and
18 identifying the specific ecological or human health endpoints to select. During Problem
19 Formulation, decision-makers also need to identify clearly the range of potential
20 decisions and management options, examine economic, political, or other constraints
21 on the options to be considered, and to characterize the scope and time frame for IED
22 implementation.
23
24 In summary, in Phase I the participants—scientists, decision-makers, and
25 interested and affected parties—seek agreement through deliberative-analytic dialogue
26 (in the meaning of NRC, 1996) on such issues as:
27
28 a) the goals for the exercise, including environmental goals to be achieved;
29 b) which environmental problems/stressors/systems will be included and which
30 will not;
31 c) effects of concern;
32 d) the spatial, temporal, and organizational dimensions of the problem;
33 e) relevant data, models and analyses;
34 f) possible approaches to data analysis;
35 g) scoping of the uncertainties involved;
1-16
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 h) research needed to significantly reduce critical uncertainties;
2 i) a first-cut on the range of options available to reduce risks;
3 j) the endpoints upon which the health of the ecological or societal systems
4 ultimately will be judged; and
5 k) the type of factors that will be considered when reaching a decision.
6
7 The intent of Phase I is to have an open, yet structured, exchange of information,
8 concerns, opinions, and values that will help to address the kinds of issues listed above.
9 Chapters 2, 3, and 6 describe approaches by which some of the more technical aspects
10 of this information might be developed, and Chapter 5 describes deliberative processes
11 that might be used to set environmental goals and incorporate values deliberation in
12 Problem Formulation. The Agency's experience with some of its stakeholder processes
13 provides insights on constructive interaction for the purposes of planning and scoping.
14
15 1.3.3 Phase II: Analysis and Decision-Making
16
17 Phase II is that portion of the decision-making process in which most of the
traditional "work" is done. Whereas in Phase I the participants formulate the problem
19 using screening level information gained from risk comparisons, goal setting, and
20 preliminary options analysis, in Phase II the technical specialists employ similar
21 approaches, but with greater specificity and data requirements, to develop the in-depth
22 technical information that helps the risk managers to reach a final decision.
23
24 In practice, the analysts take the information and general directions gained from
25 Problem Formulation and generate more detailed, more fully supported risk
26 assessments and risk reduction options. In the context of the IED framework, options
27 analysis requires consideration of risk reduction opportunities with regard to their
28 technical feasibility, overall aggregate risk reduction to be obtained (e.g., reductions in
29 "target" risks and collateral reduction in all affected risks), full economic consequences
30 of various risk reduction scenarios, and so forth (see Figure 1-5). Decision-makers may
31 also request analysis of potential options with regard to sustainability, equity, and other
32 potential
1-17
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
Figure 1-5. Options Analysis, Screening, and Selection
This component of the I ED framework is focused on identifying the best risk reduction opportunities
and is applicable to cases of either a single stressor or for multiple stressors. In its simplest form—the form
that has generally been used historically at the Agency— options analysis involves examination of multiple risk
reduction options to address a highly ranked risk so as to identify the option(s) that will be most cost-effective in
reducing that particular nsk. The IED framework expands this analysis to emphasize that ancillary reductions of
other risks should also be assessed and factored into the decision process. For example, control of fly-ash
emissions to reduce mercury emissions to the environment may not only reduce mercury but may reduce
emissions of participates and possibly other pollutants as well. In an IED approach, these ancillary benefits
should be considered explicitly when selecting a management option for a single stressor.
In the single stressor case, the approach described in Chapter 6 for designing a risk reduction
program identifies possible actions that could be taken to reduce either the stressor or its effects on the
system(s) at risk. An important aspect of this analysis is to examine a broader array of potential options than
might typically have been done in the past. Criteria are developed in consultation with decision-makers to
screen potential options, aggregate or disaggregate options, and, through an iterative process, converge on a
set of options that analysis indicates would optimally reduce the risk. Chapters 4 and 5 provide more detailed
guidance on assessing the economic and societal consequences of various options, an important aspect of
options analysis.
While this simple form of integrated options analysis can yield a broader view of the benefits after an
option or set of options has been selected, a more powerful feature of this integrated analysis results from its
application during Problem Formulation, prior to selection of risk reduction options. In such an application,
scientists should examine appropriate methods by which one can combine subsets of the ranked risks in order
to investigate management options that could impact on the combined risks. The basis for such combinations
of risk might include features such as common sources or pathways through the environment. The critical link
between the environmental risks in the subset would then be that they are all affected by a single risk reduction
option/strategy. While it may not be possible to group all risks of concern on the basis of their technical
attributes, a scientific analysis of the risks may well reveal commonalities that indicate which risks will be
affected by the same risk reduction option. The likely effect of this integrated view is that the option(s) selected
to reduce a group of risks might differ from that which would be selected to reduce the top ranked risk, if it were
to be considered in isolation.
When applied to the multi-stressor situation, the IED framework calls for an expanded analysis of risk
reduction options so as to identify those options that may simultaneously reduce, directly or indirectly, risks
posed by more than one stressor. The goal in this case is to maximize the reduction of the total aggregate nsk
from multiple stressors, rather than to maximize the reduction of risks posed by any single stressor. This
approach, requiring as it may the simultaneous consideration of risks from quite different types of stressors,
has not yet been fully utilized, and will not be trivial to implement Nevertheless, the SAB believes that its
development and implementation offer tremendous potential for improving environmental health overall.
1
2
3
4
5
decision criteria. Methods for assessing the economic and societal consequences of
potential actions are described in Chapters 4 and 5. A detailed discussion of a process
for designing a risk reduction program is contained in Chapter 6.
1-18
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 The Analysis portion of Phase II is generally more "analytic" than "deliberative"
2 (NRC, 1996) although a continued level of interaction between the participants in the
3 overall process (scientists, risk managers, and interested and affected parties) is
4 important. Phase II is also more resource-intensive than Phase I since it can involve
5 consideration of more options (e.g., various groupings of related stressors, as well as
6 different risk reduction options) and more detailed analyses (e.g., the effects of different
7 risk reduction options on multiple stressors or groups of stressors, as well as the
8 inclusion of cost-effectiveness factors).
9
10 In the Decision-Making portion of Phase II, the Agency or other decision-makers
11 utilize outputs from the analyses of risk and risk reduction options, consider widely-held
12 public values, the views of stakeholders, and legal and institutional constraints, and,
13 ultimately, make environmental decisions. Clearly, this is not a totally scientific process.
14 However, the best science should inform and contribute to decision-making. This can
15 be accomplished, for example, by making explicit a) the implications of the chosen
16 management option(s) to the health of ecological or human systems; b) the economic
17 costs and benefits associated with the option; and c) the societal values that are
affected by the decision, including both values relating to economic efficiency and
19 values relating to sustainability and equity.
20
21 Another important role for scientific and technical analysis is to make clear the
22 uncertainties associated with estimates of risk, the estimates of risk reduction that may
23 be achieved by different management options, and the economic assessments of
24 various risk management scenarios. The IED approach does not eliminate the
2& uncertainties associated with making decisions. However, by encouraging an open and
26 comprehensive examination of environmental problems in an integrated fashion, the
27 IED framework should lead to a more clear identification of the nature and extent of the
28 uncertainties associated with the available information. In any event, environmental
29 decision-making must proceed in the presence of uncertainties, and nothing in the
30 proposed framework should be construed as precluding environmental decisions
31 because uncertainties remain.
32
33
1-19
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 1.3.4 Phase III: Implementation and Performance Evaluation
2
3 Phase III of the IED framework consists of implementation of the chosen risk
4 reduction options and evaluation over time of the extent to which the risk reduction
5 measures are achieving the desired environmental outcomes. This Phase involves the
6 articulation and execution of the specific actions that must be taken to implement the
7 decision, and the establishment of a process to evaluate performance and results of
8 such action. The specific activities required to implement an environmental decision will
9 depend on the suite of management options selected for any particular problem or set
10 of problems, and thus we do not address this aspect of the IED in any detail. The
11 Agency has considerable experience with many of the risk reduction options described
12 in this report (e.g., adopting best available technology and imposing permit limits) and is
13 gaining valuable new experience with others (e.g., regulatory negotiation and National
14 Environmental Performance Partnerships).
15
16 In contrast to implementation, however, the performance evaluation process is
17 fundamentally rooted in science because it is science that can translate the public's
18 overarching goals (e.g., improved health, sustainable ecosystems) into discrete,
19 measurable components. Accordingly, science is essential in deciding what to monitor,
20 i.e., specifying the endpoints of concern for the systems at risk and identifying the
21 specific measures that need to be monitored in order to characterize the status and
22 trends for those selected endpoints with respect to the environmental goals. Further,
23 the scientific issues of spatial and temporal variability, measurement error, time lags,
24 and so on, must be explicitly addressed in order to demonstrate environmental
25 ' condition and to separate signal from noise. And finally, reference conditions and
26 benchmarks or milestones along the way to the desired system conditions must be
27 defined scientifically so that meaningful and measurable performance criteria for
28 success or failure can be defined.
29
1-20
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
Figure 1-6. Performance Evaluation/Report Cards
The Government Performance and Results Act (GPRA) created a requirement that,
beginning in Fiscal Year 2000, federal agencies report annually on measurable results of their
various programs and activities. GPRA has resulted in a flurry of activity aimed at measuring
program performance, including an effort funded by the Office of Science and Technology Policy to
design a national environmental report card.
In the IED context, an Environmental Report Card is a tool for communicating to multiple
audiences the performance of a risk reduction program in measurable terms related to environmental
outcomes. At the broadest level, the Report Card should inform all of the IED participants about how
well, in general, the ecological or societal system at risk is responding to the actions taken. At a more
detailed level, the Report Card should provide scientists with sufficient monitoring data to improve the
risk assessments and/or the risk reduction options previously selected. The need for monitoring data
to assess performance in this framework emphasizes again the importance attached to EPA's
development of monitoring programs which can measure both ecological and human health exposure
and outcomes.
The report card should contain specific milestones that can be used to measure progress
towards achieving the environmental goal(s) agreed upon by the IED participants. Each of the
selected endpomts defined during Problem Formulation should be a part of the report card, as well as
the specific measures or indicators that are monitored to characterize those endpoints. The
frequency of the reporting should be decided upon with all of the participants and should be
commensurate with the nature of the risk and the time frame for system response. Four types of
performance measures can be used in evaluating progress: measures of administrative effort, often
called process measures (e.g. number of permits issued, number of control technologies installed);
measures of stressor levels; measures of exposure; and measures of environmental outcomes (i.e.,
measures of adverse effects or condition) that report on changes in the state of the systems at risk
(e.g. hectares of wetland restored or the number of cancer cases avoided). Decreases in adverse
effects or improvements in health or environmental condition are the ultimate basis for evaluating risk
reduction programs, and environmental outcome measures (or early markers of the final outcome)
are therefore preferable. It is often necessary, however, to supplement outcome measures with
shorter term measures (e.g., early markers of the final outcome, as well as process, exposure, and
stressor measures) for purposes of program accountability and course correction.
1 The IED framework includes the use of Environmental Report Cards to
2 document performance and outcomes of risk reduction activities at several levels and
3 for different audiences (see Figure 1-6). As shown in the IED framework, the
4 performance evaluation process contains several important feedbacks associated with
5 the Report Card. One feedback loop is to the Analysis and Decision-making Phase,
6 reporting on how well the selected risk reduction options and strategies are achieving
7 the environmental goals. This feedback loop allows for adaptive management and
8 changes in implementation activities, including the possible need to identify and analyze
9 additional options to further reduce risks. A second important role for Report Card
1-21
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 information is to allow re-examination, as needed, of the initial risk rankings or other
2 aspects of Problem Formulation. As risk reduction options are put into place, for
3 example, particular risks should be reduced, and a reordering of risk rankings may be
4 appropriate. Further, there may be a shift in or redefinition of societal values over time,
5 requiring different sets of environmental goals and, therefore, different environmental
6 decisions.
7
8 In summary, the IED framework emphasizes the need to consider performance
9 information at several points in the decision-making process and to review
10 environmental decisions in light of new scientific understanding, shifts in societal
11 values, changes in stakeholder preferences and available resources, and/or responses
12 of the environment to previous decisions. The topic of performance evaluation and
13 report cards is discussed more fully in Chapter 7.
14
15
16
1-22
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 1.4 Nature of the Framework
2
3 1.4.1 Major Characteristics
4
5 Integrated environmental decisions should exhibit the following characteristics:
6
7 a) Transparency. The IED framework is designed to promote transparency
8 so that interested parties will be able to follow the process and be aware
9 of the information that was considered in reaching decisions.
10 Transparency is enhanced by the use of clearly articulated goals,
11 analytical methods, and criteria; open deliberative processes; and well-
12 documented decisions.
13
14 b) Flexibility. The IED framework can be applied in a flexible manner
15 depending on the specific circumstance; i.e., where appropriate, to permit
16 valid short-cuts, to eliminate unnecessary procedures, and so to expedite
117 the process of decision-making and implementation. Factors such as the
extent and nature of a problem, the amount and kind of information
1 g available, and the information gaps identified will influence the required
20 degree of complexity of approach and level of detail of the analyses.
21
22 c) Dynamic process design. The technical analyses required to implement
23 the framework should not be conducted in an isolated, stepwise manner.
24 For example, during the Problem Formulation Phase, problem scoping
25 and definition and preliminary analysis of options will affect the
26 development of goals, and vice versa. Some iteration is also required
27 between Problem Formulation and Analysis, since preliminary analyses
28 will often point out missing elements in the problem definition or
29 inconsistencies in goals.
30
31 d) Explicit feedback, interaction, and cooperation. The IED process
32 requires cooperation and open, continuing communication between
33 scientists, managers, members of the public, and others involved in the
34 different phases of an IED project. Examples of critical feedback
35 processes include the evaluation of performance against specified
1-23
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 environmental goals, which allows course corrections and may improve
2 future decisions, and two-way communication between policy-making
3 bodies and the public.
4
5 e) The use of information from many sources. The IED framework
6 requires use of concepts and methods originating in many different
7 scientific, technical, and scholarly fields (e.g., physical sciences, public
8 health, environmental engineering, political science, philosophy, and
9 economics), as appropriate for any given case. In addition to science and
10 scientific judgments, inclusion of public values is needed to compare
11 unlike risks, to set acceptable goals, to assess risks to quality of life, to set
12 priorities, and to reach broadly acceptable decisions.
13
14 f) A way of thinking about environmental problems. Finally, integrated
15 environmental decision-making is not just a series of methodologies, but
16 rather is a way of thinking, in a whole and complete way, about any
17 environmental decision-making case in order to maximize the efficient
18 reduction of aggregate risk to populations or ecological systems. The IED
19 framework should help to focus attention on the multiple aspects of a
20 problem, a broad range of factors that may influence the decision, and
21 important evaluation processes that add value after a decision is
22 implemented. The following section provides a brief synopsis of the
23 opportunities for integration provided by the framework.
24
1-24
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
19
20
21
22
23
24
25-
26
27
28
29
30
31
32
33
34
35
Figure 1-7. Types of Integration in the
IED Framework
1.4.2 Types of Integration
The IED framework requires that
information and viewpoints be integrated at
multiple points in the decision-making
process. Six critical types of integration are
summarized in Figure 1-7, and discussed
briefly below.
a) Integrated Risk Assessment
Focusing on a single agent, a single
medium, and a single outcome to assess
risks is not a realistic representation of the
way environmental exposures impact on
humans and ecosystems. While such an
over-simplified view was arguably
necessary in the early days of risk
assessment, we should now work to
develop tools that can determine risks from
multiple exposures and multiple outcomes
in order to more accurately represent real
world situations. As already noted, some
early, but significant, progress has been
made in this area and the SAB encourages
the Agency to continue these efforts. In
response to the Food Quality Protection Act
of 1996, for example, the Agency has
stepped up efforts to consider risks posed
by groups of pesticides exhibiting common
modes of action, rather than setting risk levels based on each pesticide individually. In-
depth exploration of the analytical challenges inherent in integrated or cumulative risk
assessment is beyond the scope of this study, but it is an area that will require
extensive thought and methodologic development.
Integrated Risk Assessment:
developing scientific data and analytical methods
for determining risks from multiple exposures and
multiple outcomes in order to more accurately
represent real world situations.
Risk Comparisons:
considenng a wide range of environmental risks
simultaneously so that the seriousness of risks can
be characterized relative to one another.
Integrated Analysis of Management
Options:
investigation of options to reduce subsets of ranked
risks, rather than considenng single risks in
isolation, to achieve greater aggregate nsk
reduction.
Integrated Analysis of Economic
Consequences:
identifying the full range of benefits and costs, both
monetized and non-monetized, associated with
reduction of multiple risks.
Integration of Performance
Information:
using performance evaluation measures to devise
course-corrections.
Integrating Multiple Disciplines and
Points of View
understanding and utilizing information from all
concerned parties in the IED process.
1-25
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 b) Risk Comparisons
2
3 The second type of integration important for integrated decision-making involves
4 the consideration of a wide range of environmental risks simultaneously so that the
5 seriousness of risks can be characterized relative to one another. Chapters 2 and 3
6 describe approaches to technical risk comparisons based on the best scientific
7 information and expert judgment. The methods use stressors as a common point of
8 departure to facilitate an understanding of the interconnections between risks to
9 humans and risks to natural systems. Environmental concerns may also be expressed
10 in terms of sources, pathways, or endpoints; these entities are interrelated so that no
11 matter which entity is used as a basis of comparison, scientific information on stressors,
12 sources, pathways, and endpoints will be required.
13
14 An important aspect of the IED framework, however, is the explicit notion that
15 technical risk comparisons are not in themselves sufficient to inform an environmental
16 decision or to set environmental priorities per se. Other factors that should be
17 considered in setting risk priorities include the availability of management options,
18 opportunities for overall aggregate risk reduction, economic impacts, and public
19 concerns.
20
21 c) Integrated Analysis of Management Options
22
23 The third type of integration in the IED framework, discussed in Section 1.3.3,
24 occurs during the analysis and selection of risk reduction options. In this process, risk
26 rankings are analyzed with regard to opportunities for risk reduction (i.e., options
26 available to address the risks) in order to determine the best approach to reduce
27 multiple risks simultaneously in the most cost-effective manner. Clustering of risks is
28 useful both during Problem Formulation (to identify a problem set that will maximize
29 cost-effective risk reduction) and Analysis (to identify the full range of risk reduction
30 benefits associated with any set of control options). A risk reduction program designed
31 to address sets of priority risks will likely differ from a program designed to address a
32 single risk, and should include consideration of the effects (positive or negative) of
33 management options on "non-target" risks.
34
35 The SAB recognizes that the Agency often faces a mandate (legislative or court-
1-26
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 ordered) to reduce the risks from certain stressors to no more than some maximum
2 level. In these situations, the Agency can use the IED framework to develop risk
3 reduction options that cost-effectively reduce "collateral11 risks from other stressors while
4 meeting the mandated risk level for a particular stressor.
5
6 The SAB also recognizes that this type of integrated options analysis is a more
7 difficult undertaking than the traditional approach and that such an exercise would
8 require many estimations regarding combined effectiveness of management options,
9 each subject to uncertainties. However, as the Agency gains experience with and
10 develops new tools for such integrated analyses, the quality of the decisions will
11 improve, and the improvement of environmental conditions should become apparent.
12
13 Chapter 6 describes a process for identifying, screening, and selecting risk
14 reduction options for environmental problems defined in terms of single or multiple
15 stressors.
16
17 d) Integrated Analysis of Economic Consequences
19 Approaching environmental problems from the standpoint of integrated risks
20 should improve our ability to identify the full range of benefits and costs, both monetized
21 and non-monetized. In theory, economic analysis is an inherently integrated exercise in
22 which all benefits, values, or gains are compared to the costs or losses from any
23 environmental intervention. Many benefit/cost analyses in the past, however, have
24 focused on the control of a single source of a chemical. This narrow focus is largely the
25 result of the way in which the "problem" has been defined. When the approach to risk
26 is broadened to examine the multiple effects of multiple chemicals stemming from a
27 source and the reduction in multiple risks that will ensue from controlling those
28 chemicals, the scope of the benefit/cost analysis will be broadened to include the
29 multiple benefits from the reductions of multiple risks. Chapter 4 describes the basic
30 principles that underlie benefit/cost analysis and Chapter 5 discusses issues
31 surrounding the assessment of difficult-to-quantify benefits.
32
33 e) Integration of Performance Information
34
35 Many decision-making schemes make provision for later evaluations of the
1-27
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 decisions that have been made. In practice, however, such after-the-fact reviews too
2 often do not occur. Once a decision has been made, there is a natural tendency to
3 move on to new problems, rather than to re-visit old ones. And yet, integration of
4 information on performance and environmental results is a critical feature of any
5 approach that includes principles of adaptive management. Today's environmental
6 problems are often so complex that, no matter how sophisticated the analysis leading to
7 an initial judgment, it is from reviewing the effects of and making adjustments to an
8 earlier decision that a risk manager becomes better equipped to make future decisions.
9
10 Performance evaluation and feedback is an integral part of the IED framework;
11 this component of the framework is summarized in section 1.3.4 and described in more
12 detail in Chapter 7.
13
14 f) Integrating Multiple Disciplines and Points of View
15
16 Each of the five types of integration described above depends upon, to varying
17 degrees, effective interactions between scientists, decision-makers, and interested and
18 affected parties; i.e., the integration of information and understanding of all concerned
19 parties in the IED process. As noted above, the IED framework requires not only the
20 use of information and perspectives from a number of scientific disciplines, but also the
21 inclusion of public values throughout the decision-making process. In practice, this
22 means that, although the in-depth technical tasks are undertaken by experts in discrete,
23 analytic exercises, the experts periodically will be informed and their results reviewed by
24 other IED participants. Although much of the emphasis in this report is on inclusion of
25* perspectives from outside the Agency, coordination and communication within the
26 Agency during the course of the various analyses will be critical as well. In this way,
27 through a series of deliberative and analytic processes (NRC, 1996), all of the
28 participants gain an understanding of the value-based concerns of the others. In fact,
29 the IED calls for recognition, inclusion, and consideration of values throughout the
30 process, it is only through explicit inclusion and integration of the values and
31 perspectives of a diverse set of participants, from within and outside the Agency, that
32 the most acceptable and effective environmental risk reduction will occur. The
33 problems and promises of this important type of integration are discussed in Chapter 5.
34
35
1-28
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 1.4.3 Building on Previous Frameworks
2
3 From the previous section, the reader can see that the goal of IED is to identify
4 appropriate risk reduction actions to address a mixture of environmental problems (both
5 human health and ecological), so as to reduce total aggregate risk, rather than focusing
6 only on single risks in isolation. Deliberative processes are one means of deciding
7 which mix of risk reduction goals should be pursued. The IED framework explicitly
8 includes a mechanism for evaluation and feedback, so that a strategy of adaptive
9 management can be easily employed. In addition, the framework includes substantive
10 interaction with interested and affected parties throughout the process so that public
11 values are reflected appropriately.
12
13 The IED framework proposed by the SAB builds upon several previous
14 frameworks, in particular the risk assessment/risk management model described by the
15 National Research Council (NRC 1983), the ecological risk assessment framework
16 (U.S. EPA 1992), and the risk characterization process described in NRC (1996), which
17 focused on the interaction between analytic and deliberative processes in decision-
making. The IED framework moves beyond these earlier efforts in three significant
19 areas:
20
21 a) Evaluation of Single Stressors
22
23 Although the SAB emphasizes that the Agency should consider multiple
24 stressors in an integrated approach to risk, it recognizes that the IED framework also
25 should enhance the decision-making process for single stressors, of the type historically
26 considered by the Agency (e.g., a drinking water pollutant). When a single stressor is
27 considered, the IED framework should expand the previous approaches by:
28
29 (1) Characterizing stress-effects relationships across all systems and
30 populations;
31 (2) Exploring a broader range of risk reduction options, some of which may be
32 qualitatively new;
33 (3) Assessing the benefits and costs of each option, including explicit
34 consideration of non-monetary benefits and costs;
35 (4) Assessing the magnitude and nature of the aggregate risk reduction
1-29
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 associated with each option;
2 (5) Involving scientists, options analysts, stakeholders, and risk managers
3 collectively at various points throughout the process; and
4 (6) Establishing a performance evaluation "report card" to characterize the
5 efficacy of the implemented risk reduction option and to signal both the need and
6 opportunity for adapting the original management decision.
7
8 b) Evaluation of Multiple Stressors
9
10 One of the most important extensions in the IED framework is that, unlike
11 previous frameworks, a primary goal of the IED process is to consider multiple stressors
12 when formulating the problem and developing possible solutions. The IED provides a
13 structured way to begin to explore multiple ecological risks, multiple human health risks,
14 and/or multiple quality of life concerns. In some cases this approach will lead to
15 consideration of combinations or groups of risks; e.g., all organophosphate pesticides,
16 all automobile emissions, or all factors leading to local loss of biodiversity.
17
18 Initial steps in the IED process are designed to produce relative rankings of risks
19 to human health, quality of life, and ecosystems, independently. While the SAB had
20 originally planned to integrate the rankings for the various types of risks, the members
21 concluded that such a merged ranking is based on values, rather than on science
22 alone. The SAB's review of risks of concern also emphasizes that health and
23 ecological risk assessments are often qualitatively different. For example, the focus of
24 the assessments is different; that is, the focus of health risk assessment is an individual
26 within a single species, while the focus of even a narrowly drawn ecological risk
26 assessment is entire populations of any/many species. More generally, ecological risk
27 assessments often address the integrated risks to a prescribed region, such as a
28 watershed. This difference is reflected in the different types of stressors of concern; cf.,
29 carcinogens for humans vs. habitat fragmentation. However, in those instances where
30 there are common stressors of concern (e.g., chlorinated pesticides, climate change) or
31 where effects of a stressor on an ecological system produce effects on human health or
32 quality of life (e.g., habitat alteration that affects the range and activity of disease
33 vectors and infective parasites, or changes in the abundance of commercially important
34 or endangered species), there is an opportunity for some merging of concerns to take
35 place. An understanding of the relationships among different types of risks/stressors is
1-30
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 critical to Problem Formulation, both in terms of setting environmental priorities and
2 goals, and for designing subsequent risk and risk reduction analyses.
3
4 In addition to ranking of risks, which allows scientists to compare the expected
5 impacts of each risk, the IED framework suggests combining risks into logical
6 groupings, e.g., those with a common source or pathway, in order to identify risk
7 reduction opportunities across stressors. In order to be successful, this analysis
8 requires open, publicly-accessible, and frequent dialogue among those who assess and
9 compare risks, those who determine methods for reducing risks, and those who make
10 the final decisions.
11
12 In summary, as the IED framework is implemented to address multiple stressors,
13 it should:
14
15 (1) Lead to a more realistic ranking of risks to humans and to ecosystems ,
16 where some of those risks may be posed by combinations of related
17 stressors;
(2) Lead the Agency to consider in a systematic fashion all of the appropriate
19 factors related to risks in a given circumstance, including aggregate risk,
20 economic factors, and societal values; and
21 (3) Lead to action that will increase the reduction in aggregate risk posed by a
22 combination of stressors in a given circumstance.
23
24 As the Agency gains experience and develops more sophisticated analytical
25 tools, the benefits of IED should become more apparent. At first, it may be easiest
26 simply to examine the total reductions in exposures. More complete evaluations of
27 outcomes are complicated by the need to estimate reductions in several types of risk
28 (e.g., health, ecological, and quality of life risks) resulting from the reductions of multiple
29 stressors affecting multiple receptors. Nonetheless, appropriate methods should be
30 developed to measure the total impact and benefits of risk reduction programs in the
31 future.
32
33 c) Considering Environmental Values
34
35 The third extension that the IED framework provides is a more explicit
1-31
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 consideration of the analytic/deliberative process described by the MAS Panel on Risk
2 Characterization (NRC, 1996). It is through such a process that societal values
3 intersect with the scientific risk characterization and risk reduction analyses. The
4 proposed I ED framework emphasizes the role and timing of stakeholder and decision-
5 maker inputs to the analytic processes. It explores more deeply the valuation of
6 environmental outcomes and risks and the need to include not only the concepts of
7 economic efficiency and wiHingness-to-pay, but also issues of environmental
8 sustainability and equity. The IED framework recognizes that societal values constitute
9 the milieu in which integrated environmental decision-making occurs forming the basis
10 for societal goals for improved social welfare, improved ecological/health conditions,
11 and long-term sustainability and equity. It is in the realm of social values that the
12 success or failure of environmental decisions will be judged.
13
14
1-32
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 1.5 Benefits and Challenges of the Framework
2
3 1.5.1 Benefits
4
5 A fully implemented IED framework should result in a new level of environmental
6 protection. The IED framework refines and moves beyond the risk ranking exercises
7 contained in the SAB's Reducing Risk report. The techniques described in this report
8 for ranking health and ecological risks offer a clear and systematic approach for using
9 the best scientific information and expert judgment and are substantially more
10 transparent than were those in the earlier effort. In addition, the IED framework goes
11 beyond the comparison of risks and risk reduction options to spotlight the as-yet
12 unrealized benefits to be gained from addressing environmental problems in a truly
13 integrated fashion - including integration of technical information on risk (including
14 information on multiple stressors, sources, and effects), risk reduction options, and
15 economic consequences; integration of values and goals into the decision-making
16 process; and integration of performance information so that needed course corrections
17 can be identified. While the Agency has made some significant first steps in some of
these areas, the full benefits are yet to be achieved.
19
20 The expected benefits of the IED framework relate directly to the characteristics
21 of the framework described in this chapter. Transparency of the decision-making
22 process should increase acceptance of the outcome by affected parties, flexibility
23 allows the process to be adapted to suit the needs of a particular situation, open and
24 inclusive dialogue should lead to broader consideration of perspectives and options,
25* and so forth. Broadly speaking, then, implementation of the IED framework should
26 result in:
27
28 a) Enhanced ability to improve human health and quality of life and to protect
29 the integrity of ecological systems;
30
31 b) Improved targeting of resources for risk reduction; and
32
33 c) Greater accountability for results;
34
35
1-33
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 1.5.2 Challenges
2
3 At the same time, the IED framework presents the Agency with significant
4 challenges. The techniques recommended for ranking risks in the IED will need further
5 refinement during their development and application. Likewise, the concept of
6 combining risks into related subsets for risk management purposes offers considerable
7 promise but introduces new uncertainties that must be acknowledged and addressed.
8 The integration of economic concepts, both for setting environmental goals and for
9 evaluating how best to meet those goals, is important but may also be controversial
10 and, therefore, must be introduced clearly and objectively to avoid unwarranted
11 criticism. The consideration of management options from the total aggregate risk vs.
12 single risk perspective, which is a major feature of the framework, could also be
13 controversial, even in the face of its technical soundness. Finally, the concept of
14 evaluating the impact of decisions and reflecting that information in subsequent
15 management actions (i.e., adaptive management) is not new; however, the challenge of
16 using this approach in a consistent and continuing way remains.
17
18 In addition to the technical and methodological issues, it is important to
19 recognize that successful implementation of the IED framework will require some
20 adjustments in the manner and degree to which the many participants interact over the
21 course of the decision-making process. Integrated environmental decision-making
22 requires the sharing of information, ideas, approaches, and management deliberations
23 to a degree now seldom practiced among individuals of very different backgrounds.
24 Although this sharing is a positive aspect of the IED framework, it may require
25' significant adaptation on the part of individual policy-makers and institutions. For
26 example, decision-makers will need to interact more extensively with scientific and
27 technical analysts and the public in the course of developing integrated approaches to
28 environmental risks. Likewise, scientific and technical experts will need to recognize
29 the role (and limitations) of science in decision-making, and also to recognize the
30 legitimate role of values in establishing environmental goals and selecting management
31 approaches. This culture change will be assisted by familiarity and experience with the
32 IED process. Experience, combined with discipline, will also be needed to apply the
33 IED framework with discretion to the depth of detail necessary for a particular problem,
34 and no more; i.e., the IED framework should not become a barrier to decision-making.
35
1-34
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Additional challenges likely to be encountered when implementing the IED
2 framework include:
3
4 a) problems of understanding arising from differences in terminology and
5 outlook imbedded in the different disciplines and backgrounds of the
6 participants in the process;
7
8 b) difficulties of using both qualitative and quantitative measures
9 concurrently in the decision process;
10
11 c) the need to compare different types of risks (e.g., health, ecological and
12 quality of life risks) within a common decision framework and to discern
13 and define the inter-relationships among risks so as to define common
14 goals across the different risk types; and
15
16 d) time-lags between implementation of risk reduction plans and the
17 detection of results and effects, which make the selection of appropriate
performance measures particularly important.
19
20 Facing and surmounting the challenges, learning how to use the IED framework
21 flexibly with discretion and care, and improving the underlying methodologies over time,
22 will lead to more effective environmental decision-making and to measurable
23 environmental progress. The SAB expects that the proposed framework and the
24 recommended improvements to the component analyses will assist the Agency to
25 develop a more comprehensive, integrated, and transparent process and culture for
26 setting environmental priorities and making and implementing decisions.
1-35
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 1.6 References Cited
2
3 National Academy of Public Administration. 1995. Setting Priorities, Getting Results:
4 A New Direction for EPA.
5
6 National Research Council. 1983. Risk Assessment in the Federal Government:
7 Managing the Process. National Academy Press, Washington, DC.
8
9 National Research Council. 1994. Science and Judgment in Risk Assessment.
10 National Academy Press, Washington, DC.
11
12 National Research Council. 1996. Stem, P.C. and H.V. Fineberg (ed.s).
13 Understanding Risk: Informing Decisions in a Democratic Society. National
14 Academy Press, Washington, DC.
15
16 Presidential/Congressional Commission on Risk Assessment and Risk Management.
17 1997. Risk Assessment and Risk Management in Regulatory Decision-Making.
18
19 Science Advisory Board. 1990. Reducing Risk: Setting Priorities and Strategies for
20 Environmental Protection (EPA-SAB-EC-90-021).
21
22 Science Advisory Board. 1995. Beyond the Horizon: Using Foresight to Protect the
23 Environmental Future (EPA-SAB-EC-95-007).
24
25 Science Advisory Board. 1997. Advisory on the Problem Formulation Phase of EPA's
26 Watershed Ecological Risk Assessment Case Studies (EPA-SAB-EPEC-ADV-
27 97-001).
28
29 U.S. Environmental Protection Agency. 1987. Unfinished Business: A Comparative
30 Assessment of Environmental Problems. Washington, DC.
31
32 U.S. Environmental Protection Agency. 1982. Framework for Ecological Risk
33 Assessment (EPA/630/R-92/001). Office of Research and Development,
34 Washington, DC.
35
1-36
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 U.S. Environmental Protection Agency. 1989. Risk Assessment Guidance for
2 Superfund (EPA/
3
4 U.S. Environmental Protection Agency. 1990. Methodology for Assessing Health Risks
5 Associated with Indirect Exposure to Combustor Emissions-Interim Final.
6
7 U.S. Environmental Protection Agency. 1993a. Draft Addendum to the Indirect
8 Exposure Document.
9
10 U.S. Environmental Protection Agency. 1993b. A Guidebook to Comparing Risks and
11 Setting Environmental Priorities (EPA 230-B-93-003). Office of Policy, Planning,
12 and Evaluation, Washington, DC.
13
14 U.S. Environmental Protection Agency. 1997. Guidance on Cumulative Risk
15 Assessment, Part 1: Planning and Scoping. July 3, 1997.
16
17 U.S. Environmental Protection Agency. 1998. Guidelines for Ecological Risk
Assessment (EPA/630/R-95/002Fa).
19
20
1-37
-------
PART II INPUTS TO ENVIRONMENTAL DECISION-MAKING:
RISK COMPARISONS
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
PART II — INPUTS TO ENVIRONMENTAL DECISION-MAKING: RISK
COMPARISONS
Preface
1
2 The process of comparing environmental health and ecological risks is a key
3 component of the Problem Formulation Phase of the IED framework. In the early
4 stages of problem formulation, a draft list of stressors of concern will be identified using
5 existing knowledge and expertise of a group of scientists (including human exposure
6 and health experts and/or ecological experts) and/or on the basis of a review of
7 available scientific literature. A comprehensive list of possible stressors is developed to
8 provide a complete set of possible comparisons. The draft list developed by risk
9 scientists would then be discussed with others in the IED process, including the options
10 analysts, decision-makers, and stakeholders. As a result of these discussions,
additional stressors might then be added or removed, some groups of related stressors
might be aggregated into a single composite stressor (e.g., "waterbome infectious
13 microbes" vs. "Cryptosporidiurrf), and some stressors might be disaggregated into
u more specific stressors (e.g., "mercury" vs. "heavy metals") for subsequent analysis.
15 Quantitative and qualitative information on the stressors would be assembled into a
16 format, such as a stressor risk data sheet, to facilitate comparisons. Important
17 information would include: stressor intensity; the adverse effects likely to be associated
18 with exposure to the stressor; observed effects in the population or ecosystem; and
19 some assessment of the causal relationships between the environmental stressor and
20 the observed effects. This latter could include assessments of the probable
21 contribution of the stressor to the observed effects of concern, the so-called attributable
22 risk. Major gaps in knowledge should also be identified as part of Problem Formulation
23 and Analysis in order to inform research planning and help set research priorities.
24
25 During the design of the IED framework, the SAB participants acknowledged that
26 technical risk rankings, in isolation, offered insufficient guidance for policy decisions.
27 Given the multitude of problems and issues to be addressed, a more comprehensive
28 and systematic framework for analyzing and reducing environmental health, ecological,
29 and quality of life risks appeared necessary. Analysis and comparison of various
n-i
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 environmental health and ecosystem risks, however, remain the fundamental basis of
2 such a framework, and the foundation for defining risk reduction priorities.
3
4 As symbolized by the interlocking circles in the IED framework, the assessment
5 and comparison of risks and the definition of environmental goals are inter-related. In
6 other words, scientific information on the nature and extent of various risks influences
7 the relative priority that society places on those risks. In addition, however, non-
8 scientific issues such as dread, previous experiences, and degree to which exposure is
9 voluntary, for example, also influence the relative priority assigned by the public to
10 environmental risks. For this reason, the ultimate risk priorities that emerge during
11 Problem Formulation will be a product of the interaction between the risk comparison
12 and goal setting processes, and (as described in Part IV) the preliminary assessment of
13 risk reduction potential.
14
15 As the IED process proceeds from the problem formulation to the analysis
16 phase, the risk assessors develop more detailed risk analyses for the priority stressors,
17 which can be used to help focus the development of options. The stressor
18 comparisons will be used to help organize thinking about potential options for reducing
19 risks and to examine opportunities for reducing several risks through a single risk
20 reduction option. The more detailed information on exposures and risks also provides
21 the basis for estimating the risk reduction achievable through various options, and for
22 determining the specific environmental outcomes that will be used to evaluate
23 performance against the risk reduction goals.
24
25' During Problem Formulation, it is important to identify linkages between health,
26 ecological, and quality of life risks and to seek opportunities to address all three types
27 of risks through a risk reduction program. For this reason, it is helpful to define the risk
28 problems using a common dimension; the SAB Subcommittees chose to define risks in
29 terms of stressors.
30
31 Stressors of concern may be chemical (e.g., ozone, heavy metals), physical
32 (e.g., radiation, changes in land use), or biological (e.g., infectious agents in drinking
33 water, introduced exotic species). Further, the stressors may cause adverse health and
34 ecological effects directly (through exposures via environmental media) or may act
35 indirectly. For example, freons released into the atmosphere do not directly influence
n-ii
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 health, but, instead, reduce the thickness of the protective ozone layer of the
2 atmosphere, consequently increasing the risks of skin cancer. A methodology for
3 comparing environmental risks to quality of life was not developed by the SAB because
4 the selection of QOL ranking criteria is more appropriately done by a broad group of
5 public or stakeholder representatives rather than by a technical panel. However, the
6 SAB recognizes the need to develop and incorporate such a process into integrated
7 environmental decision-making.
8
g Two Subcommittees of the IRP developed risk comparison approaches whose
10 results could be used in the IED process. One approach, which was applied to the
11 comparison of ecological risks, utilizes an expert group to develop and weight risk
12 ranking factors in order to produce the group's consensus judgment of the relative risks
13 associated with various stressors. The second approach, which is discussed using
u human health risk issues, involves polling experts individually for their professional
15 judgment of relative degree of risk associated with various stressors, while soliciting
16 information on which factors most influenced this judgment and the degree of
confidence each expert assigns to his/her assessments. The latter method provides
information not only on the "average" or "median" risk ranking, but also on the range of
19 expert opinions. The consensus approach or the individual polling approach could be
20 applied to either human health or ecological risk comparisons using stressors and rating
21 factors specific to each group. The methodologies also could be used to elicit
22 stakeholder or public views on risk priorities by expanding the composition of the
23 surveyed or empaneled group.
24
25 The Ecological Risk Subcommittee (ERS), as a group of experts, developed a
26 set of decision rules for ranking ecological risks. The Subcommittee identified a set of
27 stressors of ecological importance and established a systematic template for
28 considering each stressor with respect to issues of ecological importance (e.g., spatial
29 extent and duration of the stress, intensity of the ecological effect, recoverability once
so the stressor is removed, and other issues of special ecological significance). The group
31 then developed a process to characterize the co-occurrence of stressor intensity and
32 ecological consequences for an at-risk ecosystem. For each identified ecological factor
33 of importance, a multiplicative value was assigned that would adjust the risk level up or
34 down for each individual stressor. Once all the factors were considered, the initial risk
35 assignment was adjusted by each factor, leading to a summary risk value for each
n-iii
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 stressor. Stressors were then placed in narrative risk categories on the basis of the
2 numerical risk scores, which derived directly from the application of the ecologically
3 based criteria.
4
5 An important advantage of the ERS method is the transparency of the factors
6 and their assigned relative weights used for deriving the risk rankings. The dialogue
7 among the convened experts helps to ensure that there is a common understanding of
8 terminology and of the specific criteria for assigning ranking factors, a process that is
9 very difficult to accomplish when polling individuals. A potential disadvantage of this
10 approach is that the quantitative scores calculated for the risk rankings could be
11 misused by attributing greater precision to the numbers than the methodology warrants.
12
13 The Human Exposure and Health Subcommittee (HEHS) adopted a process
14 based on the use of the World Wide Web to solicit risk comparisons from a large
15 number of experts in fields related to human health risks. One reason for the choice of
16 such a methodology was the extreme breadth of disciplines pertaining to environmental
17 health, e.g., epidemiology, exposure analysis, and toxicology. The respondents are
18 provided with a list of stressors, baseline information regarding these stressors and
19 their associated outcomes, and a consistent template to provide their comparisons of
20 risks to human health attributable to exposures to the stressors. As a part of the
21 survey, a set of about ten factors that might relate to the risk comparison is provided,
22 and the respondent is asked to indicate those influencing her/his rating for each
23 stressor.
24
25' An important advantage of the HEHS methodology is that it provides information
26 on the variability in expert opinion about the relative degree of risk associated with
27 various stressors, as well as on the confidence of the experts in their risk comparisons.
28 A disadvantage of the methodology is that the opportunity for interaction and
29 information-sharing among the expert participants is lost, and the effort required to
30 prepare a common set of information for all risk areas to be compared is substantial.
31
32
33
n-iv
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
•4
J
16
17
18
19
20
21
2?
23
24
25
26
27
28
29
30
31
32
Table PII-1. Correspondence of Human Health and Ecological Risk Comparison
Factors
Human Health Risk Factors
Size of population affected
Particular subpopulations at risk
Severity and persistence of health effects
Persistence in the environment and/or
human body
Percent of attributable incidence
Potential future risk
Ecological Risk Factors
Proportion of resource at risk
Distribution of "hot spots"
Recovery potential,
Species depletion
Duration of stress-effects co-occurrence
Secondary stress induction
Special ecological significance
Despite the apparent differences in the ERS and HEHS approaches to risk
comparisons, the factors influencing the relative ranking of risks share many
commonalities (Table PII-1). The primary difference in the two approaches lies in the
manner in which the decision rules are derived. The relationship between the
methodologies is illustrated in Figure PII-1. Following from left to right in this figure, the
initial comprehensive list of stressors is identified and information concerning stressor-
effect relationships is collected. This information is provided to a group of experts,
either constituted as an expert panel (as in the ERS approach, depicted by the
feedback arrow from the Risk Comparison box to the Decision Rules box) or as
individual experts polled separately (as in the HEHS approach). In the HEHS
methodology, the risk comparisons are derived directly from the experts. The experts
are also asked to indicate the factors that influenced their risk ratings for each stressor.
A multivariate regression analysis could be used to relate the risk ratings to the various
factors that the experts indicate influenced those ratings. The factor weights from the
regression analysis then define the Decision Rules. In the ERS approach, the experts
develop the Decision Rules directly, then apply them to the stressor database to derive
the Risk Comparisons.
n-v
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
Consistent Framework for Risk Comparisons
1 One clear advantage of illustrating the two approaches in the context of a unified
2 framework is the provision for verification of the ratings generated using either
3 approach. In the case of the
4 ecological risk comparisons, a
5 larger group of individual experts
6 could be polled explicitly with
7 respect to the ecological risk
8 adjustment factors, and a
9 regression analysis on that
10 database could be compared with
11 the Decision Rules developed by
12 the expert panel. Consistency of
13 the weighting factors would
14 provide an increased confidence
15 in the overall risk comparisons.
16 Correspondingly, following the
17 HEHS process, a subsequent
18 stage might be to convene a panel
19 of human health experts with Rgure Pll-1.
20 sufficient diversity of perspectives
21 and assign them the task of weighing the information generated by the surveyed
22 experts then formulating and applying explicit decision rules to arrive at a list of health
23 risk ratings.
24
25 The two subcommittees also considered the possibility of merging human and
26 ecological risks into a single science-based ranking, but realized that the comparison of
27 health risks to ecological risks was not possible on an objective scientific basis because
28 it is fundamentally based upon considerations of societal values. Nevertheless, the
29 use of analogous methods for developing technical risk comparisons for both human
30 health risks and ecological risks, as described in the following chapters, should facilitate
31 the subsequent setting of societal values-based risk priorities during Problem
32 Formulation.
33
34 An additional contribution of the risk comparison methods proposed by the two
35 subcommittees lies in their transparent use of decision rules for comparing risks. The
n-vi
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 explicit statement of decision rules, and the relative weights or importance assigned by
2 the participants to various considerations, makes clear what assumptions, scientific
3 judgments, and values went into the final results. In this way, users of the approaches
4 can see how the results would change if the judgments about decision rules were to
5 change.
6
7 Chapters 2 and 3 describe in detail the approaches developed by the ERS and
8 HEHS, respectively, for comparing multiple environmental risks.
9
H-vii
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
CHAPTER 2: ECOLOGICAL RISKS
TABLE OF CONTENTS
2.1 Background 2-1
2.2 Objectives and Approach 2-3
2.3 EPS Ecological Risk Ranking Methodology 2-4
2.3.1 Overview and Rationale 2-4
2.3.2 Ecological Risk Characterization 2-5
2.3.2.1 Selection and Aggregation of Stressors and Ecosystems .... 2-5
2.3.2.2 Development of Ecological Risk Profiles 2-8
2.3.3 Development of Relative Ranking of Ecological Risks 2-14
2.3.3.1 Multiplicative Factors Used to Assign Stressor Risk Values .. 2-14
2.3.3.2 Sample Calculation for the Stressor Pesticides 2-21
2.3.4 Sources of Uncertainty 2-21
2.4 National-Scale Ecological Risk Ranking 2-23
2.4.1 Results of the ERS National-Scale Ecological Risk Ranking 2-23
2.4.2 Synthesis and Conclusions 2-28
2.5 An Effects-Backwards Methodology for Risk Rankings 2-32
2.6 References Cited 2-36
Appendix 2A: Ecological Risk Profiles 2-38
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 CHAPTER 2. ECOLOGICAL RISKS
2
3
4 2.1 Background
5
6 During the 1980s, the U.S. Environmental Protection Agency (EPA) initiated
7 extramural research to develop new approaches to ecotoxicology and to conducting
8 ecological risk assessments (Bamthouse and Suter, 1986; Levin et al., 1989; Harwell
9 and Harwell, 1989; Kelly and Harwell, 1990). The initial development of a methodology
10 to compare and rank ecological risks began in 1986 with the cross-agency Unfinished
1 1 Business Project (EPA 1987a, b). As part of that project, EPA convened an ecological
12 risk panel to develop and apply a methodology to rank the risks to ecological systems of
13 31 listed "environmental problem areas." The methodology and results of the
14 ecological panel were reported in Harwell and Kelly (1986) and in EPA (1987b). The
15 ecological panel decided that the list of environmental problem areas that had been
developed by EPA was influenced more by programmatic considerations than by what
an ecological system might experience. Consequently, the panel developed a
18 comprehensive list of environmental stressors, defined as those physical, chemical, or
1 9 biological changes that might affect an ecological system. Many of the stressors were
20 related directly to the list of environmental problems, but other stressors were identified
21 that were not captured by the list of environmental problems. The ecological risk panel
22 also developed a list of 16 ecological system types, defined so that differences in
23 exposures and/or responses to the stressors could be identified.
24
25 In 1 989, as part of the EPA Science Advisory Board's Reducing Risk Project, the
26 SAB established a subcommittee on ecological and welfare risks, the latter addressing
27 environmentally mediated factors affecting the quality of human life. That SAB
28 ecological subcommittee adopted and further refined the methodology of the previous
29 Unfinished Business panel and developed an updated set of relative ecological risk
30 rankings for the same set of environmental stressors (SAB, 1 990a). The summary of
31 the SAB Reducing Risk project and reports of the other subcommittees are contained in
32 the report entitled, Reducing Risk: Setting Priorities and Strategies for Environmental
33 Protection (SAB, 1990b).
2-1
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 In Reducing Risk, the SAB recommended that EPA enhance the consideration of
2 ecological risks to a level comparable to human health risks and develop guidance for
3 implementing ecological risk assessments. In response, the Agency established a
4 program in the Risk Assessment Forum to develop a new framework for conducting
5 ecological risk assessments. Two major advances were necessary: a) while chemical
6 ecological stressors are analogous to the chemicals considered in human health risk
7 assessments, the domain of ecological risk assessment was broadened also to include
8 biological and physical insults that may cause significant ecological effects; and b) while
9 human health risk assessments address only one species and often only a single
10 endpoint (human cancers), assessment of ecological risks almost always requires
11 considering multiple endpoints because of pervasive differences across species,
12 ecological systems, stressors, and organizational hierarchy (i.e., from species to
13 landscape and larger scales of ecosystems). Consequently, based on two
14 EPA-sponsored workshops (Harwell and Gentile, 1992; Fava et al., 1992), EPA issued
15 a new ecological risk assessment framework (EPA, 1992) which expanded on the NRC
16 (1983) Red Book risk assessment paradigm to include the full suite of natural and
17 anthropogenic stressors (i.e., not just chemicals) that affect the environment at
18 population, community, ecosystem, and landscape levels. The EPA ecological risk
19 assessment framework calls for identification of the at-risk components of ecological
20 systems, selection of ecological endpoints, development of a conceptual model that
21 describes the ecological system and its stress-response relationships, mutual
22 characterization of the stress regime and ecological effects in terms of the selected
23 endpoints, and integration into an overall assessment of ecological risk with explicit
24 accounting of uncertainties. Since issuing the framework, the Agency has issued
25 ' ecological risk assessment guidelines (EPA/630/R-95/002B) that expand on and modify
26 the framework document, and it has issued guidance on cumulative risk assessment
27 (U.S. EPA, 1997) that further builds on the ecological risk framework. Neither the
28 ecological risk assessment framework nor EPA's proposed ecological and
29 cumulative risk assessment guidelines, however, explicitly describe the methodology
30 needed to rank the relative importance of various ecological risks. Developing an
31 improved methodology for the relative ranking of ecological risks is the focus of this
32 chapter.
33
34
2-2
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 2.2 Objectives and Approach
2
3 As part of the SAB Integrated Risk Project (IRP), the Ecological Risks
4 Subcommittee (ERS) was formed to address relative risks to ecological systems. The
5 objectives of the ERS were to:
6
7 a) examine and refine the methodology for assessing relative ecological
8 risks;
9
10 b) develop criteria to rank ecological risks at the national level;
11
12 c) apply those criteria to perform a national-level relative ranking of
13 ecological risks;
14
15 d) develop a methodology for implementing relative ecological risk
16 assessments at regional or local scales; and
e) assist in integrating ecological risk assessments into a broader conceptual
19 framework that includes ecological, human health, and welfare risks.
20
21 To accomplish these objectives, the ERS carefully considered the relative risk
22 methodologies developed in Unfinished Business and Reducing Risk as well as the
23 more recent ecological risk assessment framework. The Subcommittee concluded that
24 a new methodological framework for evaluating relative ecological risks was needed
25 that expands upon the previous methodologies but adds transparency and clarity and is
26 adaptable to multiple scales of analysis. As detailed in the following sections, the new
27 relative ecological risk methodology focuses primarily on the stressor-effects model to
28 assess ecological risks, but it can also be adapted to apportion the relative contributions
29 of various stressors to an observed effect in the environment. As a part of the
30 methodology development process, the ERS updated the list of environmental
31 stressors and ecological systems of concern, and used a matrix approach to capture
32 information about stressor-effect relationships so that both quantitative and qualitative
33 assignment of risk rankings could be accomplished. The ERS ecological risk ranking
34 methodology was applied at the national scale, and the necessary modifications were
""3 identified to apply the methodology to regional, state, or local levels. Close dialogue
2-3
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 with the IRP Human Exposure and Health Subcommittee (HEHS), charged with
2 assessing relative human health risks, has resulted in a consistent framework for both
3 types of environmental risks and allows explicit consideration of the linkages
4 between ecological risks and human health risks.
5
6 2.3 ERS Ecological Risk Ranking Methodology
7
8 2.3.1 Overview and Rationale
9
10 The basic approach proposed by the ERS for ranking the relative significance of
11 ecological risks follows the ecological risk assessment framework (U.S. EPA, 1992),
12 particularly with regard to addressing all types of ecological stressors (physical,
13 chemical, and biological) affecting the important attributes (endpoints) of ecological
14 systems. This approach requires consideration of the two fundamental components of
15 ecological risks for each stressor: a stress or exposure regime, and a response or
16 ecological effects regime. The stressors, exposure patterns, ecological endpoints, and
17 ecological effects are aggregated as appropriate for the scale of the analysis,
18 recognizing that the goal of comparative ecological risk assessments is to characterize
19 the dominant relationships between environmental stressors and ecological effects.
20 For example, in the national-scale comparative risk assessment presented here,
21 ecological endpoints and effects generally were characterized only at the ecosystem
22 and landscape levels, even though ecological effects from a stressor may be observed
23 at the population and community levels as well. Similarly, ecosystem-specific
24 differences in response were generally noted, but the Subcommittee chose to focus on
25' those types of ecological systems considered to be most at risk.
26
27 The first step in the comparative ecological risk assessment is to determine the
28 potential ecological importance of each stressor at the ecosystem level or landscape
29 level for ecological systems at risk. This ecological risk characterization is based on a
30 stress-effects profile, i.e., a graphical representation of a stressor's exposure regime
31 with its ecological effects for each at-risk ecosystem type. The stressor exposure is
32 categorized into high, medium, and low levels, normalized to actual exposures in the
33 real world, and effects distributions associated with each stressor levels are estimated.
34 The ecological effects are assigned high, medium, and low intensity categories, based
35 on criteria of the ecological significance of the effect (e.g., high-level effects reflect
2-4
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 major disruptions to the fundamental structure or processes of an ecosystem). The
2 next step in the methodology is to transform these generic ecosystem- or
3 landscape-level stress-effect relationships into a relative ranking of risk at a specific
4 scale (e.g., regional, national, or global) by applying a series of numerical modification
5 factors that relate to the ecological significance of the effect. The resulting modified
6 numerical value constitutes the risk score for a stressor at the specified scale of
7 interest. The relative risk posed by the stressors can then be determined by comparing
8 scores across the stressors and by grouping these scores into qualitative risk
9 categories (e.g., very high, high, medium, and low relative risks). This categorization
10 reflects the semi-quantitative nature of the procedure and identifies clusters of stressors
11 with similar levels of risks. Thus, the primary import of the aggregated rankings is their
12 relative value, i.e., what stressors have relatively the same level of ecological risk, what
13 set of stressors is much higher in risk than others, and so forth. Details of the risk
14 ranking methodology and its application by the EPS to the national scale are described
15 in the remainder of this chapter.
16
17 2.3.2 Ecological Risk Characterization
19 2.3.2.1 Selection and Aggregation of Stressors and Ecosystems
20
21 The ecological risk ranking process begins by identifying the stressors that
22 potentially pose environmental risks and the set of ecosystem types for consideration of
23 the risks. For this project, lists of stressors identified by the EPS and by the HEHS
24 were merged. Table 2-1 notes those stressors, or groups of stressors, that may affect
25 ecological and/or human health endpoints. Identifying a common set of stressors
26 facilitates direct comparison of risks to humans and ecosystems. For the ecological risk
27 ranking process, the stressors were characterized at the level of detail sufficient to
28 distinguish between different types of stressor-effect relationships, but not so detailed
29 that an unmanageable number of stressors had to be evaluated. For example,
30 pesticides were evaluated as a class of stressors, rather than on a
31 chemical-by-chemical basis. Some general categories of stressors were disaggregated
32 where appropriate; for example, the stressor class of habitat alteration was divided into
33 specific types of habitat alterations (e.g., habitat fragmentation, habitat conversion).
34
35
2-5
-------
Table 2-1: Environmental Stressors of Concern
Stressor Ecological Risk Human Health Risk
I. CHEMICAL STRESSORS
Criteria Air Pollutants
CO NA /
NO, / /
O, / /
SO2 / /
Airborne Particulates • /
Toxic Inorganics
Hg / /
Other Heavy Metals / (/)
Other Toxic Inorganics (As, Se, B) / (/)
Asbestos NA /
Persistent Toxic Organics / (/)
Endocnne Disrupters / /
Pesticides, Herbicides / /
Environmental Tobacco Smoke NA /
Volatile Organics NA (/)
Toxic Anions
Nitrate, nitrite NA /
Acid Deposition / NA
Nutnents s NA
Radionuclides
Radon NA /
Others (e.g., Cs, Sr) / /
Oil/Fuel Leaks, Spills, and Use / /•
Dissolved Oxygen/BOD / NA
Acid Mine Drainage / NA
Contaminated Ground Water / NA
II. BIOLOGICAL STRESSORS
Bioaerosols/Allergens NA (/)
Human Disease Agents
Airborne Viruses, Infectious Agents NA /
Waterbome Infectious Microbes NA /
Non-human Disease/Pest Outbreaks / NA
Introduced Species / /
Genetically Engineered Organisms / /
III. PHYSICAL STRESSORS
Climate Change (global, regional, and micro-climate) / /
Noise / /
Light Pollution / /
Habitat Alteration ?
Hydrologic Alteration /
Habitat Fragmentation /
Habitat Conversion /
Physical Habitat Disruption /
Turbidity and Sedimentation /
Altered Fire Regime /
Altered Salinity Regime /
Thermal Pollution / NA
Harvesting Living Resources ?
Coastal /
Freshwater s
UV-B / /
EMF NA /
2-6
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Several factors have to be considered in selecting the appropriate level of
2 aggregation or disaggregation of stressors, including the exposure and effects data that
3 are available, differential impacts on the ecological system types of interest, potential
4 management options, and so forth. There are no simple rules for this classification
5 process. For example, the Subcommittee examined alternative ways to aggregate
6 chemicals, ranging from classifications based upon mode-of-action of toxicity to
7 structural characteristics, uses of the chemical, or nature of the ecological effects. No
8 single scheme was considered ideal for the purposes of this risk ranking activity. The
9 classifications that the Subcommittee selected for ecological stressors were based
10 primarily on the source of the chemical (e.g., criteria air pollutants), the physical and
11 chemical nature of the chemical (e.g., nutrients, radionuclides), or, in the case of
12 persistent toxic organics, the usage of the chemical (e.g., pesticides). A more detailed
13 ecological risk ranking across a more disaggregated set of chemical stressors could be
14 done following the same methodology as proposed here but using a different
15 classification scheme; that is, the list the Subcommittee developed (Table 2-1) is not
16 presented as the only or even the best way to classify chemical stressors. In fact, the
17 appropriate level of stressor aggregation for relative human health risk assessment
often differed, as noted in Table 2-1.
19
20 The set of ecological system types that were selected by ERS (Table 2-2) were
21 modified from the set of ecological system types used in the Unfinished Business and
22 Reducing Risk projects. The intent of this
23 classification was to capture the range of Table 2-2. Types of Ecological Systems
24 ecological systems that occur in the U.S.,
25 but without having an unnecessarily large set
26 of ecosystem types to consider. The specific
27 set of ecological systems to use for risk
28 rankings at less than national scales may be
29 considerably different; for example, a risk
30 ranking done on the Southeastern U.S.
31 would not need to have the category of
32 semi-arid and arid ecosystems, but it might
33 have several different categories that better
34 refine forested ecosystems (e.g., coastal hardwood forests and swamps, pinelands,
35 oak-hickory upland forests, etc.).
2-7
Forests
Lakes
Rivers and Streams
Wetlands
Grasslands
Agroecosystems (Rangeland and Cropland)
Deserts and And Systems
Tundra
Estuarine and Near-Coastal Ecosystems
Marine Ecosystems
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
2.3.2.2 Development of Ecological Risk Profiles
The foundation of the comparative risk analysis
is a set of individual ecological risk profiles (see
Appendix A) that summarize for each listed ecological
stressor the nature of the stressor and its potential
ecological effects; define what is and what is not
included within the stressor category; discuss how
reliable and certain the information base is; and
provide information on any other factors that should be
considered when the relative rankings are assigned.
Because these risk profiles are modular, they can be
revised as new information is developed, thereby
facilitating a revaluation of the risk score for that
stressor. In short, the risk profile provides not only the
information used to develop the quantitative risk score
(discussed below), but also a summary of the level of
understanding about a stressor, references for that
understanding, and a description of the uncertainties
associated with it.
An important component of each ecological risk
profile is the stress-effects profile, a visual tool for
illustrating the relationship between the stressor and
effect regimes for each stressor at the ecosystem or
landscape level. In essence, the stress-effects profile
illustrates the significance of the adverse effects on
at-risk ecosystems that are expected to occur as a
result of current exposures to the stressor.
The first element of the stress-effects profile is
the stressor profile, i.e., a frequency distribution that
depicts the observed stressor intensity affecting the
ecological systems at risk in the region of interest.
The at-risk ecosystems may differ for each stressor;
In
I
I
Habitat Fragmentation
Stressor Profile
(Profile C)
a
Low Medium High
Observed Stressor Intensity
Rgure 2-1 a: Habitat Fragmentation
Stressor Profile
Introduced Species
Stressor Profile
(Profile B)
Freque
Low Medium High
Observed Stressor Intensity
Figure 2-1 b: Introduced Species
Stressor Profile
in
Acid Deposition
Stressor Profile
(Profile A)
£
u.
(0
Low Medium High
Observed Stressor Intensity
2-8
Figure 2-1 c: Acid Deposition
Stressor Profile
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
"7
19
20
21
22
23
24
ys
26
e.g., all aquatic and terrestrial systems in the U.S. are the at-risk ecosystems for the
stressors of habitat fragmentation (Fig. 2-1 a) and introduced species (Fig. 2-1 b), but
only certain aquatic and forest systems are at risk from the stressor of acid deposition
(Fig. 2-1 c). The x-axis of the stressor frequency distribution is normalized to a relative
scale of low, medium, and high intensity of occurrence of the stressor, based on an
estimated actual distribution of the stressor in the environment. Thus, a high level of a
stressor would relate to the highest levels found in the environment, not necessarily at
the highest levels that have been tested in the laboratory. This procedure allows the
ecological risk rankings to be based on relevant levels of the stressor, and it allows
identification of the risk reduction that would be achieved if the stressor distribution in
the environment were reduced. The y-axis of the stressor profile reflects the relative
frequency of low, medium, and high occurrences of the stressor as experienced by the
at-risk ecological systems, normalized on a scale of 0 to 1. Thus, if the ecosystem at
risk is, for example, estuaries, this curve would illustrate the relative frequency or
percentage of all estuaries that are exposed to high stressor intensities, the percentage
that are exposed to medium stressor levels, and the percentage that are exposed to low
stressor levels.
In practice, the ERS concluded that there are only a limited number of types of
environmental stressor frequency distributions, such as log-normal, normal, and
skewed. Thus, in order to simplify the ecological risk ranking process, a template of
four types of stressor-profiles was developed (Fig. 2-2), and each stressor was
assigned to one of these stressor-profile types to represent its frequency distribution.
For example, stressor Profile A depicts the most common situation, in which there is the
highest frequency of low levels of exposures, and lowest frequency for high levels of
exposures; this would be typical, for example, for exposures of a chemical in the
I
Standard Stressor Profiles
IB ic
L M H
L M H L M H
Observed Stressor Intensity
L M H
27
Figure 2-2: Standard Stressor Profiles
2-9
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 environment. Profile C represents the opposite situation, in which there is highest
2 frequency of high-level of exposure; a case in point would be habitat fragmentation, in
3 which most landscapes have become highly fragmented and only a relatively few have
4 low fragmentation. Note that for a more detailed relative risk assessment, the basic
5 template of profiles developed by the ERS could be replaced by the best available
6 information on actual stressor distribution profiles. Also, note that the frequency
7 distributions for a particular stressor are location- and scale-specific; for example, the
8 frequency distribution of acid precipitation is different in the Northeastern U.S. than in
9 the Southeast.
H
LMH LMH LMH LMH
Observed Stressor Intensity Observed Stressor Intensity Observed Stressor Intensity Observed Stressor Intensity
H
M
MH
19
H
M
L
H
MH
Observed Stressor Intensity Observed Stressor Intensity Observed Stressor Intensity
Figure 2-3: Standard Effects Curves
10
11 The other half of the stress-effects profile is the effects profile, i.e., the frequency
12 distribution of ecological effects associated with each of the stressor levels. That is, the
13 effects profile shows the intensity of adverse ecological effects that would result from
14 exposure to high, medium, or low levels of a particular stressor. Again, a standard set
15 of effects profiles was developed that the ERS believes represents most environmental
16 situations (Fig. 2-3), and one profile was assigned by ERS to each stressor for the
17 national-scale ranking. For example, Profile 4 illustrates a threshold situation, in which
18 a medium-level stressor causes high-level effects, compared with Profile 10, in which
2-10
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 only exposure to a high-level stressor causes a
2 high-level ecological effect.
3
4 Effects are evaluated for the ecological systems
5 considered to be most susceptible to the stressor. The
6 intent is to focus on the most important ecological
7 endpoints for those ecological systems most at risk. For
8 the purposes of the ERS national ranking, the ecological
9 endpoints were constrained to include ecosystem- and
10 landscape-level effects on the structure, function, and/or
11 composition of the system. Examples include an
12 ecological process endpoint such as decomposition
13 rates, a community composition endpoint such as
14 species richness, and a trophic structure endpoint, such
15 as the health of a critical or habitat-creating species.
16 Because of the focus on national risks, ERS did not
1 ~* address other potential ecological endpoints, such as the
population levels of individual species. However, for
19 ecological risk ranking conducted on local or regional
20 scales, population-level or other ecological endpoints
21 might supplement the ecosystem- and landscape-level
22 endpoints.
23
24 Example effects profiles are shown in Fig. 2-4.
25 Fig. 2-4a illustrates the case of habitat fragmentation, for
26 which Profile 2 was selected to reflect that even
27 high-levels of intensity of habitat fragmentation only
28 cause medium-levels of ecological effects. This is in
29 contrast to the greater potential ecological effect that
30 could result from high levels of introduced species (Fig.
31 2-4b, Profile 1). For the acid deposition example (Fig.
32 2-4c), two profiles are shown, one for the more sensitive
33 unbuffered lakes category (Profile 7), the other for
34 forests (Profile 15), although both effects profiles relate
35 to the same stressor profile. Again, in more detailed
Habitat Fragmentation
Effect Profile
(Profile 2)
i-H
Obseived Stressor Intensity
Figure 2-4a. Habitat
Fragmentation Profile
Introduced Species
Effect Profile
(Profile 1)
-H
Obseived Stressor Levels
Figure 2-4b. Introduced Species
Effects Profile
Acid Deposition
•Effect Profiles
(Profiles 7 and 15)
/ Forests,
'
J-M «
UJ
Observed Stressor Levels
Figure 2-4c. Acid Deposition
Effects Profiles
2-11
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 relative risk ranking assessments, a suite of
2 stressor-effects profiles could be developed for each
3 and every ecosystem type of importance in the area
4 of concern, or even for each and every specific
5 ecological endpoint affected by the stressor. While
6 that is beyond the scope of the present application to
7 national relative risk ranking by the EPS, it does
8 illustrate the flexibility of the ERS methodology for
9 specific risk ranking applications.
10
11 The combined stressor-effects regime, then,
12 is depicted by overlaying the effects profile (i.e., the
13 distribution of the intensity of effects caused by
14 varying levels of the stressor) onto the stressor
15 profile (i.e., the actual frequency distribution of
16 stressor in the environment of concern) (Fig. 2-5). In
17 this combined stressor-effects profile, the effects
18 profile y-axis is scaled to reflect the intensity of the
19 ecological effect (on a scale of high, medium, and
20 low) associated with the intensities of the stressor
21 that occur in the environment. The stressor profile
22 y-axis is scaled to the frequency of occurrence (0-1)
23 of the stressor in the specific environment of
24 concern. The x-axis, however, is identical for the two
25" profiles, i.e., reflecting low, medium, and high levels
26 of stressor intensity. Thus, even a high-intensity
27 stressor exposure (defined in terms of the levels
28 actually experienced in the environment) might result
29 in only a low-level ecological effect (e.g., acid
30 deposition effects on forest ecosystems shown in Fig.
31 2-5c).
32
33 In general, the high ecological effects category
34 was reserved for those ecological changes that are
35 very significant, involving major changes to the
2-12
cr
S
u.
CO Q.
Habitat Fragmentation
Stress-Effect Profile
(Profile C2)
TH
I
i
I
I
_LL
Low Medium High
Observed Stressor Intensity
Figure 2-5a. Habitat
Fragmentation Stress-Effect
Profile
Mressor Frequency
P ~
Intn
St
(P
)duced Speci
ress-Effect Pr
rofileBI)
/
/
BS
ofile
-H j
1
LM "
M ,
!
L
*
\
\
\
\
E3
Low Medium High
Observed Stressor Intensity
Figure 2-5b. Introduced Species
Stress-Effect Profile
Acid Deposition
Stress-Effect Profile
(Profiles A7.A15)
f
89
£
to o
.--TH
I
Low Medium High
Observed Stressor Intensity
Figure 2-5c. Acid Deposition
Stress-Effect Profile
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 structure, composition, and/or function of the system; an example is a major physical
2 disturbance that removes the major structural components of an ecological system.
3 Medium-level ecological effects were considered to be significant changes in the
4 ecosystem's structure, composition, and/or function; for example, the replacement of
5 one or more native species in the community by introduced exotic species would
6 constitute a medium-level effect. Low-level ecological changes were those that, while
7 detectable (i.e., non-zero), do not entail significant changes to either the structure,
8 composition, or function of the system; an example would be the shift in the salinity
9 isopleth near the freshwater outflow from a canal. The frequency of the no (or de
10 minimus) level of ecological effects was not assessed by ERS; this category applies to
11 any ecological changes that would be non-detectable. For example, in the
12 stress-effects profile for acid deposition (Fig. 2-5c), the largest fraction of affected lakes
13 are judged to have low-level ecological consequences (i.e., detectable but not highly
14 significant changes), but there may be a larger set of lakes that do not experience any
15 detectable effects at all and therefore effects would be judged to be insignificant for the
16 ecosystem.
17
As with the stressor regime, the effects profile can be modified if and when
19 additional data are acquired. Also, the use of multiple effects profiles, such as in Fig.
20 2-4c, allows the ready identification of those ecological effects that might differ across
21 different ecosystem types, thereby highlighting differential ecological sensitivities and
22 risks. In principle, such a distinction could also be made within a particular ecosystem
23 type, showing differential effects for various ecological endpoints for that ecosystem.
24 This approach would be particularly useful for localized comparative risk assessments.
25
26 It should be noted that there are insufficient data on many types of ecological
27 responses to stressors, and much of the available information excludes major
28 constituents of ecological systems (e.g., microbes and amphibians). The
29 Subcommittee recommends using the best available information for characterizing
30 ecological effects from stressors and, if there are quite different responses among
31 different components of ecological systems, to focus on the more sensitive or more
32 ecologically important components. Further, as new information is acquired on
33 stress-response relationships, the database upon which the ecological risk profiles and
34 resulting relative risk rankings are based should be updated and new rankings done
35 periodically. Note that different stress-effects profiles might be appropriate for
2-13
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 conducting this risk ranking exercise at regional or local scales. That is, at a regional
2 scale the frequency distribution of the stressor (stress profile) may well differ from the
3 distribution at the national scale. On the other hand, since the effects profile is based
4 on the responses of a particular ecosystem type to the stressor, effects profiles are not
5 different for risk rankings at different scales. This is an important distinction, as it
6 means that an effects profile database can continually be updated as more information
7 becomes available, and that same database could be used for different risk ranking
8 exercises, but the stressor distribution database must be developed for each specific
9 area of concern, e.g., a particular region or the entire nation.
10
11 2.3.3 Development of Relative Ranking of Ecological Risks
12
13 Although the stress-effects profiles described in the previous section provide an
14 initial, visual characterization of the ecological risks for vulnerable ecological systems,
15 the ERS concluded that in order to characterize relative ecological risks, it is necessary
16 to convert each profile into a quantitative score. The Subcommittee developed such a
17 methodology that goes well beyond the relative risk ranking approaches of Unfinished
18 Business and Reducing Risk. The ERS methodology reflects the proportion of the
19 ecological resource at-risk to a particular stressor and adjusts the risk level assigned to
20 each stressor in the context of the specific scale of concern (e.g., regional, or national).
21 In order to produce a national ranking of relative ecological risks, the ERS developed
22 several specific risk modification factors that should be considered in making this risk
23 adjustment. These multiplicative factors, and the rationale for selecting their specific
24 numerical values, are discussed below and summarized in Table 2-2. For ecological
25 risk rankings done at other than the national scale, different factors, or perhaps different
26 numerical values for the same factors, might be needed.
27
28 2.3.3.1 Multiplicative Factors Used to Assign Stressor Risk Values
29
30 The multiplicative factors proposed for the national-scale ecological risk ranking
31 were developed by expert judgment, expressed through consensus of the
32 Subcommittee. The intent was to establish the weights to apply to various
33 scale-specific considerations that would lead to a higher or lower assignment of a risk
34 score and would more accurately, and more transparently, represent the ecological risk
35 of each stressor vis a vis the other stressors. In essence, each multiplicative factor
2-14
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 addresses the question of what is there about this particular stressor that makes it more
2 or less an ecological risk compared to another stressor.
3
4 At the outset, the ERS assigned a numerical value to each of the three levels of
5 ecological effects as defined above: High = 4, Medium = 2, and Low = 1. The
6 calculated risk score for a stressor is quantified by applying the appropriate
7 multiplicative factors to each of three possible effects intensities (H, M, L), then
8 summing the results, as discussed in more detail below. Since the initial effects
9 intensity scores were based on a factor-of-two basis (i.e., high = 2x medium = 4x low),
10 the multiplicative factors were scaled so that a factor that warranted a jump in the major
11 risk category (e.g., from medium to high) would be assigned a multiplicative value of 2.
12 For example, the proportion of the resource experiencing a high level of stress from
13 harvesting of coastal fisheries populations was considered so extensive that a jump of
14 the risk value one full category was assigned (Table 2-3), whereas the proportion of
15 ecological systems experiencing a medium level of effects from UV-B exposure was
16 considered so low that a factor of 0.5 was assigned, resulting in one full drop in risk
17 category.
19 Less significant factors were scaled relative to this two-fold standard. For
20 example, a multiplicative factor of 1.1 means that the risk value would be increased by
21 a small increment, and a string of about 7 such increments would be needed among the
22 various ranking factors to equal one full category jump (i.e., 1.17). The Subcommittee
23 sought consistency across the factors, so that two factors assigned the same
24 multiplicative value were considered to be relatively equivalent in terms of the risk
25 ranking. Thus, a recovery potential requiring a centuries time scale for recovery once
26 the stressor was removed (multiplicative factor 1.25) was considered equivalent to a
27 medium level of species depletion in terms of ecological significance to the ranking
28 (Table 2-3).
29
30 While these multiplicative factors represent the best judgment of the
31 Subcommittee, other considerations might result in somewhat different assigned
32 values. The opportunity to implement the integrated ecological-human health risk
33 ranking framework through an expert opinion survey (discussed in Chapter 3) could
34 provide a verification of the weights assigned by the Subcommittee for the
35 national-scale ecological risk ranking. Again, the transparency of the ERS methodology
2-15
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 readily allows examination of the assumptions and the values assigned by the expert
2 judgment process. If it were determined that the weighting factors needed adjustments,
3 it would be a simple process to revise the multiplicative values and recalculate the risk
4 rankings without having to redesign the relative risk ranking methodology.
5
6 Once the risk ranking weighting factors are established, their application to the
7 stressor-effects profiles is straightforward: First, each level of ecological effect is
8 considered separately to calculate partial risk scores. The first partial risk score for a
9 stressor focuses on the high level of ecological effects (initial value of 4). The
10 proportion of resources affected at the high intensity of effects (if any) is assessed,
11 based on the stress-effects profiles (Fig. 2-5). This is followed by each of the other
12 weighting factors described below, all multiplied together to develop a partial risk score
13 reflecting the contribution to the total risk ranking score from the high intensity of the
14 stressor. If the stress-effects profile indicates that no ecological systems experience
15 the high level of ecological effects from that stressor, even when the stressor is at the
16 high level of intensity (e.g., habitat fragmentation profile [C2] shown in Fig. 2-5a and the
17 acid deposition-forest profile [A15] shown in Fig. 2-5c), then the multiplier of zero will
18 apply for the factor "proportion of resources at risk." Consequently, the assigned partial
19 risk contribution from the high intensity effect level is zero. This illustrates an important
20 reason why the weighting factors are multiplicative rather than additive; i.e., if any factor
21 is assigned a value of 0, then that partial risk score is 0.
22
23 Similarly, the second partial risk score is calculated for the medium level of
24 ecological effects caused by the stressor (which begins with an initial assigned value of
25' 2). The third partial risk score is based on the low level of ecological effects caused by
26 the stressor (which begins with an initial assigned value of 1). In these cases, the
27 partial risk scores are calculated by applying the appropriate multiplicative factors in the
28 same manner as described for the high-level partial risk score.
29
30 Finally, the three partial risk scores (for the H, M, and L levels of effects) are
31 summed to result in the final total risk ranking score for the specific stressor, calculated
32 for the specific scale of concern (in our case, the national scale). The details of the
33 partial and total risk ranking scores developed by ERS are shown in a matrix form in
34 Table 2-4.
35
2-16
-------
Table 2-3. Multiplicative Factors for Deriving National Risk Rankings
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Factor (Assessed at H/M/L Effect Level)
1. Proportion of Resource At Risk
>60 percent
30-60 percent
10-30 percent
1-10 percent
<1 percent
effect level (H/M/L) does not occur
2. Existence of Hot Spots
Nationally distributed
Regional
Local or None
3. Recovery Potential
Irreversible or >centunes
Centuries
Decades
< Decades
4. Duration of Stress-Effects
>centunes
Decades
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 The following sections describe each multiplicative factor and its assigned
2 weighting values for national-scale rankings.
3
4 a) Proportion of Resource at Risk - The first entry into the risk scoring process
5 characterizes the proportion of the ecological resource that experiences effects at the
6 specified exposure intensity. This information is the same as shown graphically in the
7 stress-effects profile for each stressor. The question asked is what proportion of the
8 ecological systems at risk in the region of concern for the ranking (e.g., the proportion of
9 all lakes in the nation) is exposed to the stressor such that the high level of ecological
10 effects is realized. If the assigned effects distribution does not show severe ecological
11 effects at high intensities of the stressor, then this proportionality factor is assigned the
12 value of zero. If high-level ecological effects do occur, then the proportion of the
13 ecological systems of concern experiencing high ecological effects is estimated as one
14 of the following categories: <1 %, 1-10%, 10-30%, 30-60%, or >60%. Then the
15 multiplicative factor (Table 2-3) is applied based on that proportion category. For
16 example, if more than 60% of the at-risk ecological resources experience the high
17 effect, then the factor would be 2x. On the other hand, if less than 1 % experienced that
18 level of effect, then the multiplicative factor assigned would be 0.25x. Similarly, the
19 same question is asked for the proportion of the ecological resource that experiences
20 medium-level effects from the stressor, and finally the proportion at the low-level of
21 effects. For example, in the case of habitat fragmentation, shown in Fig. 2-5a, no
22 ecological systems experience high-level ecological effects, even at high levels of
23 fragmentation, so the partial risk score for the high-level effects is 0. The medium-level
24 effect results from high levels of habitat fragmentation, which the figure illustrates
25 . occurs at more than 60% of the systems, resulting in the initial value for medium effects
26 (2) being multiplied by 2x (for the >60% proportionality factor). Similarly, only low-level
27 effects occurs at medium or at low levels of habitat fragmentation, and those conditions
28 occur 10-30% of the systems, resulting in a multiplicative factor of 1 x for the low effects
29 partial risk score.
30
31 b) Distribution of "hot spots" - This multiplicative factor adjusts the risk score
32 if the stressor is distributed in localized hot spots, with disproportionate ecological
33 effects from what would be predicted from the stressor frequency distribution profile.
34 This approach acknowledges that adverse ecological effects may be significant in
35 highly localized areas and, therefore, warrant national concern even though they do not
2-18
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 affect a large proportion of the resource and thus would not show up in the profile (?).
2 This factor is assigned at the highest level of effect for which ecological effects occur
3 (i.e., at the high level if there are high effects, at the medium level if there are medium
4 but no high effects, and at the low level if there are no medium effects). If no or very
5 few hot spots occur, the multiplication factor is 1.0; if hot spots do occur within a few
6 regions of the nation, the multiplication factor is 1.25; and if they occur in most regions
7 of the nation, the factor is 1.5. For example, a 1.5 factor for hot spots was assigned to
8 the stressor heavy metals, which occur at high concentrations in localized areas across
9 the nation.
10
11 c) Recovery potential - This multiplicative factor relates to ecosystem
12 resilience, i.e., the hypothetical, relative potential of the affected ecosystem to recover if
13 the stressor were to be removed, even if such removal does not appear realistic given
14 physical, societal, or economic constraints. The factor reflects the fact that effects that
15 are irreversible or very long-lasting (e.g., over geological time) are more significant
16 ecologically than those effects that can be reversed quickly. For example, the recovery
17 factor for nutrient effects on estuaries was assigned a value of 0.75 since the ERS
estimated that once the nutrient inputs were removed, the system would recover within
19 a decade.
20
21 d) Duration of the stress-effect - This multiplicative factor takes into account
22 the time into the future that the stress could be expected to occur, so that an adverse
23 effect that is expected to last a long time would be given a higher risk assignment than
24 one that is of short duration. For example, hydrologic modification was assigned the
25 duration factor of 1.25, in part because major changes in streambeds, such as from
26 dams, are expected to persist for centuries.
27
28 e) Secondary stress induction - This multiplicative factor relates to the number
29 and strength of interactions between the selected stressor and other stressors on the
30 list, i.e., if the stressor induces or predisposes the occurrence of secondary stressors.
31 The purpose of this multiplicative factor is to note those stressors that create cascading
32 effects through the induction of other stressors. If two or more significant interactions
33 exist, the score is multiplied by a factor of 1.1; otherwise, the factor is 1.0. For
2-19
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 example, hydrologic alteration would be expected to cause increases in the nutrient,
2 sedimentation, salinity, and other stressors, so its secondary stress induction factor is
3 1.1.
4
5 f) Species depletion - This multiplicative factor relates to the potential for the
6 stressor to result in depletion of species, from local loss of a population (extirpation) up
7 to global extinction of a species. The species depletion factor is included because of
8 the importance of the loss of species and the reduction in biodiversity that has occurred
9 during the past century, significantly affecting the structure, composition, and function of
10 many ecological systems. It also has particular importance to society in that some
11 species are of special concern (e.g., threatened or endangered species). The
12 Subcommittee defined high, medium, and low-level species depletion effects as follows:
13
14 High-level effects: extinction or extirpation of many species, resulting in loss of
15 species diversity or richness at one or more ecological systems; an example of
15 high-level species depletion occurs for high intensities of habitat conversion,
17 such as when grasslands are converted to agricultural systems.
18
19 Medium-level effects: extinction or extirpation of a number of ecologically and/or
20 societally important species, but not necessarily resulting in overall decreased
21 species diversity; an example of medium-level species depletion is when
22 pesticides occurring at high intensities cause the loss of many species of insects
23 or other invertebrates in an ecological system adjacent to an agricultural field.
24
25 Low-level effects: extinction or extirpation of one to a few species; an example is
26 the loss of a single species because of a pest or disease outbreak.
27
28 g) Special ecological significance - This multiplicative factor was developed to
29 capture any special significance of the ecological effects from the stressor on the
30 affected ecosystems and/or ecological endpoints. One such modification captures the
31 exceptional importance of certain ecological attributes, i.e., ecosystems or effects that
32 are disproportionate in importance relative to their spatial extent or frequency of
33 occurrence. For example, effects that are particularly consequential to coastal wetlands
34 as a group, or a stressor that could eliminate an entire class of ecological system, such
35 as vernal ponds, or that has broad-based effects on a group of organisms (e.g., all
2-20
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 reptiles) would trigger the special ecological significance multiplier. This multiplier was
2 reserved for ecological risk considerations not captured by the other factors and was
3 only infrequently invoked.
4
5 2.3.3.2 Sample Calculation for the Stressor Pesticides
6
7 As discussed above, the quantitative step in the ecological risk ranking
8 methodology is to apply the multiplicative adjustment factors to the initial scores for
9 each of the high, medium, and low levels of effects occurrence. To demonstrate this
10 process, consider the calculation of the total risk ranking score for the stressor
11 pesticides (see Tables 2-3 and 2-4). For the high-level effects partial risk score, the
12 Subcommittee concluded that the high intensity of ecological effects (initial value 4)
13 occur only in 1-10% of ecosystems (thus, the multiplicative factor is 0.5); hot spots are
14 distributed nationally (multiplicative factor is 1.5); recovery would be expected to occur
15 in decades (multiplicative factor is 1.0); the duration of stressor-effects would likely be
16 decadal (multiplicative factor is 1.1); there are no secondary induced stressors
'7 (multiplicative factor is 1.0); species depletion is highly significant (multiplicative factor is
j 1.25); and there are disproportionate effects on higher trophic-levels and critical species
19 (multiplicative factor is1.1). Thus, the application of multiplicative factors results in a
20 modified partial risk score of 4.54 for the high-intensity effect. Similarly, calculations
21 were done for the medium-intensity partial risk score (product = 1.10) and for the
22 low-intensity partial risk score (product = 1.10). These three partial risk calculations
23 were then summed to produce a total risk ranking score of 6.74. To complete the risk
24 ranking process, this total risk score for pesticides is compared to the total risk scores
25" calculated for all the other stressors of concern.
26
27 2.3.4 Sources of Uncertainty
28
29 There are many sources of uncertainty inherent in relative ranking of risks.
30 Some causes of uncertainty are reducible through further research and better scientific
31 information, such as having an improved understanding of stress-effect relationships for
32 a specific stressor affecting a specific ecological endpoint. Other sources of uncertainty
33 are essentially irreducible, such as those caused by the intrinsic complexities and
34 variability of natural ecosystems. Uncertainty is only important here, however, if it
35 significantly changes the relative assignment of risk estimates across the stressors.
2-21
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Examples of the sources of uncertainty in the ERS national ecological risk
2 ranking process include:
3
4 a) aggregation of stressors for the analysis into a class of stressor, i.e.,
5 grouping of different specific stressors with different exposure or effects
6 characteristics - an example is the category "persistent toxic organics";
7
8 b) aggregation across specific ecological system types, i.e., inclusion of
9 different specific ecosystems with different effects characteristics;
10
11 c) aggregation across specific ecological endpoints;
12
13 d) lack of information on the exposure regime;
14
15 e) lack of information on ecological effects;
16
17 f) interactions among multiple stressors; and
18
19 g) composition and expertise of the members of the ranking panel.
20
21 Clearly, the ERS process of developing a national-scale ranking was limited by
22 the availability of information and time for the exercise, factors that could be mitigated
23 by a concerted effort within EPA or another institution to acquire and analyze the full set
24 of available information on each stressor. A major advantage of a comparative risk
25 ranking process is that many of the associated uncertainties do not make a difference
26 in the ranking, whereas a process of doing an assignment of absolute risk values would
27 be much more subject to uncertainties and much less reproducible. Moreover, when
28 the risk scores are converted to qualitative characterizations of risk, as opposed to
29 numerical risk values, the importance of uncertainties vis a vis the overall ranking is
30 even further reduced. In practice, the Subcommittee found that the quantitative
31 ecological risk scores for each stressor tended to clump into distinct ranges, allowing
32 separation of the stressors into qualitative categories of risk. In most cases,
33 uncertainties in specific multiplication factors applied to a stressor did not cause a
34 reassignment of the stressor from one major, qualitative category to another. This
35 significantly enhances the confidence that can be placed on the relative rankings. As
2-22
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 noted in the next section, however, in some cases uncertainties may be so great that
2 risk associated with a particular stressor (e.g., endocrine disrupting chemicals and
3 genetically engineered organisms) cannot be ranked.
4
5 2.4 National-Scale Ecological Risk Ranking
6
7 2.4.1 Results of the ERS National-Scale Ecological Risk Ranking
8
9 In order to apply the ecological risk ranking methodology to the national scale of
10 relevance, the ERS developed ecological stressor-effect profiles (Appendix 2A) for 32
11 of the stressors listed in Table 2-1 and scored each stressor using the partial and total
12 ecological risk scores procedures described above. The results of the ERS
13 national-scale risk ranking process are summarized in Table 2-4. The table shows for
14 each stressor the primary ecosystem type considered by ERS to be at risk. The
15 stressor-effect profile curves are next listed to indicate the stressor and effects
16 distributions that the Subcommittee considered was appropriate for that stressor at the
national scale. Then, for each effect intensity level, the partial risk scores are
calculated by taking the intensity value and multiplying by the number assigned at each
19 cell along the row, resulting in the partial scores at the right of the table. These three
20 partial scores are added and shown in the "sum score" column. The next entry in the
21 table is the ERS's judgment as to whether the trend for risk from the stressor is
22 increasing, decreasing, or remaining the same nationally; for example, the risks from
23 metals were judged to be declining, but the risks from climate change were judged to be
24 increasing. Additionally, ERS identified the potential for surprises in its assignment of
25 risk levels; for example, ERS judged there to be little likelihood of surprises for nutrient
26 additions to the environment, but a high potential for surprises from introduced exotic
27 species. Note that the latter two columns (trends and surprises) were added to
28 characterize further the Subcommittee's judgment of the stressor, but were not
29 incorporated into the relative risk rankings per se. Finally, it should be repeated that: a)
30 ERS assigned the various values based on the assumption that current environmental
31 laws and regulations continue to be enforced; b) these relative rankings apply only at
32 the national scale and might be quite different for regional or local scales; c) only
33 ecosystem- and landscape-level endpoints were considered; d) these rankings were
34 based on the understanding of the ERS members and did not entail analysis of
?K extensive data or models; and e) limitations in the available data may in some cases
2-23
-------
Table 2-4: Summary of Ecological Risk Ranking Scores
stressor
NOx
ecosystem type stress- effects
at risk effects Intensity
curve
terrestrial A15
1
2
4
proportion
of resource
0.25
0
0
hot
spots
1
recovery duration secondary species eco
stress depletion slgnlf
Induction
0.75 1 1
1 1
Partial
SCORE
019
000
000
SUM trends
SCORE
0.19 1
potential for
surprises
L
tropospheric
ozone
forests, shrubs A10
and grasslands
1
2
4
OS
0.25
0.25
1
1
1
075 11 1
075 1.1 1
0 75 11 1
1 1
1 1
1 1
041
0.41
083
1.65 =
L
tropospheric
ozone
agric A7
1
2
4
1
0.5
025
1
1
1
0.75 1 1
0.75 1.1 1
0.75 1.1 1
1 1
1 1
1 1
0.75
0.83
0.83
2.40 =
L
SO2
terrestrial A12
1
2
4
0.25
0
0
1
1 1.1 1
1 1
028
000
000
0.28 1
L
Hg
freshwater and esturarie: A1 2
1
2
4
0.25
0
0
1.25
1 11 1
1 1
034
000
000
0.34 ?
UH?
other
heavy metals
freshwater and esturartes A13
1
2
4
025
025
0.25
1
1
1.5
1 1.
1 1
1 1.
toxic
Inorganics
(As. Se. B)
freshwater, eslurarles A10
and agriculture
1
2
4
0.25
025
0.25
1
1
125
.
persistent toxic
organlcs
aquatic and terrestrial A10
1
2
4
05
0.25
0.25
1
1
1.5
f
.
.
pesticides
aquatic and terrestrial A7
1
2
4
1
05
0.5
1
1
1.5
1
1
1
1
1
028
055
165
2.48 1
H
1
1
1
1 1
028
0.55
1.51
2.34 =
H
1
1
1
1
1
1 1
0.55
0.55
165
2.75 1
H
1
1
1 .1 1
1 1
1 1
1.25 1.1
1 10
1.10
4.54
6.74 A
H
-------
Table 2-4 (continued)
w
stressor
acid deposition
ecosystem type stress- effects
at risk effects Intensity
curve
lakes A7
1
2
4
proportion
of resource
1
0.5
0.25
hot
spots
1
1
1
recovery duration secondary
stress
Induction
1 1 1
1 1.1 1
1 1.1 11
species eco
depletion slgnlf
1
1
1
1
1
1
Partial SUM trends potential for
SCORE SCORE surprises
1.00 3.31 |
1.10 V
1.21
L
acid deposition
forest A15
1
2
4
0.25
0
0
1
125 1.1 11
1
1
0 38 0.38 1
0.00 T
000
L
,
nutrients
freshwater and esturaries B19
1
2
4
1.5
1
0.25
1
1
1 5
075 1.1 1
075 11 1
075 1.1 11
1
1
1
1
1
1
1.24 4.25
1.65
1.36
L
radlonuclldes
aquallc and terrestrial A3
1
2
4
0.25
0
0
1
1 1 1
1
1
0 25 0.25 -
000
0.00
'
oil spills
freshwater, eslu rarl as A 1 1
and tundra
1
2
4
0.25
0.25
0
1
1
1 1 1
1 1 1
1
1
1
125
0 25 0.88
0.63
0.00
L
DO/BOD
acid mine
drainage
freshwater and esturarle: B1 9
freshwater A7
1
2
4
1
2
4
0.5
05
0.25
0.25
0.25
0.25
1
1
1.5
1
1
125
075
075
075
075
0.75
075
1 1
1 1
1.1 1.1
1 1 1
1 1 1
11 1
1 0.38
1 0.75
1 136
1 0.21
1 041
1 103
2.49 I
V
1.65 I
T
L
L
contaminated
groundwater
freshwater and eslurarles A12
1
2
4
025
0
0
1
075
1.25 1
1
1
0 23 0.23 ?
0.00
000
L
disease/pest
outbreaks
aquallc and terrestrial A7
1
2
4
0.5
0.5
0.5
1
1
1.5
1
1
1
1 1
1 1
1 1 1
1
1
1
1
1
1
0 50 4.80 =
100
3.30
H
-------
Table 2-4 (continued)
1
N
^
stressor
Introduced
exotic
species
climate change
noise
hydrologlc
alteration
habitat
fragmentation
habitat
conversion
physical habitat
disruption
turbidity
sedimentation
ecosystem type stress-
at risk effects
curve
aquatic and terrestrial B1
aqualic and terrestrial C7
aquatic and terrestrial A1 2
(fauna only)
aquatic and terrestrial A1 2
(fauna only)
streams, wetlands. D7
and estuaries
aquatic and terrestrial C2
aquatic and terrestrial C7
aquatic and terrrestrial A1
aquatic and wetland B1
effects
Intensity
1
2
4
1
2
4
1
2
4
1
2
4
1
2
4
1
2
4
1
2
4
1
2
4
1
2
4
proportion
of resource
1
15
1
1
1
1.5
025
0
0
025
0
0
1 5
0.5
1.5
1
2
0
0
1
1.5
1
05
025
1
1.5
1
hot
spots
1
1
1
1
1
1
1
1
1
1
1
15
1
1
1
recovery
1
1
1
1
1
1
1
1
075
1
1.25
1
1
0.75
0.75
1
duration secondary
stress
Induction
1 1
1 25 1
1.25 1.1
1.25 1
1 25 1.1
1.25 1.1
1 1
1 1
1.25 1
1.25 1.1
1.25 1.1
1.25 1
1 25 1.1
1.25 1.1
1.25 1.1
1
1.1
1.1
1
1 1
1.1
F species eco
depletion slgnlf
1
1.25
1.5
1
1
1.25
1 1
1 1
1 1
125 1
1.5 1.1
1 1
1.25 1
1.25
15
1
1
1
1
1
1
Partial
SCORE
1.00
4.69
825
1.25
2.75
1031
0.25
000
000
025
000
000
141
1.72
17.02
125
688
000
000
344
12.38
1 10
1.21
182
083
272
484
SUM trends |
SCORE
13.94 A
T
14.31 A
T
0.25 A
T
0.25 A
T
20.14 =
8.13 A
T
15.81 A
T
4.13 A
T
8.39
totentlal for
surprises
H
H
L
H
L
L
L
L
L
-------
Table 2-4 (continued)
stressor
altered
fire regime
ecosystem type
at risk
terrestrial and wetlands
stress-
effects
curve
A13
effects
Intensity
1
2
4
proportion
of resource
0.5
0.5
0.25
hot
spots
1
1
1
recovery duration secondary species eco
stress depletion slgnlf
Induction
1 1
1 1
1 1
.1 1
1 1.1
1 1.1
altered
salinity regime
freshwater, esturaries
and wetlands
A7
1
2
4
1
025
0.25
1
1
1
1
t
1
freshwater
populations at-risk
68
1
2
4
1
1
0
1
1
0.75
0.75
1
1
1.1 1
1
1
1
Partial
SCORE
055
1.21
121
SUM trends
SCORE
2.97 II
potential tor
surprises
L
1
1
1.1
1.10
055
133
2.98 A
L
1
1 1
t
1
083
1.65
000
2.48 A
L
aquatic and terrestrial
Including ocean
2
4
0.5
0
1
1
0.75 1 25 1
0.75 125 1 1
1
1
0.94
0.94
0.00
1.88 A
H
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
Table 2-5: Relative Ranking of Ecological
Risks at the National Scale
1 influence the risk rankings; for example,
2 some chemicals may rank low for ecological
3 risks because there are few data on
4 ecological effects other than for the aquatic
5 and pesticide endpoints required by law.
6 Nevertheless, the Subcommittee is
7 reasonably confident that the overall
8 ecological risk rankings accurately reflect the
9 relative risks to ecological systems in the
10 U.S. from anthropogenic stressors.
11
12 2.4.2 Synthesis and Conclusions
13
14 The results from the relative ranking
15 of national ecological risks derived by the
16 ERS using the stressor-based methodology
17 are shown in Table 2-5 and Figure 2-7. The
18 process for assigning the stressors into
19 qualitative risk categories was to rank order
20 the stressor total risk scores and identify,
21 where possible, distinct breaks between
22 groups of stressors. Inspection of Figure 2-7
23 shows the clear discontinuities that were
24 used to group the stressors into the four
25 qualitative ecological risk categories (highest
26 risks, high risks, medium risks, and low
27 risks). Note that the division between
28 medium and low risks was less distinct and,
29 therefore, somewhat more arbitrary than the
30 distinction between the highest and high
31 categories, and the distinction between the
32 high and medium categories. Other
33 aggregations or dividing lines could be
34 assigned to these numerical rankings, but
35 the important consideration is that there is a clear distinction across the stressors, so
HIGHEST ECOLOGICAL RISKS
hydrologic alterations
harvesting marine living resources
habitat conversion
climate change
introduction of exotic species
HIGH ECOLOGICAL RISKS
turbidity/sedimentation
habitat fragmentation
pesticides
MEDIUM ECOLOGICAL RISKS
disease/pest outbreaks
nutrient additions
physical habitat disruption
acid deposition (lakes)
altered fire regime
altered salinity regime
persistent toxic organics
heavy metals other than Hg
DO/BOD
harvesting freshwater living
resources
troposphenc ozone
toxic inorganics
UV-B
acid mine drainage
Low ECOLOGICAL RISKS
oil spills
acid deposition (forests)
Hg
S02
radionuclides
noise pollution
light pollution
groundwater contamination
thermal pollution
NO,
UNKNOWN BUT POTENTIALLY
IMPORTANT RISKS
endocrine disrupters
genetically engineered organisms
2-28
-------
N
i
M
Figure 2-7. Ecological Risk Ranking Scores
25
Highest Risks
20
15
tl
E
10
5 -
High
Risks
t
Medium Risks
L
Low Risks
•sum score
U.tt.u
I
I i f i i I !
i
1 ! I i
r I
I
« =
I
« 1
•8 s
E
E
Stressor
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 that, for example, the assessment of the ecological risks from hydrologic alteration was
2 clearly quite different from the assessment of ecological risks from oil spills. The
3 Subcommittee believes that its assignments of relative risks into the qualitative
4 categories in Figure 2-7 is a reasonable interpretation of the data.
5
6 Further, the Subcommittee believes this categorization is relatively robust, as
7 demonstrated by an exercise EPS conducted of substitution of alternate plausible
8 values for weighting factors assigned to specific stressors, i.e., a type of sensitivity
9 analysis. In all cases the Subcommittee examined, to shift significantly the position of a
10 particular stressor relative to the other stressors would require changes in the
11 multiplicative factors that do not seem warranted based on the information available.
12 Obviously, as new or more complete data become available, or as the exercise is
13 applied to specific regions or locales, then there may be shifts in the relative risk
14 rankings.
15
16 Consistent and at least semi-quantitative relationships are maintained in these
17 final, total risk categorizations: The initial entry value of 4.0 that was assigned to
18 high-level ecosystem effects has now been modified so that all total risk scores above
19 5.0 are assigned to the high risk category, and all total risk scores above 10 are in the
20 highest ecological risk category. Similarly, from the initial value of 2.0 for medium-level
21 effects, the total risk scores in the range of 1.0 to 5.0 now fall into the medium
22 ecological risk category. And anything less than a total ecological risk score of 1.0 was
23 considered to be in the low ecological risk category.
24
25 At the national scale, the highest level of ecological risk was assigned by ERS to
26 two classes of habitat alteration (hydrologic modification and habitat conversion), as
27 well as to climate change, introduced exotic species, and over-exploitation of living
28 marine resources through harvesting. Grouped at the high level of ecological risk are
29 turbidity/sedimentation, habitat fragmentation, and pesticides. The medium ecological
30 risk category includes disease and pest outbreaks, nutrients, physical habitat disruption,
31 acid deposition (lakes), and altered fire or salinity regimes. Note that two stressor
32 categories (endocrine disrupters and genetically engineered organisms) did not have
33 calculated risk scores and were not assigned to a narrative risk category because the
34 Subcommittee believed the potential for ecological risks may be very high but is
35 currently largely unknown. The Subcommittee, however, felt that it was important to
2-30
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 distinguish between unknown, but potentially important, risks and low risks. Further
2 research and experience with these stressors may lead at some point in the future to a
3 confident assignment of these to a risk category.
4
5 The results shown in Table 2-5 indicate that the greatest ecological risks in the
6 U.S. relate predominantly to physical and biological stressors, not to chemical
7 stressors. That conclusion is consistent with the relative ecological risk ranking of the
8 Unfinished Business and Reducing Risk projects (EPA 1990a, 1990b; Harwell et al.,
9 1992).
10
11 It should be recalled that the risk ranking process done by ERS focused on the
12 national scale of concern and explicitly considers that current environmental laws and
13 regulations will continue to be enforced. At less than national scales, the ecological risk
14 rankings might be different; for example, at a local level the most important risk might
15 be the acid drainage from a mine entering a valuable freshwater stream. The national
16 scale of ranking is meant to look more broadly, and such local-scale risks only become
•|~7 nationally important if captured by the criteria for scaling-up; that is, acid mine drainage
would have had a higher national-scale ranking if it occurred at many locations all
19 across the nation.
20
21 Another noteworthy conclusion is that many environmental problems that have a
22 high public perception of risks and/or have a high allocation of resources for
23 environmental protection and management may, in fact, constitute relatively lower
24 ecological risk as nationally ranked. Included in this are ecological risks from
25 radionuclides, oil spills, heavy metals, toxic organic and inorganic chemicals, and SO2
26 and NOX air pollution. Again, this conclusion is consistent with Reducing Risk, and
27 comparisons of its ecological risk rankings with a poll of public perceptions of ecological
28 risks (Harwell et al., 1992).
29
30 There are at least two possible reasons for this discrepancy: The first plausible
31 reason is the success that the environmental regulatory system has had over the past
32 several decades The ERS risk ranking was predicated on the assumption that existing
33 environmental laws and regulations would remain in effect and continue to be enforced;
34 the same risk ranking done in the 1960s might have given quite different results. In that
35 case, risks that have been reduced may not yet be perceived as having been reduced,
2-31
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 illustrating a time-lag in the perception of environmental successes. Second, there is
2 an anthropocentric focus of concern about environmental risk agents, i.e., historically
3 there has been primary attention to human health effects rather than to ecological
4 effects from environmental stressors, so the public has much better understanding of
5 those risks. Both of these discrepancies can be mitigated by increased education of
6 the public on the scientific bases of risk ranking.
7
8 The results of the SAB ERS relative ecological risk rankings have considerable
9 implications for the ability of EPA and other regulatory and resource management
10 agencies to reduce risks to ecological systems. If society desires to restore, protect,
11 and sustain the environment with high ecological quality, then the remaining highest
12 priority risks to ecological resources must be a focus of attention. At present, the
13 highest ranked risks are not well regulated by EPA. In addition, examination of the risk
14 rankings in conjunction with expected trends in the associated stressors (Fig. 2-7)
15 highlights the fact that, in the Subcommittee's assessment, many of the top-ranked
16 ecological risks are the very ones that are predicted to experience an upward trend.
17
18 2.5 An Effects-Backwards Methodology for Risk Rankings
19
20 In the ecological risk ranking procedure described above, the major ecological
21 effects caused by a stressor or group of stressors were analyzed. The logical
22 progression began with the stressor, then proceeded to rank the resulting ecological
23 effects. It is often useful, however, to apportion the major, multiple causes of a
24 particular adverse ecological effect, i.e., to work backwards from an observed effect to
25' estimate the relative contribution to that ecological effect from each of several
26 stressors. In this section, we briefly present the outline of a methodology for beginning
27 with an effect and determining the relative contribution of various stressors to the effect
28 at a national, regional, local, or other scale of interest.
29
30 The first step in an effects-backwards relative risk assessment is to identify and
31 characterize the ecological effect, including the spatial and temporal scale of interest
32 and the resources or habitat types affected. For example, the causes of species
33 depletion, defined as loss of native species within the exposed ecosystems for all types
34 of ecosystems nationwide, might be analyzed. A smaller-scale example could be a
35 reduction in forest ecosystem productivity for a particular watershed, which might be
2-32
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 caused by more than one stressor.
2
3 The second step in the procedure would be to identify all of the major stressors
4 for which a cause-and-effect linkage could be established or reasonably hypothesized.
5 For this purpose, the stressor list presented in Table 2-1 could be used as a starting
6 point. Depending upon the particular ecological effect being analyzed, groups of
7 stressors could be aggregated to simplify the analysis. Logical aggregations might
8 consist of stressors likely to occur in the same locations or groups of similar classes
9 (such as chemicals) whose independent effects may be small but whose cumulative
10 effects may be significant.
11
12 Once the stressor list is developed, a stress-effect profile, similar to those
13 described earlier, could be used to establish the relative strength of each stressor-effect
14 relationship. If desired, a screening-level analysis could be performed by considering
15 only those stressors that exhibit a strong relationship to the effect. Note that in most
16 cases, the stress-effect profiles generated in this context will not be the same as those
•1 "* developed for the stressor-based analysis because in the latter case the effects profile
represents an aggregation of multiple ecosystem (or landscape-level) effects, rather
19 than the occurrence of a particular effect. This is because the adverse effects on
20 structure, composition, and function of one or more ecosystem types or landscapes
21 were lumped to streamline the national analysis.
22
23 In order to complete the effects-backwards stress-effect profile, it is important
24 explicitly to define high, medium, and low-level effects. In many cases, this step will not
25 be trivial because it entails quantifying or otherwise describing the total effect in order to
26 normalize the y-axis and thereby derive the high, medium, and low effect ranges. In the
27 species depletion example, therefore, the number of species reported extirpated or at
28 significant risk of extirpation within the next decade in all types of terrestrial and aquatic
29 systems could be estimated to provide the 100% line in the normalization. High,
30 medium, and low levels of effect could then be defined accordingly. Alternatively, a
31 rough measure of species at-risk or already affected might be developed for a
32 screening-level analysis; for example:
33
34 high-level effects: the extinction or extirpation of a number of different
35 species in different types of ecological systems
2-33
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 located in many areas;
2
3 medium-level effects: the extinction or extirpation of a few species, species
4 in only a few ecological systems, or species only in a
5 few localized areas; and
6
7 low-level effects: extinction or extirpation of only a few species in a few
8 ecological systems in selected locations.
9
10 In the forest ecosystem productivity example, data may be available to assess the
11 decrease in forest productivity, defined here as the decrease in the amount of energy
12 stored by the forest ecosystem, using measurements of tree diameter, height, volume,
13 basal area, or mean annual increments of growth. High, medium, and low levels of
14 effect would then be normalized to this total.
15
16 Using the information in this stress-effect profile, the next step in the
17 effects-backwards relative risk methodology is to assign an ecosystem- or
18 landscape-level partial score and modify that score with appropriate multiplicative
19 factors as in the stressor-based methodology. Once a numerical score for each
20 significant stressor is developed, the stressor-effect scores can be summed. The
21 proportion of each stressor's partial score relative to the total score provides a measure
22 of the relative importance of the stressor in producing the effect of concern. Note that
23 the particular multiplicative factors may well be different from those presented in the
24 national risk ranking methodology above, because of the characteristics of the particular
25 effect or scale of interest. For example, for species depletion at the national scale, the
26 spatial extent, hot spot, and ecological significance modification factors might remain
27 the same. Other factors might need revision: for example, secondary stress induction
28 might be redefined to refer only to an effect on a keystone species, and duration and
29 species depletion potential would not be relevant. Additional factors, such as
30 secondary effects induction, could be added.
31
32 While the Subcommittee did not further develop these considerations into a
33 more detailed methodology nor apply an effects-backwards methodology for any
34 environmental problem, we suggest the Agency consider instituting research to
35 complete the development and application of an effects-backward methodology to
2-34
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 apportion causes to effects that are manifested in the environment.
2
2-35
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 2.6 References Cited
2
3 Bamthouse, L.W. and G.W. Suter, II (Ed.s.). 1986. User's manual for ecological risk
4 assessment. ORNL-6251. Oak Ridge National Laboratory, Oak Ridge, TN.
5
6 Fava, J. A., L.W. Bamthouse, J. Falco, M.A. Harwell, and K. Reckhow. 1992.
7 Chairpersons' Summary Report on the EPA Draft Document: Framework for
8 Ecological Risk Assessment. EPA/625/R-91/022. U.S. Environmental Protection
9 Agency, Risk Assessment Forum, Washington, DC.
10
11 Harwell, M. A. and J. Gentile. 1992. Report of the EPA Ecological Risk Assessment
12 Guidelines Strategic Planning Workshop, Miami, FL, May 1991.
13 EPA/630/R-92/002. U.S. Environmental Protection Agency, Risk Assessment
14 Forum, Washington, DC.
15
16 Harwell, M. A. and J.R. Kelly. 1986. Ecosystems Research Center Workshop on
17 Ecological Effects from Environmental Stresses. ERC-140, Ecosystems
18 Research Center.
19
20 Harwell, M.A., and C.C. Harwell. 1989 Environmental decision-making in the presence
21 of uncertainty. Chapter 18, pp 517-540. ]n: Levin, S. A., M. A. Harwell, J. R.
22 Kelly, and K. Kimball (Ed.s.). 1989. Ecotoxicology: Problems and Approaches.
23 Advanced Texts in the Ecological Sciences Series. Springer-Verlag, New York.
24 547 pp.
25
26 Harwell, M. A., W. Cooper, and R. Flaak. 1992. Prioritizing ecological and human
27 welfare risks from environmental stresses. Environmental Management
28 16(4) :451-464.
29
30 Kelly, J. R. and M. A. Harwell. 1990. Indicators of ecosystem recovery. Environmental
31 Management 15(5) :527-545.
32
33 Levin, S. A., M.A. Harwell, J.R. Kelly, and K. Kimball. (Ed.s.). 1989. Ecotoxicology:
34 Problems and Approaches. Advanced Texts in the Ecological Sciences Series.
35 Springer-Verlag, New York. 547 pp.
2-36
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 National Research Council. 1983. Risk Assessment in the Federal Government:
2 Managing the Process. National Research Council, National Academy Press,
3 Washington, DC.
4
5 Science Advisory Board. 1990a. The Report of the Ecology and Welfare
6 Subcommittee, Relative Risk Reduction Project. Reducing Risk Appendix A.
7 EPA SAB-EC-90-021A. U.S. Environmental Protection Agency, Science
8 Advisory Board. Washington, DC.
9
10 Science Advisory Board. 1990b. Reducing Risk: Setting Priorities and Strategies for
11 Environmental Protection. SAB-EC-90-021. U.S. Environmental Protection
12 Agency, Science Advisory Board. Washington, DC.
13
14 U.S. Environmental Protection Agency. 1987a. Unfinished Business: A Comparative
15 Assessment of Environmental Problems. Appendix III. Ecological Risk Work
16 Group, Office of Policy Analysis, U.S. Environmental Protection Agency,
••"• Washington, DC.
19 U.S. Environmental Protection Agency. 1987b. Unfinished Business: A Comparative
20 Assessment of Environmental Problems. Overview Report. Office of Policy
21 Analysis, U.S. Environmental Protection Agency, Washington, DC.
22
23 U.S. Environmental Protection Agency. 1992. Framework for Ecological Risk
24 Assessment. EPA/630/R-92/001. Risk Assessment Forum, U.S. Environmental
25 Protection Agency, Washington, DC.
26
27
28
29
30
31
32
33
34
2-37
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Appendix 2A: Ecological Risk Profiles
2-38
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
CHAPTER 3. HEALTH RISK RATING METHODOLOGY
TABLE OF CONTENTS
3.1 Introduction 3-1
3.1.1 Background 3-1
3.1.2 Approach 3-1
3.1.3 Role and Significance of Expert Opinion in Risk Analysis 3-2
3.2. The Environmental Health Risk Rating Methodology 3-4
3.2.1 Nature of the Methodology 3-4
3.2.2 Scope of Stressor List 3-8
3.2.3 Stressor Risk Characterization Data Sheets 3-10
3.2.4 Rating of Confidence in Relative Risk Rating of a Stressor 3-14
3.2.5 Factors Influencing Relative Risk Rating 3-14
3.2.6 Survey Design 3-15
3.3 Analysis and Reporting of Relative Risk Rating Survey Data 3-16
3.4 Implications of Ratings 3-18
3.5 A Fuzzy Logic Approach 3-19
3.6 Extensions and Refinements of the Methodology 3-25
3.7 Summary and Conclusions 3-26
3.8 References Cited 3-27
Appendix 3A. Health Risk Assessment Introduction 3-28
Appendix 3B. Instructions 3-30
Appendix 3C. Risk Characterization Data Sheets 3-33
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 CHAPTER 3. HEALTH RISK RATING METHODOLOGY
2
3 3.1 Introduction
4
5 3.1.1 Background
6
7 Adverse health effects in the U.S. population arise from a wide variety of
8 individual causes and their combinations. EPA is broadly mandated by Congress to
9 address those causes that are environmentally-mediated, i.e., to reduce risks from
10 environmental stressors or conditions that impair health.
11
12 As noted in Chapter 1, EPA initiated its first effort to assign relative risk ratings to
13 environmental problems in 1986 to improve allocation of Agency resources. This effort
14 was reported in Unfinished Business (EPA, 1987). EPA's Science Advisory Board
15 (SAB) expanded upon this work in the report Reducing Risk: Setting Priorities and
16 Strategies for Environmental Protection (SAB, 1990). For that report, the
"* Environmental Health Effects Subcommittee conceptualized health risks as a
-4 multidimensional matrix whose axes comprised variables such as exposure sources,
19 exposure routes, health endpoints, and stressors. It noted that the practical translation
20 of this conceptual structure could be achieved by a relational database design allowing
21 independent ratings of the different variables across various dimensions. At the end of
22 the rating process, each of its elements could then be evaluated independently to
23 achieve a global view of where the major problems resided and where resources could
24 most profitably be directed. This perspective, in essence, has been adopted as a guide
25 to the selection of risk reduction options, a topic discussed at length in Chapter 6.
26
27 3.1.2 Approach
28
29 To meet the requirements of the IRP, the Human Exposure and Health
30 Subcommittee (HEHS) took as its objective the development of a methodology for
31 relative rating of the human health risks from exposures to environmental stressors.
32 The HEHS also viewed the matrix structure as a useful conceptual plan, but, as in
33 1990, was compelled to consider the pragmatic question of its translation into practice.
34 It concluded that any method developed by the Subcommittee should offer, in so far as
3-1
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 possible, certain key attributes. Specifically, the method should be:
2
3 a) as simple as possible to understand, implement, and use;
4 b) capable of gathering judgments from many respondents quickly and
5 inexpensively;
6 c) responsive to the technical knowledge of individual experts and able to
7 capture some of the reasoning of the experts;
8 d) responsive to the concerns of other interested groups; and
9 e) suitable for analysis of correlations among various respondents.
10
11 The method also should explicitly utilize the best existing scientific information on
12 environmental health risks and should be a scientifically based and transparent
13 methodology.
14
15 Because evaluation of sources and source categories was to be a focus of the
16 Risk Reduction Options Subcommittee and because coordination of ecological and
17 human health risks was seen as one of the crucial aims of the current exercise, HEHS
18 adopted the Ecological Risks Subcommittee's strategy of structuring its methodology
19 around environmental stressors, i.e., chemical, biological, and physical agents released
20 into the environment by or modified through human activities. Although many
21 comparative risk projects have been structured around the ranking of problems, such as
22 drinking water contamination or urban air pollution, stressors provide a more
23 homogeneous basis for rating adverse health effects. Furthermore, reducing human
24 environmental health risks eventually requires reductions of exposure to environmental
25, agents or stressors, largely by controlling sources. The analysis of stressors can
26 provide a starting point for the development of policy options.
27
28 3.1.3 Role and Significance of Expert Opinion in Risk Analysis
29
30 Expert opinion plays a special role in the architecture of risk assessment,
31 especially when data are incomplete, contradictory, or multidimensional (Morgan and
32 Henrion, 1990). One method for incorporating expert opinion relies upon a small group
33 of respondents whose implicit or direct aim is to provide a consensus judgment. The
34 Delphi method, described in many publications (Adler and Ziglio, 1996; Linstone and
35 Turoff, 1975^, exemplifies such an approach. In striving for consensus, such a group
3-2
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 exercise may deliberately suppress significant disagreement among members of the
2 expert panel. The extent of disagreement, however, is a crucial piece of information.
3 Presenting only a consensus position is equivalent, in some respects, to presenting
4 experimental data only as arithmetic means unaccompanied by any measures of
5 variability. As noted by Morgan and Keith (1995), "...when uncertainty is high because
6 of fundamentally different views about underlying...processes, a consensus summary
7 may not best serve policy analysis needs. An alternative approach, widely used in
8 applied Bayesian decision analysis, formalizes and quantifies the judgment of individual
9 experts through expert solicitation."
10
11 A high degree of variability in risk judgments among experts chosen for their
12 specialized knowledge may indicate deficiencies in the data upon which judgments are
13 based, or disagreements in how experts interpret the data. Moreover, an aspect of
14 interpretation often ignored is the degree of confidence the judges place in their
15 assigned ratings, a judgment vector, so to speak, that largely reflects the breadth and
16 depth of the available data. It may also reflect the complexity of such data; a rich data
n set may, paradoxically, point in several directions simultaneously.
19 Any approach or methodology aimed at health risk assessment must deal
20 simultaneously with both current questions and future projections. For most currently
21 perceived problems, the primary impediments to establishing a risk ranking or rating
22 reside in the nature of the available data. Although a sparse data base is an obvious
23 source of uncertainty, it is by no means the sole one. For some agents or conditions,
24 even extensive data does not guarantee a shrinkage of uncertainty. Different observers
25 may interpret the same information from different points of view. A process that takes
26 account of both variations in data availability and interpretation is required to conform to
27 current standards of risk characterization such as those described in Understanding
28 fl/s/f(NRC, 1996).
29
30 Another feature that should be incorporated into a health risk rating methodology
31 is geographic flexibility. Although the IRP has focused primarily on national concerns,
32 particularly in its choice of stressors, the risk rating process should be able to
33 accommodate state and local concerns as well. Without some uniformity in how ratings
34 are achieved, consistency between levels of geographic integration will be difficult to
35 achieve.
3-3
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 A third aspect of desirable methodological flexibility is the capacity to adapt to
2 changes in currently perceived or mandated stressors, exposure standards, or public
3 concerns. Barriers to flexibility include factors ranging from the inability to convene
4 expert rating panels to the formidable difficulties of publishing and distributing new
5 printed documents addressing a particular risk.
6
7 The rating process itself, beyond logistics, should also be able to define its limits.
8 An assignment of low risk, for example, must be presented in context. For example,
9 although experts might assign a low overall rating to a specific stressor, such a rating
10 might have to be modified for a specific subpopulation, or when the available
11 information is tentative.
12
13 3.2. The Environmental Health Risk Rating Methodology
14
15 3.2.1 Nature of the Methodology
16
17 The HEHS developed a highly simplified system for polling and characterizing
18 expert judgements of the relative risks of a set of stressors. In the method developed
19 by HEHS, each expert, or other respondent, is asked to rate the health risks of
20 environmental stressors, from "very low" to "very high." Information is provided on the
21 effects of, and population exposures to, each stressor, although the respondent is
22 instructed to draw on his or her own knowledge and scientific judgment to interpret the
23 information provided. The respondent is also asked to identify the determinants of his
24 or her rating. Such determinants might include severity of effect, the size of the
25 affected population, risks to subpopulations, or other factors.
26
27 The final results of the rating exercise include not only the average rating of each
28 stressor, but also the distributions of these ratings, and information on the chief
29 considerations upon which different respondents based their ratings. From these
30 results, one can see not only the average rating for each stressor, but one can also
31 identify those stressors for which there is substantial disagreement and variation in
32 ratings.
33
34 To implement this approach, HEHS constructed a World Wide Web site
35 designed to collect two types of information: a) ratings of health risks, from a list of
3-4
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 environmental stressors, provided by a sample of respondents not constrained to attain
2 consensus, and b) ratings of the confidence these respondents place in their ratings.
3 The Web site design aimed for the five fundamental properties listed in the previous
4 section.
5
6 Another desirable property was the ability to identify the backgrounds of the
7 respondents and to correlate them with their responses. Although this method was
8 designed for the elicitation of expert judgements, an additional advantage of a
9 web-based process is that it can also be used by stakeholders, members of a specific
10 community or interest group, or members of the general public.
11
12 The Entry Page for the Web Survey (Appendix 3A) explains the purpose of the
13 survey and presents the instructions. The respondent then begins the process by
14 registering (Figure 3-1). The respondent enters his or her name, affiliation, and e-mail
15 address. To help ensure the integrity of the process, passwords would be issued to
16 restrict access. In addition, however, each respondent, in transmitting the ratings, also
17 automatically provides the Internet address from which they are sent. The respondent
then proceeds to the rating page, shown in Figure 3-2. The respondent chooses a
19 stressor for rating by a mouse click on the stressor name. Table 3-1 gives the current
20 stressor list.
21
22 Once a stressor is selected, the respondent enters his or her ratings of Risk
23 Rating and Confidence, which range from Very High to Very Low, on the appropriate
24 buttons. The rater can also, by selecting the Information button, access the appropriate
25 Stressor Risk Characterization Data Sheet; risk data sheets on the web site (Appendix
26 3C) summarize key available scientific information on each of the stressors, e.g.,
27 information on exposure route, populations exposed, identified health effects, and so
28 forth. Categories of information that should be presented in the risk data sheets are
29 described further in Section 3.2.3. In rating the stressors, respondents can access the
30 risk data sheet for each stressor and use all, some, or none of this information plus their
31 own expert knowledge to arrive at a risk rating.
3-5
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Stressor Risk Characterization User Name registration.
Please Register your User Name and Password:
Register yoa name ltd pauvtrd HERE. 3« compfete and nrefall
Motto's
Em*
Address:
iOSNWSc
Figure 3-1. Registration form for the risk
survey. Only valid registrants are permitted to
submit ratings.
Stressor Risk Characterization and Health Risk Ranking
Data Sheet
For
Dr. Bernard Weiss
Univ. of Rochester
weiss@envmedrocrtester.edu
If the above information is NOT correct for YOUretumto ttie login page and enter YOUR name and PASSWORD.
Vlew/Edit/Moditv Your Current Stressor Ranking Data
Select a Stressor to Rank: i
Get INFORMATION ON:
Relative
Risk Rank
Confidence in Factors innuencing
Ranking rankmg
Please rank Stressor and indicate r
your confidence to your ranking:
Review of General Instructions
VH
<• H
r M
i- L
c VL
- 7
ChKklll that Apply
f VH
r H
<• M
r L
<• VL
r •>
I r StZ(ofpopidrfonalT«atd
i cfBtiaitarsubpopiilxliora
! r S«vei9y of health effects
P«t«* of
incidence
Pmi^cnc
irtvirontTwrit
Figure 3-2. Risk rating form. The respondent selects a Stressor,
then accesses the data sheet (Appendix 2) by selecting the
Information Window.
3-6
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
.17
B
19
20
21
22
23
24
25,
26
27
28
29
30
31
32
33
34
35
A Review of the General Instructions can also be selected with a click.
Lastly, the respondent checks the appropriate boxes to indicate the factors
influencing his or her rating. One or more factors may be selected. A description of
how these factors may be interpreted is also given in the Instructions (Appendix 3B). A
respondent may also indicate other reasons for the rating and enter a specific
comment. Provision for such entries appears at the bottom of the form and is shown in
Figure 3-3.
The respondent then proceeds systematically through the list of stressors. The
rater can review his or her ratings and accept or modify them, as shown in Figure 3-4.
Finally, once all of the stressors have been rated and accepted by the respondent, the
results are sent via the Internet for analysis with the ratings submitted by other
respondents.
Other reasons for ranking;
General Comments:
1
V.fes»vS^^S!*^^H*<»^'^fe-Saft'4S
1
>';>:W^''<-"^AvSS—j*:f^.^i'i!j:^;3
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
IS
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
gW&iSS^g^tS;:
Edit/Update Risk Ranking Data
For
Dr. Bernard Weiss
Univ. of Rochester
If the above information is MOT correct for YOU return to the Login oacs and enter YOUR name and PASSWORD.
IStressor
Risk Confidence
Rank Rank
General Comments Other Reasons
5
Incidence-of poisonings jn
dtidran
.tiabQenated
Hydrocarbons
tl-Dec-1997
;09:54:16AM
NewdataonreOTductionj
E-Mail a copy of my current Entries to:
Figure 3-4. Selection summary and revision form. This
feature allows the respondent to review and modify previous
selections.
3.2.2 Scope of Stressor List
A multitude of physical, chemical, and biological stressors are associated with
environmental health risks. For the organic compounds alone one can easily
enumerate thousands of compounds distributed among air, water, soil, and food.
Clearly, the Risk Rating Methodology could be overwhelmed by undertaking to evaluate
too large a list of stressors, which would leave the process paralyzed at the stage of
acquisition and development of information data sheets. Such a broad evaluation
would come at the expense of a more thorough examination of those stressors offering
an established reason for potential or actual concern. The proposed list of stressors to
be rated might include the Criteria Pollutants (air, water), Hazardous Waste Priority
3-8
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
20
21
22
23
24
25
26
27
28
29
30
31
32
Chemicals, Persistent Organic Pollutants,
contaminants regulated by emissions or effluent
standards, radionuclides regulated by EPA and the
Department of Energy, common indoor pollutants,
pesticides, metals, indirect stressors (e.g., ozone
depleting chemicals), and emerging issues (e.g.,
hormone-disrupting chemicals).
From a large list of stressors, EPA needs to
dedicate the resources to reduce the stressors to
be rated to a manageable number, probably no
more than 100 at most. This might be done by
aggregating some of the stressors on the basis of
criteria such as commonality of toxic targets,
endpoints, and sources. Since EPA resources will
be limited, especially at the beginning, the Agency
might wish to begin with a deliberately limited list of
stressors, but one whose members intersect a
number of groupings. For example, HEHS
developed a short list of stressors with which to test
and illustrate the methodology (Table 3-1).
This list, with categories that are not always
mutually exclusive, is presented as an example and
should be modified as needed based on review and
analysis of broader lists of pollutants and other
stressors noted above. The approach should be
iterative, with stressors added and deleted over
time based upon new information about exposures,
health risks, and stakeholder concerns.
Table 3-1. List of Current Stressors
Selected for Risk Ratings
Criteria Air Pollutants
Carbon Monoxide
Ozone
PM-10
Heavy Metals
Arsenic
Lead
Mercury
Persistent Organic Chemicals
Dioxins, furans, PCBs
Hormone disrupting chemicals
Pesticides, Herbicides
Environmental Tobacco Smoke
Polycyclic Aromatic
Hydrocarbons
VOCs (Air Toxics)
Benzene
Tetrachloroethylene
Formaldehyde
Bioaerosols
House dust mites allergen and
fungi
Physical Agents
Radon
EMF
Indirect Stressors
CO2 and other greenhouse gases
Freons and other ozone layer
depleting chemicals
3-9
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 3.2.3 Stressor Risk Characterization Data Sheets
2
3 The Stressor Risk Characterization Data Sheets provide key information useful
4 in rating risks. This summary information should be based primarily on studies reported
5 in the peer reviewed scientific literature and presented succinctly. Peer-reviewed
6 comprehensive risk assessments developed by EPA, such as the those for the criteria
7 air pollutants, or those developed for environmental tobacco smoke (EPA, 1992), radon
8 (EPA, 1992), dioxins (EPA, 1995), can provide key conclusions and information for
9 inclusion in the risk data sheets. For stressors not yet subjected to a comprehensive
10 risk assessment, including a review and analysis of the literature, the information
11 included in the data sheet should be based primarily on peer-reviewed papers reported
12 in the scientific literature that vary in experimental designs and protocols. Risk data
13 sheets to be used in the rating process should themselves be subjected to peer review
14 before use and should include references to the sources of information. An example of
15 a Stressor Risk Characterization Data Sheet is shown in Table 3-2, and the full
16 collection of risk data sheets prepared by HEHS is presented in Appendix 3C. The
17 following kinds of information should be included for each risk data sheet, if available:
18
19 a) Exposure routes and pathways pertaining to the Stressor. Route of
20 exposure might be inhalation, ingestion, dermal or some combination,
21 depending upon the particular Stressor. Pathway of exposure should
22 provide some information on how the stressor travels from the source to
23 the human receptor. For example, radon is advectiveiy transported from
24 soil gases into homes and its decay progeny are inhaled. The dioxins are
25, generally emitted from combustion sources, transported through the
26 environment and accumulated in food.
27
28 b) Population exposed to the stressor, e.g., total U.S. population or a
29 special subpopulation(s). For many stressors, e.g., criteria pollutants,
30 virtually all of the U.S. population is exposed to some degree. An example
31 of an exposed sub-population of concern would be subsistence fishermen
32 and their families, who may have a very high exposure to PCBs, dioxins or
33 methylmercury.
34
35 c) Average dose to (or exposure concentration) for the exposed population.
3-10
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 Ideally, the distribution of exposures for the population would be
2 presented here. However, this information is currently available only for a
3 limited number of stressors, e.g., radon. Information on high end
4 exposures in the exposed population should also be described, if
5 possible.
6
7 d) Relevant animal toxicological data, e.g., NOAELs or LOAELs if
8 available. It is generally accepted that an agent that produces an adverse
9 effect in experimental animal studies will potentially pose a hazard to
10 humans following sufficient exposure. In many instances, animal
11 toxicology data can provide supporting evidence of adverse health effects
12 in exposed humans. In the absence of adequate data on the toxicity of an
13 agent in humans, we assume that effects in animal species are probably
14 indicative of effects in humans. This is a default assumption, and is
15 abandoned only when all relevant information on aspects of differential
16 absorption, distribution, metabolism, and toxicokinetics between species
17 have been considered and indicate that effects observed in animals will
not occur in humans. Minor adverse effects will not be discounted unless
19 there is evidence that they are not relevant to humans.
20
21 e) Health effects from environmental exposures. Information on the
22 observed or potential adverse health effects for human populations at
23 current environmental exposure levels should be summarized. Well-
24 designed and conducted epidemiological studies of various
25' sub-populations, such as highly exposed workers or the general
26 population (e.g., PM-10, ozone), which have been peer reviewed, provide
27 good evidence that exposed populations are at risk of adverse health
28 effects.
29
30 f) Biological half life in humans, if known.
31
32 g) Sizes of populations experiencing specific adverse health effects due
33 to current exposures to the stressor. This information will not be available
34 for every stressor.
35
3-11
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 h) Severity and persistence of adverse health effects. For example,
2 mortality, morbidity, and reversibility of endpoint.
3
4 i) Occupational exposure limit(s). (Optional) Although Threshold Limit
5 Values and Permissible Exposure Limits developed for the protection of
6 healthy workers in industry are much higher than would be appropriate for
7 the general population, which includes children and susceptible
8 sub-populations such as asthmatics, these limits are often very close to
9 exposures at which adverse health effects have been observed in human
10 populations. They provide a useful reference point for comparison to
11 average and high-end environmental exposures.
12
13 j) Environmental half-life. Environmentally persistent stressors, such as
14 dioxins and freons, can accumulate in various environmental
15 compartments and may be of greater concern than those which degrade
16 rapidly under environmental conditions.
17
18 k) Comments. Any additional information which might be of use in rating
19 the.relative risk of a particular stressor should be included here.
20
21 Although it will not be possible to provide these kinds of data for every stressor,
22 the data sheets should provide a space for each. Gaps in the data sheets will enable
23 the experts (and the Agency) to easily identify critical gaps in information. In addition, it
24 might be desirable to include an estimate of the fraction of disease incidence
1% attributable to exposures to the environmental stressor. A stressor for which a
26 substantial fraction of the total annual disease incidence is attributable to exposures to
27 the stressor might be rated somewhat higher than one which contributed little to overall
28 incidence. There are, however, some methodological pitfalls in making such estimates,
29 e.g., double counting across two stressors.
30
3-12
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
Table 3-2. Stressor Risk Characterization Data Sheet for PM-10
STRESSOR
Exposure Routes and
Pathways Considered
Population Exposed
Average (Potential)* Dose for
Exposed Populations
Animal Toxicological Data
Health Effects from
Environmental Exposures
Size of Populations
(estimated) experiencing
health effects:
Severity and Persistence of
Adverse Health Effects:
Estimated Fraction of
Disease Incidence
Attributable to
Environmental Exposures (if
Known):
Other Comments
PM-10
(Airborne Paniculate Matter, <10 ^m MMAD )
Inhalation of outdoor air (PM in outdoor air infiltrates into indoor
environments)
All of U.S. population
• - 1 1 m3/day X (9 - 34 ng/m3) = 99 ^g/day to 374 ^g/day
inhaled, depending upon where one lives
• Doses to patients with chronic obstructive pulmonary
disease (COPD) may be 3-times greater than for healthy
adults (U.S. EPA, 1996)
• Increased death rates in elderly from cardiopulmonary
disease
• Increased hospital admissions for COPD (emphysema,
bronchitis) patients
• Increased risk of acute respiratory disease
• Decreased lung function in children and asthmatics
• -60,000 premature deaths in the elderly annually from
cardiopulmonary disease (American Journal of Respiratory
and Critical Care Medicine, 1995)
• Relative risk values range from about 1 .02 to 1 .08.
Although RR is not very high, exposed population is so
large than incidences of health effects large.
• Premature death
• Hospitalizations due to exacerbation of COPD
• Decreased lung function in children and asthmatic is
irreversible effect
PM is estimated to account for about 60,000 premature
cardiovascular deaths annually; this is about 8% of the annual
cardiovascular deaths in the U.S.
REFERENCES:
U.S. EPA. 1996. Air Quality Criteria for Particulate Matter. EPA/600/P-95/001cF, April 1996.
U.S. EPA. 1992. Technical Support Document for the 1992 Citizen's Guide to Radon. U.S.
Environmental Protection Agency.
3-13
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 3.2.4 Rating of Confidence in Relative Risk Rating of a Stressor
2
3 Respondents are asked to indicate their level of confidence in the risk rating
4 given for each stressor. The confidence ratings are Very High, High, Medium, Low and
5 Very Low. This rating can include both the respondent's judgment of the state of
6 existing knowledge as well as the weight accorded his or her individual expertise about
7 a particular stressor. The Committee considered separate ratings for these items but,
8 after much discussion, decided to include only a single confidence rating in the interest
9 of maintaining a simple methodology. If a respondent believes the state of scientific
10 knowledge to be inadequate for a rating, this information can be supplied in the
11 comments section.
12
13 3.2.5 Factors Influencing Relative Risk Rating
14
15 For each stressor rated, the respondent is also asked to indicate the major
16 factors which influenced his or her judgment of relative risk by checking the appropriate
17 boxes. These factors are:
18
19 a) Size of the population affected. A high risk rating might be based on
20 evidence that many members of the population are experiencing adverse
21 health effects from environmental exposures to the stressor, while a low
22 risk rating based on size of population would indicate that only a smalt
23 fraction of the population is being exposed at present.
24
25 b) Particular sub-populations at risk. A check for this factor in
26 combination with a high risk rating for the stressor might arise if some
27 small sub-population(s) are subjected to very high exposures and risks.
28
29 c) Severity and persistence of health effects. A high risk rating might be
30 given for a stressor based on health effects that are life-threatening, life
31 shortening, and/or irreversible. A low risk rating might be assigned to a
32 stressor because the most significant adverse health effect is transient
33 and apparently reversible.
34
35
3-14
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 d) Percentage of attributable incidence. A respondent might assign a high
2 risk rating to a stressor because exposure to the environmental stressor
3 accounts for a significant percentage of the total annual incidence of a
4 given disease or health endpoint. For example, based on the estimated
5 annual number of lung cancers from radon exposures (EPA report on
6 radon, 1992), it can be estimated that about 1 -2 % of the annual incidence
7 of lung cancer in the U.S. is attributable to radon exposures.
8
9 e) Persistence in the environment. A respondent might rate a stressor as
10 very high or high because it is not rapidly degraded in the environment but
11 tends to accumulate and, because of this property, poses an
12 environmental health risk.
13
14 f) Potential future risk. A respondent might rate a particular stressor as
15 very high or high because the stressor is very likely to pose a significant
16 environmental health risk in the future if no actions are taken to reduce
17 anticipated exposures arising from identified sources. The respondent
might also check this factor as influencing his/her low risk rating because
19 he/she judges that no increased health risk is likely to arise in the future.
20
21 g) Other. The respondent is asked to check this if some major factor, other
22 than those provided, influenced his or her rating. The respondent is also
23 asked to identify this other factor(s).
24
25' h) Comments. Experts may have important knowledge that has not yet
26 become widely known. The "Comments" box, as well as the "Other"
27 category, provide means to alert EPA to this knowledge.
28
29 3.2.6 Survey Design
30
31 Implementation of an expert survey requires a specific description of the
32 population to be surveyed and how the sample is drawn. Even the Delphi process, or
33 any other process that relies on a select committee to make judgments or issue ratings
34 is, sometimes implicitly, a hostage to sample selection procedures and possible bias.
35 One virtue of the Web design is the enlarged possibilities it presents for explicit
3-15
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
samples. For example, a risk rating exercise might draw from the total SAB
membership; or, for some purposes, it might sample from the membership of the
Inhalation Specialty Section of the Society of Toxicology or the International Society for
Exposure Analysis; or, its sampling population might come from a specific community.
Because expert groups may be viewed as inherently biased, the ability of EPA to use
an unbiased survey technique should prove useful.
3.3 Analysis and Reporting of Relative Risk Rating Survey Data
A leading purpose of the Web survey instrument is to evaluate the diversity of
rater estimates. The importance of rater diversity is illustrated in Table 3-3, which
shows some of the patterns of rater judgments that would yield equivalent average
(mean) risk ratings.
Table 3-3. Distributions Of Risk Ratings For a Stressor That Would Yield an Average
Rating of "Medium"
Very High
oo
0
ooo
oooo
High
oo
o
oo
0
Medium
00
oooooo
oooooooooo
Low
oo
0
00
o
Very Low
oo
o
ooo
oooo
Patterns of rater responses convey important messages. A stressor whose
ratings are concentrated at the extremes reflects a high degree of polarization among
the population sampled. Similarly, the degree of consensus among raters on the
dimension of confidence also reflects the characteristics of the sample population and
its interpretation of current information. Figure 3-5 depicts four synthesized examples
of the kinds of patterns such plots might produce (based on 12 fictional respondents).
As discussed by the Subcommittee, they convey data potentially helpful to the
formulation of risk options. For example, plot (C) might characterize a phenomenon
such as global warming whose possible health impacts might be widespread and
3-16
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 severe but that currently lacks enough supporting data to elicit much confidence in that
2 evaluation.
A
f
cc.
•5
S
C
?
1
•5
£
VI
•
VI
• •*«
I
«
«
•
• •
C.
i
« •
C
H
•
•
•
atfdtrc
U
orviocn
H
»
•
•
H
;«
VM
•
VM
B
H ?
a «
L £
VI
m D
N ?
B «
L S
VI
VL
VL
I
0
I
C
il
• * •
inndcni
H
oflfldtfi
N
— —
•
H
ec
VH
• ••«
VH
nt
N
•
L
VL
nt
<
I
•
A
Figure 3-5. Synthetic plots of risk versus confidence (12 respondents): A) high
variability; B) low risk, mid to high confidence; C) high risk, low confidence; D) high
risk, high confidence
3 The Subcommittee views distributions of respondents' risk ratings and
4 confidence in those ratings as the key pieces of information provided by the rating
5 exercise. In many instances, these raw data may prove sufficient for conveying the
6, relative degree of risk and confidence associated with each rated stressor. However,
7 the Agency may apply data reduction techniques to these data; for example, many
8 readers of EPA publications will be comfortable with means and standard deviations,
9 which can be expressed by converting rater responses into numerical scores. When
10 the distributions are clearly asymmetrical, Box Plots showing medians and 10th, 25th,
11 75th and 90th percentiles will prove more informative. Other kinds of transformations
12 are also available, but are constrained by the limitation that the ratings are ordinal, not
13 ratio scales. The Agency could also use non-numerical, visual presentations of the
14 results of the survey that would provide a different way of looking at the results, and
15 might entail fewer underlying assumptions than do numerical data reduction techniques.
16
3-17
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 In essence, the Agency is encouraged to develop a set of rules governing how
2 the various combinations of risk/confidence ratings (25 in the current scheme) are
3 ordered. These are subtle judgments, and might be assigned to an expert panel for
4 resolution once the survey data have been gathered.
5
6 A second analysis of the Web data will be based on rater responses to the
7 prompt requesting the factors upon which the ratings were based (Figure 3-2). These
8 data provide considerable amounts of information, but are not readily transformed into a
9 simple score. However, they could be tallied as distributions, and when tallied, would
10 illuminate the risk ratings and provide the Agency with potential initiatives for
11 remediation or regulation, and specific goals for health research.
12
13 One major advantage of a Web survey is the ease of modification. The original
14 design sought to include sources, pathways (exposure route), and endpoints as well as
15 the basis for selection. In this first version, respondents were asked to rank primary,
16 secondary, and tertiary relationships for all variables. For example, several different
17 sources might be responsible for environmental exposure to a particular stressor (e.g.,
18 both transportation and point sources for air pollutants), or a stressor might traverse
19 different exposure routes (e.g., ingestion and inhalation for arsenic), or might affect
20 more than one endpoint (e.g., brain development and blood pressure for lead).
21 Although HEHS deemed this ranking form too complex for its current purposes, it
22 illustrates the flexibility afforded by electronic survey techniques. Some of the
23 modifications that might be incorporated in future versions, at the cost of increased
24 complexity, include provision for rater choices of endpoint, default hazard
25 identifications, default q1* and RfD values, graphical data, and the submission of the
26 rater's own quantitative risk figures. Building in such modifications, however, would
27 constitute a major software design project.
28
29 3.4 Implications of Ratings
30
31 The Relative Risk Rating Survey will provide information from a variety of experts
32 and stakeholders that can yield ratings for a stressor(s) that reflects the category of risk,
33 and assignment of the confidence that the responders to the survey made about the
34 rating. Currently, assignment of a stressor to a priority category has typically been
35 formulated by a single number or classification. Separating risk and reliability estimates
3-18
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 provides a regulator, policy maker, scientist, or other interested party, with the types of
2 information necessary to begin establishing a more meaningful priority for the particular
3 stressor. Based upon the n'sk and confidence ratings, actions or potential responses
4 could range from no additional actions at the time, to new research, or to
5 implementation of new exposure and risk reduction strategies. These should be
6 considered, at the outset, as the initial priorities, and be revisited periodically to retain or
7 to modify the initial risk assignment. At the same time, new stressors can be added to
8 the list based upon new environmental release or formation, exposure, or health data.
9
10 3.5 A Fuzzy Logic Approach
11
12 An important element in the process of soliciting probabilistic judgments about
13 risks from experts is obtaining their views of the uncertainties bounding their estimates.
14 Morgan and Keith (1995) describe how they tried to obtain both mean and standard
15 deviation judgments about climate change from a small group of authorities and
16 Morgan and Henrion (1990) describe several ways by which such information might be
17 presented graphically. The survey technique described in this chapter deals with
uncertainty by eliciting the judge's evaluation of the confidence he or she places in the
19 health risk rating. A low degree of confidence corresponds, in this instance, to wide
20 uncertainty and a high degree to narrow uncertainty.
21
22 Different combinations of risk and confidence ratings, as depicted in Figure 3-5,
23 can help illuminate which stressors deserve closer examination or which might be
24 assigned high or low priorities. Modifications of the survey protocol could be designed,
25 however, to procure indices of uncertainty more directly. Fuzzy logic approaches could
26 prove useful for doing so.
27
28 Fuzzy logic is the term invoked by Lotfi Zadeh, the originator of the discipline, to
29 designate a tactic, in essence, for dealing with uncertainty. As noted by Morgan and
30 Henrion (1990), even experts find it more comfortable to deal with uncertain
31 probabilities by applying verbal labels such as "likely" and "unlikely" rather than by
32 issuing numerical estimates. Fuzzy logic, rather than representing uncertainty in the
33 conventional sense, as in soliciting estimates of variance, deals directly with ambiguities
34 such as those inherent in the terms above. Conventional logic treats every propositon
35 as either true or false. Conventional mathematical definitions of set membership also
3-19
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
rely on binary definitions. Fuzzy systems view truth definitions and fuzzy set
membership as possessing any value in the range [0.0,1.0].
Consider a speaker's description of the weather as "hot." It conveys
considerable information although the adjective is imprecise and subjective and is a
function of the speaker's personal experience and environment. Figure 3-6 is a Fuzzy
depiction of "hot." Almost everyone would consider an ambient temperature of 40°C to
be hot, so that an ambient temperature that high might be assigned a truth value of,
say, 0.90. In Zadeh's terminology, that value corresponds to its membership in the set,
"hot." Below that temperature, your definition might depend largely on where you live.
Some inhabitants of Rochester, New York, would classify a reading of 25°C as "hot,"
but residents of Miami might consider such a reading as somewhat cool. At 25°C,
membership in the set, "hot," might elicit a truth value of 0.15. In traditional, Boolean,
set theory, "hot" would be assigned a crisp definition such as greater than 30°C and
symbolize a clear dividing line certain to evoke debate between Rochesterians and
Floridians.
20
30 40
Temperature (°C)
Figure 3-6. Membership in the class, "Hot," is not fixed,
but depends upon individual definitions. The shading is
not necessary, but is designed to show that, in this
instance, the proportion of judges applying the adjective
is less dense at the lower temperatures.
3-20
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
19
20
21
22
23
24
2§r
26
27
28
29
30
31
32
33
34
35
Fuzzy Rating Patterns
(Best Guess and Bounds)
-StressorA
—StressorS
—StressorC
VL
H
VH
Risk Rating
Figure 3-7. The vertical line, situated there by a
hypothetical rater, can be said to represent divided
membership in the risk classes, "Medium" and
"High." In place of a numerical value, the
respondent's rating can be interpreted as an estimate
of a truth or membership value in the fuzzy set,
"Medium" of about 0.7 and a membership value in
"High" of 0.3.
One way to display the elements of fuzzy logic classification is shown in Figure
3-7. Here, the usual numerical scale has been translated into verbal labels from Very
Low to Very High, in conformity with the categories used earlier. The figure depicts the
set (or rating) of "Medium," for example, as extending from Low to High, with maximum
membership (or truth) value at Medium and minimum membership values at Low and
High. Suppose a judge were asked to locate the point on the abscissa corresponding
to his or her best estimate of risk and marked it at the position designated by the
vertical line. The position of the line indicates majority membership in "medium," and
what might be called minority membership in "high." Roughly translated into
percentages, the line coincides with about 70% membership in "medium" and 30% in
"high."
3-21
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
The estimate defined by the vertical line in Figure 3-7 provides no indication of
the rater's uncertainty about her judgment. As we noted earlier, a rater's uncertainty
can be translated into what statisticians term "confidence limits." Morgan and Keith
(1995) secured analogous information from their judges by asking for standard
deviation estimates. With a fuzzy logic approach, a similar kind of estimate can be
obtained more simply.
Figure 3-8 depicts fuzzy risk ratings of three hypothetical stressors. The triangle
labeled "Stressor A" reflects the rater's judgment that this agent presents a Low risk
that is unlikely to be much lower or higher. The fuzzy rating for "Stressor B" is centered
at Medium, and is asymmetric to reflect the judgment that it could be as small as Low or
as great as Very High. The third triangle, centered between High and Very High, shows
the rater inclined towards the latter, but also locating the lower bound between Low and
Medium.
M H VH
Risk Rating
Figure 3-8. Hypothetical ratings for three different
stressors. The peak represents the rater's best guess
of where along the abscissa the risk belongs; that is,
where membership is believed to be 1.0. The lower
comers of the triangle are set to reflect the rater's
judgment of the lower and upper bounds of risk. To
summarize the judgments of many raters about a
particular Stressor, the abscissa locations are defined
by a scale, say, from 0 to 100 (Figure 9) and the heights
summed at each integer position from 0 to 100.
3-22
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
S
9
10
11
12
13
14
15
16
IT.
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
An example of the kind of visual display that can be used to elicit such
judgments from raters is shown in Figure 3-9. The rater views three horizontal bars,
each equipped with a slider that can be manipulated by a computer mouse. As the
slider is moved, its corresponding value on the abscissa, which ranges from 0 to 100,
appears above the bar. The uppermost bar is used to set the location of the rater's
best guess about the risk posed by the particular stressor. The left bar allows the rater
to set her estimated lower bound, defined as, "Unlikely to be any lower than this." The
right bar sets the location of the estimated upper bound, defined as, "Unlikely to be any
higher than this." Respondents tested so far find it relatively easy to manipulate the
sliders. The triangles plotted as a result may be asymmetric, as shown in Figure 3-7.
For summarizing the responses of a specified sample of raters, the abscissa is divided
into, say, 100 units, and the summed ordinate values then calculated.
Stressor Risk Characterization Ranking Form
Best guess
64
Lower bound Upper bound
Figure 3-9. Depiction of a visual display presented to a
rater asked to indicate where she would position her best
guess of the health risk posed by a specific stressor and
where she would place the lower and upper bounds of her
estimate. With the computer mouse, the respondent
locates the slider at the appropriate numerical index. For
this display, the scale runs from 0 to 100, but a second
parallel scale, could show the corresponding verbal labels
from, say, Very Low to Very High. These numerical
positions can be converted into the kind of visual display
shown in Figure 3-8.
3-23
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 Fuzzy logic can also be applied to the criteria upon which the raters have based
2 their risk estimates. For any particular stressor, they could be asked to also submit
3 their degree of reliance on these criteria. That is, they could be asked to supply the
4 weight accorded them, from Very Low to Very High, in making their risk judgment. For
5 example, the rating for a particular stressor might have been based largely upon the
6 size of the population affected while another stressor rating might have been based
7 largely on outcome severity, and its impact on a particular subpopulation such as
8 children. The weights given these criteria and the associated uncertainty about their
9 relevance can be translated into the kind of responses depicted in Figure 3-8.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
3-24
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 3.6 Extensions and Refinements of the Methodology
2
3 The Health Risk Rating Methodology was developed by the Human Exposure
4 and Health Subcommittee to provide a unique tool for assigning relative risk to selected
5 stressors. Its design was governed by three determinants that are seen as lacking in
6 many risk evaluation exercises. First, it explicitly recognizes uncertainties by requesting
7 raters to provide estimates of their confidence in the available data and in their own
8 grasp of this information. Second, it offers the Agency a mechanism with which to
9 secure risk ratings from designated populations not limited by the constraints of time
10 and geography imposed by the usual committee deliberations. Third, it permits
11 improved estimates of rater variability, a key policy index that often is not available or is
12 explicitly concealed by focusing on consensus.
13
14 For this prototype exercise, the selection of stressors and the characterizations
15 of each stressor in the data sheets were based upon the experiences and professional
16 judgment of the members of the Subcommittee. Because of the limited resources
17 available to it, the Subcommittee was unable to extensively test and refine the
methodology. In practice, the EPA would likely convene a working group and identify
19 resources for conducting the described activities.
20
21 These resources will provide the basis for developing an implementation plan for
22 the proposed rating methodology. This must include: a) selection of the stressors to
23 submit for risk rating at the present time, b) development of a sampling program for
24 expert selection and completion of data sheets, c) statistical methods for combining the
25 risks and the confidence to obtain a weighed priority, d) methods that link the data
26 sheets to basic resources or sources of information or exposure, effects, and other
27 variables, e) approaches for peer review of data sheets prior to implementation of a
28 expert survey, f) activities to promote the availability and use of the Stressor Risk
29 Characterization Data Sheets, ratings and prioritizations of individual stressors by
30 stakeholders, e.g., on the Internet, and g) identification of a timetable for
31 implementation of the project, and completion of respondents on the initial data sets
32 and newly identified stressors.
33
34
3-25
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 3.7 Summary and Conclusions
2
3 a) Risk ratings can inform environmental decision-making, but in and of
4 themselves cannot be the sole basis of these decisions.
5
6 b) Risk ratings are not purely scientific since value judgments concerning the
7 relative importance of a number of factors must be used to arrive at a
8 rating for each stressor. Furthermore, when information on a particular
9 stressor is limited, judgments must be made about the uncertainties and
10 the degree of protectiveness that is appropriate.
11
12 c) Human health risk ratings, however, can be scientifically based, justifiable,
13 well-defined, and transparent.
14
15 d) Risk ratings are traditionally the product of a committee instructed to
16 reach a consensus. Elicitation of the opinions of many individual experts,
17 through a formal process, can provide an alternative approach to rating
18 environmental health risks, especially when data are incomplete,
19 contradictory, or multidimensional. This approach also provides the
20 means to formally address uncertainties in the ratings, and to elicit
21 information on the factors that influenced the ratings.
22
23 e) Development of a single, merged ecological and human health risk list of
24 ratings requires value judgments. Ecological and health risk ratings can
25 be developed within a consistent conceptual framework, with many
26 commonalities in the factors used to rate risks.
27
3-26
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2 3.8 References Cited
3
4 Adler M. and E. Ziglio. 1996. Gazing into the Oracle: The Delphi Method and its
5 Application to Social Policy and Public Health. [London: Jessica Kingsley
6 Publishers]
7
8 tinstone H. A., and Turoff, M. 1975. The Delphi Method: Techniques and Applications.
9 [London: Addison-Wesley Publishing Co.]
10
11 Morgan M. G. and M. Henrion. 1990. Uncertainty: A Guide to Dealing with Uncertainty
12 in Quantitative Risk and Policy Analysis. Cambridge: Cambridge University
13 Press.
14
15 Morgan M.G. and D.W. Keith. 1995. Subjective judgments by climate experts. Environ.
16 Sci. Technol. 29:468A-476A.
17
National Research Council. 1996. Stem, P.C. and H.V. Fineberg (Eds).
19 Understanding Risk: Informing Decisions in a Democratic Society. National
20 Academy Press, Washington, DC.
21
22 Science Advisory Board. 1990. Reducing Risk: Setting Priorities and Strategies for
23 Environmental Protection (EPA-SAB-EC-90-021). U.S. Environmental Protection
24 Agency, Science Advisory Board. Washington, DC.
25
26 U.S. Environmental Protection Agency. 1987. Unfinished Business: A Comparative
27 Assessment of Environmental Problems. Office of Policy Analysis, U.S.
28 Environmental Protection Agency, Washington, DC.
29
30 U.S. Environmental Protection Agency. 1992. RADON REPORT??
31
3-27
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
i Appendix 3A. Health Risk Assessment Introduction
2
3 Thank you for agreeing to participate in this survey. Its purpose is to inquire about how you
4 rate the health risks of a variety of environmental stressors. It also asks about the degree of
5 confidence you place in your ratings or in the available scientific data. We are also interested in
6 the reasons for your risk ratings. This information will be incorporated into a report for EPA's
7 Integrated Risk Project, an undertaking of the Science Advisory Board.
8 Here is how it works. At the bottom of this page, you select Go to the Register Form. An
9 entry form will appear. To register using the Registration Form, enter the information
10 requested and choose a password, as indicated. You can also practice using the system by
11 entering test and fesf for the username and password, respectively. When you select Login,
12 the Stressor Risk Characterization and Health Risk Rating Data Sheet will appear. On the
13 same page, you can select Review of General Instructions, which describes your task. It also
14 contains descriptions of some of the criteria (Factors) you may have used to assign your
15 ratings. After reviewing the instructions, you Exit the page (or click on Go to Stressor Risk
16 Characterization Data Sheet) and proceed to rate the stressors. When you choose Select a
17 Stressor to Rate, a list of stressors will appear. Once you have made a selection, you are
18 asked to make a rating of Health Risk from Very High to Very Low. Your will also mark your
19 confidence in your rating. "Confidence" is a way of dealing with the largely subjective nature of
20 such ratings, and permits us to join combinations of risk and confidence ratings for the
21 assignments of priorities or ratings. For each of the stressors, we have tried to assemble the
22 most relevant information in a highly compressed format. You access this information by
23 selecting Open Information Window. After you have reviewed this information, you return to
24 the rating form either by selecting Close Information Window (if you have the rating form
25 open concurrently) or by exiting the information page.
26 Once you have rated the first Stressor, you then go on to the next Stressor in the list. If you do
27 not wish to rate a particular sstressor either because of lack of information or because of the
28 nature of the Stressor, there is no reason for you to do so; that is important information in itself.
29 We have provided opportunities for you to offer additional reasons for your ratings and for
30 general comments.
31 We have confined this demonstration to a limited number of stressors. The forms can be
32 expanded. Try it so we can see how it works and how we might want to proceed.
33 When you have finished making your selections, check your answers and then push the "Send
34 Results" button. That will mail the answers to me (Bernie Weiss). For technical details or
35 problems e-mail Geoff Inglis (inglis@envmed.rochester.edu).
36 You may send additional comments on these pages to Bernie Weiss. Your comments will be
37 posted for the group to read on the "comments page"
3-28
-------
IRP Integrated Draft Report-Peer Review Draft, May 3, 1999— Do Not Cite or Quote
Go to the Register Form
If \/mi arp alrpaHv rociistpfffri' ___ —
3 Go to the Login Form
4
5
6
7
3-29
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
i Appendix 3B. Instructions
2
3 For each environmental stressor, please indicate:
4 Your expert judgment of the relative environmental health risk that it poses to the
5 U.S. population, either current or potential future risk. Assume that all current
6 controls and regulations remain in place for the rating. Please rate as:
7 Very High (VH), High (H), Medium (M), Low (L) and Very Low (VL) or (?).
8
9 We have provided some information on each stressor in a "Risk Data Sheet." For
10 the relative risk rating, please use your expert judgment, and all, some, or none of
11 the information provided on the risk data sheet.
12 If you feel that you cannot assign a relative risk rating to any given stressor, for
13 any reason, please indicate this with a " ?" in the entry for rank.
14 Your level of confidence in your rating.
15 Very High (VH), High (H), Medium (M), Low (L) and Very Low (VL) or (?}.
16 Your confidence in the relative health risk rank that you assign to a given
17 stressor can include both your judgment of the state of the existing scientific
18 knowledge about the health risks of the stressor plus a judgment of your own
19 expertise regarding any particular stressor.
20 Please indicate also (by checking) the major factor or factors that influenced
21 your relative risk rating. For example, one particular factor might have been
22 decisive. Alternatively, a collection of several factors might have determined your
23 rating.
24 Check all factors that apply.
25 Size of population affected: A high risk rating might be based on evidence that
26 many members of the population are experiencing adverse health effects due to
27 environmental exposures to the stressor. Conversely, a low risk rating might be
28 based on evidence that few, if any, members of the population are experiencing
29 adverse health effects as a result of exposure to the stressor.
30 Particular subpopulations : Although most of the population may not be exposed
31 to levels of the stressor great enough to cause adverse health effects, a rating of
32 high risk might arise if some smaller sub-populations are subjected to very high
33 exposures and risks; or, the stressor is might be viewed as acting on an
34 especially susceptible subpopulation. A rating of low risk might arise if little
3-30
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 evidence points to high exposures or risks for any subpopulations or if extreme
2 sensitivity has not been identified in any subpopulations.
3 Severity of health effects: A high risk rating might be based on the presence of
4 pronounced, irreversible, or life-threatening adverse health effects irrespective of
5 the size of population affected. In parallel, a low risk rating might be based on
6 mild, transient health consequences arising from exposure to the stressor
7 especially if they occur in a limited number of individuals.
8 Percent of attributable incidence: A high risk rating might arise if a significant,
9 though small, proportion of the total U.S. incidence of a health effect is seen to
10 arise from exposures to the environmental stressor; e.g., it is estimated that
11 about 1-2% of the annual lung cancer incidence in the U.S. is due to radon. In
12 contrast, a low risk rating might be based on lack of evidence that a significant
13 proportion of the incidence or prevalence of an identified health effect can
14 reasonably be attributed to the stressor.
15 Persistence in the environment: A high risk rating might follow if the stressor is
16 believed not to degrade rapidly in the environment but, instead, tends to
accumulate over time in one or more environmental compartments. A low risk
rating might follow if the stressor is believed to degrade relatively rapidly in the
19 environment.
20 Potential future risk: A high risk rating might be based on predictions of a
21 significant potential for environmental health risks in the future if no actions are
22 taken to reduce anticipated exposures arising from identified sources. A low risk
23 rating might be based on the assumption that no rise in or relatively minor
24 increases in exposures to the stressor are likely to be seen in the future.
25 Other. Please describe briefly any other reason for your ratings.
26 Comments: Provide, if you wish, comments further amplifying or describing the
27 reasons for your choice of ratings.
28 To Summarize: Please be sure that you have responded to all three questions:
29 3Your relative risk rating
30 3Your confidence in your rating
31 SThe factors determining your responses
32
33
34
3-31
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 Go to Stressor Risk Introduction page.
2
3 Go to Stressor Risk Characterization Ranking Sheet
4
5
6
7
3-32
-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
I
2 Appendix 3C. Risk Characterization Data Sheets
3-33
-------
PART III INPUTS TO ENVIRONMENTAL DECISION-MAKING:
ECONOMICS AND VALUATION
-------
IRP Draft integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 PART III—INPUTS TO ENVIRONMENTAL DECISION-MAKING:
2 ECONOMICS AND VALUATION
3
4 Preface
5
6
7 Integrated environmental decision-making at its heart is about how people,
8 acting as individuals or through their government, can make themselves best off
9 through actions that affect the environment directly and indirectly. Judging which
10 among alternative actions will produce the most "well being" requires information not
11 only on risks, discussed in the previous chapters, but also on the definitions of "well
12 being", i.e., goals, and the relationships among goals and among alternative choices or
13 actions. Just as the IED framework requires that we look at the full range of risks, it is
14 also necessary to consider the full economic consequences of decisions of whether and
15 how to address certain environmental risks or sets of risks. An understanding of the
16 tradeoffs implied by these choices is crucial both during Problem Formulation and
Analysis and Decision-Making.
18
19 The Economic Analysis Subcommittee (EAS) was given the task of describing
20 the ways in which benefit/cost analysis can be used to frame choices about different
21 uses of societal (public and private) resources. In addition, at the request of the Deputy
22 Administrator and in recognition of the fact that some environmental changes are
23 difficult to monetize and value, a complementary subcommittee, the Valuation
24 Subcommittee, was formed to focus explicitly on improved methods for assessing
25 ecosystem values and incorporating that information into decision-making. The VS was
26 a multi-disciplinary group that included ecologists, economists, and other social
27 scientists. The work of these subcommittees is contained in the following two chapters,
28 which taken together provide a framework within which choices about the environment
29 and its protection can be understood and implemented to best effect.
30
31 As discussed in previous chapters, environmental decision- making requires an
32 understanding of the physical and biological relationships between stressors-
33 substances and conditions that can cause harm-and organisms or ecological systems
34 of concern, including the relationship between alternative levels of these stressors (or of
Ill-i
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 substitutes for them) and their effects. This is the province of science, and with other
2 factors included, is known by the term "risk assessment."
3
4 By itself, however, such information is inadequate to guide decisions. The
5 reason is that there are a near-infinite number of substances and conditions that can
6 and do impose harm, and it is impossible and unwise to act to eliminate or even to
7 reduce them all. The ability to lessen the risks they pose is limited by the resources of
8 labor, capital, knowledge (technology), and physical endowment that is devoted to the
9 task. And these same resources are those on which people depend to provide the
10 other things desired and indeed necessary for life, including food, shelter, fiber,
11 manufactured goods, amenities, and investment to provide for future populations and
12 for increases in ability to satisfy wants in the future. A critical task during Problem
13 Formulation, therefore, is to choose which of the risks (potentials for harm) or
14 combinations of risks are to be acted upon, and to what degree; i.e., to set
15 environmental or risk reduction goals. Intimately related, of course, is the choice of the
16 most effective and efficient mechanisms and instruments to be used in actually
17 accomplishing the task; these aspects of options selection are discussed in Part IV.
18
19 The premise our democratic society has selected as the fundamental building
20 block for such choices is that they should be guided by the individual goals and desires
21 of people as expressed directly and in and through their governmental and other
22 institutions. Logic then suggests that the task of the decision-maker is to maximize the
23 attainment of these goals and desires with respect to both environment and all the other
24 things that people want. These include, of course, things material and non-material,
25 and also people's goals for future generations and in fulfilling their stewardship
26 responsibilities for natural systems. Put directly, then, it follows that the task of
27 environmental protection activities is to provide-for now and for the future-the
28 healthiest, safest, most ecologically secure set of conditions that the American people
29 are willing to pay for (in terms of other things they must give up to get them).
30
31 Chapters 4 and 5 provide a framework to guide decision-makers to deliver on
32 this task.
33
34
l-n
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Chapter 4 adopts the commonsense framework for decision-making of
2 comparing the advantages (benefits) and disadvantages (costs) of the range of
3 plausible actions presented by a situation. It first addresses the difficult question of how
4 to sort out the true benefits and costs of actions designed to reduce risks and improve
5 the environment, and in the process identifies the many ways in which unsophisticated
6 efforts to assess economic consequences both underestimate and overestimate some
7 elements of benefits and costs, and totally ignore others. It then discusses techniques
8 for measuring and eliciting the benefits and costs of environmental actions, and
9 describes the requirements for and difficulties and uncertainties of such techniques.
10
11 Chapter 4 demonstrates that the development of benefit/cost measures in terms
12 useful to decision makers is not a simple task. It requires disciplined treatment of a
13 number of factors. For one thing, the treatment of a stream of benefits and costs as
14 they are realized over time is crucially important. The chapter concludes that
15 discounting of those streams in a consistent way to show current decision-makers the
16 full implications (present and future) of alternative choices is essential to provide
17 meaningful information. It also recognizes that this discounting process can be
controversial, and that it cannot be undertaken blindly.
19
20 Another matter of concern in presenting benefit/cost information is how to treat
21 the fact that while analysis can demonstrate which risk management scenario provides
22 the greatest net benefits, it cannot take account of the fact that the distribution of the
23 costs and benefits is not uniform. Sometimes the costs are borne by some people and
24 the benefits gained by others. More frequently, some may lose a little and others a lot,
25 while at the same time others are benefiting disproportionately or not at all.
26 Complicating all of this is the fact that individuals affected by the environmental action
27 are unequally endowed in wealth and income. This influences both the benefits and
28 costs as they enter the evaluation framework, and the sense of fairness that pervade
29 the results. Chapter 4 recognizes that these distributional issues cannot be handled
30 within the framework of benefit/cost analysis, but it asserts that this makes them no less
31 important in the choice of actions to be taken. Consequently, distributional issues are
32 seen as matters on which the analysis can shed light, but which must be taken into
33 account outside the analysis at the final decision stage.
34
35 Chapter 4 conveys a mixture of confidence and humility. The framework of
Ill-iii
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 benefits and costs is seen as a consistent, coherent, transparent, robust tool that
2 provides decision-makers with a firm basis on which to proceed. When well done, the
3 products of specific benefit/cost analyses are seen as important inputs for decisions,
4 sufficient to the task to be relied upon with confidence most of the time, always limited
5 by data and sometimes by methodology, and sometimes sufficiently uncertain as to
6 provide only indicative information—which nonetheless can be useful.
7
8 As with risk assessment and comparison, however, even the best and most
9 complete benefit/cost analyses are also shown to be inadequate in themselves to yield
10 "answers." Other factors always are present and must be taken into account. In
11 essence, integrated environmental decision-making requires consideration of the broad
12 implications of actions taken to improve the environment across stressors and across
13 options to relieve them. The framework outlined in Chapter 4 provides a
14 methodological approach to such decision-making.
15
16 Chapter 5 is a vital complement to Chapter 4. It addresses the important issue
17 of how the benefits (and sometimes costs) of changes in ecological outcomes can be
18 properly determined and incorporated into the decision process. By their nature,
19 changes in ecological conditions are often not easily observed in quantifiable terms that
20 can be rendered in the monetary units that most often are used to compare possible
21 outcomes in a benefit/cost analysis. For this reason, it is often thought that the benefits
22 received from ecological protection are under-estimated, and consequently, that
23 ecological systems are under-protected as compared to other goals and desires of
24 people.
25,
26 Chapter 5 examines the issue of ecological valuation in detail. In doing so, it
27 starts with an examination of ecological values and concludes that in principle they are
28 not different from other values, and that they need not enter into the decision process in
29 any unique way. At the same time, however, measuring and incorporating the values
30 ascribed to anthropogenic changes in ecological conditions do present serious
31 difficulties that require that special care be taken.
32
33 For one thing, citizens doing the valuing often have insufficient knowledge of how
34 changes in ecological factors affect the things they care about. They need expert
35 scientific assistance in making the important connections. For another, many of the
Ill-iv
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 benefits from ecological systems are subtle and do not enter into conscious
2 consideration in the normal course of events, as do market goods and services. For
3 these, special techniques for eliciting preferences are required. In addition, some of the
4 values ascribed to ecological outcomes arise in a social context and may not surface
5 from commonly used individual preference measures. Equity, sustainability, and
6 stewardship are often cited as examples of such values.
7
8 Most importantly, because ecological services often do not enter into markets or
9 enter incompletely, it is difficult if not impossible to provide monetized measures of the
10 benefits they deliver. Often it is not possible to even determine quantitative measures
11 for differences in outcomes. Qualitative measures of benefits and costs, therefore,
12 must be arrived at and then incorporated into decision processes. In short, the
13 valuation of ecological costs and benefits is prone to error because elements valued by
14 people may be omitted or incorrectly specified and because measurement is inherently
15 more difficult than with goods and services for which market and market-like measures
16 are available.
17
Chapter 5 offers suggestions to practitioners of benefit/cost analysis and
19 decision-makers that will improve the incorporation of ecological matters into decisions.
20. It goes further, however, in making two recommendations. The first is that there be an
21 expanded use of deliberative processes to assure that all relevant elements are
22 included in decisions, and that they are valued properly. While deliberative processes
23 are seen to have an important role, Chapter 5 also notes that such processes should be
24 used discriminately and be tailored to the situation. When this is done, deliberative
25, processes will not only generally lead to outcomes that are more satisfactory and
26 robust, but also will not cause inordinate delay and indeed may even speed the delivery
27 of ecological protection because they lessen post-decision controversy.
28
29 The other key recommendation of Chapter 5 is for further research on and
30 experimentation with additional approaches to valuing benefits of environmental
31 systems. Existing approaches are seen to be inadequate in their treatment of such
32 values as fairness and sustainability. They also have difficulty in incorporating the
33 systemic benefits of such matters as biodiversity, and are incomplete in their treatment
34 of dynamic responses to change. More holistic approaches that take account of the
35 web of interactions involved in decisions surrounding ecological systems and their
Ill-v
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 interactions with other production and consumption activities would be helpful.
2
3 The overall message from Chapters 4 and 5 is that integrated environmental
4 decision-making requires a framework within which the decision-maker can meld the
5 results of science and the goals and values of the people served and formulate
6 acceptable decisions. The further message of these chapters is that such a framework
7 exists, but that its use requires skill and artistry, and indeed that in some respects the
8 framework itself remains a work in progress.
9
Ill-vi
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
CHAPTER 4. BENEFIT/COST ANALYSIS FOR INTEGRATED RISK
DECISIONS
TABLE OF CONTENTS
4.1 Introduction 4-1
4.2 Fundamental Questions in the Economic Analysis of Risk 4-3
4.3 The Benefits of Risk Reduction 4-8
4.3.1 Revealed Preference 4-13
4.3.2 Stated Preference 4-15
4.4 Costs of Environmental Protection 4-17
4.5 Comparing Total Benefits and Total Costs 4-22
4.5.1 Calculating Net Benefits 4-22
4.5.2 Present Values and Discounting 4-22
4.5.3 The Net Benefit Criterion and the Scope of Public Projects 4-24
4.5.4 Uncertainty in the Measurement of Benefits and Costs 4-26
4.6 Distributional Considerations 4-27
4.7 Conclusions 4-28
4.8 References Cited 4-31
Endnotes 4-33
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 CHAPTER 4. BENEFIT/COST ANALYSIS FOR INTEGRATED RISK
2 DECISIONS
3
4
5 4.1 Introduction
6
7 No society, no matter how wealthy, has available the resources to "solve" every
8 problem that confronts it. Schools will never be as good, nor food, clothing, shelter and
9 access to medical care as abundant as we would all like. Similarly, no society-modem
10 or otherwise - will ever be free of all environmentally mediated risks to human health
11 and ecosystems. Thus, it is of great importance to be able to identify the most serious
12 threats to health and the environment, and also to target certain of them for possible
13 mitigation.
14
15 Identifying what we might call the total risks associated with air and water
16 pollution, solid and hazardous waste disposal, drinking water contamination and other
activities is largely the province of risk assessment. Determining these risks requires us
to ascertain which pollutants are extant in the ambient environment, the route(s)
19 through which humans or ecosystems are exposed to these pollutants and the duration
20 of these exposures, and the resulting consequences. At a minimum, this requires the
21 skills of engineers and chemists; atmospheric, terrestrial, and aquatic scientists;
22 toxicologists, epidemiologists and biostatisticians, and clinical health specialists; and
23 ecologists (including botanists, forest and fisheries scientists and others). Some risks
24 may be described quantitatively, occasionally with some statistical precision. Often, we
25 can do no more than suggest that a particular pollutant may have adverse effects on
26 human populations, or on ecological systems. As described in Part II, assessment and
27 comparison of risks associated with different stressors or activities is an important part
28 of both formulating "the problem" and analyzing possible risk reduction options.
29
30 Economics typically concerns itself with the way in which people allocate scarce
31 resources among many competing needs so as to make themselves-broadly
32 speaking-as well off as possible. As such, the possible contributions of economics to
33 Integrated Environmental Decision-making lie largely-though not exclusively- in
34 deciding which risks it makes sense for society to make incremental changes in
4-1
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 (sometimes called "marginal" changes) once they have been characterized, in deciding
2 to what degree those risks ought to be reduced, and in identifying appropriate policy
3 instruments that can be employed to achieve such risk reductions.
4
5 Economics, along with other behavioral sciences, can also help in understanding
6 risks by characterizing the extent to which changes in behavior can effect exposure to
7 stressors. For example, although many people spend time outdoors on a clear day, it
8 would be incorrect to count them as the population at risk on a very polluted day - for
9 the simple reason that some of those people would alter their behavior on account of
10 the pollution. Similarly, contamination of an aquifer used for drinking water may well
11 increase some people's risk of acute or chronic illness, but may not affect all users if
12 everyone were informed about the risk and most chose instead to drink bottled water.
13 In both of these cases, economics would suggest that there were "damages" associated
14 with the pollution, but these damages can take forms other than just increased risk of
15 illness.
16
17 It is very important to note the distinction drawn above between risk posed by a
18 certain exposure level and the possible "changes" in risk that would result from control
19 actions that reduce either stressor or exposure levels, or both. Economic analysis is
20 especially useful in determining which changes in risk level make sense to pursue via
21 policy measures and which do not.
22
23 The focus of this chapter is on the role of economic analysis, particularly
24 benefit-cost analysis (or BCA), in helping society to decide which environmental risks to
25 * address first and which to leave for a later time. The IED framework recognizes that
26 such analysis should be an important component of decision-making, but not the sole
27 consideration. In fact, there are instances where benefit-cost considerations do not
28 play a direct role in decision-making. For instance, although it would be most unusual,
29 decisions about which risks to address and how to address them could be determined
30 in a national plebiscite in which voters directly established risk management priorities;
31 voters in California often do make decisions in statewide referenda (including-decisions
32 on important environmental matters) that are elsewhere made by legislatures, or even
33 by administrative agencies to which legislatures have delegated power. As suggested
34 above, such decisions are more frequently made by elected officials, who sometimes
35 not only specify which environmental risks are to be controlled, but also to what level of
4-2
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 residual risk and using which control technologies or techniques.
2
3 Most commonly, however, risk management decisions are made by appointed
4 officials and the civil servants who work for them, acting under statutes that delegate to
5 federal, state or local regulatory agencies the power to establish priorities and
6 determine who must take actions to reduce particular risks. Their decisions can be,
7 and generally are, informed by a variety of types of analysis, of which economic
8 analysis is one. The IED framework is intended to help ensure that the broadest range
9 of analyses are applied in an integrated fashion.
10
11 4.2 Fundamental Questions in the Economic Analysis of Risk
12
13 Once risks have been characterized and compared, at least two important
14 questions remain. How much, if at all, should each of the risks be reduced, and in what
15 order? How should these reductions be accomplished? Economics offers insights on
16 the answers to both questions. Let us address the second question first, because it is
17 perhaps the easier of the two. Because resources are scarce relative to human wants
even in the richest of societies, economists answer the second question - how shall
19 risk reductions be accomplished? -- quite directly. Generally, their answer is, "In the
20 least expensive way possible." For, by reducing as cheaply as possible the risks
21 associated with indoor radon, heavy metals in aquatic ecosystems, or radioactive
22 contamination in landfills, to pick but three examples, more of society's resources are
23 available for education, national defense, health care or the private goods and services
24 that people desire. So long as we accomplish what we set out to do (in risk reduction),
25 why spend more than we have to?
26
27 In fact, there is a "cousin" of BCA - called cost-effectiveness analysis (CEA)~
28 that deals explicitly with this question. If, and this is a big "if," the only output of a
29 regulatory program is to reduce cancer incidence, say, CEA would rank cancer
30 prevention programs on the basis of "cost per cancer case avoided." Then the least
31 expensive option would be pursued first, followed by the next most attractive one, and
32 so on until a decision had been made to no longer pursue reductions in cancer risk.
33 The point at which this process stops is the subject of a subsequent section. This
34 approach would maximize the reduction in cancer incidence for the amount of money
35 spent. Other patterns of spending the same amount would prevent fewer cancers.
4-3
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 There are several problems with CEA that bear mention here. First, it only works
2 for "apples vs. apples" comparisons. If one program reduces cancer cases and also
3 improves the visibility of the air, while another reduces cancer and protects aquatic
4 ecosystems, too, they cannot be compared on a simple cost per cancer case avoided
5 basis. Most environmental regulatory programs have multiple outputs (that is, they
6 provide more than one kind of beneficial effect). Thus, CEA is often not of much
7 practical assistance in risk ranking and management. Another problem with CEA was
8 alluded to above. While it is easy to see the common sense in starting with the
9 cheapest control opportunities first, and working our way up the list, CEA provides no
10 obvious "stopping rule" unless Congress, for instance, has limited the size or cost of the
11 cleanup effort. That is, it does not tell us when we have reached a cost per cancer
12 case avoided that is too much from society's standpoint.
13
14 Still another problem with CEA is that it treats cases of cancer - or other
15 endpoints - as being homogenous, when in fact they may not be. For example,
16 suppose the cheapest way to prevent cancer cases is to pay the cost of smokers'
17 participation in smoking cessation programs that they would not otherwise attend.
18 Some members of society, perhaps many, might object to such a program on the
19 grounds that these risks are bome voluntarily - at least more voluntarily, say, than
20 those that arise from airborne exposures to other carcinogens. Even though the latter
21 might be more expensive to reduce, society might prefer to address them first. In other
22 words, we care about dimensions of risk other than sheer statistical magnitude. In
23 addition to the voluntary or involuntary nature of the risk, these other dimensions may
24 include such things as the degree of "dread" associated with the risk; for instance, for
25, whatever reason, people seem to fear radiation-induced cancers more than those
26 associated with other causes. Despite this and its other limitations, CEA has been a
27 very useful tool in assisting regulatory officials interested in rationalizing risk reduction
28 programs.
29
30 One final word about doing things for least cost. The attractiveness of this idea
31 is the reason why economists and others are often enthusiastic about what have come
32 to be called market-based or economic-incentive approaches to environmental policy.
33 These include such things as taxes on pollution, tradable discharge permits, and
34 deposit-refund schemes, among others. The attraction of these approaches is that they
35 are able not only to provide incentives to meet an overall environmental goal (mercury
4-4
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 emissions are reduced to x tons per year, for instance), but also that the costs of
2 complying with this limit are minimized across society. Furthermore, the burden of
3 identifying the least cost options falls on the private sector, which is especially good at
4 reducing costs. This is not the place to go into the reasons why this is so, though it is
5 worth pointing out that such approaches have been used quite successfully in the
6 United States (for the phase out of leaded gasoline in the 1980's, and for the control of
7 sulfur dioxide in the 1990 amendments to the Clean Air Act) and in a number of
8 countries in Europe and elsewhere around the world.
9
10 This brings us to the question of how much reduction we should seek for the
11 various risks that society faces. There are several ways this question can be answered.
12 One approach would be to direct the regulatory authorities to set standards for the risky
13 pollutants so as to provide a margin of safety against adverse health effects - an
14 approach premised on thresholds. This is the directive Congress has given to the EPA
15 in important parts of the Clean Air Act, the Clean Water Act, the Resource Conservation
16 and Recovery Act, and in other environmental statutes as well.
17
This seemingly attractive approach often founders when the underlying science
19 suggests - as it often does - that there is no "safe" level; that is, a threshold
20 concentration cannot be found that guarantees the health of affected individuals or
21 ecosystems. This issue has arisen recently in the context of the revision of the National
22 Ambient Air Quality Standard for ozone, and the establishment of a new standard for
23 particulate matter of 2.5 microns and less. In both cases, the Clean Air Science
24 Advisory Committee has indicated that no clear "bright line" can be found such that
25- concentrations below that level can be regarded as being safe.
26
27 There is another problem with the threshold approach. It might sometimes be
28 the case that even if a safe level could be found, society might deem it too expensive to
29 provide, especially if a somewhat less protective level costs substantially less. A
30 safe-standard approach denies the legitimacy of such tradeoffs. Once again, this issue
31 has arisen in the context of air quality standard setting, with the EPA having publicly
32 acknowledged that setting the ozone standard at too low a level could have significant
33 adverse economic repercussions.
34
35 Another way to decide the question of how much protection to provide is to tie
4-5
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 the answer to affordability. That is, one could decide that any risk reductions that
2 regulated entities (whether corporations, lower levels of government, or individuals) can
3 "afford" are worth undertaking. This is the approach often taken when the EPA writes
4 discharge standards for new sources under the Clean Air or Clean Water Acts. Indeed,
5 they have been directed to do so by Congress, which has drafted statutes in which
6 "economic achievability" is one of the relevant criteria. While this approach may seem
7 to deal satisfactorily with all economic concerns, this is not so. First, controls on some
8 sources might be desirable even if the source could not afford them - in some
9 instances, completely shutting down a plant may turn out to be in the best interests of
10 society if there is no other alternative measure, and if that plant discharges substances
11 that poses serious threats to human health and/or the environment.
12
13 Second, some controls may be affordable but nevertheless not worth pursuing.
14 This would be the case for relatively trivial risks that would be expensive to ameliorate.
15 Even if the sources of these risks were quite profitable, many people would be
16 uncomfortable spending a lot of money on not-so-serious problems. The problem with
17 requiring successful firms to do more than those that are just breaking even is that it
18 creates exactly the wrong incentives for firms: do well and you will be more heavily
19 regulated, do poorly and we'll take it easy on you. That is neither the right
20 environmental nor the right economic signal to be sending.
21
22 There is, of course, a third way to make decisions about which risks to control
23 and by how much. This approach involves an effort to strike a balance between, on the
24 one hand, the good that will be done for human health and the environment when one
25 , or more risks are reduced and, on the other, the costs or other adverse consequences
26 associated with taking such actions. The criticisms above of both the threshold and the
27 affordability approaches go right to the heart of this matter.
28
29 Benefit-cost analysis is the technique most commonly used to assist in balancing
30 the favorable effects of risk reductions (the benefits) with the adverse consequences
31 (the costs). Underlying BCA is the notion that it may be possible to measure both the
32 good and bad effects of a policy change (a measure to reduce one or more risks in this
33 case) using a common denominator - money. If so, one can then ascertain whether
34 the gains to the gainers outweigh the losses to the losers, and thus determine, on net,
35 whether society as a whole is made better or worse off as a result of the change. A
4-6
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 whole host of possible risk reduction policies could then be compared on the basis of
2 which ones do the greatest net good and which do little or no net good.
3
4 We offer several observations about BCA before moving on to a discussion
5 about how one might value benefits and costs, make comparisons among them, and
6 take into account effects that are felt at different points in time.
7
8 First, and perhaps most importantly, most economists think of BCA as no more
9 than a tool to assist in decision making. This is important because most
10 non-economists (and perhaps some economists) believe that BCA must be used as a
11 decision rule. That is, they believe that the use of BCA means that only those policy
12 changes that pass a benefit-cost test should be put in place. To repeat, most
13 economists do not hold this position, although all economists would argue that the
14 information in a well-done BCA can be of great value in helping to make decisions
15 about risk reduction policies (Arrow et al., 1996).
16
17 Second, when trying to translate benefits and costs into dollar terms, it is the
preference of individuals that count. That is, we look to each individual to determine
19 how to value a given benefit or cost.
20
21 This is often a hard pill for non-economists to swallow. Some people reason that
22 if the public knew what they themselves know about a particular problem, surely
23 everyone would ascribe greater benefits to solving the problem than they do currently.
24 And sometimes it is the case that willingness to pay to reduce a particular risk is small
25 solely because affected parties do not understand the consequences. Often, however,
26 they understand the problem quite well, but simply do not care as much about the
27 effects as someone else. It is unclear how one would attempt to value the
28 consequences of a policy change to someone without attempting to determine that
29 person's considered valuation of the change.
30
31 Another difficulty with BCA has to do with its reliance on individuals' valuations,
32 however they might be revealed. When willingness to pay is used as the monetary
33 value of a change, as it generally is, the values that individuals reveal are conditioned
34 by their incomes. Thus, an individual will be willing to pay no more each year for a risk
35 reduction than the total amount of money he/she has available after taxes. Accordingly,
4-7
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 the social benefit (as counted in a BCA) of reducing the risk of premature mortality for a
2 wealthy person will generally be greater than doing so for a poor person (because the
3 former will be able to pay more for the same statistical reduction in risk than the latter).
4 This is a problematic aspect of BCA for almost all economists. From time to time,
5 people have attempted to derive weights that might be used to inflate the preferences
6 of the poor so that they count for more in BCA. The problem with this approach is
7 obvious to all - who shall be delegated the responsibility of assigning the weights, and
8 on what basis should these weights be derived? The final section of this chapter takes
9 up such questions related to the distribution of benefits and costs.
10
11 We turn now to a discussion of the benefits of risk reduction policies, and how
12 they can be estimated and expressed in dollar terms.
13
14 4.3 The Benefits of Risk Reduction
15
16 Economic valuation of environmental resources has been characterized by some
17 other disciplines as being anthropocentric and utilitarian, and inattentive to the intrinsic
18 value of these resources. To some degree this is true. If an environmental change
19 never matters in any way to any human - today or in the future - then it will not, even in
20 principle, show up in any economic valuation or assessment. As Freeman (1997)
21 points out, however, it is important to distinguish between ecosystem functions (e.g.,
22 photosynthesis, absorption, and dispersal) and the environmental services produced by
23 ecosystems that are valued by humans. The range of these services is great. They
24 include obvious environmental products such as food or fiber and services such as
25 flood protection, but also include the quality of recreational experiences, the aesthetics
26 of the landscape, and such desires (for whatever reasons) as the protection of marine
27 mammals. Humans may "passively" value environmental services. For example,
28 "existence value" reflects human recognition of the so-called "intrinsic" value of an
29 ecosystem. If an environmental measure produces a change that matters to humans,
30 for whatever reasons, then it is an item to be counted in the assessment, at least in
31 principle. In practice, there are, however, numerous problems. Table 4-1 presents
32 examples of environmental and resource service flows for which value measures might
33 be desired (Freeman, 1993).
34
35 Policies for environmental protection often affect the functioning of ecosystems.
4-8
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 The social benefits of this protection depend on the link between these functions and
2 some associated service flow that is valued by humans. In most cases an
3 environmental protection measure will enhance environmental services marginally, and
4 it is this change in services that economists seek to value. Before the benefits of
5 environmental protection can be calculated, therefore, it is necessary to predict the
6 actual consequences on natural systems from failing to engage in protective activities.
7 Good science is essential to competent environmental benefit-cost analysis.
8
9 In some cases the physical consequences of pollution are clear-cut and easy to
10 quantify. For example, a single "point source" of water pollution upstream renders
11 cooling water for a downstream factory sufficiently impure that special pre-treatment
12 equipment must be installed before the water can be used. In other cases, however,
13 the connection is far less clear. Deteriorating air quality due to pollutants from a wide
14 range of sources (vehicles, factories, and perhaps even natural vegetation) may reduce
15 local average life expectancy by several months for susceptible individuals. About the
16 most that can be claimed is that the poor air quality changes the probability that any
17 given individual will live to be, say, at least 70 years of age. But so many other
1 behavioral and genetic factors influence the individual's actual life-span that it is
19 impossible to say for sure what would be any individual's actual benefits from reducing
20 air pollution. Often, it is not even scientifically clear what effect reducing pollution from
21 one source (say vehicles) will actually have on overall ambient pollutant levels.
22
23 In still other cases, such as the dramatic alteration of ecosystems through
24 destruction of wetlands or through global warming, scientists are uncertain even as to
25 the nature of the outcome. In these cases, economic valuation of potential gains from
26 environmental protection measures falls short. This is not because benefit-cost
27 methods ignore environmental services that do not produce market goods, but because
28 the pathways by which humans are ultimately affected cannot be well articulated.
29
30 It is obviously very much easier to value the benefits of environmental protection
31 when the effect is straightforward, as in the first example, but increasingly difficult as
32 the scientific uncertainty about the physical effects increases. Even in the intermediate
33 case, where the environmental service effects of the policy in question can be
34 articulated no more clearly than as a change in average life expectancies or a change
35 in the probability of respiratory disease, the task of ascertaining individuals' values of
4-9
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 these changes is somewhat complicated.
2
3 From an economic perspective, the environment can be viewed as a form of
4 natural asset that provides service flows used by people in the production of things like
5 recreation, agricultural output, the assimilation of pollutants, or even an amorphous
6 'good" such as "quality of life." This is analogous to the manner in which real physical
7 capital assets (e.g., factories and equipment) provide service flows used in
8 manufacturing production. Like real physical capital, if the natural environment (as a
4-10
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
TABLE 4-1. Examples of Environmental and Resource Service Flows
Resource or environmental media
(Air, Water, Fishery, Forest)
Effects
Direct impacts on humans
Human health (morbidity/mortality associated with pollution)
Odor, visibility, visual aesthetic
Ecosystem impacts (biological mechanisms)
Impacts on the economic productivity of ecological systems
Agricultural productivity
Forestry
Commercial fisheries
Other ecosystem impacts
Recreational uses of ecosystems-fishing, hunting
Ecological diversity, stability
Impacts through nonliving systems
Materials damage, soiling, production costs
Weather, climate
Economic channels
Market values
Changes in income to producers
Changes in the availability/price for marketed goods/
services to consumers.
Non-Market Values, i.e., changes in availability of:
Health
Environmental amenities
Visibility
Opportunities for recreation
Adapted from Freeman, 1993, pp. 13-14
productive asset) is allowed to deteriorate, this lessens the flow of services it is capable
of providing.
A decision to devote society's scarce resources to environmental protection,
however, means that these resources will not be available to be used for other
purposes. These other purposes might be more food produced today or greater current
manufacturing output, or it might be more investment in productive capital equipment or
research to enhance output in the future, all of which are also valuable to society. It is
4-11
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 not surprising that poor countries spend little on environmental protection. Banning
2 DDT, or other harmful but cheap and effective pesticides, has little appeal for people
3 whose current battle to protect themselves from immediate illness and to feed
4 themselves today depends on these chemicals. The concept of valuation of
5 environmental goods and services is couched in terms of society's willingness to make
6 trade-offs between competing uses of limited resources, and by aggregating over
7 individuals' willingness to make these trade-offs.
8
9 Economists' tools of valuation were originally developed in a more limited
10 context, one in which policy changes mostly caused changes in an individual's income
11 and/or prices that he faced in the market. Over the last twenty years, however, these
12 ideas have been extended to accommodate changes in the qualities of goods, to public
13 goods that are "shared" by individuals, and to other nonmarket services such as
14 environmental quality and human health. Many environmental goods and services are
15 considered by economists to be public goods because one person's enjoyment of them
16 does not diminish the ability of others to enjoy them also.
17
18 We can think of people as having preferences among alternative "bundles" that
19 include both market and nonmarket goods. Some of these nonmarket goods include
20 environmental services. Typically people's preferences allow substitution - if the
21 quantity of one good in a bundle is decreased, the quantity of some other good would
22 have to be increased in order to leave the individual just as well off as before the
23 change. This substitutability between goods establishes the trade-off between pairs of
24 goods that both matter to people. If one of the goods has a monetary value, then this
25 willingness to make a trade-off reveals the monetary value of the other good. It is not
26 necessary that money be the metric for comparison, it is merely convenient.
27
28 When the alternative 'bundle" includes something dramatic, such as loss of life
29 or loss of a loved one, the individual will be understandably unwilling to make trade-offs
30 with other goods. However, environmental actions rarely, if ever, imply such dramatic
31 outcomes with certainty. An environmental protection measure is more likely to alter
32 slightly the probability of illness or loss of life, much like wearing a seat belt affects the
33 probability of death, and people are observably willing to trade off money and time to
34 alter this probability.
35
4-12
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Rarely is it necessary to determine the value of the environment as a whole. The
2 typical challenge is to ascertain the social value derived from maintaining some type of
3 ecosystem function at its current level through pollution prevention or increasing its
4 functions somewhat through more stringent regulations. Failing to implement the policy
5 would compromise the ecosystem's functioning to some extent, thereby lessening the
6 quantity of services the environment is capable of providing for human enjoyment.
7
8 When it is possible to observe freely functioning markets for goods whose
9 availability is affected by policy, economists are fairly comfortable with the
10 appropriateness of their standard tools for estimating the benefits or loss of benefits
11 resulting from that policy. If a policy results in higher incomes or lower prices for them,
12 people are generally made better off. Higher incomes or lower prices mean an
13 increased difference between what the individual is willing to pay for a good and the
14 amount that he is required to pay. This difference is called a "surplus," and modem
15 welfare economics recognizes that welfare gains (or losses) as a result of some policy
16 are approximated well by increases (or decreases) in this surplus that result from the
17 policy's implementation.
19 The same logic still holds despite the absence of standard markets for many
20 environmental services. Standard markets make the task easier because consumers'
21 decisions as to how much of a good to purchase at different market prices helps reveal
22 something about the surplus they gain at any given price. With non-market
23 environmental goods, it is necessary to infer this willingness to trade off dollars for
24 additional quantities of environmental services using other techniques. Environmental
25 economists have developed a repertoire of techniques over the past thirty years that fall
26 roughly into two categories: a) indirect measurement (revealed preference), and b)
27 direct questioning (stated preference). The following is a brief survey.
28
29 4.3.1 Revealed Preference
30
31 Economists have always preferred to measure trade-offs by the observed
32 decisions of consumers in real markets whenever possible. These are called "revealed
33 preference" methods, because the consumers' actions reveal something about their
34 willingness to trade-off one good for another. In other words, we get to see them put
35 their money where their preferences are. The task is made easier if the researcher can
4-13
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 exploit relationships that might exist between the non-marketed (environmental) good
2 and a good that has a market price. For example, suppose a change in the
3 environmental good only matters to a particular person if he is also a purchaser of
4 some good that has a price (e.g., water quality is important to those who buy fishing
5 rods). Then the environmental good and this market good can be seen to be
6 complements of sorts. The individuals' behavior with respect to the marketed good,
7 when the environmental quality changes, will reveal something about how he values
8 that environmental change. Likewise, if the individual can mitigate the negative
9 consequences of the decline in some environmental service by purchasing some
10 market good (buying bottled water when aquifers are contaminated, for instance), then
11 the environmental good and the market good are substitutes of some sort. We can
12 leam something about the lost benefits of the environmental service from these
13 individuals' actions with regard to this substitute market good.
14
15 First, the Travel Cost Method (TCM) is perhaps the oldest of these revealed
16 preferences methods for non-market valuation and relies on the complementary
17 relationship. An individual will decide whether to take a trip (generally a recreational
18 trip) to consume an environmental good if the perceived benefits of the visit exceed the
19 costs of the visit. In this case, environmental service might be viewed as a quality
20 characteristic of the trip that matters oniy if the individual visits the relevant site. By
21 observing how people trade off such things as distance traveled to access sites of
22 better environmental quality, researchers can leam something about the value people
23 place on differences in this quality.
24
25 Second, economic methods that model individuals' decisions to avert or mitigate
26 the consequences of environmental deterioration may shed light on how people value
27 other types of changes in environmental quality. These techniques are most applicable
28 where the failure to institute a given policy would increase risk of illness or loss of life,
29 although they have other applications as well. For example, if a household installs filter
30 systems or purchases bottled water to avoid contaminants in the groundwater that they
31 drink, then they are attempting to avert or mitigate the consequences of water quality
32 deterioration. By analyzing this type of behavior, economists can at the least establish
33 bounds on the willingness of individuals to pay for improved water quality. In this case,
34 bottled water is a substitute for clean groundwater in altering the risk of illness from
35 drinking water.
4-14
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Third, in some cases, individuals reveal their preferences for environmental
2 goods in the housing market Realtors know that the price paid for a property varies
3 with the characteristics of that property. Additional bathrooms or bedrooms or a better
4 view will show up in the transaction price of a dwelling, but local environmental
5 characteristics will show up as well. For example, researchers have found that housing
6 prices vary with air quality in many cities. Home-buyers' willingness to pay for this
7 additional environmental quality reveal something about their willingness to trade-off
8 money for this type of environmental improvement. This approach is called the hedonic
9 property value method of valuation.
10
11 Fourth, wages for similar jobs in different locations may vary with the local level
12 of environmental quality. Higher wages may have to be paid to induce workers to take
13 jobs in areas with more pollution, greater health risks, or fewer environmental
14 amenities. In fact, the values used by the EPA and other regulatory agencies for
15 reductions in mortality risk come almost exclusively from studies linking compensation
16 to the riskiness of the job. Through the hedonic wage method, it is sometimes possible
1"7 to attribute differences in wages to differences in the quality of the environment, and
thereby indirectly to value variations in environmental quality.
19
20 The above methods are well established for measuring the conceptual trade-offs
21 that economists consider as the basis of environmental valuation. However they are
22 applicable only in special cases. For the first method - the travel cost approach - to be
23 complete, individuals must care about changes in the environmental good only if they
24 visit a particular site. For the second to be complete, (that dealing with complementary
25 purchases), individuals' values for changes in environmental goods must depend solely
26 on their ability to produce some ultimate service which can be produced as well using
27 market goods, but at some increased cost. In all of these methods, individuals must be
28 able to discern the environmental consequences and the researcher must be able to
29 find a marketed good with a special relationship to the environmental good that can be
30 exploited. These conditions hold only in a subset of the cases in which environmental
31 protection measures need to be evaluated.
32
33 4.3.2 Stated Preference
34
35 Economists generally mistrust trade-off information that has not been observed
4-15
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 (even indirectly) in real markets, but for some changes in environmental goods and
2 services, we simply cannot observe behavior that reveals their values to people. This is
3 particularly true when the value is a passive one. For example, an individual may value
4 a change in an environmental good because he wants to preserve the option of
5 consuming it in the future (option demand) or because he desires to preserve the good
6 for his heirs (bequest demand). Still others envision no current or future personal use
7 by themselves or their heirs, but still wish to protect the good because they believe it
8 should be protected or because they derive satisfaction from knowing it exists in a
9 protected state (existence demand). With no standard market trade-offs to observe,
10 economists often resort to surveys in which they construct hypothetical markets. Value
11 information elicited this way is described as "stated preference" (as opposed to
12 revealed preference) information.
13
14 In this approach (often called contingent or hypothetical valuation), survey
15 respondents are presented with a well-defined scenario that requires them,
16 hypothetically, to trade-off something (generally money) for a change in the
17 environmental good or service in question. While rarely asked in such a direct way,
18 researchers seek answers to questions such as: "How much would you be willing to
19 pay to prevent an additional 5% of the population of sea otters from being killed due to
20 a hazardous waste spill?" (Or, alternatively, what compensation would you require in
21 order to willingly accept a reduction in this environmental amenity?) Early attempts to
22 use this method were fraught with problems. As might be imagined, the values that
23 result are susceptible to manipulation through the manner in which the valuation
24 questions are posed. Protocols have been continuously refined, however, and the
25 reliability of these methods is improving. (For further discussion of the problems
26 associated with elicitation of values, see Chapter 5).
27
28 An alternative but related method is to infer values from individuals' hypothetical
29 behavior (as opposed to hypothetical payments). This approach can be particularly
30 useful when paying to prevent environmental degradation seems too implausible or
31 subject to bias. Respondents might be asked "If visibility were x miles and travel costs
32 were $y, how many trips would you take to site z?" In yet another approach (contingent
33 choice models), respondents are given descriptions of several scenarios and asked
34 which one they would prefer. One characteristic of each scenario is a level of prices or
35 net income and another is environmental quality, allowing the researcher to infer
4-16
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 something about respondents' willingness to trade dollars for variations in the level of
2 environmental amenities.
3
4 The types of methods described briefly here can sometimes be used in
5 conjunction with one another, either to measure different types of values that are
6 generated by a change in an environmental resource or to better estimate one type of
7 value through added information. Whatever methods are used, they are least
8 successful when the protection measure to be valued has little-understood, amorphous,
9 or very long-term consequences. In these cases it is difficult to articulate to individuals
10 the way in which the policy will affect society.
11
12 4.4 Costs of Environmental Protection
13
14 The task of estimating the costs of specific environmental protection efforts may
15 seem straightforward, compared with the conceptual problems and empirical difficulties
16 associated with estimating environmental protection benefits. To some degree, this is
17 true. In most cases, it is easier to develop cost estimates than benefit estimates to a
commensurate degree of precision and reliability. But as we move towards developing
19 more precise and reliable cost estimates, significant conceptual and empirical issues
20 arise.
21
22 The economist's notion of cost, or more precisely, "opportunity cost," is linked
23 with- but distinct from-everyday usage of the word, "cost." In this part of the chapter,
24 we explore what this concept means in the context of environmental protection efforts,
25 and we briefly examine the ways in which it can be made operational through
26 quantitative, empirical analysis.
27
28 Conceptually, there are four steps required to appraise the cost of an
29 environmental-protection measure. First, we need to identify the specific policy
30 instrument that is associated with the measure. This is because the same target, such
31 as a given reduction in ambient pollutant concentration, may be achieved at very
32 different total costs with different policy instruments. For example, it is well known that
33 under a variety of circumstances, a market-based policy instrument - such as an
34 emission tax or a tradeable permit system - can enable a regulated sector to achieve
35 an ambient target at relatively low aggregate cost, compared with a conventional
4-17
-------
1RP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 standard that has the effect of requiring all sources to adopt the same abatement
2 technology (Baumol and Gates 1988). This example leads to the second conceptual
3 step: identifying the specific actions that sources will take to comply with the statute or
4 regulation as implemented with the given policy instrument. Some of these actions may
5 involve the adoption of a new piece of equipment, but others may involve a change in
6 process. Third, it is necessary to identify the true cost of each action, which - as we
7 emphasize below - requires much more than assessing required monetary outlays in
8 an accounting sense. Fourth, it is necessary to aggregate these costs across society
9 and over the relevant time frame.
10
11 Economists take cost to be an indication of what must be sacrificed in order to
12 obtain something (through purchase, exchange, or production). The concept of
13 opportunity cost provides a measure of the value of all of the things that must be
14 sacrificed if one or more risks are to be reduced. Opportunity costs typically do not
15 coincide with an accountant's measure of costs, namely monetary outlays. This may
16 simply be because out-of-pocket costs do not capture all of the explicit and implicit
17 costs that have been incurred, such as the cost of time associated with waiting in line
18 for a vehicle inspection. Or it may be because the monetary prices of the resources
19 required to produce a good or service - in our case, environmental quality - may
20 themselves provide inaccurate indications of the opportunity costs of those resources,
21 because of failures by markets that result in transaction prices not fully reflecting social
22 values. Likewise, it is important to distinguish between the private costs of some good
23 or service to producers or consumers of that good or service, and the social costs
24 imposed on society as a whole. The cost concept that is relevant for environmental
25 policy analysis refers to overall social opportunity costs. Thus, the costs of
26 environmental protection are essentially the forgone social benefits due to employing
27 scarce resources for environmental protection purposes, instead of putting these
28 resources to their next best use.
29
30 Thus, costs and benefits are two sides of the same coin. Environmental benefits
31 are created by taking some environmental policy action, while other (presumably largely
32 non-environmental) benefits are thereby foregone. Hence, in keeping with the definition
33 of benefits established above, the cost of an environmental protection measure may be
34 defined as the (gross) decrease in consumer and producer surpluses1 associated with
35 the measure and with any price and/or income changes that may result (Cropper and
4-18
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Dates, 1992).
2
3 Environmental protection measures are intended to have desirable
4 consequences, labeled "benefits," and are likely also to have undesirable
5 consequences, labeled "costs," but there are many situations in which there may be
6 some ambiguity regarding whether a specific regulatory consequence should be
7 counted as an increased cost or a decreased benefit. Take, for example, an
8 environmental regulation that provides direct protection of human health (benefits) and
9 requires expenditures of capital and labor (costs) to achieve this purpose. But the
10 positive health impacts of the regulation may have the effect of increasing labor
11 productivity, thereby decreasing the costs of producing other goods and services. Is
12 this a decrease in the costs of the regulation or an increase in its benefits? To avoid
13 this sort of confusion, economists argue for the use of net benefits (benefits minus
14 costs), as opposed to benefit-cost ratios as an evaluative criteria (see below).
15
16 With these definitions in mind, we provide in Table 4-2 a view of the costs of
17 environmental protection measures, beginning with the most obvious and moving
towards the least direct.2 First, many policy makers and much of the general public
19 would identify the on-budget costs to government of administering (monitoring and
20 enforcing) environmental laws and regulations as the cost of environmental regulation.
21 Most economic analysts, on the other hand, would identify the capital and operating
22 expenditures associated with regulatory compliance as a substantial portion of the
23 overall costs of regulation, although a considerable share of compliance costs for some
24 regulations fall on governments rather than private firms - a good example being the
25' regulation of contaminants in drinking water, a Federal regulation, the cost of which is
26 borne primarily by municipal governments. Additional direct costs include legal and
27 other transaction costs, the effects of refocused management attention, and the
28 possibility of disrupted production.
29
30 Next, there are what have sometimes been called "negative costs" (in our
31 conceptual framework, non-environmental benefits) of environmental regulation,
32 including the beneficial productivity impacts of a cleaner environment and the potential
33 innovation-stimulating effects of regulation.3 "General equilibrium" or multi-market
34 effects associated with product substitution, discouraged investment4, and retarded
35 innovation constitute another important layer of costs5, as do the transition costs of
4-19
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29.
30
31
32
33
34
35
36
37
38
39
TABLE 4-2. Costs of Environmental Regulation
Government Administration of Environmental Statutes and Regulations
Monitoring
Enforcement
Private Sector Compliance Expenditures
Capital
Operating
Other Direct Costs
Legal and Other Transactional
Shifted Management Focus
Disrupted Production
General Equilibrium Effects
Product Substitution
Discouraged Investment
Retarded Innovation
Transition Costs
Temporary Unemployment
Obsolete Capital
Social Impacts
Loss and Change of Jobs for Some Workers
Economic Security Impacts
Source: Jaffe et al., 1995
real-world economies responding over time to regulatory changes. Finally, there are
other potential social impacts that are given substantial weight in political forums,
including those involving jobs and economic security.
Within the category of direct compliance costs, business expenditures for
pollution abatement in the United States represent, on average, about 61 percent of
total direct abatement costs (Rutledge and Leonard, 1992); personal consumption
abatement, 11 percent; government abatement, 23 percent; government regulation and
monitoring, 2 percent; and research and development, 3 percent. These averages
must be taken with the appropriate grain of salt, since - as suggested earlier -
4-20
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 measuring even direct costs is by no means a trivial undertaking. For example, there is
2 significant variation between cost estimates produced by the U.S. Environmental
3 Protection Agency (1990) and the U.S. Department of Commerce (1993) for these
4 activities.
5
6 There are a number of potential problems of interpretation associated with any
7 cost data. For example, the questionnaire used by the U.S. Department of Commerce
8 (1993) to collect data for its Pollution Abatement Costs and Expenditures (PACE)
9 survey6 asked corporate and government officials how capital expenditures compare to
10 what they would have been in the absence of environmental regulations. This creates
11 two problems.7 The first involves the determination of an appropriate baseline. Absent
12 any regulation, firms might still engage in some - perhaps a great deal of - pollution
13 control to limit tort liability, stay on good terms with communities in which they are
14 located, maintain a good environmental image, etc. Should such expenditures be
15 included or excluded in the no-regulation baseline?
16
17 Second, when additional capital expenditures are made for end-of-the-pipe
abatement equipment, respondents have relatively little difficulty in calculating these
19 expenditures. But when new capital equipment is installed, which has the effect of both
20 reducing emissions and improving the final product or enhancing the efficiency with
21 which it is produced, it is far more difficult to calculate how much of the expenditures
22 are attributable to environmental standards.8
23
24 In the give-and-take of environmental policy debates, it has frequently been the
25 case that abatement costs of proposed regulations have been over-estimated
26 (Hammitt, 1997). This may be due partly to the adversarial nature of the environmental
27 policy process, but it is also a natural consequence of employing short-term cost
28 analyses that do not take into account potential, future cost savings due to
29 technological change, some of which may be endogenous to the regulatory regime.
30
31 As we said at the outset, the task of estimating the costs of environmental
32 protection efforts is relatively straightforward, at least compared with that of estimating
33 environmental protection benefits, but producing high-quality cost estimates still
34 requires careful analysis.
35
4-21
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 4.5 Comparing Total Benefits and Total Costs
2
3 After the streams of benefits and costs associated with a pollution control
4 program have been monetized, benefit-cost analysis involves evaluation according to
5 the discounted sum of the net benefits it provides. This section discusses how net
6 benefits are computed, and then aggregated over time. It also discusses in greater
7 detail what it means for a program to "pass a benefit-cost test."
8
9 4.5.1 Calculating Net Benefits
10
11 The net benefits of a program in a given year (year T) are the difference
12 between the total benefits of the program in that year and its total costs in that year. In
13 the case in which most of the costs of the program are incurred in the early years,
14 possibly in the form of capital investment, while the benefits of the program are spread
15 over time, net benefits may initially be negative, but eventually become positive. In a
16 case where the annual costs and benefits of the program are constant over time, net
17 benefits will also be constant.
18
19 Two caveats are in order. When net benefits are computed in each year, it is
20 customary to express them in constant dollars (in real terms). That is, any increases in
21 costs and benefits due to inflation are subtracted out, so that costs and benefits in each
22 year are expressed in, for example, 1997 dollars. A second caveat concerns
23 comparing constant dollars at different points in time. When adding the dollar value of
24 net benefits over several years, it is important to realize that people are not indifferent
25 between receiving a dollar in 1997 and a dollar in the year 2007. The dollar received in
26 the year 1997, if invested at 5%, will grow to $(1.05)10 = $1.63, by the year 2007.
27 Equivalently, a dollar received in the year 2007 is worth only $1/(1.05)'° = 610 today,
28 since this is all one need invest today to obtain $1 in the year 2007. For this reason,
29 economists express all future net benefits in terms of the present year's dollars; that is,
30 they compute the present value of net benefits in each year before aggregating net
31 benefits.
32
33 4.5.2 Present Values and Discounting
34
35 The present value of a dollar of net benefits occurring 10 years from today is the
4-22
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 amount that must be invested today to yield one dollar in ten years. Generally, the
2 present discounted value of $1 received in t years is $1/(1+r)', where V is the annual
3 social rate of discount.
4
5 How should the annual social rate of discount (V) be determined? Many
6 economists would argue that T should represent the real rate of return on a riskless
7 capital investment. By real return, we mean the return actually paid minus the rate of
8 inflation. To illustrate, in 1980 the rate paid on a Treasury-Bill was about 11.6%, but the
9 rate of inflation (as measured by the GDP deflator) averaged 9.4%. The real return on
10 government bonds was therefore 2.2%. Since we assume that benefits and costs will
11 be measured in real terms (adjusted for the rate of inflation), discount rates should be
12 measured in real terms also. The reason for using a riskless rate of return (such as the
13 rate of return on Treasury-Bills) is that the rate of return applied to public projects
14 should not include a risk premium. This is because public investments are financed by
15 all members of society, thus spreading the risk of the investment.
16
17 There is, however, another way to estimate the social discount rate. This is to
say that the costs and benefits in year T should be discounted to the present by a
19 consumption discount factor - the rate at which people are willing to trade consumption
20 in year T for consumption today. In a world of perfect markets, the consumption
21 discount factor should equal the discount factor 1/(1+r)' where V is the return on capital
22 investment. There are, however, cases where the two may differ, and this leads to
23 complications.
24
25 For simplicity, suppose that we measure the annual social rate of discount ("r")
26 by the real rate of return on riskless investments. In the U.S., this has ranged
27 historically from 0% to 4%. This suggests that the appropriate social discount rate lies
28 in this interval. Whether a rate of 0% or 4% is used, however, will often make a great
29 deal of difference to the magnitude of the present value of net benefits. A higher
30 discount rate, other things equal, will reduce the net present value of a program. A
31 discount rate of 2% for example, implies that one dollar received in 20 years is worth
32 670 today, whereas it is worth only 460 at a discount rate of 4%. Furthermore, the
33 discounted sum of the net benefits of a program will be more sensitive to changes in
34 the discount rate the longer the horizon over which benefits and costs extend and the
35 greater the disparity in the time profiles of benefits and costs.
4-23
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 While the concept of discounting has a sound rationale, it can lead to
2 conclusions that many people, including economists, find unpalatable. Even at a
3 discount rate of 2% per annum, a dollar of benefits received 200 years from now counts
4 for less than 20 today. This leads to the conclusion that the present value of net
5 benefits will sometimes be negative for projects whose costs are concentrated in the
6 present but whose benefits stretch far into the future. To avoid rejecting such
7 future-oriented programs, some people have called for not discounting future costs and
8 benefits at all, especially for projects with long horizons. At first, this might seem to be
9 a course of action that would favor future generations. In an important sense, however,
10 it does not. If, by using a zero discount rate, we adopt programs that do not pay off
11 until the distant future, we are implicitly passing up opportunities to invest in other
12 projects that yield higher rates of return. Since these projects may increase the capital
13 available to future generations, it is not clear that we have made them better off by
14 using a zero discount rate.
15
16 4.5.3 The Net Benefit Criterion and the Scope of Public Projects
17
18 Suppose that we have calculated the present value of the net benefits of an
19 environmental project in each year of the project's life. Adding these together will
20 produce the discounted sum of net benefits of the project.9 What can be done with this
21 information? As stated at the beginning of this section, positive net benefits imply that,
22 at its current level, the project increases social welfare, in the sense that the gainers
23 could, in principle, reimburse the losers and still have something left over for
24 themselves (see Section 4.5.I). This does not, however, mean that the size of the
25 project is optimal. Figure 4-1 shows how the present value of the marginal benefits and
26 marginal costs of an environmental project vary with the scope of the project. In the
27 illustration, the scope of the project is the reduction in pollutant emissions from a given
28 baseline. The optimal reduction in emissions, according to the diagram, is "A" ~ the
29 point where the increase in benefits (or marginal benefits) from increasing the percent
30 reduction, just equal the additional cost of the reduction. Beyond this point, the
31 additional cost of further reductions exceeds the marginal benefits of the reductions.
32
33 The total benefits of a given pollutant reduction, however, exceed the total costs
34 of the reduction at program levels that are much larger than optimal. The total benefits
35 of the reduction - the area under the marginal benefit curve - equal or exceed the total
4-24
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
8
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Marginal
Cost
Marginal
Benefit
B
Reduction in Pollutant Emissions
Figure 4-1. Marginal costs and marginal benefits associated with an
environmental project.
costs of the reduction - the area under the marginal cost curve - for all reductions from
0 to "B". Thus, even programs that propose reductions greater than optimal still pass a
benefit-cost test.
The lesson of Rgure 4-1 is that a program that passes a benefit-cost test at its
current level of implementation may still be suboptimal, in the sense that a similar
program at a different level of implementation could yield larger net benefits. A
corollary to this result is that a program that consists of many components, for example,
a regulation that covers many sources, may, overall, pass the benefit-cost test, even
though individual components of the program (regulations on individual sources) may
4-25
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 not pass the test. In the context of IED, one objective of integrated options analysis is
2 to improve net benefits by careful selection of risks to be addressed and risk reduction
3 options to be pursued.
4
5 4.5.4 Uncertainty in the Measurement of Benefits and Costs
6
7 The state of the art of measurement of the benefit and costs of reducing risks is
8 not sufficiently developed to produce exact measures of economic value. This leads to
9 the question, must policy makers wait for further research to produce exact measures
10 before they can use benefit and cost information as an aid in decision making? If not,
11 how should they interpret the ranges of values that current research has produced?
12
13 To counsel waiting for exact measures is equivalent to saying that in many cases
14 value measures should never be used. The state of the art cannot be expected to
15 advance to the point of producing exact values for all kinds of environmental change.
16 This is because of the inherent uncertainty and imprecision in measurement
17 techniques, and because of uncertainties about which models are appropriate in
18 specific circumstances. So how are policy makers to proceed in the face of continued
19 and inherent uncertainty about economic values?
20
21 A simple approach with a basic intuitive appeal is to perform sets of calculations
22 with the upper and lower bounds of the range, and perhaps with the midpoint of the
23 range as well. This is the essence of the approach for valuing reductions in mortality
24 risks outlined in the EPA guidelines for performing regulatory impact analyses (U.S.
25 EPA, 1983).
26
27 Clearly, if the benefits of a policy calculated with the upper end of the ranges are
28 less than the lower end of the range of estimated costs, the policy is unlikely to be
29 justifiable on economic grounds; and if the benefits calculated with the lower end of the
30 range exceed the upper end of the range of costs, the economic case for the policy is
31 quite strong. This simple-minded approach is a step in the right direction; but it does
32 not make use of all of the relevant information contained in the set of available
33 estimates. This range reflects only the information contained in the two estimates
34 yielding the highest and lowest values; it ignores the information on the quality of these
35 two estimates; and it ignores the information contained in the other estimates that yield
4-26
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 intermediate values.
2
3 It is possible to make use of ail available estimates and to incorporate judgments
4 about the quality of each of these estimates. This can be done by viewing probabilities
5 as statements about the degree of confidence held about the occurrence of some
6 possible event. The approach involves assigning probabilities to all of the values
7 produced by the available estimates, where a higher probability reflects a greater
8 degree of confidence in that estimate. For example, the assignment of a probability of
9 unity to a particular estimate means we are certain that this study has produced the
10 correct value. Once the probabilities have been assigned, various useful summary
11 statistics can be generated. For example, the expected value of the parameter in
12 question (the probability distribution) can be calculated and used for benefit-cost
13 calculations. The variance of the distribution can be used to determine confidence
14 intervals on the value to be used, thus preserving for policy makers information on the
15 uncertainty about economic values.
16
17 4.6 Distributional Considerations
i
19 This discussion of benefits and costs, as well as the way the two are compared,
20 glosses over an important point - one that has been raised increasingly in discussions
21 about which risks the government ought to be addressing. Specifically, BCA is silent
22 about the distributional implications of risk reduction measures. This has given rise to
23 concerns about the absence of environmental equity or environmental justice in the
24 decision making processes of regulatory agencies.
25-
26 To illustrate this point, suppose hypothetical^ that emissions from a series of
27 small auto paint shops in Los Angeles were resulting in slight increases in ambient
28 ozone levels in Beverly Hills. This in turn was slightly increasing the risks of eye
29 irritation among wealthy joggers there. Because the latter have very high incomes, the
30 calculated benefit from reducing this risk is very great - the joggers would be willing to
31 pay substantial amounts to be free of this annoyance. Suppose further that closing the
32 paint shop, and in the process rendering unemployed some of the unskilled people who
33 work there, is the only way to reduce the risk - control equipment could not be afforded.
34 In this case, the calculated benefits of the risk reduction could be well in excess of the
35 costs, but because all the benefits go to the wealthy while the costs are borne by the
4-27
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 poor, one might be quite uncomfortable with enacting the controls.
2
3 In other situations, the situation is reversed. That is, the physical benefits of a
4 risk reduction program would go to a poor population - say, those who live around a
5 Superfund site - while the costs of the cleanup would be shared much more broadly
6 and, in general, by those with higher incomes. But because the willingness to pay of
7 those affected by the site is quite low, the benefits as calculated in a BCA would be
8 small relative to the costs. Thus, such a cleanup might fail a benefit-cost test even
9 though it would have what many would regard as favorable distributional
10 consequences. Simply put, BCA counts a dollar's worth of benefit or cost to a wealthy
11 person the same as a dollar's worth of benefit to a pauper. This can result in some
12 efficient programs (programs for which the net benefits are positive) having unattractive
13 features. Economists have long acknowledged that efficiency and equity are not
14 always compatible objectives.
15
16 In the introductory section, we mentioned this concern with BCA, and suggested
17 that much thought has been given over the years to the incorporation of distributional
18 weights for BCA that would attempt to incorporate such considerations into
19 determinations of efficiency. We believe this is a poor idea, not because we are
20 unconcerned with the distribution of income, but rather because we see no way to
21 derive any kind of consensus on what the weights should be. It seems better to
22 estimate the benefits and costs as best we can, and also provide as much information
23 as possible about who will gain and who will lose, and let decision makers take
24 distributional considerations into account in whatever way they see fit.
25
26 4.7 Conclusions
27
28 What are we to conclude from this discussion? Several points bear repeating.
29 First, and perhaps most importantly, knowing the risks that arise from current and future
30 exposures to pollutants in the environment is certainly important, but it is an insufficient
31 basis, on its own, for planning future regulatory or other forms of action. That is, it does
32 not naturally follow that the best thing to do once one knows which risks pose the
33 biggest threats to human health or the environment is to take actions to reduce the
34 biggest risks first. While perhaps superficially counterintuitive, it may make much more
35 sense to attack problems farther down on the list.
4-28
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2 This is because a more sensible basis for action is asking which problems most
3 lend themselves to amelioration. That is, we ought to ask where the resources we have
4 available for environmental improvement (be they financial, human, or other) can best
5 be deployed. This in turn requires that we know how much risk reduction we can get
6 and for how much money in each of the risk categories. This is no more than the
7 environmental application of "the biggest bang for the buck" aphorism that originated
8 when analytical thinking was applied to defense spending.
9
10 Although economics can provide only parts of the answer to the question, "How
11 can we best spend our environmental dollars?", those are important parts nonetheless.
12 Economics can tell us, for instance, how much it would cost to attempt to reduce
13 average ambient ozone concentrations from 0.10 parts per million to 0.08 parts per
14 million. It can tell us how much it might cost to reduce heavy metal deposition into
15 aquatic ecosystems by 30 percent. And it can tell us what we could expect to spend to
16 reduce pesticide residues on fresh fruits and vegetables.
17
Economics can do more than inform us about the costs of various types of
19 control actions. It can also illuminate the value that individuals place on these and
20 other types of risk reductions, and how these values compare to the costs of taking
21 actions. This latter role of economics-benefit-cost analysis-is controversial, poorly
22 understood, and often misused. This is no reason, however, to throw the baby out with
23 the proverbial bathwater. If carefully used, with its basic assumptions clearly and
24 openly described, benefit-cost analysis can be an invaluable aid to informed decision
25 making.
26
27 If economics is used carefully, it can contribute significantly to the nation's
28 environmental protection efforts. Specifically, along with quantitative risk assessment
29 and other analytical tools, it can help us identify places where we can accomplish a lot
30 for a little, and avoid spending a great deal to do relatively little good. It is harder
31 making decisions in cases where both the benefits and the costs seem large, or those
32 cases where both are on the small side. Here, finer distinctions must be drawn. This
33 means in turn that the assumptions that underlie a benefit-cost assessment may make
34 the difference between a favorable and an unfavorable outcome. So long as those
35 assumptions are made explicit, however, and so long as distributional and other
4-29
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 considerations are also factored into decision making, BCA can be a most valuable
2 input into the formulation of an appropriate environmental policy.
3
4
4-30
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 4.8 References Cited
2
3 Arrow, K., M.L Cropper, G.C. Eads, R.W. Harm, L.B. Lave, R.G.Noll, P.R. Portney, M.
4 Russell, R. Schmalensee, V.K. Smith, and R.N. Stavins. 1996. Is there a role
5 for benefit-cost analysis in environmental, health, and safety regulation?
6 Science. 272:221-222.
7
8 Baumoi, W.J. and W.E. Gates. 1988. The Theory of Environmental Policy. Second
9 Edition. Cambridge, United Kingdom: Cambridge University Press.
10
11 Cropper, M.L and W.E. Gates. 1992. Environmental Economics: A Survey. Journal
12 of Economic Literature 30:675-740.
13
14 Freeman, A. M., III. 1993. The Measurement of Environmental and Resource Values:
15 Theory and methods. Resources For the Future. Washington, DC. 516 pp.
16
17 Freeman, A. M., III. 1997. On Valuing the Services and Functions of Ecosystems.
Chapter 11 in Simpson, R. D. and N.L. Christensen, Jr. (Eds), Ecosystem
19 Function and Human Activities: Reconciling Economics and Ecology, New York.
20 Chapman and Hall.
21
22 Hammitt, J. 1997. Are the Costs of Proposed Environmental Regulations
23 Overestimated? Evidence from the CFC Phaseout. Working Paper, Harvard
24 Center for Risky Analysis, January.
25
26 Jaffe, A.B., S.R. Peterson, P.R. Portney, and R.N. Stavins. 1995. Environmental
27 Regulation and the Competitiveness of U.S. Manufacturing: What Does the
28 Evidence Tells Us? Journal of Economic Literature 33:132-163.
29
30 Palmer, K., W.E. Gates, and P.R. Portney. 1995. Tightening Environmental
31 Standards: The Benefit-Cost or the No-Cost Paradigm? Journal of Economic
32 Perspectives 9:Number 4, pp. 119-132.
33
34 Porter, M.E. and C. van der Linde. 1995. Toward a New Conception of the
35 Environment-Competitiveness Relationship. Journal of Economic Perspectives
4-31
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 9:Number4, pp. 97-118.
2
3 Rutledge, G.L. and M.L. Leonard. 1992. Pollution Abatement and Control
4 Expenditures, 1972-1990. Survey of Current Business, June. Pp. 25-41.
5
6 Schmalensee, R. 1994. The Costs of Environmental Protection, jn Balancing
7 Economic Growth and Environmental Goals. Mary Beth Kotowski (Ed).
8 Washington, D.C.: American Council for Capital Formation Center for Policy
9 Research. Pp. 55-75.
10
11 U.S. Congressional Budget Office. 1992. Environmental Regulation and Economic
12 Efficiency. Washington, D.C.: U.S. Government Printing Office.
13
14 U.S. Department of Commerce. 1993. Pollution Abatement Costs and Expenditures,
15 1991. Economics and Statistics Administration, Bureau of the Census.
16 Washington, D.C.: U.S. Government Printing Office.
17
18 U.S. Environmental Protection Agency. 1983. Guidelines for performing regulatory
19 impact analyses. Washington, DC.
20
21 U.S. Environmental Protection Agency. 1990. Environmentallnvestments: The Cost
22 of a Clean Environment. Washington, D.C.
23
24 von Winterfeldt, D. and W. Edwards. 1986. Decision analysis and behavioral research.
25 Cambridge University Press.
26
27
4-32
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2 Endnotes
3
4 'See definitions of consumer surplus and producer surplus, above, in our discussion of
5 environmental benefits.
6
7 2For a useful decomposition and analysis of the full costs of environmental regulation, see:
8 Schmalensee (1994).
9
10 'The notion that environmental regulation can foster economic growth is a controversial one
11 among economists. For a debate about this proposition, see: Porter and van der Linde, 1995;
12 and Palmer et aL, 1995.
13
14 "For example, if a firm chooses to close a plant because of a new regulation (rather than
15 installing expensive control equipment), this would be counted as zero cost in narrow
16 compliance-cost estimates, but it is obviously a real cost.
17
18 5Note that it is extremely difficult to measure "retarded innovation," because the degree of
19 innovation that would have occurred in the absence of the environmental-protection measure is
20 essentially unobservable.
21
00 6For over twenty years, the PACE survey provided a unique source of information on pollution
control costs, since it was the only annual, representative survey to cover all of U.S.
&4 manufacturing. In 1996, the collection of data was terminated for budgetary reasons. Given
25 that the annual costs of the survey were approximately $1 million, this decision itself ought to be
26 questioned on benefit-cost grounds. In April 1999, EPA announced that it would reinstitute the
27 annual PACE survey with the Bureau of the Census.
28
29 7For detailed discussion of environmental compliance cost measurement problems, see U.S.
30 Congressional Budget Office (1985).
31
32, 8An irony of the movement of environmental policy towards more reliance on performance
33 standards (both uniform and market-based) and less on end-of-pipe technological standards is
34 that it is becoming increasingly difficult to measure even the narrowest notion of pollution
35 abatement costs (Jaffe et al., 1995).
36
37 9 Formally, if bt denotes benefits in year t and ct costs, the discounted sum of net benefits over
38 a horizon of T years is given by:
39
40 T
41 I (bt-ct)/(1+r)t
42 t=0
4-33
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 CHAPTER 5. ASSESSING THE VALUE OF NATURAL RESOURCES
2
3 TABLE OF CONTENTS
4
5
6 OVERVIEW 5-1
7
8 5.1 Introduction 5-4
9 5.1.1 Background 5-4
10 5.1.2 Objectives and Approach 5-5
11
12 5.2 Valuation and the Decision Context 5-6
13 5.2.1 The Valuation Process in Regulation 5-7
14 5.2.2 The Decision Context 5-9
15
16 5.3 The Nature of Values 5-14
5.3.1 Introduction 5-14
5.3.2 Values 5-15
19
20 5.4 The Economic Valuation Framework 5-20
21 5.4.1 The Concept of Economic Value 5-20
22 5.4.2 Economic Value and Benefit-Cost Analysis 5-21
23 5.4.3 Economic Valuation of the Functions and Services of Environmental
24 Systems 5-22
25 5.4.4 Issues and Problems 5-23
26 5.4.5 Conclusions 5-25
27
28 5.5 The Importance of Deliberative Processes to Valuation 5-26
29 5.5.1 Introduction 5-26
30 5.5.2 Aspects and Recommendations 5-28
31
32 5.6 Additional Approaches to Valuation of Environmental Systems 5-37
33 5i6.1 Introduction 5-37
34 5.6.2 Findings 5-37
"<;
5.7 Summary and Conclusions 5-47
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2 5.8 References Cited 5-52
3
4
-------
IRP Draft Integrated Report-Peer Review Draft, May 3, 1999-Do Not Cite or Quote
1 CHAPTER 5. ASSESSING THE VALUE OF NATURAL RESOURCES
2
3
4 OVERVIEW
5
6 In its 1990 report, Reducing Risk: Setting Priorities and Strategies for
7 Environmental Protection, the SAB recommended that the Agency "develop improved
8 analytical methods to value natural resources and to account for long-term
9 environmental effects in its economic analyses" because "traditional methods of
10 economic analysis tend to undervalue ecological resources and fail to treat adequately
1 1 questions of intergenerational equity." During 1996, Deputy Administrator Fred Hansen
12 urged the SAB to address the need for improved methods of measuring environmental
13 benefits.
14
15 The Valuation Subcommittee was established as a multidisciplinary group to act
16 on this request. The specific charge was for the Subcommittee to define better the full
range of relevant questions that must be considered in ecological and human health
•i w valuation. It was further to identify conditions where existing economic methodologies
19 seem not to address or monetize adequately ecological endpoints that may be
20 important contributors to the value society places on ecological resources. The
21 Subcommittee was challenged to use its broad diversity of expertise and perspective to
22 scope out a more complete framework for valuation and to identify the types of
23 methodological developments or research needed to implement the framework.
24 Although it was requested to do so in the charge, the Subcommittee was not able fully
25 to consider valuation of human health issues, and the Subcommittee's conclusions may
26 not apply directly to human health.
27
28 The group discussed the charge and worked on its response in public meetings
29 on two separate occasions. Also, in a three-day workshop during April, 1997 the
30 concept of environmental valuation was discussed from five perspectives: 1) the
31 environmental management decision context for valuation; 2) the nature of value and
32 values; 3) the economic concept of value; 4) the importance of deliberative processes
33 to valuation; and 5) additional approaches to environmental valuation.
34
5-1
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Many stakeholders feel that existing economic analysis methods undervalue
2 ecological resources because they miss important issues. The Subcommittee agrees
3 that valuation of ecological services is a complex process and a still-evolving field of
4 research and that improvements in both approach and practice are needed. This report
5 discusses the components of the process that should be considered in valuation efforts.
6 It concludes that there is no single approach to valuation that can be used in all
7 situations. The Subcommittee's work confirms that an interdisciplinary perspective is
8 required to meet the challenges and complexities of valuation.
9
10 One of the basic premises of welfare economics is that economic values are
11 based on individuals' preferences and that people know their preferences. Where
12 individuals are ignorant of the roles of ecological functions in contributing to valued
13 service flows, it may be necessary to use experts' knowledge of the functioning of
14 environmental systems as an input in the valuation process. In principle, the economic
15 valuation framework can be utilized to define and measure the economic values of
16 changes in the functions and services of environmental systems that affect individuals'
17 welfare either directly or indirectly. However, this framework may be difficult to
18 implement in practice where the relationship between the function and the service flow
19 to individuals is indirect or subtle. Economic approaches to valuation are not
20 mechanisms for producing "the answer" since they may be incomplete, may include
21 some elements which are difficult or impossible to estimate, and may employ
22 preference elicitation processes that are incomplete.
23
24 Not all benefits or costs can be easily quantified, much less translated into dollar
25 terms. There is a need for qualitative methods to help decision makers understand the
26 hard-to-define values that are important in finding solutions to complex ecosystem
27 problems. When integrating the results from different methods, care must be taken to
28 assure that quantitative factors do not dominate important qualitative factors. To be
29 most useful, valuation issues and approaches should be made as explicit as possible
30 and should involve assembling the appropriate available information, clearly stating the
31 assumptions and uncertainties, and ensuring that the application of methods is
32 transparent. These analyses are best used to inform but not dictate, decisions related
33 to environmental protection policies, programs, and research.
34
35
5-2
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 There is much uncertainty in our knowledge and understanding of many of the
2 factors necessary in the environmental decision-making process. Uncertainty will
3 persist, yet we cannot wait for certainty in our analytical inputs to make decisions. The
4 process of adaptive management, discussed in this report, is attractive because it
5 allows decisions to be made and management to proceed in the face of uncertainty. In
6 doing so, it allows for feedback from experience gained during implementation of
7 environmental management practices, the development of new knowledge by the
8 research community, and for revisiting and revising past decisions on the basis of the
9 new insights provided by this research and experience.
10
11 General themes that emerged during the Subcommittee's discussions included:
i
1 1) For decision-making purposes in a governmental context, ecological valuation
1 is an anthropocentric exercise (people's wishes count; there is no external set of
1 values watting to be discovered for application to decision-making).
1 .
1 ' 2) The value of anything reflects its contribution toward the achievement of
some goal. The process of valuation cannot be separated from the need to
j i reach agreement on goals.
J i
3) Environmental valuation requires a diverse and interdisciplinary process
involving interaction and deliberation among scientists, decision makers, and
! other stakeholders to identify goals and to define endpoints to characterize those
goals.
i 4) Existing economic approaches, broadly considered, are consistent and
coherent frameworks for valuation because they organize a system of trade-offs.
I However, they are not mechanisms for producing "the answer" because they
} may omit trans-economic values that may be important, may include some
) elements that are difficult or impossible to estimate, and may employ preference
I elicitation processes that are incomplete.
3 5) An expanded, rich, and complex process using multiple approaches is
I required to fully encompass ecological valuation.
5-3
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 The Subcommittee's work confirms the challenges and complexities of
2 environmental valuation exercises, and the resultant environmental management
3 actions based on those values. The Subcommittee recommends that expanded, rich,
4 and complex processes be employed to fully characterize environmental values. This
5 process will involve interaction and deliberation among scientists, decision makers, and
6 other stakeholders in order to identify goals, define endpoints to characterize those
7 goals, and to implement approaches to achieve those goals. This process must be one
8 of on-going dialogue and adjustment and it should consider 1) why people care for the
9 things they do-their preferences; 2) the appropriate use of deliberative processes to
10 elicit preferences and the rationale for them; 3) economic valuation frameworks to
11 define and measure the economic value of changes in environmental systems functions
12 and services affecting individual welfare; and 4) presentation of available physical or
13 other quantitative measures, or qualitative descriptions of the effects of alternative
14 actions when costs and benefits are not fully captured by monetary measures. The
15 Subcommittee recognizes that environmental valuation remains a craft embedded in
16 political processes.
17
18 5.1 Introduction
19
20 5.1.1 Background
21
22 In its 1990 report, Reducing Risk: Setting Priorities and Strategies for
23 Environmental Protection, the U.S. Environmental Protection Agency's (EPA) Science
24 Advisory Board (SAB) recommended that the Agency "develop improved analytical
25 methods to value natural resources and to account for long-term environmental effects
26 in its economic analyses." This recommendation was based on the view that
27 "traditional methods of economic analysis tend to undervalue ecological resources and
28 fail to treat adequately questions of intergenerational equity." In February 1996, Deputy
29 Administrator Fred Hansen met with the SAB Executive Committee and urged the SAB
30 to address the need for improved methods of measuring environmental benefits as part
31 of its project to update Reducing Risk.
32
33 In response, the Executive Committee agreed to establish a Valuation
34 Subcommittee as part of the Integrated Risk Project. The Valuation Subcommittee's
35 deliberations were intended to complement those of the Economic Analysis
5-4
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Subcommittee so that the final Integrated Risk Project report would include information
2 relevant to the economic assessment of the benefits and costs of various risk reduction
3 options, as well as recommendations for improving future assessments of the value of
4 ecological resources to society.
5
6 5.1.2 Objectives and Approach
7
8 The Valuation Subcommittee was charged to define better the full range of
9 relevant questions that must be considered in ecological and human health valuation.
10 The prevailing feeling was that existing economic methodologies, which are used to
11 quantify the monetary value of certain ecological resources, do not address or monetize
12 adequately important ecological endpoints (ecological processes, environmental
13 structures, and many other characteristics) that may be important contributors to the
14 value society places on ecological resources (e.g., intergenerational issues,
15 stewardship issues, sustainability options, and time scale issues). Similarly, many
16 environmentally mediated quality of life issues of great importance to people and
1** society are difficult to quantify (e.g., aesthetic, culture, religion, security, equity).
Consequently, a new look at the overall approach to valuation was requested: one that
19 considered the utility of current economic methodologies as a point-of-departure, but
20 which went beyond these and identified other approaches to supplement them and
21 provide a more comprehensive basis for valuing ecological systems and their
22 relationships to human health and quality of life. The Subcommittee was challenged to
23 use its broad diversity of expertise and perspective to scope out a more complete
24 framework for valuation and to identify the types of methodological developments or
25 research needed to implement the framework.
26
27 In order to meet this charge, a diverse group was assembled with individuals
28 from disciplines such as public policy, economics, ecology, philosophy,
29 communications, psychology and sociology. Institutional affiliations represented in the
30 group were also diverse; for example, members came from academia, industry,
31 independent research institutions, and consulting firms. Some members were formerly
32 associated with government organizations.
33
34 The Subcommittee discussed the charge and planned for its response in public
35 meetings on two separate occasions. These meetings led to a three-day workshop in
5-5
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Baltimore, Maryland during April, 1997. At this workshop, the members were assigned
2 to one of five panels and led Subcommittee discussions about valuation from five
3 perspectives. The discussions and recommendations resulting from these activities are
4 the basis for this chapter.
5
6 The Subcommittee discussions fell into the following broad areas: 1) the
7 environmental management decision context for valuation; 2) the nature of value and
8 values; 3) the economic concept of value; 4) the importance of deliberative processes
9 for eliciting values and the goals on which they are predicated; and 5) additional
10 approaches to environmental valuation. Although requested to do so in the charge, the
11 Subcommittee was not able to consider valuation of human health issues.
12 Consequently the Subcommittee's conclusions may not apply directly to human health.
13
14 The focus of Subcommittee activity was the consideration of alternatives and
15 enhancements to economic analysis techniques that would allow the Agency to
16 characterize the benefits and costs of environmental protection actions more
17 comprehensively and accurately and to ensure that environmental decision-making
18 takes account of the values that people attach to ecological systems, and system
19 functions and components.
20
21 5.2 Valuation and the Decision Context
22
23 The need to conduct environmental valuation exercises arises because people,
24 as they exercise their prerogative of free choice, sometimes carry out activities that
25' result in changes to environmental systems that can be detrimental. For some of these
26 situations, society has established laws to control the behavior that leads to the
27 detrimental effects. Government agents are entrusted with the responsibility and
28 authority to evaluate the detrimental changes that meet or exceed triggering-criteria
29 contained in these laws, and to decide whether action is needed to prevent or control
30 these changes. These decisions require, in turn, a basis for comparing alternative
31 outcomes, and this leads to the need to place a value on the ecological changes
32 envisioned.
33
34
35
5-6
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 5.2.1 The Valuation Process in Regulation
2
3 When a statutory criterion is triggered or a problem is otherwise identified, a
4 chain of activity is initiated that results in a decision on the need for, and if appropriate,
5 the type of intervention. As discussed in previous chapters, the I ED framework
6 provides a way of considering information (both data and professional judgments) on
7 risks, control options and practices, economic impacts, and social and quality of life
8 issues. Equally important is the consideration of the value ascribed to alternative
9 outcomes. Methods to estimate such values are a critical component of the IED
10 framework.
11
12 As discussed in Chapter 4, value is routinely assessed during the environmental
13 management decision-making process using economic analysis techniques. The usual
14 approach is to apply economic analyses to estimate the costs and benefits of each of a
15 suite of potential environmental management options. The costs and benefits are then
16 compared to determine the net benefit associated with each option, and this
117 information, along with other decision criteria specified by the decision-maker, serves
as the basis for deciding whether to take action, and if so, on the most suitable action to
19 implement.
20
21 Figure 5-1 reflects a prototypical relationship between increasing degrees of
22 environmental stressor control and the societal costs and benefits that might be
23 associated with a specific control option. The relationship demonstrated in the figure
24 suggests that at first, toward the origin, the initial marginal costs of control are minimal
25 but gradually increase because a higher degree of effort is needed to bring about
26 increasing reduction in the stressor. At the same time, the associated marginal benefits
27 are high with the initial reduction in the stressor but decline with additional reduction.
28 The key point shown by the figure is that typical economic analysis techniques, which
29 compare costs and benefits of control options (usually in dollar terms) reveal a point in
30 the stressor reduction process at which the associated costs of additional control are
31 not "worth it" because the benefits gained are less than those costs. Further reductions
32 in the stressor will bring benefits, but these are won only at costs (expenditures of
33 resources with alternative uses) that on net, in the judgment of the party doing the
34 evaluation, make the overall outcome worse. In a public policy context, the government
35 decision-maker stands as the agent for the people as a whole and acts on the
5-7
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
($) A
D
o
I
I
a
r
s
Percent Reduction in Stressor
100%
Figure 5-1. Typical relationship of marginal costs and marginal benefits associated
with a proposed environmental control option.
1' valuations that they place on stressor reduction and on the associated use of resources
2 to accomplish it. An example is the distribution of the heavy metal, lead in the
3 terrestrial environment at a hazardous waste site. Initially high concentrations of lead
4 pose significant environmental risks which can be sharply reduced during a site clean
5 up program. These initial reductions in lead concentrations can be achieved at
6 relatively low costs; however, a point of excess control can be reached where the costs
7 of additional lead removal exceed the marginal benefits and associated reduction in
8 risks realized.
9
10 While accepting the stylized approach to trade-offs embodied in this diagram, the
11 Subcommittee noted that its effectiveness in practice as a guide to action depended on
5-8
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 the completeness and correctness of the elements included in the calculation of the
2 prototypical curves. As illustrated in the discussion below, with respect to
3 completeness, the Subcommittee found that in practice some elements valued by
4 people could be systematically omitted, others could be poorly quantified, and others
5 might evolve only in a social context and not surface from commonly used individual
6 preference measures. With respect to correctness, the Subcommittee noted that
7 selection of the elements to be valued was determined by the social goals presumed to
8 be relevant, and that those goals were a complex mixture that included matters
9 sometimes overlooked such as fairness, long-term sustainability, and stewardship.
10 Further, the Subcommittee noted that in practice, estimation of the values that
11 individuals placed on these elements is subject to error, and that aggregation of such
12 values involves methodological uncertainties. These considerations formed the basis
13 for Subcommittee discussions of enhancements to economic analysis techniques to
14 allow a more comprehensive and accurate characterization of environmental values.
15 The results of the discussions are captured in the sections that follow.
16
1"7 5.2.2 The Decision Context
19 Many environmental regulations have been positively received by the public and
20 these actions have produced noticeable improvements in the quality of our air, land,
21 and waters. In addition, the regulation of certain chemical products and residuals have
22 brought improvements in-human health and the state of certain wild species. However,
23 some environmental decisions have met with concern and conflict and have led to an
24 erosion of trust in governmental institutions. Common criticisms of environmental
25 decisions are that they are yielding inadequate returns for the resources used, often
26 focus on perceived, rather than actual, risks, and define problems too narrowly (e.g.,
27 see Sexton, 1996). In addition, there is a growing sense that decision-making should
28 be more open, allowing greater understanding and participation by stakeholders and
29 the public.
30
31 In considering the source of controversy in environmental management decision-
32 making, Brown (1997) suggested that "...disputes about environmental policy, while
33 often seeming to be about the facts, are-at least as much-about an underlying value
34 framework that legitimates a particular decision." Dietz et al. (1989) noted that
35 participants in risk policy debates differ in their perceptions about the role of facts and
5-9
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 values in such conflicts. The norms and goals of society, in effect, provide the
2 dominant contextual grounding for environmental decision-making.
3
4 The values that people place on certain health and environmental states,
5 conceptions of their rights and obligations, and outcomes associated with
6 environmental perturbations and responses, reflect their norms and social objectives.
7 The values underlying these debates and decision-making processes are not always
8 clear. These values may be complex and in competition with other values, including not
9 only the values of other people, but even values held by the same person. Because
10 value frameworks are so diverse and complex, there is a need to refine methodologies
11 for eliciting and comparing values quantitatively and qualitatively, both interpersonally
12 and individually.
13
14 "Values" as used in the two preceding paragraphs is a term which describes the
15 set of underlying factors that, taken together, cause people to hold the opinions that
16 they hold and to make the choices that they make when presented with real situations.
17 Unfortunately, the English language also uses the term "value" and its derivatives in an
18 operational sense as a descriptor of the "worth" of outcomes, measured in terms of
19 what would be willingly sacrificed in exchange. Thus the confusion in terms between a
20 person "valuing" health in the abstract, the first definition of the term, and a person
21 placing an inferred 'Value" of so many dollars, but no more, on avoiding a day of limited
22 activity and discomfort due to a cold. The exchange of economic "value" is therefore
23 derived from the vector of ail the underlying abstract "values" held by the individual in
24 concert with the situation presented. When it comes to making environmental (and
25 other) decisions, "value" is used in an operational sense as a measure of what one
26 outcome is worth in comparison with alternatives. This is the sense in which values are
27 reflected in, for example, benefit/cost analysis. It is with this meaning that "value" is
28 used in most appearances in this Chapter. Other uses of the term also occur, however,
29 and are made clear by the context.
30
31 The Valuation Subcommittee discussed the issues of scientific uncertainty,
32 competing values, and the completeness of individual valuation approaches. It
33 concluded that all of the approaches used to develop summary information on risks,
34 costs, and benefits for decision-makers reflect to some degree the disciplinary
35 background and values held by those individuals who conduct the assessments. This
5-10
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 highlights the importance of knowing who participates in data collection and evaluation
2 and the methods they use to do so. It further suggests the importance of making sure
3 that individuals and groups having a stake in the outcome of decision-making
4 participate in the process of describing and valuing changes in environmental systems,
5 and in estimating the costs and benefits of intervention to control or redress these
6 changes. Some form of interaction and deliberation among agencies and among those
7 who hold direct or indirect interests in decisions is often recommended as a way of
8 identifying and addressing the disparate values held by individual stakeholders (NRC,
9 1996; Risk Commission Report, 1997). The Subcommittee has endorsed this
10 recommendation.
11
12 Valuation and value judgments are important in all phases of the IED framework
13 (see Figure 5-2). Initial definition of values, and the goals to which they give rise, is
14 required in the early problem formulation phase to help define the scope of the
15 problem, its context, and the breadth of evaluation required in the analysis phase.
16 Value judgments are also made during the analysis and characterization of risks.
17 During the Analysis and Decision-making Phase, valuation input is again needed as
alternative management or control options (e.g., technologies and practices) are
19 evaluated to help identify and measure the relative costs and benefits of different
20 degrees of risk reduction and of alternative ways of accomplishing it.
21
22 Values considered during environmental management decision-making are
23 those associated with changes to both human dominated (e.g., farms, forests) and
24 natural environmental systems. Substantial variation exists within natural ecosystems
25 and to be socially significant, a change caused by a stressor must exceed the natural
26 variability of the environmental system and involve impacts affecting significant social
27 values. However, since no precise standards exist for measuring stressor reduction
28 costs and associated societal benefits at the system level, it is inevitable that the
29 decision on which values to weigh most heavily will require deliberation and negotiation.
30 Therefore, it is clear that the articulation of these values, and the resultant regulatory
31 actions based on them, must necessarily involve deliberation including scientists, the
32 public, and policy makers in a process of on-going dialogue and adjustment.
33
34
35
5-11
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2F
26
27
28
29
30
31
32
33
34
35
Environmental Valuation
Environmental Valuation
Figure 5-2. Valuation in the IED Process
Although current valuation techniques can deal satisfactorily with some
environmental system changes (e.g., the impact of air pollution on managed agricultural
systems), they may deal inadequately or not at all with others. When one thinks of
considering the benefits of environmental systems, it is important to recognize that
some things can be quantified and some not. Of those that can be quantified, some
can be monetized and some not. For example, it is possible to quantify concern with
biodiversity without being able to monetize it, or otherwise express it in units that can be
compared to, say, changes in unemployment. Therefore, environmental system
5-12
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 valuation efforts must include both quantitative and qualitative perspectives in an
2 integrated approach to capture all components of things valued.
3
4 Quantitative methods include those approaches designed to measure human
5 attitudes, preferences (i.e., choices when presented with alternatives, desires,
6 operational proximate goals) and values. As discussed in the previous chapter, these
7 methods include: 1) direct measures of preferences and values using survey methods;
8 2) inferring preferences and values from observed behavior and choices; 3) inferring
9 preferences and values from choice experiments; and 4) using estimates of values of
10 similar commodities that were derived from using one of the above methods, or what
11 economists refer to as benefits transfer.
12
13 Since not all environmental costs and benefits can be easily quantified for use in
14 benefit-cost analyses, there is a need for qualitative methods to help decision-makers
15 understand the hard-to-define values that are important in finding solutions to complex
16 environmental system problems. Qualitative methods must involve at least the detailed
17 narrative description of effects in complete terms such that comparisons with other
1 ecological attributes can be made. Although this qualitative information cannot be
19 directly compared to quantitative information in common terms, qualitative information
20 should be clearly described and it should be readily available to assist during decision-
21 making (e.g., see Florida Risk-Based Priority Council, 1996).
22
23 This information is used to guide decisions. At times, the path forward seems
24 clear and uncomplicated, and decisions are straight forward. At other times, however,
25 uncertainties exist — in how the interventions chosen will work, how the systems to be
26 altered will react, and in how the outcomes will be valued. The process of adaptive
27 management has developed as an innovative technique to assist in environmental
28 decision-making in the face of such uncertainty. Rather than waiting until complete
29 data are available that allow participants to understand all aspects of the at-risk system,
30 adaptive management allows action to proceed on the basis of the best understanding
31 initially available. The response to action is monitored, and the information gained is
32 used to modify or design the next stage in the program. In adaptive management there
33 is direct feedback between science and management such that policy decisions can
34 make use of the best available scientific information in a sequential fashion. This idea
35 has been well-developed in many contexts for nearly three decades (Campbell, 1969)
5-13
-------
IRP Draft Integrated Report-Peer Review Draft, May 3, 1999-Do Not Cite or Quote
1 and has been recently articulated around environmental problems by Lee (1993). As
2 the IED framework implies, valuation is not a one-time activity, it is an iterative part of
3 the risk management process responding to the development of new information over
4 time. That is, changes in the perception of the value of actions toward reaching goals is
5 part of adaptive management as well.
8
9 In the following sections we discuss four topics relevant to environmental
1 0 valuation: 1 ) the nature of values; 2) the traditional economic valuation framework; 3)
1 1 the importance of the deliberative process to valuation; and 4) alternatives for the
12 determination of environmental system values. Each of these sections includes a
1 3 discussion and provides some recommendations for the consideration of valuation
14 issues during environmental decision-making.
15
1 6 5.3 The Nature of Values
17
18 5.3.1 Introduction
19
20 Controversies surrounding environmental decisions are often expressed in terms
21 of disagreements about values. These controversies are difficult to resolve in part
22 because the word "values" is used in different ways and with different meanings.
23 "Values" can refer to different goals of environmental decisions or to different means for
24 achieving those goals. But "values" can also be used to express concerns that are not
25 about goals or outcomes at all. Such concerns can be about the appropriateness of
26 different procedures for reaching decisions, the expressive functions of different
27 actions, or other things. People sometimes use the word "values" to refer to the
28 strength of preferences for different outcomes, and it may be possible to interpret these
29 values quantitatively, in monetary terms, without distortion. But people also use
30 "values" to refer to more qualitative and difficult-to-quantify aspects of environmental
31 outcomes. The challenge for the IED concept, then, it to include appropriate means for
32 both interpreting values and incorporating them into the decision-making process.
33
34 Valuation exercises should be designed to explore the nature, meaning, and
35 importance of these different environmental values. They should help to reveal what
5-14
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 people value, why, and (at least for some values) how much. They should challenge
2 those people to think critically and clearly about their value judgments and the reasons
3 for making them. And they should provide satisfactory ways of bringing together
4 different kinds of values or concerns, while taking account of the costs and benefits of
5 different alternatives, in ways that help parties reach agreement on values, where this is
6 possible, or agreement on decisions where disagreements about values persist. The
7 section that follows explores some aspects of the nature and meaning of environmental
8 values, and it examines a few of the complex factors that must be understood and
9 reconciled in the process of environmental decision-making.
10
11 5.3.2 Values
12
13 a) EPA should base its decisions on human interests and values. In
14 doing so, it is important that EPA realize that people care for things
15 for different reasons and in different ways. Valuation must be
16 sensitive to these features of human values.
17
1«. Philosophers have debated whether moral values are fundamentally
19 anthropocentric (i.e., whether values refer to or are based solely on human needs and
20 interests) or whether some values may be understood independently of human nature.
21 EPA must base its decisions on human interests and an understanding of the reasons
22 why people care for the things they do, and for this reason it must focus on human
23 values. At the same time, it is important to realize that people care for things for
24 different reasons and in different ways. Some things are valued for themselves, or as
25 ends, and some things instrumentally or as means for the satisfaction of other ends.
26 People care about some things simply because they desire them or because they
27 contribute to their happiness. But other things have value because people judge them
28 to be good or important in their own right, and therefore care about these things
29 because they believe them to be valuable.
30
-31 The valuation of environmental resources must be sensitive to the basic
32 character of human values. Most people value the aspects of nature and
33 environmental systems both instrumentally and as ends. They want to protect the
34 environment as a thing to be valued in itself, as well as managing it for the other
35
5-15
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 services it provides. Environmental management involves multiple goals, which differ in
2 kind as well as in intensity. The decision problems of environmental management are
3 thus multi-attribute problems. They involve bringing together different kinds of values
4 that may conflict or be incompatible in a number of ways. Values from all these
5 sources are involved in forming preferences.
6
7 b) Preferences need to be examined. Some preferences may be more
8 important than others in ways that are not captured by
9 measurements of strength of preference alone. Preferences should
10 be elicited in ways that reveal which preferences individuals endorse
11 in light of information and the reasons that bear on them.
12
13 Preferences refer to a person's ordering or ranking of alternative states, or in the
14 economic sense, alternative bundles of market and non-market goods and services.
15 Preferences are the basis for the choices that people make; and preferences may be
16 revealed by these choices. As described in the previous chapter, economists have
17 developed two categories of techniques for assessing values: indirect measurement
18 (revealed preference) and direct questioning (expressed preference). The discussion
19 and recommendations that follow describe a supplementary approach, wherein
20 stakeholders are directly involved in the deliberative process to collectively define
21 ecological values.
22
23 EPA needs to examine the preferences people express for different alternatives
24 and their reasons for holding those preferences. People can be asked to reveal their
25 preferences, but the process often should not end there. Some preferences may be ill-
26 considered; some may be based on erroneous or insufficient information; and some
27 may be inconsistent with some other deeply held values. And some preferences may
28 be less important to some individuals than they are to others-tor clear and convincing
29 reasons. The process of revealing preferences needs to explore and examine these
30 reasons, and it should lead people not only to express their preferences but to endorse
31 the preferences they have expressed after they have been examined in the light of
32 information and reasons that bear on them. Further, the process can lead others to
33 better appreciate reasons they may not have considered or to formulate opposition to
34 some expressed preferences that may form the basis for rational deliberation and
35 consensus-building.
5-16
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 EPA's objective, therefore, should not be simply to elicit preferences but to try to
2 uncover or construct critically examined preferences. The process for doing this often
3 must be deliberative and not be simply limited to gathering and assembling
4 preferences.
5
6 c) Some important preferences for protecting and managing
7 environmental resources are constructed in the elicitation process,
8 so it is important to be sensitive to the dependence of preferences
9 on the elicitation method and the way the alternatives are framed.
1 o This fact has some important consequences: 1) the assessor is
11 necessarily more an active participant than a neutral observer in this
12 process; 2) people need to be provided tools to help them think
13 clearly about their preferences; and 3) the valuation process should
14 be guided by some operating principles, e.g., questions should not
15 be posed in only one way.
16
1~ In many cases, values are revealed by the choices people make in their daily
1 v lives. It might be said that the values underlying these choices were formed by the
19 individual's life experiences. In other cases, the values may not be pre-formed or
20 revealed in the choices people have already made, and they must be constructed
21 during the valuation process: The preferences expressed in the context of a valuation
22 of changes in environmental resources are often constructed at the time the
23 preferences are elicited and cannot be separated from the elicitation process. This fact
24 does not negate the importance of the information gained from preferences constructed
25 in this way. People should continue to be asked to express the values they hold,
26 however they are formed. But we must be extremely sensitive to the dependence of
27 preferences on the process of elicitation and the way the alternatives are framed.
.28
29 One consequence of the constructed values perspective is the need to be
30 sensitive to the reasons and causes of the preferences that are expressed. A second
31 consequence is the need to understand the sensitivity of expressed values to
32 methodological factors. A third consequence is an appreciation that the assessor is not
33 a neutral elicitor but necessarily an active participant in the process of revealing
34 preferences.
35
5-17
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Decisions that are made by the individual being questioned in the course of a
2 valuation exercise will include important values and factors that are not independent of
3 how information is framed and how preferences are measured. EPA needs to go
4 beyond some traditionally accepted economic approaches to the valuation of natural
5 resources. Simply providing respondents with information may not yield expressions of
6 coherent and consistent values. People need to be provided with tools to help them
7 think about their values. Examples of tools that might help people indicate which of
8 their desires or preferences they would regard as important, and which might be merely
9 fleeting, are to be found in multi-attribute value analysis and various forms of structured
10 deliberation processes that encourage people to think and reflect upon what they are
11 saying, and to consider whether they really mean what they say.
12
13 Finally, the constructive nature of valuation processes should lead EPA to
14 exercise caution and to take care, where possible, to pose a question in more than one
15 way. EPA should also be explicitly aware of framing and elicitation effects in order to
16 avoid the biases that they create. Approaches should be developed to attempt to
17 decrease biases from framing, anchoring, etc. EPA might also consider sponsoring
18 research to determine how accurate direct expressions of preference might be, under
19 the best deliberative conditions, as indicators of citizens' true environmental values.
20
21 d) Because of the qualitative dimensions of values, the valuation
22 process must emphasize deliberative as contrasted to mechanistic,
23 predetermined components. Valuation must allow people to express
24 the different qualities and kinds of values they regard as important,
25 and it should challenge them to think clearly and usefully about
26 these values.
27
28 Structuring valuation processes to emphasize deliberative as contrasted to
29 mechanistic or algorithmic ways should help EPA to achieve another important goal,
30 which is to be sensitive to the different qualities and characteristics of values that
31 people often regard as important. Especially in the areas of environmental protection,
32 many people regard some values as "protected" in ways that lead them to resist
33 conceiving what is important to them as an exchangeable commodity. Also, many
34 people think there are some important social values involved that cannot be reduced to
35 an aggregation of individual preferences.
5-18
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2 The valuation process should allow room for people to express values of these
3 sorts. Again, however, even these preferences are not simply to be accepted at face
4 value. These values, too, need to be challenged and explored. A sensitivity to the
5 different characters of values, as well as to the need to explain and defend them with
6 reason, should become part of the deliberative process.
7
8 The nature of values and the choices they support are important topics of
9 economics, but these subjects have also been studied in depth by other social science
10 disciplines, especially psychology and sociology, and by philosophy. The participation
11 of these other disciplines in the design of valuation instruments/surveys should be
12 encouraged.
13
14 e) The valuation process should be transparent and explicit. Any
15 valuation process is subject to possible abuse. Rather than trying to
16 deny this fact, EPA should guard against possible abuse or bias by
t"7 making the process explicit and open to review.
19 Any valuation process is subject to possible abuse, and that is certainly true for
20 the kind of more deliberative and constructive process that we are recommending be
21 considered along with more traditional approaches. Given the nature of environmental
22 values, the agent or elicitor of those values is involved in the process in a way that
23 cannot be detached. Rather than deny that fact, EPA should aim instead at protecting
24 against the potential for abuse in other ways. These involve making as much of the
25 process explicit and open as possible, so that it will be transparent to participants and
26 observers alike.
27
28 Eliciting values is a difficult process in that there are no hard and fast formulas
29 that guarantee success. However, there are better and worse ways of doing these
30 things, and it is important to try to become aware of the lessons that have been learned
31 by able practitioners and to support a culture in the EPA that will encourage Agency
32 officials to develop and use the requisite skills.
33
34
35
5-19
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 The next section discusses the concept of economic value, economic value and
2 benefit-cost analysis, economic valuation of the functions and services of environmental
3 systems, and some issues and problems associated with economic valuation.
4
5 5.4 The Economic Valuation Framework
6
7 5.4.1 The Concept of Economic Value
8
9 a) The economic value of an environmental change is the increase in
10 human economic well-being or welfare that it produces. It is
11 measured by the amount of some other good (usually money, for
12 convenience) that can be taken away from those individuals affected
13 by the change without reducing their welfare below the pre-change
14 levels.
15
16 The instrumental value of anything stems from its ability to contribute to the
17 achievement of some goal (Costanza and Folke, 1997, p. 49). One kind of value is
18 economic value. The economic value of a thing depends on its contribution to the
19 economic well-being of people. This economic concept of value has its foundation in
20 neoclassical welfare economics. Each individual's welfare depends not only on that
21 individual's consumption of private goods purchased in markets and of goods and
22 services provided by the government, but also on the quantities and qualities of
23 nonmarket goods and service flows each receives from the environment. Thus the
24 basis for valuing changes in the flows of goods and services from the environment is
25 their effects on human welfare. This anthropocentric focus of economic valuation does
26 not preclude individuals from having concerns for the survival and well-being of other
27 species. Individuals can value the survival of other species not only because of the
28 uses they make of them (for example, for food and recreation) but also because of an
29 altruistic or ethical concern. The latter can be the source of nonuse, passive use, or
30 existence values.
31
32 The economic value of a good, that is, its contribution to the welfare of an
33 individual, is measured by how much of some other good is required to make up for its
34 loss. The economic value of a good could in principle be measured by the quantity of
35 any other good that can substitute for its loss. If the substitute good has a market price,
5-20
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 then value can be expressed in terms of the money required to purchase the necessary
2 quantity of the substitute good. Since in principle any other market good can be a
3 substitute for the good in question, the economic value of the good can be defined as
4 the amount of money or purchasing power required to substitute for the loss of a unit of
5 the good, that is, to restore the individual to his/her original level of welfare.
6
7 b) Economic values can be revealed by the choices that people make
8 as they trade off one good thing for another.
9
10 Since the source of economic value is the preferences of individuals, the choices
11 that people make reveal something about these preferences and values. A variety of
12 methods have been developed to infer the values placed on environmental goods and
13 services from the choices that people make. The trade-offs that people make as they
14 choose less of one good and substitute more of another good reveal something about
15 the values people place on these goods. An example is the trade-off between money
16 and air quality that people consciously or unconsciously make because houses in
17 clean air areas have higher prices, other things held equal, as has been demonstrated
in a number of studies. For further discussion of economic valuation methods, see, for
19 example, Braden and Kolstad (1991), Freeman (1993), and Chapter 4.
20
21 5.4.2 Economic Value and Benefit-Cost Analysis
22
23 a) The economic values of changes in environmental systems should
24 be included in benefit-cost analyses of environmental policies. Such
25 benefit-cost analyses will be useful inputs into public policy
26 decision-making.
27
28 The purpose of economic valuation of the services of environmental systems is
29 to provide decision-makers with information that will help them make choices about
30 policies toward specific environmental systems in the face of scarcity and opportunity
31 cost. In most cases the question is not in the form of "either/or," but about how far to
32 go. For example, the question usually would not be whether to protect an
33 environmental system or do nothing, but instead what level or degree of protection or
34 other management intervention should be attempted. Information on the values of
35 environmental systems should take a form that is helpful in these cases. This means
5-21
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 that the relevant question is the value of the change in the services of environmental
2 systems that an action would produce, not the total value of all services (Toman, 1997).
3
4 If society wishes to make the most (in terms of individuals' well-being) of its
5 endowment of resources, it should compare the values of what its members receive
6 from any environmental change or use of a resource (that is, the benefit of the change)
7 with the values of what its members give up because of the change (that is, the costs).
8 If the sum of all benefits exceeds the sum of all costs, the proposed change passes the
9 "easy" benefit-cost test. When there is more than one proposal or project or degree of
10 change under consideration, only one option will pass the more restrictive benefit-cost
11 test and yield the highest aggregate net benefits. A society could choose to make
12 public policy choices solely on the basis of comparisons of aggregate benefits and
13 costs (maximizing net benefits). However, there may be other things besides economic
14 benefits and costs that a society might want to consider when making policy choices
15 (see, e.g., Arrow et al., 1996).
16
17 5.4.3 Economic Valuation of the Functions and Services of Environmental
18 Systems.
19
20 a) The economic value of a function or service of an environmental
21 system can take the form of a direct use value (the value of a service
22 provided to people), indirect use value (when a function indirectly
23 supports a service used by people), or nonuse or existence value
24 (when an individual values an environmental system even though
25 he/she does not make any direct or indirect use of it).
26
27 The functions of environmental systems include photosynthesis, decomposition,
28 nutrient recycling, and so forth. The services of environmental systems are the
29 materials and other services that they provide that enhance human welfare and are
30 therefore valued by people. Examples of service flows include wood and fiber from
31 forests, and amenities such as scenic vistas and wildlife observation. The values of
32 these services are often called direct values (Brown, 1990) or direct use values
33 (Goulder and Kennedy, 1997) to distinguish them from the nonuse or existence values
34 mentioned above and the indirect use values discussed below.
35
5-22
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 The functions of an ecological system can also have an economic value if they
2 help to support some service flow to people. If they do support a valuable service flow,
3 the contribution made by these functions is an indirect use value (Brown, 1990;
4 Perrings et al., 1995; Goulder and Kennedy, 1997). The indirect use value of a function
5 can be derived from the change in the value of the service flow that it supports. For
6 example, if an increase in the rate of photosynthesis in an ecological system results in
7 an increase in the flow of economically valuable food or fiber, the economic value of the
8 increase in photosynthesis is the increase in the value of the food or fiber it supports.
9 What is required to measure this indirect use value is knowledge of the link between the
10 function of the ecological system and the economically valuable service that it supports.
11
12 Sometimes the connection between a function of an environmental system and
13 an economically valuable service flow may be quite direct, as in the case of
14 photosynthesis producing useful plant material. But in many cases, the connection can
15 be indirect and quite subtle. For example, photosynthesis by wild flowers may help to
16 support a population of wild bees that also pollinate commercially valuable fruits. The
17 basic point is that the economic theory of value accommodates changes in the
functions of ecological systems that affect the well-being of individuals indirectly.
19 However, these values can enter into the decision process only if the links between
20 functions and services are known.
21
22 The functions of environmental systems may also have intrinsic worth or value in
23 the eyes of some. This intrinsic value may be a form of nonuse value. Such nonuse
24 values have the same standing as direct and indirect use values because they reflect
25 the preferences of individuals. While quantification of nonuse values is difficult, they
26 are of no lesser importance for this reason.
27
28 5.4.4 Issues and Problems
29
30 a) Aggregation of Individual's Values — Benefit-cost analyses are
31 based on the summation of individuals' positive (benefits) and
32 negative (costs) values. This simple summation raises issues of
33 equity or fairness. For this and other reasons, most economists
34 suggest using aggregate net benefits as only one of several possible
35 considerations in public policy decision-making.
5-23
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 As discussed in Chapter 4, the simple summing of individuals' values to calculate
2 net benefits is not without its problems. In summary, assessment of net benefits offers
3 no guidance on the desirability of the resulting distribution of benefits and costs (i.e.,
4 some projects with positive net benefits may produce benefits for some people but at
5 the expense of reducing the well-being of others). Another problem is that the
6 economic values of individuals reflect not only their preferences, but also their
7 economic circumstances, especially their wealth or income. One who judges the
8 present distribution of income or wealth to be inequitable has reason also to question
9 the economic values that emerge from that distribution as a basis for making public
10 policy decisions.
11
12 For reasons such as these, many writers advocate using aggregate net benefits
13 as only one input into the decision process and allowing consideration of other factors,
14 such as distributional equity, as well (e.g., Arrow et al., 1996).
15
16 b) Preferences, Values, and Knowledge — One of the basic premises of
17 welfare economics is that economic values are based on individuals'
18 preferences and that people know their preferences. Where
19 individuals are ignorant of the roles of ecological functions in
20 contributing to valued service flows, it may be necessary to use
21 experts' knowledge of the functioning of environmental systems as
22 an input in the valuation process.
23
24 A basic premise of welfare economics is that economic values are based on
25 individuals' preferences and that people know their preferences. This premise is
26 problematic when it comes to applying welfare economics to the valuation of the
27 services and functions of environmental systems because an individual might act as if
28 he/she placed no value on a function if the individual were ignorant of its role in
29 contributing to a valued service flow from the ecological system (Dasgupta, 1990;
30 Goulder and Kennedy, 1997). In the practice of environmental system valuation, there
31 are ways to work around the ignorance of individuals about ecological systems. Value
32 measures can be based on ecologists' knowledge of the relationship between the
33 functions of ecological systems and valued services that they provide. Many examples
34 exist in the environmental valuation literature where technical knowledge of physical
35 and biological relationships coming from experts is incorporated into the process by
5-24
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 which individuals' preferences are determined. For example, even though people may
2 not know the relationship between concentrations of air pollutants and the incidence of
3 respiratory disease, expert knowledge of this relationship can be combined with
4 peoples' revealed values for avoiding disease to calculate the benefits of improved air
5 quality. However, when values are elicited through stated preference methods, the
6 provision of information by experts is more problematical, as is discussed in section
7 5.3.2 above.
8
9 c) Replacement Cost as a Measure of Value — The cost of replacing an
10 ecological function with a human engineered system can be used as
11 a measure of the value of the natural system, but only if substantial
12 conditions are met.
13
14 Some authors have used estimates of the cost of replacing a function of an
15 ecological system with a human engineered system as a measure of the economic
16 value of the function itself. For example, Gosselink et al. (1974) used an estimate of
17 the cost of a tertiary sewage treatment system as the economic value of the nutrient
removal function of a wetland. Replacement cost can be a valid measure of economic
19 value only if three conditions are met (see, for example, Shabman and Batie, 1978): 1)
20 the human-engineered system provides functions that are equivalent in quality and
21 magnitude to the natural function; 2) the human-engineered system is the least cost
22 alternative way of performing this function; and 3) individuals in aggregate would in fact
23 be willing to incur these costs if the natural function were no longer available.
24
25 5.4.5 Conclusions
26
27 In principle, the economic valuation framework can be utilized to define and
28 measure the economic values of changes in the functions and services of ecological
29 systems that affect individuals' welfare either directly or indirectly. In principle, the
30 framework can be used to formulate the choices required to incorporate a variety of
31 factors into social decisions. This framework may be difficult to implement in practice
32 where the relationship between the function and the service flow to individuals is
33 indirect or subtle. Further, effective application of the framework requires that all
34 relevant matters be included, which sometimes suggests the use of processes including
35 deliberation, discussed immediately below, to assure that this happens. There is now a
5-25
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 growing number of examples of application of the economic framework to
2 environmental valuation (see, e.g., Pairings et al., 1995; Freeman, 1997).
3
4 * * » • »
5
6 The next section discusses the need for deliberation in eliciting values. This
7 reflects the Subcommittee's interest in finding ways to determine people's values
8 regarding specific environmental problems, in the face of the shortcomings that are
9 suggested as being associated with traditional environmental valuation exercises.
10
11 5.5 The Importance of Deliberative Processes to Valuation
12
13 5.5.1 Introduction
14
15 Deliberation among interested parties, though not required in all decision-making
16 cases, is often useful for eliciting information on important goals, on "best judgment" on
17 certain relationships, on measures for judging goal attainment, and on value tradeoffs
18 that must be made in decision-making and implementation. A variety of types of
19 deliberation are available, depending upon the sufficiency of knowledge available to
20 address the problem at hand, and the extent of agreement that might exist on the types
21 of values involved in the issue.
22
23 Appropriate types of deliberation can help EPA select suitable analyses and
24 reach decisions that incorporate broadly acceptable tradeoffs of efficiency, fairness,
25" and sustainability. A substantial literature related to environmental decision-making,
26 including a number of recent reports (NRC, 1996; Presidential Commission of Risk
27 Assessment and Risk Management, 1997), stress the importance of deliberation to
28 Agency decision-making.
29
30 By deliberation we mean..."any formal or informal processes for communication
31 and for raising and collectively solving problems" (NRC, 1996; p. 73). Deliberation is
32 seen as the process of bringing together those people whose expertise or perspective
33 is useful in order to craft management and policy options. Deliberation can examine
34 findings, interpretations, and/or economic and non-economic values in light of each
35 other to achieve greater clarity and to construct and consider management options.
5-26
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Thus, deliberative processes differ from but supplement analytic processes which use
2 "systematic application of specific theories and methods" (NRC, 1996; p. 215) such as
3 those of science, social science, and economics. Some types of deliberation primarily
4 involve disciplinary experts, while others require involvement of stakeholders who bring
5 disparate perspectives and knowledge to the table. In some cases, stakeholders may
6 be disciplinary experts who may have different interpretations of the state of knowledge
7 than experts in the Agency. Deliberative processes need not be reserved for
8 interactions with non-federal stakeholders, but may also be used to facilitate intra-
9 Agency or interagency discussions on an issue.
10
11 Some of the reasons given in the NRC and Presidential Commission reports for
12 using deliberative processes include:
13
14 Clarifying and potentially advancing resolution of issues of fairness. Issues
15 such as distributional equity (Who benefits?) and procedural equity (Who
16 decides?) can be raised explicitly and viewed in the specific context in which an
*7 environmental decision needs to be made (see, e.g., Vaughan, 1995). (When
we use the word "fairness" we are referring to both distributional and procedural
19 equity, two issues that have been concerns of the Environmental Justice
20 movement.)
21
22 Informing multi-dimensional tradeoffs among efficiency, fairness,
23 environmental sustainability, and other concerns. In situations that involve
24 conflicts about such tradeoffs, deliberation can explore potential tradeoffs and
25 craft various options that might reduce the amount of conflict (see, e.g., NRC,
26 1996).
27
28 Increasing credibility. By discussing various perspectives, the credibility of
29 information and Agency decision-making can be enhanced, although agreement
30 with Agency process does not necessarily result in agreement with Agency
31 decisions (see, e.g., Rosener, 1981; Mazmanian and Nienabur, 1979).
32
33 Informing priorities for research. Studies suggest that more data are not
34 necessarily better for organizational decision-making. Individuals and
35 organizations have limits on the amount of information they can assimilate. In
5-27
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 addition, some information may be peripheral to the key dimensions of the
2 decisions to be made and do little to clarify or resolve conflicts (Dietz, 1988).
3 Deliberation can help determine the research that is most likely to be key to
4 decision-making and distinguish it from research that may not reduce uncertainty
5 sufficiently to inform the decision at hand (NRC, 1996). Thus deliberation can
6 conserve resources.
7
8 5.5.2 Aspects and Recommendations
9
10 a) We recommend that the context (characteristics, scope, and
11 implications) of an Agency decision be taken into account in the
12 selection of deliberative processes.
13
14 This concept is expressed by our typology of deliberative processes with
15 stakeholders and experts (Figure 5-3). The type of deliberation, the selection of
16 specific deliberative techniques [e.g., scientific meeting, citizen advisory panel, citizen
17 jury, informal meeting, etc.), and the selection of participants all vary by the situation
18 (English etal., 1993).
19
20 There is no one-size-fits-all approach. The Agency should consider two
21 questions when determining the type of deliberation that is most appropriate:
22
23 /. To what extent is the agreement on values (e.g., fairness, sustainability,
24 efficiency, etc.) and on appropriate tradeoffs among them sufficient to
25 reach a decision?
26
27 The relative importance of efficiency, fairness, sustainability and other
28 concerns may vary among experts and stakeholders. When Agency
29 decisions require tradeoffs among and along these dimensions and are
30 likely to lead to conflict, agencies are often forced to make judgments that
31 cannot be based solely on knowledge. The selection of an appropriate
32 type of deliberation will depend on where the extent of agreement falls on
33 the continuum between high and low.
34
35
36
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
STATE OF
KNOWLEDGE
Insufficient
DELIBERATION TYPE
Sufficient
EXPERT
DELIBERATION
INTEGRATED
DELIBERATION
OVERSIGHT
DELIBERATION
STAKEHOLDER
DELIBERATION
STATE OF VALUE AGREEMENT
Figure 5-3. Typology of Deliberation Processes With Stakeholders
and Experts (adapted from Chess, Dietz, and Shannon, 1998)
//. To what extent is the state of knowledge sufficient to address the
problem at hand?
By knowledge we mean information and understanding from the biological
and physical sciences, engineering, economics, and the other social
sciences. The answer to this question depends on the extent of
knowledge about information critical to making a particular decision. In
many situations, knowledge from environmental sciences, such as
information about environmental system processes and the nature of
potential threats, may play a decisive role. In others, knowledge about
economic costs and benefits may be decisive. Social science knowledge
may also be decisive, for example by providing an understanding of
communities where demographics, ethnicity, or racial composition must
5-29
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 be given serious consideration. Agency diagnosis of a situation, and the
2 form of deliberation needed, will depend, in part, on its assessment of the
3 extent to which available information is adequate for making a decision.
4 When the state of knowledge is insufficient or controversial, or when there
5 is lack of agreement about the state of knowledge, more extensive
6 deliberation will be needed.
7
8 The Subcommittee appreciates that Agency decisions may be constrained by
9 regulations, resources, and court decisions. These limitations should be made clear to
10 participants in any deliberative process. Yet, situations in which agencies have no
11 latitude are rare (Pflugh and Shannon, 1990). Without reflection on the above two
12 questions, agencies may fail to appreciate the likelihood that a decision, made without
13 appropriate deliberation, might become hopelessly stuck in environmental gridlock.
14
15 b) When agreement about values (economic and non-economic) is high
16 and the state of knowledge (relevant science, economics, and social
17 science) is sufficient and/or non-controversial, Agency decision-
18 making is likely to be routine. Deliberation will only be needed
1 g periodically, if at all, for oversight (Oversight Deliberation).
20
21 Most Agency decisions are routine administrative ones that conform to existing
22 regulations and policies. Such decisions may include non-controversial permitting,
23 changes in labeling, and minor shifts in administrative procedures. In such situations,
24 oversight deliberation, the periodic conferring of experts to assess a program and
25 potential modifications, is appropriate. However, if conflict develops around multi-
26 dimensional tradeoffs or the state of knowledge, the type of deliberation will need to
27 move toward another quadrant.
28
29 c) When agreement about values is low, but the state of knowledge is
30 sufficient and/or non-controversial, Agency decision-making will
31 require multi-dimensional tradeoffs based on knowledge.
32 Stakeholder Deliberation is needed.
33
34 In such situations, the state of knowledge is sufficient to inform multi-dimensional
35 tradeoffs, but there is little agreement about which tradeoffs to make. Because the
5-30
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 conflict is usually based fundamentally on values, not knowledge, the deliberation can
2 involve primarily stakeholders who will evaluate tradeoffs in light of their priorities.
3 Stakeholders, informed by available knowledge, can craft options with varying tradeoffs.
4 Experts may provide information about the potential impacts of various options, but they
5 need not be as extensively involved as in expert deliberation, as noted in the following
6 section.
7
8 d) When agreement about values is high, and the state of knowledge is
9 insufficient and/or controversial, Agency decision-making is likely to
10 be experimental and iterative. Expert Deliberation is needed.
11
12 In such situations, making decisions is difficult primarily because of the state of
13 knowledge. For example, there may be limited knowledge about the impact of human
14 management on a particular environmental system. Expert deliberation — on-going
15 conferring among experts (often from different disciplines) — can be needed to develop
16 appropriate monitoring processes and to interpret results. Based on the results, expert
•f deliberation may result in recommendations for changes in management of the
environmental system; i.e., adaptive management with expert deliberation at intervals
19 determined by the nature of the experiment. For example, monitoring of the impact of
20 reducing water flow to an environmental system may require experts to confer at regular
21 intervals to review monitoring data and determine if the water flow should be changed.
22 However, if value-based conflict arises over the results of such iterative decision-
23 making, the situation will require integrated deliberation, involving experts and outside
24 stakeholders working together to make multi-dimensional tradeoffs on the basis of
25 limited knowledge.
26
27 e) When agreement about values is low and the state of knowledge is
28 insufficient and/or controversial, Agency decision-making is likely to
29 require multi-dimensional tradeoffs based on insufficient knowledge.
30 Integrated Deliberation involving both experts and outside
31 stakeholders is needed.
32
33 These decisions are usually the most difficult for agencies because there is little
34 confidence in the state of knowledge about the impacts of tradeoffs on economic
35 efficiency, fairness, sustainability and other concerns. In such situations integrated
5-31
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 deliberation may be needed. By integrated deliberation we mean on-going interaction
2 among experts and stakeholders during problem formulation, collection of information,
3 and development of options (NRC, 1996; Commission on Risk Assessment and Risk
4 Management, 1997). Integrated deliberation may also be needed during
5 implementation and performance evaluation, and may take the form of adaptive
6 management with stakeholders and experts reviewing the results and suggesting
7 iterative changes. The nature of integrated deliberation depends on the situation, but,
8 in general, the greater the conflict (or potential conflict), the more extensive the
9 deliberation needed.
10
11 f) EPA can identify stakeholders by asking the following questions
12 (adapted from Chess and Hance, 1994):
13
14 -Who has information and expertise that might be helpful?
15 -Who has been involved in similar decisions before?
16 -Who has wanted to be involved in similar decisions before?
17 -Who may be affected by the decision?
18 -Who may reasonably be angered if not included?
19
20 The process for selecting stakeholders should be perceived as fair, lead to a
21 broad representation of stakeholders, and should ensure that representatives of
22 stakeholder groups are perceived by the groups as acceptable representatives (NRC,
23 1996; English, 1993).
24
25 Stakeholder involvement processes need not preclude involvement by a greater
26 number of participants. In fact, in conflicted situations it may be useful to complement a
27 stakeholder involvement process with survey data (see, e.g., Kathlene and Martin,
28 1991). Researchers are also exploring forms of deliberative processes such as citizen
29 juries, which involve random selection of citizens (Brown et al., 1995; Crosby et al.,
30 1986).
31
5-32
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 g) Involving stakeholders before extensive development of Agency
2 proposals may generate constructive options and trust. Integrated
3 and stakeholder deliberation should include stakeholder involvement
4 in problem formulation and options generation.
5
6 Early involvement facilitates agencies' ability to define problems and craft
7 solutions that may meet satisfactorily the goals of both agencies and stakeholders.
8 When participation is on-going, the Agency can identify problems before they become
9 crises (Thomas, 1995). As the Government Accounting Office (1994) has suggested in
10 the context of Superfund, earlier outreach can improve participatory efforts and
11 minimize dissatisfaction.
12
13 h) We recommend that the existing literature on group processes,
14 ongoing research on deliberation, experienced practitioners, and
15 stakeholder input be used to provide guidance on the selection of
16 deliberative methods.
.»^
Just as selection of the type of deliberation is determined by the situation, the
19 selection of deliberative methods is dependent on the context (English, 1993). There is
20 no single guide that can cover all situations, and care must be taken in designing a
21 process so it is appropriate to the problem being addressed. As with all policy analysis
22 tools, high quality results depend on careful application of the methods that are
23 grounded in both theory and practical experience. The sources of guidance for
24 selection of deliberative processes include:
25
26 Existing Literature. There is a substantial and sophisticated body of
27 research on small group processes and a smaller but important body of
28 research on deliberative processes. Key entry points to this literature
29 include Mazmanian and Nienaber, 1979; Rosener, 1981; Fiorino, 1990;
30 Dietz, 1994; Renn et al., 1995; NRC, 1996; and U.S. Department of
31 Energy, 1996.
32
33 Ongoing Research. Increasing effectiveness in the use of deliberation
34 will require ongoing research on deliberative processes and their use in
35 policy contexts. Some research of this type is proceeding, including a few
5-33
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 projects funded by the recent EPA/National Science Foundation joint
2 initiative. But compared to other policy analysis methods, very little has
3 been invested in research to support deliberatiorv(NRC, 1996).
4
5 Experienced Practitioners. Like many methods of policy analysis, the
6 translation of theory and empirical research into practice is a craft that
7 depends on practitioners who are both knowledgeable and experienced.
8 Such practitioners can also be an important source of systematic
9 information. However, deliberation rarely can simply be turned over to a
10 consultant. Both integrated deliberation and stakeholder deliberation
11 require Agency involvement.
12
13 Stakeholder Input: Those participating in a deliberation are a vital source
14 of information on how to proceed. Further, systematic post-hoc studies of
15 stakeholder perceptions are critical to building cumulative knowledge (see,
16 e.g., Balch and Sutton, 1995; and NRC, 1996). For further examples of
17 such studies, see also "Existing Literature" above.
18
19 i) EPA should build additional institutional capacity for deliberation.
20
21 Effective decision-making depends upon institutional capacity appropriate to the
22 challenges posed by the situations we have described. That institutional capacity is
23 inadequate, both in the larger community and in EPA, and the Agency should move to
24 develop it in both arenas. Research and training support, for example, would deepen
25 the resources available to EPA, while greater use of deliberative processes would call
26 forth additional interest in the outside community. EPA can utilize external resources to
27 make more effective use of deliberation, but it also needs to increase its internal
28 capacity by adding expertise and making other organizational changes such as:
29
30 Expertise. Deliberative processes depend on experienced, trained
31 practitioners and knowledge and skills provided by those with familiarity
32 with social science research in areas such as psychology, sociology,
33 anthropology, political science, management, and economics. EPA may
34 need to hire personnel with this expertise.
35
5-34
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Organizational Structure. Currently, community relations, public
2 information, and other personnel serve as technicians (carrying out
3 instructions) rather than being part of organizational decision-making
4 (Grunig, 1992). Thus, key decisions may be made without the presence
5 of a manager with expertise in stakeholder values and concerns. We
6 suggest that those with expertise in deliberation should serve not only as
7 technicians, but as both mid-level and senior managers, who can assist
8 the Agency to develop appropriate deliberative processes and to make
9 critical decisions that involve conflicts over values.
10
11 Organizational Learning. To improve deliberative processes, agencies
12 must leam from experience (Chess et al., 1995). Mechanisms are
13 needed to collect, analyze, and interpret data on deliberative processes.
14 Just as records of environmental monitoring or risk assessments help to
15 improve those practices, learning about deliberation will require record
16 keeping and means to institutionalize what is learned.
17
j) Because environmental systems are arranged in hierarchies,
19 effective deliberation requires institutional arrangements that reflect
20 this structure, with appropriate authority, institutional capacity, and
21 information available at local, regional, and national levels.
22
23 Such institutional arrangements allow citizens who are directly affected by
24 tradeoffs among fairness, economic efficiency, and sustainability to deliberate over their
25 resolution and to bring to bear critical information about local conditions, particularly
26 those involving spatial and temporal distribution of resources.
27
28 Relationships among institutions at the same level and with those at other levels
29 should be relatively flexible, allowing stakeholders and experts the leeway to match
30 institutional capacities with the specific ecological problem. To enable community-
31 based institutions to participate in making multi-dimensional tradeoffs, care should be
32 taken to avoid severe imbalances among different levels of institutions.
33
34 When community-based institutions have already developed management of
35 resources and made multi-dimensional tradeoffs, other levels of government should
5-35
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 recognize the local management systems that are in place and, to the extent possible,
2 work with rather than supplant those systems (see, e.g., Schlager, 1994). While
3 community institutions can draw on local knowledge about distribution of resources over
4 space and time, they are often at a disadvantage in involving additional scientific
5 expertise. Conversely, higher level institutions have greater capacity and resources to
6 draw on scientific expertise but may be at a disadvantage in their appreciation of local
7 knowledge about the functioning of an environmental system or of a resource within it.
8 Thus, successful resolution of sustainability problems and multi-dimensional problems
9 may be more likely if scientific capacity of local institutions is increased and higher level
10 institutions draw on local knowledge.
11
12 k) Deliberation should be protected from undue manipulation by
13 special interests.
14
15 Conflicts among interest groups vying for power is inherent in democracy.
16 Deliberative processes can be protected from undue manipulation through: a) open and
17 transparent processes, b) rules and mechanisms to ensure due process; c) sufficiently
18 broad participation so that deliberation is not dominated by special interests, d)
19 structured group processes that reduce the possibility of manipulation, and e) expertise
20 of those leading the processes so that manipulation is seen quickly and constrained.
21
op *****
23
24 The following section presents information on additional approaches to valuing
25' benefits of environmental systems. The additional approaches are discussed in
26 response to the need to improve this ability to value environmental benefits in ways that
27 address the limitations of traditional economic analysis techniques. The following
28 considerations, and others, need to be the subject of continuing research as society's
29 needs to address environmental issues become increasingly more complicated and
30 exceed the adequacy of current methodologies.
31
32
33
34
35
5-36
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 5.6 Additional Approaches to Valuation of Environmental Systems
2
3 5.6.1 Introduction
4
5 "Valuation ultimately refers to the contribution of an item to meeting a specific
6 goal. A baseball player is valuable to the extent he contributes to the goal of the team's
7 winning. In ecology, a gene is valuable to the extent it contributes to the goal of survival
8 of the individuals possessing it and their progeny. In conventional economics, a
9 commodity is valuable to the extent it contributes to the goal of individual welfare as
10 assessed by willingness to pay. The point is that one cannot state a value without
11 stating the goal being served. Goals are not necessarily independent, i.e. increasing
12 activity toward one goal may result in tradeoffs which decrease other goals" (Costanza
13 and Folke, 1997)
14
15 5.6.2 Findings
16
17 a) The value of something stems from its ability to contribute to the
achievement of some goal.
19
20 b) Fairness, sustainability, and other values, as well as economic
21 efficiency, are goals that should be recognized and explicitly
22 addressed in environmental resource decision-making. Such goals
23 can be compatible and mutually reinforcing; i.e., sustainability could
24 be enhanced by improving the efficiency of resource use and
25 reducing ecological stressors, and fairness can enhance
26 acceptance of and compliance with environmental decisions.
27
28 Conventional economic value is based on the goal of maximizing individual
29 economic welfare, which includes all the things people want including environmental
30 quality and sustainability of ecological systems. But explicit recognition of such non-
31 private goals, and of their importance, is crucial in environmental valuation. For
32 example, sustainability is critical to future generations. It is important to properly
33 ascribe value to outcomes that contribute to achieving sustainability, social equity, or
34 other ends that may be deemed important, as well as the value placed on the goals of
35 producing things people want directly. This broadening is particularly important if the
5-37
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 goals are potentially in conflict. The narrow view of environmental decision-making as a
2 choice of one or another action, each representing widely differing value-driven
3 viewpoints, outcomes, benefits, and costs, is rarely the case. Rather, environmental
4 decision-making is an activity that involves trade-offs among a number of goals that are
5 the product of different value choices. Decisions must reflect some combination of
6 these disparate values. Goals that may be important in environmental decision-making
7 include those relating to aesthetics, services of nature, recreation, consumption,
8 culture, biodiversity, and human health. Because of the variety of values-driven goals
9 that enter the decision-making process, there is a need for a number of ways to
10 conduct valuation exercises that are used to inform decision-makers.
11
12 c) Ecological systems provide diverse benefits to society, including
13 marketed goods and services, nonmarket direct and indirect use
14 values, and nonuse or existence values. Ecological services have
15 values that can be identified and at least partially quantified.
16
17 d) Appropriately characterizing the societal value of the full range of
18 environmental services requires approaches in addition to traditional
19 individual preference techniques. Such techniques would inform the
20 democratic processes that can be used to determine and then
21 implement policies which enhance the flow of environmental
22 services that would not be provided by the operation of market and
23 quasi-market fulfillment of individual choice.
24
25 e) An important example of an element that may not be adequately
26 reflected in valuation exercises is biodiversity. Biodiversity is
27 positively related to ecological services and is declining principally
28 because of anthropogenic- induced land use changes and other
29 habitat alteration. Biodiversity has benefits, and its loss is a cost that
30 should be factored into decision-making.
31
32 Biodiversity is one of the important attributes of environmental systems which
33 clearly emerges when aggregates of species and populations are considered. The term
34 biodiversity is commonly used to describe the number and types of organisms in a
5-38
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 given area, as well as the variability (genetic and phenotypic) within populations of
2 organisms. In addition, biodiversity includes "the ecological roles that these species
3 play, the way that composition of species changes as we move across a region, and the
4 groupings of species (ecosystems) that occur in particular areas (such as grassland or
5 forest), together with the processes and interactions that take place within and between
6 these systems. It also covers the diversity of ecosystems in landscapes, of landscapes
7 in biomes, and of biomes on the planet." (Heywood et a!., 1995).
8
9 Because biodiversity reflects the complex interactions among species in an area,
10 and between living organisms and the physical environment, it is reasonable to
11 suppose that the services provided by a specific environmental system are affected by
12 the various levels of diversity contained within the system. Therefore, when conducting
13 valuation exercises, it is important to consider the biodiversity of the system.
14
15 According to Heywood et al. (1995), The values placed on biodiversity are
16 strongly linked to the human influences on it and their underlying social and economic
17 driving forces. They are also dependent on some degree of knowledge of the scientific
role of particular elements or processes of biodiversity in the functioning of our
19 ecosystems and societies.... [W]hile it is undoubtedly true that the multiple values of
20 biodiversity are not adequately captured in its market value, if we want to commit and
21 prioritize resources to its conservation and sustainable use, then applying economic
22 measures to its evaluation is unavoidable."
23
24 It is likely that the environmental values most often included in economic analysis
25 are those associated with environmental services flows having direct market values
26 (e.g., timber, Pharmaceuticals, fish, wildlife, etc.). However, the importance of
27 biodiversity to the provision of these, and many other service flows, has not yet been
28 adequately recognized or explored. Therefore, we need to identify critical functions and
29 to focus research efforts on these so that important measures reflective of the variety of
30 factors associated with biodiversity can be completely reflected in the analysis of
31 benefits from environmental management.
32
33 f) Environmental systems are organized in the form of nested
34 hierarchies, having spatial, temporal, and organizational dimensions.
35 For this reason, decision-making authority should be organized in a
5-39
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 way that mirrors or parallels the natural systems. A major benefit
2 would be reduced information costs and analytical complexity.
3
4 There is a need to manage events in the human/environmental system at a
5 range of different scales (Wilson and Dickie, 1995). In the management of natural
6 resources, the spatial scale inherent in the regulatory apparatus often influences the
7 nature of the questions that are addressed. For example, marine fisheries are
8 managed primarily on a regional scale (by the U.S. regional fishery management
9 councils), and this management has tended to focus on questions of how many fish to
10 harvest rather than on ecological factors such as habitat and population structure that
11 occur at smaller scales. In addition, large scale management oversimplifies local
12 phenomena that are important to regulatory goals; ecological interactions are complex
13 and the outcomes of human intervention are often uncertain. Because region or
14 ecosystem-specific differences can lead to large transaction costs in centralized
15 management schemes, an allocation of decision-making authority toward a
16 decentralized hierarchical regulatory system can bring timely and accurate information
17 to bear on both the regulatory and valuation problems.
18
19 With respect to valuation, the ideal is to exercise decision-making authority at the
20 scale where individual incentives are most closely aligned with societal objectives. Put
21 another way, individuals, acting in their own self interest, may over-exploit open access
22 resources (often referred to as the "tragedy of the commons"-Hardin, 1968).
23 Individuals acting as parts of groups or communities may adopt norms or rules of
24 behavior that serve to avoid such over-exploitation (Ostrum, 1990; Schlager et al.,
25 1994).
26
27 Another benefit of decentralized management is to improve the equity and
28 credibility of decision-making processes. These are important characteristics that can
29 improve the potential for stakeholder "buy-in" and compliance with decisions.
30
31 g) Ecosystem management is a promising new approach for making
32 choices in the face of the uncertainties associated with the risks and
33 benefits of management decisions, especially in the areas of fairness
34 and environmental sustainability.
35
5-40
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Many of the more straightforward environmental problems that have been the
2 focus of the last few decades of environmental protection in the U.S. (e.g., air and water
3 pollution by toxic chemicals) have been addressed in EPA's and other agencies'
4 environmental decisions and policies, resulting in improved environmental conditions.
5 However, the human activities that now cause higher risks for ecological systems are
6 more physical than chemical, involving such stressors as habitat alteration and
7 hydrologic modifications that cause effects at regional scales. The next generation of
8 advances in environmental protection, then, must deal with regional-scale interactions
9 of humans with natural ecological systems. Ecosystem management (also termed
10 community-based environmental protection [CBEP] by EPA) is an emerging concept
11 that addresses such regional issues and seeks to develop mutually dependent
12 sustainability for human and ecological systems (U.S. MAB, 1994; Christensen et al.,
13 1996; Harwell et al., 1996).
14
15 Ecosystem management uses a deliberative process for society to make explicit
16 choices about what level of sustainable ecological quality is desired at each specific
17 location within a region. This deliberative process requires extensive and recursive
interactions among scientists, decision-makers, and stakeholders to identify spatially
i9 explicit environmental goals, define the ecological and societal endpoints that
20 characterize those goals, and implement management approaches to achieve the
21 goals. This requires an approach to governance that is commensurate with the spatial,
22 temporal, and organizational hierarchies of the ecological systems. Further, ecosystem
23 management involves making decisions with explicit consideration of the uncertainties
24 associated with environmental risks and benefits.
26
26 Sustainability is a central desired tenet of ecosystem management.
27 Consequently, the time horizon for decisions is by necessity inter-generational in
28 human terms and, thus, issues of sustainability and inter-generational equity (i.e.,the
29 distribution of ecological services over generations of humans) are intimately linked.
30 Since ecosystem management is goal-driven, it is essential that science properly inform
31 the societal goal-setting process by defining what is sustainable and what constraints
32 the natural systems impose on management options. Just as importantly, however, it is
33 up to society, not scientists alone, to set the sustainability goals. Once these goals are
34 established, then science can help by translating the societal goals into specific
35 ecological, hydrological, and societal conditions and identifying the specific measures
5-41
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 that need to be monitored in order to evaluate how well the goals are or are not being
2 met. The comparison of how ecological systems respond to decisions and policies with
3 specified performance criteria constitutes the critical element needed for adaptive
4 management, which is a key component of ecosystem management. Consequently, for
5 regional-scale environmental protection, ecosystem management constitutes a driving
6 force both linking societal values with scientific understanding and guiding the
7 environmental decision-making and adaptive management processes, forming the
8 context in which the science and society, analytic/deliberative process develops (MAS,
9 1996).
10
11 h) Scientific analyses and models that explain the principles of
12 operation and the interconnections among socioeconomic and
13 ecological systems are important to environmental management.
14 Not all environmental problems require the same level of scientific
15 analysis and deliberation. Complex issues may need to emphasize
16 deliberation and use judgment based on qualitative analyses.
17
18 i) During deliberations and presentations to the Subcommittee, several
19 interesting directions emerged that, with further research, may
20 improve our ability to measure some of the emergent values that
21 environmental management approaches, and environmental systems
22 in general, confer on the economic system. Results of such
23 research should help to determine how best to align the regulatory
24 process with the opportunities for minimum cost compliance.
25 Research is also needed to determine how dynamic adjustments to
26 regulatory interventions can be incorporated better into the process
27 of estimating the prospective costs and benefits of possible actions
28 so that realistic information is given to decision-makers. For
29 example, research may be fruitful in the following areas:
30
31 1) The relationship between biodiversity and environmental
32 service flows of importance to humans:
33
34 As discussed above, it seems a good working hypothesis that
5-42
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 biodiversity of an ecosystem is related to the level of goods and services
2 of value to humans from that system.
3
4 2) Systems analysis of the relationship between economy and
5 environment:
6
7 A number of recent empirical studies in the United States have
8 shown that states with better environmental conditions generally have
9 better economic conditions. For example, Meyer (1992) demonstrated a
10 positive statistical correlation between higher state environmental rankings
11 and better job conditions, e.g., lower unemployment and higher
12 productivity. Similarly, Templet and Farber (1994) demonstrated that
13 surrogate measures of environmental risk (e.g., the ratio of chemical
14 releases, as reported by the Toxic Release Inventory (TRI) data base, to
15 jobs created in the discharging industrial sector) are inversely related to
16 measures of economic welfare (e.g., unemployment rates and disposable
17 income per capita), and Cannon (1993) showed that higher economic
growth rates prevail where environmental rankings are higher. Hall (1994)
19 developed state rankings based on economic and environmental
20 conditions that turned out to be remarkably similar and correlated well.
21
22 These results do not establish the fact or direction of causation.
23 While these results suggest that the economy could be enhanced if the
24 environment is improved, it is also possible that citizens in higher income
25 states have a higher demand for environmental quality and convince their
26 governments to deliver it. It also may be that the initial endowment of
27 natural and man-made resources in some states led to both economic
28 and environmental conditions being better.
29
30 A positive correlation between broad, general measures of
31 environmental conditions and measures of economic production is
32 frequent in circumstances where rampant industrial growth has occurred
33 without consideration of the damage done by residuals, a condition that is
34 familiar in some developing countries, and indeed, occurred historically in
35 the United States. When such externality conditions are presented in
5-43
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 specific situations (i.e., industrial water discharges damaging a fishery),
2 existing regulatory and tort remedies are readily available. The issue is,
3 however, whether more subtle and general levels of environmental stress
4 cause effects such that actions can be taken which would improve both
5 the environment and economic production. Research on this issue could
6 be designed to determine whether such circumstances exist, and if so,
7 what remedies are available that would improve environmental conditions
8 at a cost that did not exceed the improvement of economic production and
9 the increased value ascribed to the environmental improvement itself.
10
11 The complementarity interpretation can be understood in relation to
12 a simple model (Figure 5-4) which places the economic system within the
13 environmental system and reliant on it to provide energy, materials and
14 information and to accept waste. If the environmental system is
15 negatively impacted by disruption or residuals, it may contribute less of
16 these services to the economic system and, as a consequence, the
17 economy may become less productive and contribute less to public
18 welfare. In this view, the economy could be enhanced if the environment
19 is improved and vice versa.
20
21 3) Biophysical measures as a surrogate for the value of
22 ecological systems:
23
24 One alternative method for estimating ecological values assumes
25 a biophysical basis for value (e.g., Costanza, 1980; Cleveland et al., 1984;
26 Costanza et al., 1989). This theory suggests that in the long run, humans
27 come to value things according to how costly they are to produce, and
28 that this cost is ultimately a function of how organized the goods are
29 relative to their environment. To organize a complex structure takes
30 energy, both directly in the form of fuel and indirectly in the form of other
31 organized structures like factories. For example, a car is a much more
32 organized structure than a lump of iron ore, and therefore, it takes much
33 energy (directly and indirectly) to organize iron ore into a car. The amount
34 of solar energy required to grow forests can therefore serve as a measure
35
5-44
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
id
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
ENVIR ON WENT
Energy
Information*
Materials
ECONOMIC SYSTEM
Purchase Price (S)
\
r
Goods & Services
l
r
PRODUCTION \ /^
& ) ( CONSUMERS
DISTRIBUTION i \
Reconstituted
Materials (low
entropy)
t
Labor
Wages ($)
Stressors, e g.. Residuals,,
Disruption (high entropy)
ENVIR ON MENT
Figure 5-4. The Relationship of Economy to Environment
(modified from Templet, 1995)
of their energy cost, their organization, and hence, according to this
theory, their value. Various methods have been suggested in the
literature for estimating these direct and indirect energy costs ranging
from methods based on input-output analysis (Costanza, 1980; Cleveland
et al., 1984; Costanza et al., 1989; Ulanowicz, 1986; Mageau et al., 1995)
to methods based on embodied energy (or "emergy") analysis (Odum,
1996). For a comparison of input-output based and "emergy" based
analysis, see Brown and Herendeen (1996). Of these, the input-output
approach, borrowed from economics, appears to offer the most promise
5-45
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 for estimating direct and indirect energy costs.
2
3 4) The efficiencies and environmental benefits that may result
4 from the implementation of industrial ecology principles:
5
6 If Figure 5-4 is treated as the basis for a simple model, inputs of
7 energy, information and materials are seen to yield outputs from the
8 economic system in the form of goods, services, and residuals. Improved
9 processes for converting inputs into desired outputs can yield lower
10 energy use and/or lower production of residuals. Both of these results
11 tend to reduce environmental stresses. Recognition of these interactions
12 has led to a focus on "industrial ecology," which emphasizes a holistic
13 approach to organizing production processes as a substitute for the
14 piecemeal, sub-process decisions which often ignore potential beneficial
15 connections across the enterprise and even outside of it. Recycling of
16 residuals into useful products, use of waste heat from some operations as
17 input into others, changes in production processes to reduce residuals or
18 to yield residuals that are more easily recycled or used, and other means
19 of more nearly "closing the loop" have been implemented. The implication
20 of all this is that the relationship between energy and other inputs and
21 outputs, including residuals that stress the environment, is not immutable.
22 Hence, environmental policy actions that penalize the production of
23 residuals need not lead to proportionate reductions in output of desired
24 goods. Instead, in reaction to regulatory requirements, the system may,
25 and most likely will, adjust at multiple points and these adjustments will
26 lower costs (compared to a response that just dealt with the residual "at
27 the end of the pipe") and may have other beneficial effects. Therefore,
28 regulatory instruments should encourage private sector decision-makers
29 to look for means to lower harmful residuals while leaving them with
30 maximum flexibility in how this is accomplished.
31
32 Recent experience with industrial ecology suggests that a priori
33 estimates of the social costs of reducing residuals that may harm the
34 environment are often excessive. As the tenets of industrial ecology
35 become more widespread and widely adopted, the promise is that
5-46
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 improved environmental quality can be bought at a lower cost in lost
2 goods and services production than previously thought. This suggests
3 that estimates of the costs of reducing harmful residuals should
4 incorporate realistic opportunities for systemic adjustment. Further,
5 estimates of benefits from such action should be holistic as well. For
6 example, they should take into account any expected reductions in
7 environmental harm from input or process changes that would result from
8 the regulatory action.
9
10 The conclusion is that the optimal level of environmental quality
11 may be underestimated if too narrow and static a picture of response to
12 regulatory pressures is taken as the basis for regulatory decisions.
13 Consequently, opportunities for increases in social welfare may be lost.
14 Research on the efficiencies and cost-savings that may result from the
15 implementation of industrial ecology principles would improve our ability to
16 estimate accurately net benefits for different management scenarios.
17
In principle, the economic valuation framework can be used to formulate the
19 choices required to incorporate ecological and other factors into social decisions.
20 However, traditional economic analysis techniques have limitations of methodology and
21 completeness. This section has identified some of these limitations and sketched some
22 of the research directions, that if successful, may relax them. A key theme of this
23 section is that provision of a rich and diverse set of information to decision makers can
24 contribute to improved decision making.
25
26 5.7 Summary and Conclusions
27
28 Because many scientists, decision-makers, and other stakeholders feel that
29 economic analysis methods undervalue environmental resources, the EPA Science
30 Advisory Board established a Subcommittee as part of its Integrated Risk Project to
31 explore ways to advance the dialogue that would lead to a more complete framework
32 for estimating the benefits of environmental management actions. In meeting its
33 charge, the Valuation Subcommittee considered: 1) the environmental management
34 decision context for valuation; 2) the nature of value and values; 3) the economic
35 concept of value; 4) the importance of deliberative processes to valuation; and 5)
5-47
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 additional approaches to environmental valuation. This chapter is the result of the
2 Subcommittee's discussions on this issue. The Subcommittees conclusions for each of
3 the areas considered suggest the following:
4
5 Context is critical to valuation. Important contextual issues include: 1) the
6 type of environmental management action being considered; 2) societal norms
7 and objectives that underlie the action being considered; 3) the spatial and
8 temporal scope of the issue; 4) the identity of stakeholders; and 5) the range of
9 technological and behavioral options available to control the problem.
10
11 The norms, principles, and objectives of society represent its values,
12 provide the reason for addressing environmental problems, and provide
13 guidance in identifying the goals of environmental management actions. The
14 benefit of an action reflects its contribution toward the achievement of those
15 goals. Determining benefits cannot be separated from the need to reach
16 agreement on goals.
17
18 Fairness, sustainability, and other values, as well as economic efficiency,
19 form goals that should be recognized and explicitly addressed in environmental
20 resource decision-making.
21
22 For decision-making in a governmental context, environmental valuation is
23 recognized as an anthropocentric exercise. People care for things for different
24 reasons and in different ways. Things are valued in themselves (as ends) and
25 instrumentally (as a means for the satisfaction of other ends). People care about
26 some things because they desire them or because they contribute to their
27 happiness, and other things have value because people judge them to be good
28 or important in their own right.
29
30 Some analysis techniques reveal people's preferences for one or another
31 thing through empirical study. Other techniques attempt to elicit preferences
32 through interaction with individuals. Careful and unbiased interactions with
33 individuals permit us to leam the qualitative dimensions of their value
34 statements.
35
5-48
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 A basic premise of welfare economics is that economic values are based
2 on individuals' preferences and that people know their preferences. Where
3 individuals are ignorant of the roles of ecological functions in contributing to
4 valued service flows, it may be necessary to use experts' knowledge of the
5 functioning of environmental systems as an input in the valuation process. In
6 principle, the economic valuation framework can be utilized to define and
7 measure the economic values of changes in the functions and services of
8 ecological systems that affect individuals' welfare either directly or indirectly.
9 However, this framework may be difficult to implement in practice where the
10 relationship between the function and the service flow to individuals is indirect or
11 subtle.
12
13 Economic approaches, while consistent and coherent frameworks for
14 valuation, are not mechanisms for producing "the answer" since they may omit
15 trans-economic values that may be important, may include some elements that
16 are difficult or impossible to estimate, and may employ preference elicitation
17 processes that are incomplete.
1
19 Not all benefits or costs can be easily quantified. There is a need for
20 qualitative methods to provide valuation measures for decision-makers to use in
21 solving complex ecosystem problems. Care must be taken to assure that
22 quantitative factors do not dominate important qualitative factors.
23
24 To be most useful, valuation approaches, methods, and information
25 should be made as explicit as possible. These analyses are best used to inform,
26 but not dictate, decisions related to environmental protection policies, programs,
27 and research.
28
29 Deliberation among interested parties, though not required in all or
30 perhaps most decision-making situations, is often useful for eliciting information
31 on important goals, on "best judgment" on uncertain relationships, on measures
32 for judging goal attainment, and on value tradeoffs that must be made in
33 decision-making and implementation. Depending upon the sufficiency of
34 knowledge available to address the problem at hand, and the extent of
35 agreement that might exist on values involved in the issue, one of the following
5-49
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 types of deliberation may be useful during decision-making: oversight
2 deliberation, stakeholder deliberation, expert deliberation, or integrated
3 deliberation.
4
5 The relationship between energy and other inputs and outputs (including
6 residuals that stress the environment) is not immutable. The optimal level of
7 environmental quality may be underestimated if too narrow and static a picture of
8 response to regulatory pressures is taken as the basis for regulatory decisions.
9 Consequently, opportunities for increases in social welfare may be lost.
10 Research is needed to determine how best to align the regulatory process with
11 the opportunities for minimum cost compliance. It is also needed to determine
12 how dynamic adjustments can be incorporated better into the process of
13 estimating the prospective costs and benefits of possible actions.
14
15 Decision-making must proceed in a timely manner. Even though there is
16 much uncertainty associated with the factors important to decision-making,
17 decision-makers cannot wait for certainty to make and implement decisions. A
18 process of adaptive management can allow decisions to be made and
19 implemented in the face of uncertainty. Knowledge gained from implementation
20 experience and research provides feedback with which to revisit and revise past
21 decisions.
22
23 It is important that the total U.S. investment in environmental protection be
24 applied effectively and efficiently. Rather than simply address stressors in order
25 of some scientifically derived risk ranking, multiple risks should be considered
26 simultaneously, using both quantitative and qualitative factors, to determine the
27 optimal way to address the full suite of problems that are in need of attention.
28
29 The Subcommittee's work confirms the challenges and complexities of
30 environmental valuation exercises and the environmental management actions based
31 on those values. The Subcommittee recommends that expanded, rich, and complex
32 processes be employed to characterize environmental values more fully. These
33 processes should involve interaction and deliberation among scientists, decision
34 makers, and other stakeholders in order to identify goals, define endpoints to
35 characterize those goals, clarify important but uncertain relationships, and to implement
5-50
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 approaches to achieve those goals. Processes should include on-going dialogue and
2 adjustment, and they should consider: 1) why people care for the things they do; 2) the
3 appropriate use of deliberative processes to elicit preferences and the rationale for
4 them; 3) economic valuation frameworks to define and measure the economic value of
5 changes in environmental systems functions and services affecting individual welfare;
6 and 4) presentation of available physical or other quantitative measures, or qualitative
7 descriptions of the effects of alternative actions when costs and benefits are not fully
8 captured by monetary measures. The Subcommittee recognizes that environmental
9 valuation remains a craft embedded in political processes and that much additional
10 research is needed in all areas that are important to estimating the benefits and costs of
11 environmental management action.
12
13
5-51
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 5.8 References Cited
2
3 Arrow, K., M.L Cropper, G.C. Eads, R.W. Hahn, LB. Lave, R.G. Noll, P.R. Portney, M.
4 Russell, R. Schmalensee, V.K. Smith, .and R.N. Stavins. 1996. Is there a role
5 for benefit-cost analysis in environmental, health, and safety regulation?
6 Science. Vol. 272:221-222.
7
8 Balch, G.I. and S. Sutton. 1995. Putting the first audience first: Conducting useful
9 evaluation for a risk-related government agency. Risk Analysis. 15(2): 115-125.
10
11 Braden, J.B. and C.D. Kolstad, eds. 1991. Measuring the Demand for Environmental
12 Quality. Amsterdam, The Netherlands: North-Holland.
13
14 Brown, G.M., Jr. 1990. "Valuation of Genetic Resources," jnG.H. Orians, G.M. Brown,
15 Jr., W.E. Kunin, and J.E. Swierzbinski, eds., The Preservation and Valuation of
16 Biological Resources. Seattle, WA: University of Washington Press.
17
18 Brown, M.T. and R.A. Herendeen. 1996. Embodied energy analysis and EMERGY
19 analysis: a comparative view. Ecological Economics. 19:pp. 219-35.
20
21 Brown, P. 1997. Analytical and ethical frameworks for thinking about the environment
22 and soil. Presentation at a Workshop on Human Perceptions of the Soil during
23 the Inaugural Meeting of the International Center for Soil and Society. Univ. of
24 Maryland. July 1997.
25
26 Brown, T.C., G.L Peterson, and B.E. Tonn. 1995. The values jury to aid natural
27 resource decisions. Land Economics. 71 (2):250-260.
28
29 Brulle, R. J. 1995. Environmental discourse and environmental movement
30 organizations: A historical and rhetorical perspective on the development of U.S.
31 environmental organizations. Sociological Inquiry. 65.
32
33 Campbell, D.T. 1969. Reforms as experiments. Amer. Psychologist. 24:409-429.
34
35 Cannon, F. 1993. Economic Growth and the Environment, Economic and Business
5-52
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Outlook, Bank of America Economics-Policy Research Department, (415)
2 622-3215.
3
4 Chess, C. and B.J. Hance. 1994. Communicating environmental risk: Ten questions
5 managers should ask. New Brunswick: Center for Environmental
6 Communications. Rutgers University.
7
8 Chess, C., M. Tamuz, and M. Greenberg. 1995. Organizational learning about
9 environmental risk communication: The case of Rohm and Haas' Bristol Plant.
10 Society and Natural Resources. 8:57-66.
11
12
13 Chess, C., T. Dietz, and M. Shannon. 1998. Who should deliberate when? Human
14 Ecology Review. 5(1):45-48.
15
16 Christensen, N.L., A.M. Bartuska, J.H. Brown, S. Carpenter, C. D1 Antonio, R. Francis,
17 J.F. Franklin, J.A. MacMahon, R.F. Ross, D.J. Parsons, C.H., Peterson, M.G.
Turner, and R.G. Woodmansee. 1996. The report of the Ecological Society of
19 America committee on the scientific basis for ecosystem management.
20 Ecological Applications 6(3): 725-747.
21
22 Cleveland, C.J., R. Costanza, C.A.S. Hall, and R. Kaufmann. 1984. Energy and the
23 United States Economy: A Biophysical Perspective. Science. 255:890-897.
24
25 Commission on Risk Assessment and Risk Management. 1997. Framework for
26 environmental health risk management. Presidential/Congressional Commission
27 on Risk Assessment and Risk Management.
28
29 Costanza, R. 1980. Embodied energy and economic valuation. Science. 210:1219-24.
30
31 Costanza, R., S.C. Farber, and J. Maxwell. 1989. The valuation and management of
32 wetland ecosystems. Ecological economics. 1:335-361.
33
34 Costanza, R. And C. Folke. 1997. Valuing ecosystem services with efficiency,
35 fairness, and sustainability as goals. Pp. 49-70 in: G. Daily (Ed.) Nature's
5-53
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Services: Societal dependence on natural ecosystems. Island Press.
2 Washington, D.C. 392 pp.
3
4 Crosby, N., J. Kelly, and Schaefer. 1986. Citizen panels: A new approach to public
5 participation. Public Administration Review.
6
7 Dasgupta, P. 1990. "Commentary," in Gordon H. Orians, Gardner M. Brown, Jr.,
8 William E. Kunin, and Joseph E. Swierzbinski, eds., The Preservation and
9 Valuation of Biological Resources. Seattle, WA: University of Washington Press.
10
11 Dietz, T. 1987. Theory and method in social impact assessment. Sociological Inquiry.
12 57:54069.
13
14 Dietz. T. 1988. Social impact assessment as applied to human ecology: Integrating
15 theory and method. Pp. 207-227 in Human Ecology: Research and Applications.
16 Ed. R. Boarden, J. Jacobs, and G.R. Young. College Park, Maryland. Society for
17 Human Ecology.
18
19 Dietz. T. 1994. What should we do? Human Ecology and Collective Decision-making.
20 Human Ecology Review. 1:301 -309.
21
22 Dietz, T. and A. Pfund. 1988. An impact identification method for development program
23 evaluation. Policy Studies Review. 8:137-145.
24
25 Dietz, T., P.C. Stem, and R.W. Rycroft. 1989. Definitions of conflict and the
26 legitimation of resources: The case of environmental risk. Sociological Forum.
27 4:47-70.
28
29 English, M., A.K. Gibson, D.L. Feldman, and B.E. Tonn. 1993. Stakeholder
30 involvement: Open processes for reaching decisions about the future uses of
31 contaminated sites. Knoxville: Waste Management Research and Education
32 Institute.
33
34 Fiorino, DJ. 1990. Citizen participation and environmental risk: A survey of institutional
35 mechanisms. Science, technology and Human Values. Pp. 226-243.
5-54
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 Freeman, A. M., III. 1993. The Measurement of Environmental and Resource Values:
2 Theory and Methods. Washington, DC: Resources for the Future.
3
4 Freeman, A. M., III. 1997. "On Valuing the Services and Functions of Ecosystems," in
5 R. David Simpson and Norman L Christensen, Jr., eds., Ecosystem Function
6 and Human Activities: Reconciling Economics and Ecology. New York: Chapman
7 and Hall.
8
9 Gosselink, J.G., E.P. Odum, and R. M. Pope. 1974. "The Value of the Tidal Marsh,"
10 Center for Wetlands Research, Louisiana State University, Baton Rouge,
11 LSU-SG-70-03.
12
13 Goulder, L.H. and D. Kennedy. 1997. "Valuing Ecosystem Services: Philosophical
14 Bases and Empirical Methods," in Gretchen C. Daily, ed., Nature's Services:
15 Social Dependence on Natural Ecosystems. Washington, DC: Island Press.
16
1"7 Government Accounting Office. 1995. Superfund: EPA's community relations efforts
could be more effective. Washington. GAO B-247753.
19
20 Grunig, J. 1992. Excellence in public relations and communication management.
21 Hillside: Lawrence Erlbaum.
22
23 Hall, B. 1994. Gold and Green, Institute for Southern Studies. P.O. Box 531, Durham,
24 N.C. 27702.
25
26 Hardin, G. 1968. The tragedy of the commons. Science. 162:1243-1248.
27
28 Harwell, M.A., J.F. Long, A.M. Bartuska, J.H. Gentile, C.C. Harwell, V. Myers, and J.C.
29 Ogden. 1996. Ecosystem management to achieve ecological sustainability: the
30 case of South Florida Environmental Management 20(4) :497-521.
31
32 Hettige, H., R.E.B. Lucas, and D. Wheeler. 1992. The Toxic Intensity of Industrial
33 Production: Global Patterns, Trends, and Trade Policy, American Economic
34 Review, V.82, no. 2, May, pp. 478-81.
35
5-55
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Heywood, V.H., I. Baste, and K.A. Gardner. 1995. Introduction, in: Global Biodiversity
2 Assessment V.H. Heywood, Exec. Ed. Cambridge Univ. Press. Published for the
3 United Nations Envir. Programme. New York.
4
5 Kathlene, L. And J. Martin. 1991. Enhancing citizen participation: Panel designs,
6 perspectives, and policy formation. Journal of Policy Analysis and Management.
7 10:46-63.
8
9 Kelman, S. 1981. Cost-Benefit Analysis: An Ethical Critique. Regulation, vol. 5, no. 1,
10 pp. 33-40.
11
12 Mageau, M.T., R. Costanza, and R.E. Ulanowics. 1995. The development and initial
13 testing of a quantitative assessment of ecosystem health. Ecosystem Health.
14 1(4):201-213.
15
16 Mazmanian, D.A. and J. Nienaber. 1979. Can organizations change? Washington,
17 D.C. Brookings Institution.
18
19 Meyer, S.M. 1992. Environmentalism and Economic Prosperity: Testing the
20 Environmental Impact Hypothesis, Project on Environmental Politics and Policy,
21 Mass. Inst. of Technology, Bldg./Room E38-628, Cambridge, MA. 02139.
22
23 National Research Council. 1996. Understanding Risk: Informing Decisions in a
24 Democratic Society. P.C. Stem and H.V. Fineberg (Eds). Washington, D.C.
25 National Academy Press.
26
27 Odum, H.T. 1996. Environmental Accounting: EMERGY and decision-making. John
28 Wiley, New York, NY.
29
30 Ostrum, E. 1990. Governing the Commons. The Evolution of Institutions for Collective
31 Action. Cambridge Univ. Press. New York. Pp. 280.
32
33 Pearce, D. And D. Moran. The Economic Value of Biodiversity. Earthscan Publications
34 Ltd. London. Pp. 172.
35
5-56
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1 Perrings, C., et al. 1995. The Economic Value of Biodiversity. Jjr V. Heywood, ed.
2 Global Biodiversity Assessment. Cambridge, UK: Cambridge University Press.
3
4 Renn, O., T. Webler, and P. Wiedemann. 1995. Fairness and competence in citizen
5 participation: Evaluating models for environmental discourse. Dordrecht: Kluwer
6 Academic Publishers.
7
8 Rosener, J.B. 1981. User-oriented evaluation: A new way to view citizen participation.
9 Journal of Applied Behavioral Science.
10
11 Sagoff, M. 1988. The Economy of Earth: Philosophy, Law, and the Environment. New
12 York: Cambridge University Press.
13
14 Schlager, E., W. Blomquist, and ST. Tang. 1994. Mobile flows, storage, and self-
15 organized institutions for governing common pool resources. Land Economics.
16 70(3):294-317.
17
Sen, A. 1995. Rationality and Social Choice. American Economic Review, vol. 85, no.
19 1.PP-24.
20
21 Sexton, K. and B.S. Murdock, Eds. 1996. Environmental policy in Transition. Making
22 the Right Choices. The Minnesota Series in Environmental Decision-making.
23 Volume I. Center for Environmental and health Policy. School of Public Health.
24 University of Minnesota. Minneapolis, MN. 109 pp.
25
26 Shabman, LA., and S.S. Batie. 1978. The Economic Value of Natural Coastal
27 Wetlands: A Critique. Coastal Zone Management Journal, vol 4., no. 3, pp,
28 231-237.
29
30 Templet, P.M. 1995. The Positive Relationship between Jobs, Environment and
31 Economy: An Empirical Analysis and Review. Spectrum. Spring 1995 Issue,
32 pp.37-49.
33
34 Templet, P.M. and S. Farber. 1994. The Complementarity Between Environmental and
35 Economic Risk: An Empirical Analysis, Ecological Economics. 9:153-165.
5-57
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Thomas, J.C. 1995. Public participation in public decisions: New skills and strategies for
2 public managers. San Francisco: Josey Bass.
3
4 Toman, M.A. 1997. "Ecosystem Valuation: An Overview of Issues and Uncertainties,"
5 in R. David Simpson and Norman L. Christensen, Jr., eds., Ecosystem Function
6 and Human Activities: Reconciling Economics and Ecology. New York: Chapman
7 and Hall.
8
9 Ulanowicz, R.E. 1986. Growth and Development: Ecosystems Phenomenology.
10 Springer-Verlag, New York.
11
12 U.S. Department of Energy. 1996. Site-specific advisory board initiative: Evaluation
13 survey results. USDOE.
14
15 U.S. Environmental Protection Agency. 1990. Environmental Investments: The Cost of
16 a Clean Environment. Washington, D.C.
17
18 U.S. Man and the Biosphere Program. 1994. Isle au Haut Principles: Ecosystem
19 Management and the Case of South Florida (HDS 002), US Department of
20 State, US MAB, Washington, DC.
21
22 Vaughan, E. 1995. The significance of socioeconomic and ethnic diversity for the risk
23 communication process. Risk Analysis. 15(2): 169-179.
24
25 Wilson, J. and L. Dickie. Parametric Management of Fisheries: An Ecosystem Social
26 Approach, 1995. in Property Rights in a Social and Ecological Context, eds.
27 Susan Hanna and Mohan Munasinghe, The Beijer Institute of Ecological
28 Economics and the World Bank, Stockholm and Washington, pp. 153-166.
5-58
-------
PART IV INPUTS TO ENVIRONMENTAL DECISION-MAKING:
RISK REDUCTION APPROACHES
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
PART IV—INPUTS TO ENVIRONMENTAL DECISION-MAKING: RISK
REDUCTION APPROACHES
Preface
Previous chapters have discussed approaches for assessing and comparing
environmental risks and the implications of various risk reduction goals and choices on
economic or other measures of welfare. These analyses are important tasks in the
Problem Formulation and Analysis and Decision-making phases of the I ED framework.
An additional set of analyses required by the framework, however, relates to the overall
design and selection of risk management scenarios to address "the problem." This
latter set of analyses, including identification and selection of specific risk management
approaches, is the focus of the following chapter. The IED framework calls for
preliminary consideration of risk reduction options during Problem Formulation, but the
most in-depth consideration of options occurs during Phase II when the environmental
problem, or set of problems, and associated environmental goals, have been defined.
Options are analyzed with regard to their potential to reduce single and multiple risks of
concern, associated costs, sustainability, equity, and other criteria specified by the IED
participants.
The Risk Reduction Options Subcommittee (RROS) was composed of
individuals with broad experience in the various engineering and non-engineering
approaches to risk reduction so that a wide range of management tools would be
considered. The group was charged with developing a process for identifying the most
effective risk management approaches for a variety of types of risk problems that might
confront a decision-maker. To meet this charge, the RROS formed subgroups to
consider options analysis from three different perspectives: 1) for a stressor-based
problem (e.g., ozone); 2) for a geographically based problem (e.g., risks associated with
an urban area); and 3) for a media-based problem (e.g., contaminated groundwater).
The resulting 10-step method for identifying, screening, and selecting risk
reduction options is the subject of Chapter 6, and is summarized in Figure 6-1. The ten
steps are:
Step 1: Define the problem
IV-i
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
Step 2: Develop background information
Step 3: Identify the spectrum of risk reduction options
Step 4: Establish screening criteria
Step 5: Screen potential risk reduction options
Step 6: Evaluate remaining risk reduction options
Step 7: Refine the options
Step 8: Select an option
Step 9: Document the process
Step 10: Quantify option effectiveness
In relation to the I ED framework, Steps 1 through 4 are primarily the province of the
Problem Formulation Phase, whereas Steps 5 through 9 relate primarily to the Analysis
and Decision-making Phase. As noted in the I ED framework, however, some iteration
is inevitable between Phases I and II. The final step, Step 10, occurs during the
Implementation and Performance Evaluation Phase of the IED framework.
The first step in the 10-step method is to define the problem. The chapter
articulates the importance of a clear problem statement, including specific goals for
what/whose risk is to be reduced and by how much. Clear environmental goals, with
explicit statement of the relationships and the potential tradeoffs among goals, are the
foundation from which objectives for the risk reduction program can be derived. In
addition to information on the nature of the risks (e.g., location, severity, specific
subpopulations or endpoints affected, and exposure media), the chapter discusses
other apects of Problem Formulation that are particularly critical for the design of a risk
reduction program. For example, the desired state or goal must be stated in
measurable terms and methods for measuring improvements toward the goal (i.e.,
indirect measures such as stressor levels or environmental outcome measures such as
reduction in a specific adverse health effect) must be specified. It is also essential to
identify likely constraints on possible solutions (e.g., budget, time, jurisdictiona!
authority, and legality) that will help establish screening criteria for the risk reduction
options.
The chapter then goes on to describe the remaining 9 steps in the methodology,
from development of background information and identification and screening of
potential risk reduction options, to selection of an option that best meets the screening
IV-ii
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
and selection criteria and documentation of the decision. Various decision analysis
methods are discussed briefly, including benefit/cost analysis (described more fully in
Chapter 4), matrix qualitative ranking, and multi-attribute decision procedures. The last
step in the RROS method, quantifying the effectiveness of the risk reduction program,
is analogous to the Performance Evaluation phase of the IED framework. The topic of
performance evaluation and report cards is discussed further in Part V of this report.
In addition to describing a comprehensive decision process for selecting risk
reduction options, the chapter places special emphasis on the types of environmental
problems best or least suited to different risk reduction tools.
Although the chapter often refers to "the problem solver," integrated
environmental decision-making requires that many types of participants communicate
and contribute to problem formulation and solution. In addition, much of the chapter is
written around the assumption that the problem and the constraints surrounding its
solution have already been specified and that the task at hand is to choose among
approaches for risk reduction. It is worth emphasizing, however, that the IED
framework recognizes the important role of preliminary options analysis during Problem
Formulation, so that problems are not just defined on the basis of risk and our interest
in reducing risk, but also take into account our ability to reduce risk within likely
constraints.
The methodology for developing and selecting risk reduction options is
discussed primarily in terms of reduction of a particular risk, e.g., one associated with a
particular chemical stressor or source. The chapter notes that although it is often best
to address risks with a combination of risk reduction tools, this is often not done
because of inadequate information on the multiple sources of a stressor and their
relative contribution to total risk. Extending the methodology to an integrated problem
set containing risks from multiple stressors (e.g., those experienced by a particular
community) will further increase the complexity of the analysis. Thus, it will be
important to aggregate or disaggregate the problem set (e.g., using "root cause" or
"common source/common pathway" analysis) so that analysis of risk reduction options
is more manageable. The chapter also notes that the complexity of the analysis,
including the screening and selection of options, increases greatly with an increasing
number of criteria against which options are to be "optimized."
IV-iii
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
In summary, when applying the 10-step process to an integrated problem set,
consisting of multiple stressors, sources, or endpoints, the following issues arise:
1) Aggregation is helpful in identifying multiple stressors that may have a
common set of risk reduction options with the objective of selecting a set of
options that will provide the most risk reduction for the set of risks being
analyzed. Practically speaking, it will not be possible to optimize risk reduction
over all stressors of concern considered at once. For this reason, screening and
aggregation of stressors into manageable subsets should be driven by analysis
of common aspects of stressors, roots causes, and/or activities.
2) In order to compare risk reduction across sets of options and sets of risks,
and to evaluate risk reduction per unit cost, it is critical to have a common
measurement of risk or common denominator for all risks. In many cases,
comparisons of risks and risk reduction under different scenarios will involve
unlike risks (e.g., cancer risk in humans vs. chronic health effects vs. effects on
wildlife populations), even where those risks have a common "root cause" (e.g.,
a single stressor or source).
3) Uncertainty of the analysis is likely to increase as a broader set of options is
considered for more than one stressor. Sources of this uncertainty include the
relative contribution of different stressors/sources to the total aggregate risk; the
effectiveness of combined options for reducing aggregate risk; and the
benefit/cost, equity, or other tradeoffs involved in addressing groups of risks.
Some types of options are easier to implement when uncertainty is high (e.g.,
communication and education and environmental management systems, rather
than regulations/enforcement); the extent of uncertainty associated with a
decision may affect the balance of options selected.
IV-iv
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
CHAPTER 6. RISK REDUCTION OPTIONS
TABLE OF CONTENTS
6.1 Introduction and Approach 6-1
6.2 Define the Problem 6-6
6.3 Develop Background Information 6-12
6.4 Identify the Spectrum of Risk Reduction Options 6-16
6.4.1 Communication/Education 6-18
6.4.2 Enforcement 6-21
6.4.3 Engineering 6-23
6.4.4 International and Intergovernmental Cooperation 6-23
6.4.5 Management Systems 6-23
6.4.6 Market Incentives 6-23
6.4.7 Regulation 6-26
6.5 Establish Screening and Prioritization Criteria 6-33
6.6 Screen and Prioritize Potential Risk Reduction Options 6-40
6.7 Evaluate the Remaining Risk Reduction Options 6-48
6.8 Optimize the Options 6-54
6.9 Select an Option 6-59
6.10 Document the Process 6-65
6.11 Quantify Option Effectiveness 6-65
6.12 References Cited 6-71
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 CHAPTER 6. RISK REDUCTION OPTIONS
2
3
4 6.1 Introduction and Approach
5
6 The focus of this chapter is on the identification and selection of risk reduction
7 options to solve an environmental problem or set of problems. The methodologies
8 and thinking in this chapter are a product of the Risk Reduction Options
9 Subcommittee (BROS) and are articulated in the 10-step decision process that forms
10 the backbone of the discussion.
11
12 EPA, state, and local regulators have embraced the concept of risk-based
13 environmental decision-making, including the need to prioritize environmental
14 problem-solving efforts. Nonetheless, the RROS recognized that regulatory decision-
15 makers face many constraints in determining risk priorities and in selecting risk
16 management options, including constraints imposed by implementing statutes and by
Congress. The Subcommittee defined its goal as helping regulatory officials maximize
18 the risk reduction achieved for any fixed amount of resources, whether those
19 resources are public or private. This chapter, therefore, suggests a methodology for
20 achieving improved risk reduction outcomes, rather than recommending risk reduction
21 solutions for specific environmental problems.
22
23 The goal of the RROS was to develop a methodology that would: (1) identify an
24' appropriate set of options for any given human health or environmental problem, (2)
25 create a reproducible and documented decision process where risk reduction
26 decisions are optimized based on explicit decision-maker policy choices, and (3)
27 produce risk reduction decisions transparent to interested parties.
28
29 At the first meeting, the RROS explored different approaches to risk reduction
30 available to regulatory or community decision-makers. The approaches fall into one of
31 the following seven categories which may not be completely independent.
32
33 a) Communication and education approaches, including information or data
34 dissemination to consumers and communities, options that provide technical
6-1
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 assistance and technology transfer to emission producers, and options that
2 provide information to govemment(s);
3
4 b) Enforcement approaches, including government-implemented options to
5 obtain improved compliance with existing regulations as well as options that
6 involve non-government parties in enforcement activities;
7
8 c) Conventional and innovative engineering approaches, including options
9 where technology for achieving pollution prevention, waste minimization, or end
10 of pipe control is communicated, encouraged, or mandated;
11
12 d) International and intergovernmental cooperation, including both formal and
13 informal cooperation with other U.S. and non-U.S. government agencies and
14 non-government organizations;
15
16 e) Environmental management systems, including options that focus on
17 communication, regulation, and/or market incentives for improved management
18 systems which will reduce risks/liabilities and improve compliance.
19 Management systems cover the areas of setting performance expectations,
20 defining roles and responsibilities, identifying and managing risks, ensuring
21 communication and training, developing implementing procedures, and
22 assessing/measuring performance.
23
24 f) Market incentives, including a broad range of options that can be
25 implemented independently or in combination with regulation, communications,
26 or enforcement options. These options can include marketable permits,
27 deposit/refund systems, fees and taxes, subsidies and tax credits, differential
28 regulatory requirements based on performance, government procurement
29 approaches, etc.; and
30
31 g) Regulation including mandated technologies, performance levels, risk levels,
32 specific product composition, specific activity requirements, specific information
33 disclosure requirements, and so forth.
34
6-2
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 Knowledgeable experts in each of the seven areas were invited to share their
2 experience with these potential risk reduction tools. Each expert was asked to
3 address where the option has been used, what lessons have been learned about
4 where it works well and where it doesn't, existing obstacles to its use, biggest
5 advantages and disadvantages to the approach, the future directions for the
6 approach, and risk areas that could optimally benefit from the approach. These
7 presentations and opportunities for interaction provided Subcommittee members with
8 a common base of information for methodology development.
9
10 In order to pursue methodology development, the RROS began by dividing into
11 three subgroups. Each subgroup selected a different approach to a human
12 health/environmental problem. One subgroup selected ozone, a stressor-based
13 approach. Another subgroup selected Elizabeth, New Jersey, an urban
14 geographic-based or location based approach. The third group selected groundwater,
15 a media-based approach. All three groups had the same goal; to develop a risk
16 reduction decision methodology that policy-makers could utilize and interested parties
would find transparent. By selecting examples of three very different types of problem
18 sets, it was anticipated that a universal methodology that could work for any human
19 health/environmental problem would result.
20
21 Once the three subgroups completed their work, a single common ten-step
22 methodology was jointly developed by the full RROS. That methodology was tested
23 against a fourth environmental area of concern, namely, "How can EPA best
24 implement the tolerance-setting requirements laid out in the Food Quality Protection
25 Act of 1995?"
26
27
6-3
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 The overall risk reduction methodology devised by the RROS contains ten
2 steps, summarized in Figure 6-1, and discussed in detail in subsequent sections. The
3 ten steps are:
4
5 1. Define the problem,
6 2. Develop background information,
7 3. Identify the spectrum of risk reduction options,
8 4. Establish screening and prioritization criteria,
9 5. Screen and prioritize potential risk reduction options,
10 6. Evaluate the remaining risk reduction options,
11 7. Optimize the options,
12 8. Select an option,
13 9. Document the process, and
14 10. Quantify option effectiveness.
6-4
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
2
3
4
5
6
7
8
9
10
Define Problem:
Devetpp Background Information
External
^Information^
Identify All Risk Reduction Options
u
Experience]
Establish Screening Criteria
.Screen and Prioritize Options
U
Evaluate Remaining Options
Optimize Options
Rgure 6-1. Risk Reduction Option Selection Methodology
The RROS methodology has multiple applications. First, if a decision-maker is
faced with a specified environmental area of concern, he/she can apply the
methodology to obtain the best risk reduction outcome for that problem. A single area
of environmental concern may be expressed in terms of: a) individual stressors, such
as heavy metals or particulates; b) activities, such as energy use or transportation; c)
specific industries; or d) geographic areas, such as the Gulf of Mexico or the greater
Chicago area. The RROS methodology provides a formal approach for evaluating
6-5
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 various risk reduction options so that the decision-maker can select an outcome that
2 maximally achieves goals while considering existing constraints.
3
4 Another application of the methodology is to select a single risk reduction
5 outcome that will maximize cost-effective risk reduction for a number of different
6 environmental areas of concern. For example, the decision-maker may want to
7 identify which type of risk reduction outcome would maximize risk reduction for a set of
8 stressors. Or, the decision-maker may want to identify which type of risk reduction
9 option would maximize risk reduction for certain activities, such as energy utilization,
10 transportation, and agriculture.
11
12 A third application of the methodology is to compare the potential risk reduction
13 outcomes for a several different environmental areas of concern and select the one
14 risk reduction option that maximally achieves goals while considering constraints.
15
16 The comparison of risk reduction options is contingent upon the ability to
17 quantify, or measure, risk. Qualitative risk evaluations do not work well in this
18 methodology.
19
20 6.2 Define the Problem
21
22 Before attempting to solve a problem it is necessary to develop a clear
23 definition of the problem. People often minimize the importance of defining a problem
24 for which they seek a solution. They desire to get right to problem solving, without first
25 making sure that they understand completely and precisely what needs to be
26 corrected or improved. In the risk reduction context, failure to adequately define the
27 problem can result in a sub-optimal solution or possibly a solution that does not work
28 at all.
29
30 The assumption is that at this stage the problem has been identified. The
31 challenge is simply to choose among various approaches to reducing risk for a
32 specific problem. It would be easy, therefore, to jump directly into problem solving or
33 to a solution.
34
6-6
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 This section describes the need for a clear articulation of the problem. It also
2 discusses the four components of a clear problem statement, including the use of dis-
3 aggregation (breaking the problem apart), the need for stakeholder input, and the
4 need for coordination among and within governments.
5
6 Today's environmental problems are extremely complex and interrelated. For
7 example, exposure to lead may come from many sources, some of which are
8 interdependent. Inhalation of volatilized lead in air, ingestion of lead-contaminated
9 drinking water, and ingestion by children of paint chips containing lead are but a few of
10 the sources of lead exposure. Exposure to pesticides presents a similar range of
11 sources: workplace exposure experienced by manufacturing workers, formulators and
12 applicators; incidental exposures in lawns, golf courses or vegetable gardens;
13 exposure to grocery produce; and consumption of groundwater contaminated with
14 pesticides.
15
16 Given this complexity, pursuit of a solution without understanding precisely
what needs improvement will be ineffective. For example, does one seek to reduce
18 the lead burden overall, or is the interest in reducing (at least) the lead exposure of
19 children, who are believed to be more susceptible to its adverse effects?
20 Understanding these issues allows one to focus on the issues that will yield the
21 optimum results. It will allow the identification of the root cause of the problem so that
22 solutions can be tailored accordingly.
23
24 Problem definition is also essential in understanding and articulating the
25 objectives, or improvement goals, for the risk reduction program. These aspects of
26 the issue must be clearly stated as part of the problem definition as well. It is a
27 tautology to say that the over-arching objective of any risk reduction project is to
28 reduce risk and leave it at that. In any program designed to reduce a risk, or set of
29 risks, there will be complementary and competing objectives. All must be articulated,
30 and competing goals prioritized, before one can design a program to achieve them.
31
32 The program objectives must also be measurable. For example, an objective
33 for a lead reduction program might be: reduce the blood lead level in children in the
34 U.S. to a level at which adverse effects are not likely to occur. Likewise, a consumer
6-7
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 risk reduction program for residual pesticides might seek to reduce the level of
2 residual pesticides to a level that minimizes adverse effects based upon science's
3 current definition of what level is acceptable.
4
5 Without a clear understanding of the objective and the process for measuring
6 progress, it will be impossible to determine whether the program is effective. In other
7 words, these factors are crucial to defining success. Congress, environmental groups
8 and others frequently berate EPA for failing to show positive results from its programs.
9 This stems, in large part, from a lack of adequate criteria for defining program
10 success, or, in the case of some programs, using the wrong measurement criteria.
11 The problem-solver must clearly understand the objectives of the program if he/she
12 hopes to design a successful solution. Without this, developing a solution becomes a
13 game of "pin the tail on the donkey," in which the donkey's location is not known until
14 the tail (the solution) has been pinned on and the blindfold removed.
15
16 The overall objective of the problem definition phase is to develop a specific,
17 narrowly focused statement of what is wrong with the current state, and to provide a
18 clear understanding of the desired state, or goal. The objective of this phase is to
19 move from a version of reality that defies resolution to a version along the pathway to
20 success.
21
22 To accomplish this, one needs to proceed through the six-step process set forth
23 below.
24
25 a) Identify the elements of the problem context.
26
27 b) Understand and, if possible quantify, the magnitude of the harm.
28
29 c) Identify the goal of the risk reduction program or desired state after
30 implementation.
31
32 d) Identify any constraints, or hard limitations on the solution.
33
34 e) Define the affected population.
6-8
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 f) Narrow the problem to its essence, or core elements
2
3 The first step is to identify the elements or variables of a problem context. This
4 might also be characterized as understanding the nature of the problem. What is it
5 that needs to be improved? Is it reducing environmental harm, improving health, or
6 improving visibility, for example.
7
8 The second step involves an attempt to quantify, or at least understand
9 qualitatively, the magnitude of the harm. In addition to these fundamental questions,
10 the problem-solver must understand the characteristics of the problem context. Is the
11 problem site-specific, such as a waste site; or is it broad-based, such as nitrate runoff
12 into estuaries. What are the sources of the risk and the nature of those sources? Do
13 they involve multiple environmental media, or are they limited to a particular
14 environmental media? Are they large quantity, discrete sources or small quantity,
15 diffuse sources? What are the relationships, if any, among these variables? Answers
16 to these questions will significantly affect the structure of an effective risk reduction
program. For example, large quantity, discrete sources are inherently more
18 manageable through direct regulatory approaches than are small diffuse sources.
19
20 Third, while gathering information about the nature of the problem, one should
21 also gather information about the desired outcome. Specifically, by looking beyond
22 the solution to the desired state, the goal of the risk reduction program can be
23 identified. This understanding must come early in the process if the problem is to be
24 defined properly. In addition, to be able to measure progress toward the goal, the
25 desired state must be articulated in measurable terms.
26
27 The fourth step is to identify and articulate any constraints on the solution.
28 These are necessary to establish the screening criteria for risk reduction options.
29 Some likely solution options constraints include: budget, time, jurisdictional authority of
30 the implementing agency and the legality of the proposed solution. For example, if it
31 is desired to achieve the risk reduction within 5 years, that should be identified as a
32 constraint so that solutions that will take more than 5 years will be eliminated in the
33 screening stage. Jurisdictional authority is one screening criterion that must be used
34 cautiously. One of the key recommendations of this section is that governmental
6-9
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 agencies coordinate and work more closely with one another. Therefore, one would
2 not want to screen out options simply because multiple agencies must be involved.
3 There are, however, circumstances in which Congress will not have delegated
4 adequate authority to any agency to implement a particular solution. In those cases, it
5 will be necessary to limit the risk reduction options considered to those within the
6 scope of?? [DFO note: some text missing here?]
7
8 The legality of the proposed solution could become an issue in a variety of
9 ways. The proposed solution might not be capable of receiving the necessary
10 licenses or permits, or there might be antitrust concerns, particularly in the context of
11 market-based solutions.
12
13 The fifth step in the process is to define the affected population. For risks that
14 affect human health, this requires a clear understanding of the receptor population.
15 Are the risks imposed on a population in a specific geographical area, near a waste
16 site, for example? Or, do the risks cross geographical boundaries, such as in the case
17 of pesticides? For risks that only affect the environment, what is the affected
18 ecosystem? Who are the users of the resource?
19
20 Properly defining the problem requires divergent thinking, to insure that all of
21 the variables are explored. This requires solicitation of inputs from diverse
22 stakeholders and experts, including the affected population and the users of the
23 natural resource, other governments (multi-national and state) and other agencies,
24 and the potentially regulated community.
25
26 One must balance the desire for broad stakeholder involvement with the
27 recognition that there is an associated process efficiency cost. Allowing for public and
28 other inputs to any process necessarily takes time. For problems to which there is a
29 need or desire to move quickly, it may be necessary to allow limited, or in some cases
30 no, public input. For example, cleaning up an oil spill, or addressing an imminent and
31 substantial endangerment under the Superfund Program, requires expeditious action
32 and therefore may not allow for public input because a lack of action may exacerbate
33 the problem. Even in the case of problems for which longer-term solutions are
34 appropriate, the cost in time and resources must be considered in determining how
6-10
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 extensive a stakeholder input process to devise. In general, however, appropriate
2 stakeholder input strengthens the overall decision-making process. Important
3 considerations for selecting the type and extent of deliberation for a particular decision
4 are discussed more fully in Chapter 5.
5
6 Beyond simple stakeholder input, it is critically important that the agency
7 leading the risk reduction effort coordinates with and involves other agencies that may
8 have overlapping or related jurisdiction over the problem. It is so often the case today
9 that one federal agency is not fully coordinated with others when developing programs
10 to solve problems. When that happens, at best sub-optimal solutions are the
11 outcome. At worst, the program is a failure. Therefore, in the course of defining the
12 problem, the lead agency should consult at multiple points in the process with any and
13 all relevant agencies. This should include not only federal agencies, but state
14 agencies, other national government agencies and international bodies as well.
15
16 After developing a broad array of elements of the problem, the problem solver
must proceed to the sixth step: identifying those elements that constitute the essence
18 of the problem--tf?e rea/prob/em. At this stage, it is necessary to narrow the problem
19 definition so that it becomes one of manageable dimension and to ask if it is effective
20 to deal with many contributing problems, or if they should be separated into multiple
21 problems. Moreover, one must consider whether the problem should be addressed at
22 a higher level.
23
24 This presents squarely the issue of aggregation and dis-aggregation of
25 problems, terms describing the process of combining common or hierarchical
26 problems, and, alternatively, separating them into smaller parts. Aggregation requires
27 that one deal with multiple problems, usually at a higher level. This may be
28 appropriate, particularly if the root cause(s) of the problems are common or related.
29 Alternatively, it may be better to dis-aggregate problems to better understand the root
30 cause(s) of each problem more precisely. Dis-aggregation may also be necessary to
31 create a problem that is resolvable, as opposed to simply a jumble of confusing
32 issues. Problem aggregation and dis-aggregation may need to be an iterative
33 process. The problem-solver may find that after he/she defines the problem and
34 begins preliminary work on solutions, the problem needs to be dis-aggregated. This
6-11
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 takes him/her back to the problem definition stage, where he/she must formulate a
2 workable problem definition once again.
3
4 Problem definition may require the problem solver to circle through the 6-step
5 process repeatedly. The result, however, will be an optimum problem definition.
6
7 The output of this process should be a clear, focused statement of the problem
8 that contains the following elements. First, it should contain a clear statement of the
9 current situation. This should also contain a clear articulation of the problem, or what
10 is wrong with the current state. Second, it should contain an explanation of the
11 desired state. This must be stated in measurable terms. If possible, the problem
12 definition should also describe the method or methods for measuring improvements
13 toward the desired state.
14
15 6.3 Develop Background Information
16
17 It is prudent to consider the larger problem in which the risk reduction
18 opportunity is imbedded before thinking about information needed to consider the
19 merits of a specific risk reduction option. For example, if the risk reduction option
20 under consideration is subsidizing the construction of radon resistant homes, the
21 process should begin with an analysis of the radon exposure problem. What has
22 been done to date with the problem within the Agency or within other governmental
23 and non-governmental organizations? Why hasn't the problem been solved or at least
24 substantially reduced? If previous attempts have been made to reduce the risk, how
25 much did different management options contribute toward meeting the goals? Did
26 these options take longer than anticipated to accomplish the task? How much did
27 scientific uncertainty contribute to undermining management options? If it is a new
28 problem, why has it only recently come to attention? What are the key elements of
29 the problem that need to be addressed before a solution can be devised? What might
30 be learned about the problem in the near future that may change the acceptability of a
31 risk reduction option?
32
33 After considering these contextual questions, the next action is to consider what
34 specific information about toxins, engineering, economics, geography, psychology,
6-12
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 law, and other subjects are necessary to start a process that will lead to assessing the
2 utility of a specific policy or implementation option. Does this information exist? If the
3 answer is yes, who has the data? How can it be obtained, and in what forms? Is it
4 reliable? If it is not highly reliable because of the way it was gathered or recorded,
5 and/or if the scientific basis supporting it is shaky, are there available time and
6 financial resources to gather high quality data? Or are there other data that can
7 substitute for the best data? Can professional experience or the judgment of informed
8 members of the public substitute for what is uncertain, or at least reduce the level of
9 uncertainty in the data?
10
11 At the end of the process, one should know:
12
13 a) How the problem came to exist,
14
15 b) How it has changed over time and across space,
16
c) Who has taken policy or implementation actions to date and what their
18 results have been,
19
20 d) What are the alternative risk reduction options,
21
22 e) Where the most promising places in the casual sequence to introduce risk
23 reduction options exist,
24
25 f) What the opportunity is for a multi-stage hybrid risk reduction option, and
26
27 g) What forces are likely to constrain each potential risk reduction option.
28
29
6-13
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 There are two general categories of information needed to screen options.
2
3 Risk Management Information
4 Spatial and temporal scale of the problem.
5 Risk reduction objective.
6 List of identified options.
7 Screening factors and constraints.
8 Decision-maker(s) identified and committed to the project
9 Prospective implementer of selected options.
10
11 Options Performance Information
12 Chance of success of an option within a given set of constraints.
13 Chance of success of composite options within a given constraint (including
14 synergistic effects)
15
16 Information on risk management scenarios would be collected during the
17 stages that precede screening. This information is integrated into the work done on
18 problem definition, development of key background information, and identification of a
19 full range of risk reduction options.
20
21 Options performance information is required to estimate whether an option is
22 likely to be successful within a given set of constraints. Such a determination can be
23 based on one or more of the following information bases; expert opinion, previous
24 experience, and predictive methods including mathematical modeling. Using a site-
25 specific case as an example, one may consider various treatment options (nested in
26 the engineering category of options) for reducing contaminant migration source terms
27 at a contaminated site. As determined by the kinetics of physicochemical/biological
28 phenomena and other implementation factors that apply to a given technology, a
29 minimum remediation time is usually required. Thus, the use of quantitative methods
30 to assess whether or not the time for implementing the technology meets the time
31 requirement stated in the risk reduction objective would be useful. Quantitative
32 methods for such assessments can be deterministic or probabilistic, but they can be
33 structured to estimate the reliability of an option.
34
6-14
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
6-15
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 6.4 Identify the Spectrum of Risk Reduction Options
2
3 The protections delivered by the U.S. environmental management system
4 derive largely from technology-based or harm-based emission limitations although the
5 system also relies, to a lesser extent, on other tools such as market incentives and
6 taxation. The past five years have seen a contentious debate about the structure and
7 composition of this system. Critics of the system charge that the tools commonly used
8 for emission limitations, "command and control regulations," are inefficient from the
9 standpoint of delivering benefits in an optimally cost-effective way. Specifically, the
10 system has been charged with:
11
12 a) focusing too much on end-of-the pipe;
13
14 b) ignoring the fact that some dischargers can reduce pollution more cheaply
15 than others and that some areas have more serious pollution problems than
16 others;
17
18 c) blocking opportunities available to achieve emission limitations using
19 cheaper alternatives for compliance; and
20
21 d) imposing more stringent controls on new plants than on existing ones, a
22 dichotomy that creates incentives to keep older, less efficient, higher polluting
23 plants in service (Glicksman and Chapman, 199?).
24
25 "Command and control" approaches have also been faulted with causing a
26 serious misallocation of resources, in that the regulations may be directed at trivial
27 risks while more significant risks sometimes remain unaddressed (FN85 in
28 Glicksman?).
29
30 Critics of conventional regulatory approaches have generally argued for the
31 substitution of economic incentives, such as marketable permits, effluent taxes or
32 charges, or subsidies, to deliver more efficient pollution control (citatation in
33 Glicksman, ??). Although these economic tools have been used occasionally, in the
34 past EPA has been charged with being generally biased against new approaches,
6-16
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 such as market based approaches, that could conceivably achieve better reductions
2 more cost effectively (Report on Strategic Options Subcommittee, Appendix C, RR,
3 1990).
4
5 There is little dispute that there are severe inherent limits in the extent of
6 protection that the existing regulatory structure can deliver. The large number of
7 processes and substances that must be regulated, the burden of proof to justify
8 proposed regulations, the information requirements to design the regulations, the time
9 and cost of issuing and enforcing permits all conspire against solving the still daunting
10 environmental problems that we face with conventional approaches. Many are
11 concerned that the system leads to excessive delay in addressing important
12 environmental hazards.
13
14 To date, the debate about delivering environmental protection has focused
15 nearly exclusively on the strengths and weaknesses of specific risk reduction tools,
16 such as technology-based standards or economic incentive systems. It has not
focused on the sometimes sharp differences in the nature of specific environmental
18 problems to be solved using these tools. These differences among types of problems,
19 once properly considered, suggest that no single risk reduction tool is superior across
20 the board in addressing all types of environmental risks. No tool so frequently holds
21 superior promise that the regulator should assume it can be applied to the problem at
22 hand without considerable analysis. From this perspective, then, other approaches
23 must be considered, including market based incentives, non-technology-based
24 standards, more public disclosure, or less "command and control" - without
25 references to the types of problems remaining for the Agency to address - do not
26 engage the debate where the problem really lies, which is in matching specific
27 environmental risks with the best possible risk reduction options.
28
29 The Subcommittee firmly believes that each environmental risk should first be
30 examined for certain critical elements prior to selecting a suitable set of risk reduction
31 options. In addition, our review of the application of various risk reduction options
32 reveals that the problems EPA has experienced in effectively controlling some
33 environmental risks may stem more from insufficient information regarding types and
34 number of sources of emissions contributing to the problem than from inherent
6-17
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 limitations of the option selected. This inadequate understanding of the sources can
2 result in a sub-optimal or ineffective solution. For example, Amoco Oil Company did a
3 study of its refinery in Yorktown, Virginia that led to a widely publicized critique of the
4 limits of the conventional air emissions regulatory system (reference??). The major
5 finding of the Amoco study was a very large source of fugitive benzene emissions that
6 had been unknown both to the regulators and Amoco prior to the investigation.
7 Without complete understanding of the sources to be addressed at this facility,
8 however, no risk reduction option in EPA's toolbox would effectively address
9 emissions at this plant. Thus, a complete understanding of the sources to be
10 regulated is at least as important, if not more important, than the ultimate selection of
11 the risk reduction option.
12
13 Many environmental risks are best addressed with a combination of risk
14 reduction tools. Under such an approach, certain contributors to the total risk might be
15 best controlled under one approach, while other sources contributing to the same risk
16 could be best addressed by another. Because the types and magnitude of
17 contributors to certain risks are so poorly understood, complex, or diverse, the
18 application of a mixture of risk reduction tools to decrease an environmental hazard is
19 uncommon.
20
21 This section describes the full range of risk reduction options currently available
22 to achieve environmental and public health protection. Unlike several other reviews
23 (OTA: Environmental Policy Tools, 1995; McGarity 46 Law and Contemp. Probs 159,
24, Summer 1983) addressing this topic, this approach places special emphasis on the
25 types of environmental problems (or elements of problems) best or least suited to a
26 particular risk reduction tool wherever possible. The chapter attempts to identify the
27 core problems where the tool works well, and where the tool is less effective. Table 6-
28 1 at the end of the section summarizes this information.
29
30 6.4.1 Communication/Education
31
32 Information reporting is a communication/education instrument that requires
33 regulated entities to provide specified types of information to an agency which in turn
34 makes it available to the general public. Typically, the information concerns facility
6-18
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 emissions or operating characteristics, e.g. spill plans, pollution prevention plans, etc.
2 Information reporting (for public use, as opposed to compliance use) is based on the
3 theory that disclosure of polluting activities by regulated entities will raise public
4 concern, which will then lead them to respond to this concern by changing behavior.
5 The scheme takes advantage of a regulated entity's desire to be a good neighbor and
6 responsible corporate citizen, as well as their fear of adverse publicity or possible loss
7 of sales. In addition, because the public's heightened awareness of polluting activities
8 increases the possibility of new regulations, it provides another type of incentive for
9 regulated entities to pursue pollution reduction strategies.
10
11 Information reporting provides less direct assurance than many other tools that
12 goals will be met, because it does not mandate overall pollution limits or place an
13 explicit price on pollution. Furthermore, information reporting will change the behavior
14 of different regulated entities differently. Even if a high percentage the regulated
15 community reports, it is hard to know how many will change behavior to reduce
16 pollution and whether there will be some locations with disproportionately few
reductions.
le
19 Information reporting programs are being used with increasing frequency. They
20 were sparked by the Bhopal, India disaster, which alerted many in the United States to
21 the need to know more about the chemicals used and stored at industrial facilities.
22 Now, the major federal disclosure requirements are found in Section 313 of the
23 Superfund Amendment and Reauthorization Act (SARA), the Toxics Release
24 Inventory (TRI), which calls for certain manufacturing facilities to submit annual reports
25 on the amounts of listed toxics chemicals released (routinely or accidentally) into the
26 environment. At the state level, California has created Proposition 65, California's
27 safe Drinking Water and Toxic Enforcement Act. It requires public warning of the
28 potential cancer or reproductive effects of 542 listed chemicals either emitted or
29 present in products. This policy has resulted in the installation of warning signs
30 throughout the State of California, including in grocery stores and restaurants. There
31 remain questions as to whether all the signage has truly reduced exposure of
32 Califomians to toxins.
33
34 Information reporting is most likely to be effective in situations where:
6-19
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 It is possible to convey the information in a simple and accurate manner
2
3 The regulatory target is a product where consumer preferences are driven by
4 environmental concerns, as opposed to exclusively price, quality, performance,
5 and convenience or other resource(s) for purchasing the product;
6
7 The link between the product and environmental damage is clear;
8
9 The risk, though relatively easy to mitigate, is not on a scale to have been
10 prioritized by government programs; or
11
12 Hazards are not due to multiple exposures or toxic hot spots, or where hazard
13 is due to combined factors.
14
15 Technical assistance, a communication/education approach, consists of
16 government programs designed to educate private entities and the public to make
17 better environmental choices. It seeks to achieve environmental goals by increasing
18 the understanding of pollution problems and potential solutions. It may take the form
19 of manuals and guidance, training programs and materials, information
20 clearinghouses, hot lines, facility evaluations, and technology R&D (OTA, p. 138).
21 Most technical assistance services are provided at no, or minimal, cost to the user,
22 and participation in programs is typically voluntary.
23
24 Some technical assistance programs have been developed in response to
25 congressional mandates, while EPA and other agencies have initiated others. Section
26 507 of the Clean Air Act, for example, requires states to establish Small Business
27 Stationary Source Technical and Environmental Compliance Assistance Programs.
28 Others have been initiated by EPA or other agencies to help implement mandated
29 environmental programs. For example, section 319 of the Clean Water Act calls for
30 states to manage diffuse non-point sources of water pollution. EPA and the U.S.
31 Dept. of Agriculture have developed extensive guidance documents describing best
32 management practices that non-point sources might use to control their pollution.
33 Other programs do not respond directly to statutory mandates but are derived from the
34 general objective of improving environmental quality. For example, the EPA Green
6-20
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 Lights Program conducts energy audits and makes specific recommendations for
2 energy efficiency in exchange for an agreement from participants to install the
3 recommended lighting systems.
4
5 Technical assistance is most effective where polluting entities are not
6 knowledgeable about the environmental consequences of their actions or alternatives
7 available to improve their current environmental practices.
8
9 6.4.2 Enforcement
10
11 Liability provisions require that polluters pay for the damage they cause by
12 requiring compensation for environmental damage. It is enforced two ways: by
13 common-law theories such as negligence or nuisance, or by statute, particularly the
14 Comprehensive Environmental Response, Compensation, and Liability Act
15 (CERCLA). For either type of liability, a successful claim typically requires an
16 established causal link between the harm and the pollution, which has been traced
back to its source.
18
19 Liability differs from regulation in that it engages the responsible parties
20 after-the-fact. Liability rules establish the price of environmental damage, while
21 regulations seek to control the activities that create such damage. As such, liability is
22 a crude instrument for affecting behavior; by itself it cannot ensure that the appropriate
23 amount of care is exercised in all circumstances (Shavell, 1984, in Percival, p. 134).
24 Despite this, the specter of potential liability is widely thought to encourage pollution
25 prevention and responsible waste management ethics, because the dollar amounts
26 involved can be huge, much larger than the costs of preventive behavior in the first
27 instance. As of September 1994, for example, Exxon had already spent $3.4 billion to
28 clean up the spill and settle suits from the Prince William Sound oil spill.
29
30 Liability also provides incentives for environmental auditing and other
31 self-appraisals, in order to gauge the potential financial exposure and correct
32 problems before they grow (OTA, p. 125). As a policy tool, liability is particularly
33 attractive when a private actor is in a better position than the government to assess
34 the risks of its activity and to determine the level of care to exercise. On the other
6-21
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 hand, regulation is more efficient than liability for preventing accidents caused by
2 those likely to escape responsibility for their actions; this factor usually weighs in favor
3 of regulation as a means for deterring environmental damage because of the difficulty
4 of proving causation under common law.
5
6 Several statutes provide liability schemes to address environmental problems.
7 CERCLA provides for strict retroactive liability for cleanup of hazardous waste dumps.
8 The Oil Pollution Act provides strict liability for natural resource damages, and third
9 party damages caused by petroleum spills. The Clean Water Act makes responsible
10 parties liable for cleanup costs for a spill of hazardous substances into surface waters,
11 capping liability at $50 million unless the discharge was the result of negligence or
12 willful misconduct.
13
14 Liability is an effective incentive for environmentally beneficial behavior only to
15 the degree that the decision-maker is concerned that he/she will be held liable.
16 Difficulties involved in proving a causal link between a particular action and damage to
17 a particular plaintiff has limited the effectiveness of this tool for addressing many
18 environmental hazards. Liability works best where:
19
20 The contamination is traceable back to one source (harm that is widely
21 dispersed or difficult to trace will not spark adequate incentive to reduce risks);
22
23 There is sufficient information on the extent of harm that is caused by a
24 substance or activity to establish a causal link between the harm and the
25 pollution;
26
27 The impacts of the contamination are sufficiently concentrated to make a claim
28 worthwhile to the injured party; or
29
30 The party responsible for the damage has the capacity to provide
31 compensation for the full amount of harm the actions produce. (If an activity
32 can cause more damage than the actor is capable of repaying, fear of liability
33 will not provide sufficient incentive for private investment in an efficient level of
34 precautions.)
6-22
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 If liability insurance is available (as a matter of fact, it is often required to ensure
2 that compensation will be available for victims of environmental damage), the
3 existence of this insurance can reduce responsible party incentive to prevent damage
4 in the first instance.
5
6 6.4.3 Engineering
7
8 Engineered solutions are often the first ones considered. The propensity to
9 jump to this conclusion and the proven versatility of engineered solutions has led to
10 the proliferation of end-of-pipe treatment technologies. Only the recent development
11 of pollution prevention strategies has lessened their popularity. Engineered solutions
12 are often very costly and take years to implement.
13
14 6.4.4 International and Intergovernmental Cooperation
15
16 International and intergovernmental cooperation lends itself to those issues that
1 have large geographical impact. Global warming is an excellent example whereby the
18 cooperation of many world organizations is required to control the emission of
19 greenhouse gases that eventually leads to global warming.
20
21 6.4.5 Management Systems
22
23 Management systems integrate options to create effective solutions. It starts
24 with establishing organizational principles and ends with the attainment of corporate
25 goals. This option is gaining favor as entities realize that it establishes the basis for
26 getting the most from existing resources.
27
28 6.4.6 Market Incentives
29
30 Under tradable emissions , a market incentive approach, the government first
31 sets a level of aggregate emissions over a specified time period, consistent with
32 environmental goals by issuing only the number of permits corresponding to that level.
33 The total allowable emissions are then allocated to individual sources through
34 government-issued permits. Unlike conventional permit systems, however, each
6-23
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 regulated entity can buy and sell permits from others. An entity might choose to do so
2 if the relative costs of emissions control make it less expensive to buy (or profitable to
3 sell) a permit from another entity. In theory, trading would continue until the cost of
4 controlling another pound of pollution is the same for all entities and is equal to the
5 cost of a permit.
6
7 The most well known applications of emissions trading include: Sulfur dioxide
8 (SO^ reduction for acid rain control and oxides of nitrogen (NOJ reduction for urban
9 ozone in Los Angeles. It was considered but rejected to control volatile organic
10 carbon (compounds?) (VOC) emissions (ozone precursors) in the Los Angeles basin.
11 Trading has also been proposed under the NPDES system for effluent discharge
12 trades, in particular for phosphorus reductions in the Dillon Reservoir (Colorado) and
13 the Tar-Pamlico River Basin (North Carolina) and for biochemical oxygen demand
14 (BOD) reduction in the Fox River (Wisconsin). These watershed trading programs are
15 under-utilized to date because there are not disproportionate costs across sources
16 associated with the reductions. Wetland mitigation banking, where agencies authorize
17 distribution of wetlands in return for promise of future enhancement of other wetlands,
18 is also another trading venue. A lack of certainty that the appropriate regulatory
19 agency will approve a "trade" is thought to be the "rate limiter" to trading.
20
21 The reductions must be quantifiable to an acceptable level of certainty. This
22 generally means that the regulatory agency must have the capacity to measure
23 pollutant levels in question at the source or as they apply across the system. This
24 includes measuring the baseline pollution level and changes from baseline to allow a
25 source to generate tradable credits. Because establishing actual emission reductions
26 for mobile sources is more difficult than for most stationary industrial sources, it is less
27 clear that emissions trading schemes will achieve intended results where mobile
28 sources are involved.
29
30 There must be a large number of sources with significant variations in control
31 costs. If everyone's cost to control emissions is the same, then emission credits will
32 not be a bigger bargain than reducing emissions at the facility. The pollution problem
33 must not involve hot spots. Otherwise, the trading system may result in inadvertently
34 disproportionate exposures. The program must be evasion proof. Actions that
6-24
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 produce credits must be enforceable. The program is especially valuable where
2 expected control costs are very high; this will enhance interest in a more
3 cost-effective approach. The emissions should have a well-defined relationship to the
4 problem being solved.
5
6 With pollution charges, a regulated entity is required to pay a set dollar amount
7 for each unit of pollution emitted or disposed. Sources are free to choose whether to
8 emit pollution and pay the charge, or to pay for the installation of controls to reduce
9 the charge. Pollution charges do not set a limit on emissions or production. However,
10 if they are set high enough, and if they vary according to the amount of pollution
11 produced, they can provide significant financial incentives to reduce or eliminate
12 environmentally harmful behavior.
13
14 Although charges have commonly been levied against sources as a
15 revenue-raising measure, such as with permit fees, pollution charges set at a level
16 sufficient to change behavior have not been used extensively in the United States.
Two exceptions are solid waste management and the charge that was levied on
18 chlorofluorocarbons (CFCs) during their phaseout. Under typical solid waste
19 schemes, companies pay charges that rise as waste volume rises, while most
20 households pay flat fees unrelated to the amount of waste generated. Volume-based
21 charges have been applied to household waste in about 100 jurisdictions across the
22 country as well, however, and they are believed to have been successful in reducing
23 volumes of waste generated in many of these instances (OTA, p. 123).
24.
25 During the CFC phaseout, users were required to pay a charge per pound of
26 CFC, multiplied by an ozone-depleting factor. The charge was increased in
27 subsequent years. CFC production decreased much more rapidly than originally
28 anticipated, and many attribute this rapid decline to the effectiveness of the CFC tax
29 (OTA, p. 122). Because the tax was used in conjunction with a mandatory phase-out,
30 it is difficult to ascertain the relative importance of each tool in achieving the overall
31 success of this program. Clearly, however, a combined approach of charges and
32 phase-out can be effective in reducing environmentally harmful behavior. It is
33 important to note that the CFC taxes collected by the U.S. government went to general
34 revenue and were directly allocated to environmental programs/improvements.
6-25
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 Subsidies are policy instruments that provide financial assistance to entice
2 entities to change their behavior or help entities having difficulty complying with
3 requirements. Generally, subsidies take the form of grants or low- no-interest loans to
4 municipalities and public entities and preferential tax treatment for private entities.
5 The Clean Water Act, for example, had a long-standing construction grant program
6 that enabled municipalities to comply with Clean Water Act requirements for
7 wastewater treatment plants. This grant program was phased out by the 1987
8 amendments to the Act and replaced with a state revolving loan fund. The Clean Air
9 Act also authorizes several grant programs. For example, section 105 provides for
10 grants to state and local governments to implement air pollution control programs.
11
12 For many years, private companies have been allowed to take accelerated
13 depreciation on investments aimed at reducing water pollution.
14
15 6.4.7 Regulation
16
17 Harm-based standards are established on the basis of what is required to
18 achieve health or environmental protection goals. EPA typically establishes these
19 standards by determining the amount of the pollutant in the ambient environment that
20 will meet a health or environmental goal set by Congress. This determination involves
21 scientific judgments as to the extent to which different concentrations of the pollutant
22 cause harm. After the Agency establishes an acceptable concentration, a
23 mathematical model is used to calculate an overall allowable pollution load for the
24 region with which this acceptable concentration will not be exceeded. EPA or the
25 state then apportions an acceptable pollutant concentration or loading among the
26 individual sources that it has identified. Harm-based standards are expressed in
27 facility permits as emission rates for the source (mass per unit time period), as a
28 concentration of pollutant in a source's discharge, or as a percentage reduction in
29 emissions from a source.
30
31 As a practical matter, EPA's ability to craft a policy that meets a harm-based
32 goal is contingent on a thorough understanding of the sources contributing to the
33 pollution load from the facility or region. Often, the Agency has insufficient knowledge
34 of the number, types, and magnitude of contributors to the problem. Bubble emission
6-26
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 and integrated permitting initiatives have been developed to overcome the lack of
2 knowledge in this area by providing flexibility to regulated entities to alter control levels
3 for various sources within a facility as long as overall limitations are achieved. (The
4 3M integrated permit for its Hutchinson, Minnesota facility, for example, allows the
5 company to shift emission controls among sources within the facility as long as
6 aggregate VOC control levels are satisfied.) However, neither bubbling not integrated
7 permitting have been widely used to date, again because of the inadequate
8 knowledge of sources to be "bubbled" or because of monitoring and accountability
9 difficulties.
10
11 Harm-based standards are notoriously difficult and time-consuming to set
12 because of analytical uncertainties and gaps in available data about both the scientific
13 basis for concern and the demography of the sources contributing to the problem.
14 Uncertainties inherent in predicting the effects of different patterns and levels of
15 environmental releases also pose big problems. Harm based standards also require
16 extensive data on existing ambient pollutant concentrations and health effects, which
often are not available.
18
19 Nonetheless, harm-based standards, along with technology/design standards,
20 are the most heavily used environmental policy tools in the U.S. Often, the two are
21 used together. The Clean Water Act provides harm-based standards for "water
22 quality limited streams", where industrial sources must comply with stricter pollution
23 control than would otherwise be necessary in order to achieve improved water quality
24, in a highly negatively impacted area. The Clean Air Act National Ambient Air Quality
25 Standards (NAAQS), also exemplify harm-based standards.
26
27 In contrast to harm-based standards, technology-based standards derive from
28 the level of control technology Congress expects pollution sources to implement, such
29 as "reasonably available control technology." Technology based standards can be
30 viewed as the opposite of health-based standards, because instead of asking what is
31 needed to protect health or the environment, they ask what is possible to do (Percival,
32 p. 146). Most commonly, these standards are set as a performance level for the
33 regulatory target to meet, without specifying how this should be done (performance
34 standards). They also can specify how a certain facility or piece of machinery should
6-27
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 be designed and engineered (design standards). EPA develops both design and
2 performance standards, although some of its performance standards become de facto
3 design standards. This can occur where EPA's performance regulation is established
4 on a specific technology and, in responding to the target, regulated entities decide the
5 safest course to compliance is to install a specific design technology. EPA's New
6 Source Performance Standards under the Clean Air Act and its various effluent
7 limitations under the Clean Water Act are examples of performance standards, but
8 they are often criticized for limiting acceptable technologies in practice to those that
9 EPA has used as a basis for developing the performance standards (Percival, p. 150).
10
11 Disadvantages technology-based standards are that they: a) may cause the
12 expenditure of more or less resources than necessary, b) assume state of technology
13 is knowable, and c) discourage innovation.
14
15 Fixed emissions standards based on current assessments of technology are
16 widely faulted for the absence of incentives for further innovation. Laws that force
17 industry to develop innovative solutions by adopting standards more stringent than
18 those attainable by current-available technology overcome this important limitation.
19
20 Both the Clean Water Act and the Clean Air Act call for technology-forcing
21 regulations. The effluent limitations in the Clean Water Act, which are based on "best
22 available technology economically achievable," look to transfer state-of-the-art
23 practices in one industry to another in order to achieve reductions beyond what is
24 currently in practice. The Clean Air Act Amendments? of 1990 contains a broad effort
25 to force technology developed for mobile sources, including better tailpipe controls,
26 cleaner fuels, and cleaner engines than existed at the time of passage of the Act.
27
28
29
30
31
32
33
34
35
36
"Congress's initial approach to technology-forcing regulation was to engage automakers in a
high-stakes game of chicken. Title II of the 1970 Clean Air Act mandated emission standards
for automobiles that were "a function of the degree of control required, not the degree of
technology available today.' The Act required automobile manufacturers to slash vehicle
emissions by 90 percent in order to be able to continue selling cars in the United States after a
specified deadline, subject to a one-year extension. Not surprisingly, the automobile
manufacturers fought hard to convince EPA to extend the deadline on the ground that the
necessary technology for reducing vehicle emissions was not yet available. After EPA
Administrator William Ruckelshaus shocked the industry by refusing to extend the deadline,
6-28
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
the auto industry convinced a court to order him to reconsider. In May 1973, EPA agreed to
relax the standards for two years. Meanwhile, GM had a breakthrough in the development of
catalytic converter technology, which promised to be a boon to fuel efficiency and
performance but would also meet the emission standard required in the Act. Six weeks after
EPA extended the deadline, GM announced that all of its 1975 models would be equipped
with the devices. "I came out of the whole exercise with catalytic converters as a
technological optimist in what the industry can do," said Walsh, the EPA's top auto pollution
expert in the 1970's and now a private consultant. "When you give them a challenge, they
meet it." (Percival, p. 168-171)
1
2
3
4
5
6
7
8
9
10
11 Disadvantages of technology-forcing regulations are that they a) may compel too
12 much monies to be spent in a narrow area and b) require a tremendous amount of
13 information.
14
15 A product ban either prohibits or restricts the manufacture, distribution, or use
16 of a substance identified as a problem to human health or the environment. It is most
17 seriously considered where product use is sufficiently damaging that zero use is a
1" desirable outcome. Product bans may be imposed prior to the product's sale and use
in commerce, through various pre-market product approval programs that seek to
20 prevent excessively risky products from reaching the marketplace. More publicly
21 visible bans concern themselves with substances already on the market, such as lead
22 or asbestos. Whereas the burden of proof lies with the manufacturer to show that a
23 product is safe for pre-market approval, the burden of proof lies with EPA to show that
24 a product should be banned once it is already on the market.
25
26 Product bans and restrictions focus on the commodity itself rather than on
27 polluting byproducts from its manufacturing. As a result, they are used primarily
28 where the hazard is the commodity itself. The U.S. has banned lead in gasoline and
29 paint, asbestos, PCBs, CFCs, and numerous pesticides, including DDT, Aldrin, and
30 Dieldrin.
31
32 Product bans are most effective where a single substance or activity is causing
33 a particular environmental problem. They are not effective where the environmental
34 problem has multiple or complex causes, such as low dissolved oxygen
35 concentrations in urban rivers.
6-29
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 A product ban is best used where all of the uses of a product pose
2 unacceptable risks. Otherwise, considerable analytical resources must be set aside to
3 enforce a narrower ban or prevent unauthorized use.
4
5 Product bans have superb potential to induce technological innovation by
6 stimulating rapid research on substitutes. However, they are risky when the
7 consequences of a failure to develop an alternative are very high (such as, if we were
8 to contemplate the banning of the internal combustion engine to address urban smog).
9 Their success hinges on the development of genuinely safer alternative products;
10 otherwise the substitute will pose its own hazards.
11
12 With a challenge regulation, responsibility for solving a problem is shifted from
13 government decision-makers to the sources themselves. Under this approach, the
14 government establishes targets necessary to solve an environmental/public health
15 problem, with a timetable for implementation. The targets are defined for multiple
16 sources, at the industry sector rather than the individual facility level. These sources
17 are given the collective responsibility for designing and implementing a program that
18 meets the targets. The government specifies an alternative program or sanction that
19 will be triggered if progress towards the targets is not achieved.
20
21 Challenge regulations have not yet been extensively adopted in any country
22 and have not been used to date in the United States. EPA's voluntary 33/50 Program
23 (reference?), however, is an example of a pilot program to test this approach. The
24 most widespread use abroad has been in establishing producer responsibility for
25 various forms of wastes to encourage source reduction and recycling. In the German
26 Green Dot program, the government established a regulatory approach outlining
27 industry's obligations to take back packaging from customers. However, it then gave
28 industries the opportunity to establish an alternative recycling program of their own for
29 meeting the targeted rates.
30
31 Since industry designs the implementation plan, this tool is not well suited to
32 problems with localized concerns about the effects of the pollution. Thus, it should not
33 be used to reduce pollutants or solve problems for which exposures vary widely
34 across locations.
6-30
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 The challenge is unlikely to be met if it is difficult to fairly allocate responsibility
2 across the target group. Competition among firms may make it difficult to develop an
3 implementation plan.
4
5 Some identified options may not be based on technology. They could be policy
6 options. Performance data are rather sparse on non-technological options, and
7 models constructed to estimate very high uncertainties and controversies may
8 characterize their reliabilities within a given constraint. Issues related to the
9 performance of a variety of both technological and policy-level options have been
10 discussed by Davies and Chou (1992), Levin (1990), IINEPA (1996), Norman and
11 Keenan (1996), C and EN (1996), CHEM NEWS (1996), OTA (1995), USEPA (1990)
12 and UN (1993). Performance information on composite options (i.e ) is very
13 sparse.
14
6-31
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1
2
3
4
Table 6-1. Conditions Under Which Options Are Most and Least Effective
5
6
7
8
9
10
11
12
13
14'
15
16
17
18
19
20
21
22
• Communication/Education
SaSiiKi
-Tech. Assist.
• Regulated entity is not
knowledgeable
Engineering
Management Systems
Non-materiel issues
B SiS^^^K^^^^Sl
Weak organizational structure
—££._*- ,—^y^, g .*K..-f.irv^-i/.ffs. ^ ^..-...j^-.^.
-Tradable Emissions
! •"^•<*S-;vj:r"--. • •- ^v~^ly;~'™~"~JVT"*Ts*''i?L"*v-'""";^'~;*! •^-*~^T-y^^*-v^,t?;T:yy" T* ^*)yt' "' ly^T^^^~^^ff?TS^'*^y|*^
• High control cost
• Multiple sources w/ variable
control capability
• Quantifiable emissions
• Fixed sources
• Enforceable
• Single substance is cause
> All uses have unacceptable
outcome
• Multiple sources
• Few or no alternatives
6-32
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 6.5 Establish Screening and Prioritization Criteria
2
3 The simplest way to evaluate risk reduction options would be to reduce the
4 consequences of specific options to a single measure of effectiveness. (In fact, this is
5 precisely what a classic cost-benefit analysis of a regulatory option seeks to do.)
6 However, it is not easy to achieve consensus on reducing to a single measure the
7 wide range of views and legitimate perspectives that come to bear in formulating risk
8 policy. An alternate approach is to reduce the consequences of a risk reduction
9 option to performance on as small a set of criteria as possible. The Subcommittee
10 concluded that five basic criteria could be used to evaluate risk reduction options:
11 environmental effectiveness, cost, equity, workability, and flexibility. With regard to
12 the first two criteria, we have concluded that it is more transparent to separately
13 consider environmental benefits and environmental costs, rather than to combine
14 them into a single measure of net benefits. Ideally, performance in each of these
15 groupings should be summarized along a single dimension, such as lives-saved or
1P dollars-of-cost.
18 Environmental effectiveness encompasses the variety of criteria that have to do with
19 the environmental impacts/physical benefits. For the most part, the issue here is to
20 what extent does a regulation accomplish the environmental protection goals that
21 originally motivated the regulation? The focus in on the meaningful environmental
22 consequences of a regulation, rather than simply meeting some arbitrary ambient
23 concentration goal. Another consideration in this category is the likelihood that the
24 environmental goals will be met. As an example, the NAAQS and associated State
25 Implementation Plans for urban ozone would suggest that environmental goals are
26 well served by this regulatory approach. In reality, however, in many cities the goals
27 are often not met. The economic version of this criterion (discussed in Chapter 4) is
28 efficiency, defined as that level of risk reduction which maximizes the surplus of
29 benefits over costs. Therefore, an aggregate measure of environmental benefits is
30 desirable to help compare the performance along this dimension with performance on
31 other evaluative criteria.
32
33 Cost is the second grouping and includes a number of factors. Of primary interest are
34 the direct and indirect costs associated with a regulation, excluding the monetized
6-33
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 value of the environmental benefits that may result from the regulation (which are
2 accounted for under the previous category?).
3
4 Equity deals with the distribution of costs and benefits from the regulation; i.e., who
5 bears the cost and who benefits. This might involve the distribution of costs and
6 benefits to economic groups (e.g., rich vs. poor), ethnic groups (e.g., blacks vs.
7 whites), or different regions of the country (e.g., Midwest vs. west). Clearly any
8 regulation that involves concentrated costs and benefits will have very different
9 political characteristics than a regulation with more dispersed costs and benefits.
10
11 Workability deals with how easy it might be to initiate and implement a regulation.
12 Some regulations have a great deal of administrative burden. Some regulatory
13 approaches are so new as to make their success highly uncertain. Some regulatory
14 approaches are very familiar to regulators and polluters so as to be very workable.
15
16 Flexibility refers to how easy a regulation is to change when new information emerges
17 or to adapt to new circumstances.
18
19 There are two purposes for the evaluative criteria discussed here. One is to
20 help fine-tune a given option. For instance, if an emission charge has been proposed,
21 there are many choices as to exactly how it will be implemented, what pollutants will
22 be included, what the level of the charge will be, and what kinds of enforcement will be
23 included. Using the five criteria, the option in question can be adjusted, tailored, and
24 honed to do as well as possible against each. This results in an option that performs
25 as well as possible.
26
27 The second purpose for the five evaluative criteria is to compare several
28 different approaches to solving a particular environmental problem. Each approach
29 must be fine-tuned. The only question remaining is which approach is the top
30 candidate for implementation.
31
32 In order to facilitate the comparison of options, it is important to summarize
33 performance for the criteria along as few dimensions as possible. This may be most
6-34
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 easily done for cost, using a dollar metric. It can also be done for the other criteria,
2 though perhaps not so easily. For instance, the performance of an option on equity
3 grounds could be measured on a scale of 1 to 10, with guidelines to govern the use of
4 particular ratings. On environmental effectiveness, results could be summarized in
5 terms of statistical lives saved or the willingness-to-pay of the population to achieve
6 the level of environmental performance, relative to some baseline.
7
8 A more extensive discussion of each of the five evaluative criteria follows. This
9 discussion is not intended to cover all aspects of each but to give an idea of the
10 breadth of issues that are embodied in the criteria.
11
12 Environmental effectiveness summarizes how well a risk reduction option
13 performs in terms of its environmental outcomes, which obviously is an important
14 aspect. It is preferable to define environmental performance in physical or health
15 terms (lives saved, species protected), rather than in terms of administrative process
•"* measures (e.g., the likelihood of meeting a regulatory target such as no more than
three violations of an ambient standard in a year). Outcome measures are not always
18 possible, however, since the physical effects may be difficult to quantify.
19
20 Consider the following issues with respect to environmental effectiveness:
21
22 a) How does expected performance compare to environmental goals? If the
23 goal is to reduce morbidity and mortality by a certain amount, has this goal
24 been achieved? If the goal is to reduce stressor concentrations by a certain
25 amount, has this goal been achieved? Often there will be no explicit or
26 quantitative environmental goal. One may simply wish to reduce pollutant
27 levels as much as possible (or justified). In this case, there is no explicit end-
28 point.
29
30 b) What are the quantifiable physical gains from the regulation? For example,
31 what is the reduction in morbidity and mortality? What are the goals in
32 recreation opportunities (with cleaner rivers, for example)? What are the
33 maintenance goals from a cleaner ambient environment? What is the nature of
6-35
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 species preservation or ecosystem health? The point is not to measure gain in
2 arbitrary units such as how much BOD reduction has been achieved, but in
3 more meaningful terms (if possible).
4
5 c) What is the value of the environmental gains? It is sometimes, though not
6 always, possible to quantify the physical gains. This should be in a single
7 metric such as dollars or the number of statistical lives saved and should
8 summarize the diverse environmental consequences of the regulation.
9
10 d) What is the probability of success in obtaining the expected environmental
11 gains? Some risk reduction options are more proven than others.
12
13 e) Can performance be monitored? One must be able to conclude whether
14 aggregate environmental goals are being met. It is not always easy to monitor.
15 For instance, VOC emissions from small businesses may be very difficult to
16 monitor.
17
18 f) When can results be expected? Is the regulation designed for immediate
19 results or will it take a long time to be effective?
20
21 g) What are the associated risks? For instance, regulations to reduce sulfur
22 dioxide emissions may contribute to accelerating global warming.
23
24 h) Is pollution prevented rather than abated? Although the environmental
25 effects may appear to be the same, it is more desirable to reduce the
26 generation of pollution through process change rather than simply appending
27 an abatement technology (end-of-pipe treatment).
28
29 The criterion of cost includes all of the non-environmentaf costs and benefits of
30 a risk reduction option. The specific way in which costs are measured is discussed in
31 Chapter 4. Questions relevant to this criterion include:
32
6-36
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 a) What are the direct costs of implementing the risk reduction option? This
2 would include the costs incurred by a polluter in reducing emissions, which
3 might be costs of abatement equipment or costs from adopting more costly
4 processes. Also included would be administrative costs - costs associated
5 with applying for permits, participating in review procedures, and costs
6 associated with uncertainty of what might emerge from the regulatory process.
7 A third direct cost would be the consumer losses associated with higher product
8 prices (because of higher production costs). It is also important to include the
9 regulatory costs, incurred by the EPA. These would include the costs of
10 formulating and administering the regulation as well as enforcement and
11 monitoring costs.
12
13 b) Are costs of residual damage reflected in product prices? In order for
14 consumers to receive an adequate signal regarding how damaging the
15 production of goods might be, the price of those goods should reflect all
16 environmental costs, including the costs of abatement but also including the
environmental damage associated with residual pollution. If consumers see
18 product prices that are excessively low, this will generate too much
19 consumption of the polluting product and consequent inefficiencies.
20
21 c) Are there incentives to innovate? Some options provide little incentive to be
22 creative and innovate. Often, the option spells out what to do and innovation
23 only results in tightening the requirements. It is important that an option provide
24 incentives for innovation by polluters and by equipment vendors. Often this
25 requires that all or part of the gains from the innovation accrue to the innovator.
26
27 d) How much regulatory commitment is involved with the option? If there is little
28 commitment, does that increase costs substantially? When a regulator
29 commits to a risk reduction option, and cannot retreat, then polluters make
30 least cost investments. Without commitment, polluters will take more flexible,
31 but also more costly, approaches to meeting regulations. A lack of commitment
32 on the part of regulators can only increase costs, often with no associated
33 environmental gain.
34
6-37
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 Equity: Problems arise when any identifiable group disproportionately bears
2 the costs of a selected option. One might also be concerned with benefits falling
3 disproportionate on particular groups; there seem to be fewer political problems when
4 small groups benefit from a governmental action than when small groups bear costs
5 (substantiating reference?).
6
7 Basically, what is at issue here is the distribution consequence of a risk
8 reduction option. Several issues arise:
9
10 a) What is the incidence of costs? Do any groups disproportionally bear the
11 costs? For example, if a benefit is distributed equally in a population and the
12 cost is applied equally to each individual, lower income families will pay a
13 higher percentage for equal benefit.
14
15 b) What is the incidence of benefits? This is a very similar issue as the
16 incidence of costs, although perhaps not as important. Nevertheless, it is
17 important to identify any groups in the population that may disproportionally
18 gain the benefits.
19
20 c) Do benefits and costs accrue to the same group? If one group
21 disproportionally bears costs but also reaps benefits, there is less concern than
22 if different groups bear the costs than receive benefits.
23
24 d) Are the regulations confiscatory? Polluters may be more opposed to a
25 marketable permits system where permits are auctioned off initially than a
26 system with permits freely distributed. Any regulation that is viewed as
27 excessively confiscatory is more likely to be opposed in the political process.
28
29 e) Are benefits and costs concentrated or diffused? The more diffuse the costs
30 and benefits, the less likely it is that specific groups will emerge to oppose the
31 regulation in the political process.
32
6-38
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 Some risk reduction options work better than others. Workability addresses
2 how easy it will be to implement an option.
3
4 a) How large is the experience base with a particular risk reduction option? If
5 an option is a variant of one that has been used many times before, then both
6 the EPA and the polluter will have experience in making it work. In contrast, a
7 new option may be more difficult to implement and work out the problems.
8
9 b) How much information does the government need in order to implement the
10 option?? The primary criticism of command-and-control regulation is that the
11 EPA requires an unreasonable amount of information.
12
13 c) What are the demands on government? An option that has less demands
14 on government is to be preferred to one that has heavy demands on
15 government, all other things being equal.
* -
,. d) Is the public involved in the decision process? The more the public is
18 involved as a risk reduction option is implemented, the less likely it will be that
19 unexpected problems develop. Often when the public is excluded, factors that
20 would have been brought to light through public hearings end up causing
21 significant problems during implementation.
22
23 e) Are enforcement and monitoring easy? Clearly some regulations are easier
24 to enforce than others. Product bans are straightforward, provided leakage
25 from unregulated areas is not a problem. If there is significant reliance on
26 unobserved actions by polluters (e.g., maintenance), then it will be difficult to
27 ensure that these actions have been taken.
28
29 f) Are the financial burdens on government excessive? Some options are
30 cheaper than others to administer. The larger the burden on government, the
31 more difficult to sustain.
32
6-39
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 g) Are options viewed as excessively intrusive? Some options are politically
2 unpopular because they are viewed as too intrusive. For instance, proposals to
3 use intelligent vehicle systems to automatically charge for road use have been
4 met with privacy objections. Some citizens do not like the idea of their
5 movements being centrally monitored.
6
7 No matter how well an option is designed, changes will undoubtedly be needed
8 making flexibility important. It is important to recognize the ease with which an option
9 can be modified to tailor to specific conditions. Several issues emerge.
10
11 a) Is the option adaptable to new information? If new information requires that
12 Congress re-draft the appropriate legislation, the option will be more difficult to
13 modify than if changes are automatic. The more readily an option can adapt to
14 changed circumstances, the better it will perform.
15
16 b) Does the option promote innovation and diffusion of innovation? This issue
17 was mentioned in the cost-effectiveness grouping. It is important that an option
18 be flexible enough to allow innovation (this is not always the case). It is even
19 more important that the option reward innovation. Innovation is a primary
20 means for increasing environmental effectiveness while reducing costs.
21
22 6.6 Screen and Prioritize Potential Risk Reduction Options
23
24 Options screening is the process by which a broad range of risk reduction
25 options is reduced to a set of unranked options for more detailed comparative
26 analyses. An option can be single or composite. Screened-in options are those that
27 can be implemented to meet the risk reduction objective within all constraints specified
28 in the problem formulation. A constraint is a specified magnitude of a screening
29 factor. For example, if the budget limitation is $1,000,000, then budget is a screening
30 factor, and the constraint is $1,000,000.
31
32 An important utility of the screening stage is that it enables the assessment of a
33 wide array of identified options. The probabilities of option targeting and arbitrary
6-40
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 elimination of options are minimized. Option targeting is a common problem in
2 remedy selection. For example, engineering remedies are more likely to be selected
3 as a remedy for a contaminated site than communication options, regardless of their
4 relative attributes. The screening process provides a systematic method for
5 determining whether each identified option meets specified performance criteria.
6 Depending on the constraints, both categories of risk reduction options may be
7 screened-in for more detailed comparative analysis during subsequent stages of the
8 options selection process.
9
10 The application of multi-criteria, multi-objective analyses to a large number of
11 identified options may unduly complicate the options selection process. The
12 screening step provides an opportunity for the identification of potentially unfeasible
13 options, thereby eliminating these options from further consideration. The number of
14 options suitable for comparative analyses in the post-screening stages of the options
15 selection process is then reduced to a more manageable level. In essence, data
16 requirements on options screened-in are proportional to the number of options. The
generation of a prioritized (and reduced) list of options for comparative analyses
18 reduces the costs and time required for acquiring relevant data. For example, if the
19 direct financial cost constraint specified for reducing ecosystem risk due to
20 contaminated sediments at a site is $1,000,000 and an effective treatment technology
21 can not be implemented under $3,000,000, that option would be screened out as
22 written. This would eliminate the need to acquire additional data that are required to
23 compare it with other options in subsequent analysis. It should be noted that this
24 option may be restructured and re-screened later in the process.
25
26 The screening process satisfies the need to give assurance to stakeholders
27 that a narrow set of options was not pre-targeted. Transparency is an important issue,
28 especially in cases where the implementation of an option is associated with shared
29 risks and costs. Transparency is particularly important when different parties within
30 the impacted communities prefer different options, some of which are rejected for not
31 meeting specified selection criteria. In the selection of a site for a waste management
32 facility, the potential impacts on the surrounding communities vary both spatially and
33 temporally. Consequently, communities proximal to selected sites may feel unfairly
34 treated if the site selection process is not transparent. Because the issue of
6-41
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 environmental equity has gained importance within the past decade, several
2 systematic techniques for site screening have been developed (Wright et al., 1993;
3 Smith, 1996; Frannis, 1993; Knight, 1985; Schwartz et al., 1989; Zeiss and Paddon,
4 1992). The screening process provides the opportunity for decision-makers to show
5 that no party or segment of the community had been pre-selected to bear a
6 disproportionate hardship.
7
8 Some options that are individually undesirable may exhibit increased utility in
9 combination with other options. The screening process provides the decision-
10 maker(s) with the first opportunity to identify options that could be potentially
11 combined.
12
13 The specificity of an option is an important determinant of the uncertainties
14 associated with the potential performance of the option in risk reduction. Diffuse and
15 imprecise options have high uncertainties. A distinction should be made between a
16 composite option that comprises distinct options and a genetically stated option. The
17 generic option can be regarded as a category of options. Specific options within the
18 category may screen in while others may screen out. Thus, the performance
19 uncertainty is higher for the option category than for any of the options within it. As an
20 example, Table 6-2 shows various option categories and specific options that are
21 nested within them.
22
23
24
6-42
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1
2
3
9
10
11
12
13
14
15
16
17
18
19
Table 6-2. Examples of option categories and specific options (option categories have
greater performance uncertainties than specific options).
ENGINEERING
H^
• Site cleanup
• Waste containment
• Technology development
'•*•; fGbmmunfty-
••'^Format education
REGULATION
Control of production process
Banning of products
Facility siting controls
Facility operational controls
MARKET INCENTIVES
• Tax relief
• Relaxation of controls
• Subsidies
• Direct purchase of friendly products
ENVIRONMENTAL
MANAGEMENT SYSTEMS
Application of technology and techniques to mitigate
risks, and direct use communities for operations and
planning
If an identified option is merely described as "Engineering," high uncertainty is
associated with its effectiveness as a risk reduction option. If, however, a specific
engineering measure is identified, the expected performance can be stated with a
higher level of confidence. A specific option such as site cleanup, can be further
6-43
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 desegregated into specific cleanup technologies: electrokinetics, pump-and-treat,
2 incineration, etc. Generically stated options should be refined when possible.
3
4 Another important factor is the relaxation or tightening of constraints. As
5 discussed in the next section, the amount of money or other resources allocated in the
6 budget for use on a risk reduction project is a constraint. Some potentially effective
7 options may screen out because they cannot be implemented within the budget
8 (constraint). If identified options are to be left unmodified, one alternative is to relax
9 the constraint by increasing the budget resources. The greater the gap in the
10 professional hierarchy between the project implementer and the budget allocators, the
11 more difficult it is for the risk reduction project implementer to relax a budgetary
12 constraint (i.e., it may not be within the implementer's authority to do so).
13
14 The objective of the risk reduction project needs to be defined as explicitly as
15 possible. The objective is a condition or an acceptable level of service at which any of
16 the identified options is considered potentially effective. The objective should not be
17 confused with a constraint. Examples of objectives are:
18
19 a) Reduction of particulate matter in the ambient air in Washington, DC by
20 30%;
21
22 b) Reduction of metals concentration in River Brown to Drinking Water Levels;
23 or
24
25 c) Reduction of cancer risk due to radon in U.S. homes by 50%.
26
27 The operational rule for options screening is that if an option is deemed
28 impossible to implement within limits specified on critical factors, that option is
29 eliminated from further analysis unless it is refined or combined with other options to
30 satisfy the limits set by the project constraints. These critical factors are herein called
31 screening factors. Examples of screening factors are:
32
33 a) Budget,
6-44
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 b) Time,
2
3 c) Jurisdictional Authority, and
4
5 d) Legality.
6
7 If a minimum or maximum level allowed on each screening factor is specified,
8 then this threshold level is considered as a constraint. For example, a regulatory
9 agency may seek to reduce paniculate matter in the ambient air in Washington, DC by
10 30% with a maximum budget of $13 million within the next 3 years. For this risk
11 reduction objective, budget and time are screening factors. More specifically, for any
12 option to screen in, it must be implementable within budget and time constraints of
13 $15 million and 3 years, respectively.
14
15 Essentially, a "yes" or "no" answer is required, and each screening factor is
16 taken one at a time. It may also be possible to estimate the level of confidence
associated with the "yes" or "no" answer. For example, members of a city council may
18 state: "We are 80% confident that increasing parking fees in a central business district
19 of Washington, DC will reduce air pollution by 40% within a 3-year period, because car
20 pooling or greater use of the metro system will occur at the new parking fee level."
21 Thus, "increase in parking fees" would screen in as an air pollution risk reduction
22 option at the level of confidence of 80%. The confidence level estimated is an
23 expression of uncertainty. It is worth noting that an increase in parking fees is not
24 necessarily the best option or set of options. Such a determination is not the focus of
25 the options screening stage of analysis. Rather, the purpose of this effort is to
26 determine how many options or combinations are available for implementation within
27 the constraints established.
28
29 Risk reduction objectives, screening factors and associated constraints have a
30 hierarchical but interwoven structure. The decision-maker on constraints depends on
31 the scenario. The scenario could be general or site-specific with respect to the spatial
32 coverage of the desired results of actions associated with risk mitigation objectives.
33 An example of a general scenario is an effort to implement measures that are
6-45
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1
2
3
4
5
6
7
intended to reduce skin cancer in the United States. The problem may not be
confined to a particular location with respect to causes and/or effects. An example of
a site-specific scenario is groundwater contamination beneath a gasoline storage tank.
The relative level of involvement of the decision-maker as a function of the scope of
the scenario is illustrated in Figure 6-2.
Site Specific Coverage
g- High
*
|0 >
Private Corporations (ABCD)
B
E
Ti
G IK D
t
State/Cou
Agencies (EF
International Agenc
p
nty
-GH)
'M
es (IJKL)
O H
N
National
Regulatory
Agencies (MNOP)
* 3!
L: ^
/ High
Decision Making Role
8
Figure 6-2. Decision-Makers on Constraints on Risk Reduction Objectives Under
Various Scenarios
9
10
11
12
13
14
National regulatory agencies such as the EPA tend to be involved as decision-
makers, mostly in general-coverage scenarios. In some cases, they may also be
involved in risk reduction programs at a site. Private companies are involved as
decision-makers almost exclusively at specific sites. Generally, the decision-maker
specifies the overriding constraints. Sometimes, the decision-maker may not be the
eventual implementer of some of the options that have high screen-in potential. The
6-46
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 roles of regulatory agencies and private companies in constraint specification is
2 illustrated by the following examples.
3
4 a) A regulatory agency identifies constraints for its risk reduction policy
5 development approaches.
6
7 b) Regulatory agencies usually specify constraints (generally time and
8 performance level) associated with a requirement that the private sector
- 9 implement risk reduction projects.
10
11 c) A private company specifies constraints for various screening factors that it
12 has selected for evaluation of internally generated options for solving an
13 environmental issue.
14
15 It is important to note that options screening involves the use of one screening
Y (evaluation) factor at a time. Constraint levels on other screening factors are held
1. constant. An option is considered as being fully screened after it has been evaluated
18 with respect to each of the set of screening factors. Screened-in options are those
19 that are likely to meet the specified risk reduction objective under the constraint levels
20 specified on each of the screening factors. During screening an effort is not made to
21 rank the option(s) relative to others. In contrast, subsequent analysis (broad analysis
22 and optimization) of screened-in options involve explicit ranking. Furthermore, pooled
23 or aggregate ratings or scores (with or without weights) are not used to express the
24 utility of each option as in the screening stage. The sequential steps in the options
25 screening methodology are illustrated in Figure 6-3.
26
6-47
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
Information Sources
Methodology
Input from Problem
Definition Stage
Input from Option
Identification Stage"
STEP1: Identification of Specified
Objectives and Constraints
STEP 2: Identification or Refinement
of Screening Factors and Options
Input from Collected
Background Information.
Expert Opinion.
and Models
STEP 3: Compilation of General and
Performance Information on Each
Individual or Composite Option
STEP 4: Analysis of Options with Respect to
Constraints on Each Screening Factor
Output
Itemized List of Scenario
^Factors: Objective. Screening
Factors, Constraint Levels.
and Identification Options
Summary of Input Data to
be Used in Analysis of
Each Option
ListofScreened-ln
Options for
Subsequent Analysis
STEP 5a: Refinement
of Screened-Out Option
STEP 5b: Inclusion of
Options into Screened-ln List
Figure 6-3. A Screening Methodology for Risk Reduction Options
2
3
4
5
6
7
8
6.7 Evaluate the Remaining Risk Reduction Options
The primary focus of this effort is to develop a common base of information for
comparing and selecting among the various sets of options, however different they
may be. For example, this might include comparing the use of consumer education
programs on pesticide residues, to the imposition of an environmental charge on
certain pesticide uses, to combining use restrictions with a trading program, all as
6-48
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 options for reducing the consumption of pesticide residues in food. A secondary, but
2 extremely important, purpose is to continue to lay the groundwork for explaining the
3 selection process to the wide range of interested parties.
4
5 The first concern is to be clear about the inputs to the assessment process. In
6 addition, data will be necessary on:
7
8 a) the past or current conditions relevant to understanding and controlling the
9 identified risks (patterns of pesticide use, likely alternatives, patterns of food
10 consumption);
11
12 b) the performance of individual risk reduction options in other, hopefully
13 similar, circumstances (consumer labeling, liability schemes); and
14
15 c) the characteristics of each option relevant to the evaluation criteria.
- ^
., Some of this data will have been generated on the way to this point in the
18 process.
19
20 The most problematic set of inputs, however, are the estimates of the future
21 results of implementing any particular risk reduction option. For example, estimating
22 the reduction in exposure levels resulting from a food labeling program is extremely
23 difficult. No matter how much data on past experience has been collected, these
24 estimates will reflect a significant degree of subjective judgment by the estimating
25 parties. This judgment will reflect their past experience, as well as their approach to
26 the available data.
27
28 As such, it is important to identify clearly the people who are providing the
29 estimates, hence, making these judgment calls. They can include:
30
31 a) the parties selecting the priority risks to be addressed;
32
6-49
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 b) outside experts on the characteristics of those risks or individual risk
2 reduction options (including those from other program offices in EPA or other
3 regulatory agencies);
4
5 c) the assessors themselves; and
6
7 d) other interested or affected parties (e.g., businesses, workers, consumers,
8 and environmental groups).
9
10 The likely accuracy of their estimates on any particular factor (along with their
11 approach to the underlying assumptions) should be the focus when assigning tasks
12 among these different parties.
13
14 Since application of all the assessment criteria requires that estimates of
15 performance be made, all include a degree of subjectivity. That fact tends to be
16 obscured, however, when an estimate is expressed in numbers, rather than in words.
17 Many people have a tendency to accord greater weight and certainty to quantified
18 data (whether justified or not). The task for decision-makers is to recognize and
19 account for this effect during the evaluation process.
20
21 The effort to be consistent across evaluation criteria is further complicated by
22 the fact that some measures more readily lend themselves to quantified expression.
23 Examples include: expected reductions in pesticide use; expected reductions in
24 pesticide residues; and the costs of applying alternative methods of pest control.
25
26 At the same time, all of the evaluation criteria can draw from quantitative data.
27 Population statistics are a key part of equity considerations (in what regions are the
28 pesticides of concern used and by whom; how do patterns of food consumption vary
29 across different population groups). Results of prior experience with particular risk
30 reduction options includes percentage changes in baseline conditions (such as
31 through the lead phase-out program under the Clean Air Act) and the levels of
32 incentives created in the past (such as the barriers to innovation which can arise from
33 too rigid an application of technology based controls).
6-50
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 In fact, all of the measures and results of the assessment process can be
2 expressed in either a quantitative or qualitative manner. For example, the result of an
3 assessment of environmental effectiveness for any particular option can be expressed
4 either as a number (1,2 or 3) or a word (low, medium or high). Results can also be
5 aggregated across criteria for any particular option in either qualitative or quantitative
6 terms. It should be remembered that, even though numbers can be applied, the
7 accuracy of the analysis is the same as using low, medium, and high. However, the
8 general belief is that using numbers makes the decision more accurate or quantitative.
9 Avoid falling into the "analysis trap."
10
11 Choosing a method of expression requires consideration of a number of
12 competing factors, most of them going to the question of how important to the
13 decision-maker is consistency in the options' scores across particular criteria.
14 Converting all of the estimates into the same scale of numbers, or words, helps the
15 comparison of different options across the evaluation criteria, but may obscure
16 important bits of data relevant to an option's score on any particular criterion (such as
hiding a specific estimate of the costs to be borne by particular populations in an
18 overall score of "high" or "3" on the equity criterion). Similarly, aggregating a particular
19 option's scores across all criteria make for easier comparisons among risk reduction
20 options, but obscure the relative strengths and weaknesses of any particular option
21 (an overall score of "medium" or "2" might result from either scores of "medium" on all
22 criteria, or a mix of only "lows" and "highs" across the criteria).
23
24- However expressed, the basis for the results of individual assessments should
25 be explained. The goal in doing so is to illuminate the policy choices and assumptions
26 that are behind the analysis.
27
28 The analysis of risk reduction options includes considerable amounts of
29 uncertainty. This is true in both the data that are available and the estimates that are
30 made. Data may be lacking because they have not been or cannot be collected
31 (actual pesticide residues in food consumed in all different regions of the country, with
32 any seasonal variations). Theoretically, policy makers operating under time or cost
33 constraints (experience of other regulators with trading programs) may not have
34 reviewed available data. Data may be sought on processes that are inherently
6-51
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 variable and incapable of accurate measurement or prediction (actual levels of illness
2 resulting from chronic exposure to pesticide residues in the U.S.).
3
4 At the same time, the manner in which these uncertainties are expressed
5 affects peoples' decision-making in unusual ways. As with quantified expression, the
6 more certainty that people believe is associated with a piece of data or an estimate,
7 the more weight they give to it, whether it deserves that weighting or not.
8
9 Decision-makers have several options for addressing these uncertainties. Least
10 productive is to abandon the assessment process for particular risk reduction options
11 just because the relevant inputs are uncertain. Another approach is to identify the
12 data needed, along with the time and costs necessary to obtain them, before deciding
13 how best to proceed.
14
15 The most useful approach is explicitly to reflect the uncertainty in the
16 assessment process. Any predictive process faces differing levels of uncertainty.
17 Policy action often needs to be taken in the face of those uncertainties. To the extent
18 policy makers can be explicit about the level and type of uncertainty in the different
19 parts of their decision-making process, their judgments can be understood and
20 evaluated directly by concerned parties. This is true not only for those criteria
21 expressed in a quantitative manner, but for each criterion and the assessment as a
22 whole.
23
24 Having collected the information and considered the issues described above,
25 the process of analyzing particular risk reduction options against the evaluation criteria
26 is a relatively straightforward process of applying the decision-maker's best judgment.
27 The estimated measurements developed for any particular option are reviewed and
28 scores determined for each criterion. For example, an option would have a high score
29 for cost-effectiveness if its direct implementation costs were low and a high level of
30 risk reduction was expected to be achieved.
31
32 One question that does arise is who should conduct the assessment. While the
33 information collection process and the explanation of the results should reach the
6-52
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 widest possible audiences, optimally, the assessment process itself should be
2 conducted by the ultimate decision-maker(s). Such an approach means that less
3 specific expertise is brought to bear on evaluating any particular criterion than would
4 be the case if a group of specially retained experts were asked to undertake the
5 assessment and report their results. It has the advantages, however, of providing
6 greater consistency to the review across criteria and options, as well as making the
7 policy judgments clearer and more readily explained.
8
9 Two major products result from completion of the assessments for each risk
10 reduction option. One is a series of scores across the evaluation criteria for any
11 particular group of options. As discussed above, these scores can be expressed in
12 qualitative or quantitative terms. The level of uncertainty associated with the score for
13 any particular criterion should be considered and reflected appropriately.
14
15 A major issue here is whether to aggregate the separate scores on different
16 criteria into a single, overall score for each risk reduction option. The major advantage
of doing so is to make comparison of the results easier across different options. For
18 example, if one option has an aggregate score of "high" and another an aggregate
19 score of "medium", one could arguably have a basis for choosing the first over the
20 second without further analysis.
21
22 The major disadvantage of an aggregated score is that the total hides the
23 relative strengths and weaknesses of particular options on particular criteria. For
24 example, the aggregated score of "high" for one option may include a lower score on
25 equity or flexibility than the aggregated score of "medium" for another option.
26 Assigning different weights to different criteria as part of the aggregation process can
27 make some adjustment. For some policy makers, this will not be sufficient and the
28 entire matrix should be used to compare the strengths and weaknesses of different
29 options.
30
31 The second major product of the assessment process is the identification of
32 issues that should be fed back into earlier steps in the methodology, including:
33 problem definition; data collection; risk reduction options identification; as well as the
6-53
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 choice, design and measures of the evaluation criteria. Within the time available, the
2 use of such feed-back loops will improve considerably the ultimate outcome.
3
4 The final step in the assessment of risk reduction options is to document the
5 process used. What were the inputs and how were they assembled? What measures
6 were used and how were they evaluated/assessed? How were scores assigned and
7 by whom?
8
9 With the baseline of assessment results in hand for a number of different risk
10 reduction options, the decision-maker can then move to the next steps of optimizing
11 the remaining option sets and selecting one or more for implementation.
12
13 6.8 Optimize the Options
14
15 Any given set of risk reduction options is but some small subset of the virtually
16 infinite set of possible options. For this reason, we propose that any subset that has
17 passed a preliminary screening and back-of-the-envelope evaluation should be refined
18 further before full scale option evaluation is undertaken. Thus risk reduction option
19 refinement begins with a set of options that have been at least cursorily evaluated on
20 multiple criteria and screened on one or more criteria. Some of the options
21 considered have already been redefined in this process before this stage.
22
23 The goal of this step in the analysis is to improve good alternatives, with the
24 subsidiary objectives of improving and reducing the option set under consideration.
25 Given that the options have been evaluated on a set of criteria, ideally refinement will
26 improve their ratings on each of these criteria.
27
28 The simplest way to view refinement is to think of improving a risk reduction
29 option relative to a single criterion. For example, if quality adjusted life years (QALYs)
30 lost were the criterion and the number of QALYs lost could be reduced without
31 changing other attributes of the option, the option that resulted in fewest QALYs lost
32 would be preferred, ceteris paribus. When the criteria are quantifiable and it is easy to
33 agree on appropriate measures, such refinements could, in theory, be treated as
6-54
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 optimization problems. One could think of the task as that of maximizing risk
2 reduction subject to constraints imposed by other criteria.
3
4 The choice of criteria will obviously determine what changes are made in
5 options. Less obviously, how criteria are interpreted and considered will also influence
6 how options are refined. Some criteria will be candidates for maximization. Risk
7 reduction, flexibility, and technological innovation are examples of these. Others, such
8 as cost, will be candidates for minimization. Depending on how criteria are
9 formulated, assessment of an option relative to a criterion may be qualitative or
10 quantitative. It is assumed that most will be subject at least to some kind of ranking.
11 When this is not the case, it is likely that criteria incorporate or reflect multiple values
12 that might usefully be discussed separately (see Table/Figure??-list of proposed
13 criteria, including explicit consideration of competing risks: ref, OTA report)
14
15 Treating risk reduction option refinement as an optimization problem introduces
16 several problems, however, because many criteria that might be used are not easily
quantifiable on comparable scales. For example, flexibility and technological
18 innovation could be difficult to measure on a scale that all interested and affected
19 parties would agree is comparable to any risk reduction measure chosen (e.g.,
20 QALYs). Measuring such attributes on subjective scales is a possibility (e.g., asking to
21 what extent the interested and affected parties (Commission on Risk Assessment and
22 Risk Management, 199?) agree or disagree that the option promotes technological
23 innovation). Such an approach would introduce respondent variability that could be a
24 basis for discussing further option refinements, but might be difficult to predict and use
25 in any consistent approach to optimization.
26
27 Not only will evaluation of an option on any given criterion pose a challenge, but
28 predictions of changes on multiple criteria subject to specific refinements in an option
29 or option set are likely to be highly uncertain and difficult to quantify.
30
31 Optimization generally assumes that a single, summary, quantitative measure
32 can be used to capture and represent the overall value or effect of the option. As
33 discussed elsewhere, we do not recommend such summary measures, for procedural
34 reasons among others. They involve establishing specific tradeoffs between criteria,
6-55
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 and, by representing the results as a summary measure, can conceal changes vis-a-
2 vis specific criteria. Hence optimization may be a useful way of conceptualizing partial
3 evaluations of options (e.g., relative to cost, or time), but we do not recommend it as a
4 way of thinking about the entire refinement process.
5
6 A wide range of policy tools can be used to define an option, including various
7 forms of incentives, education, regulation, enforcement (add to list) and mixes
8 thereof. Because intervening at many points in a hazardous process can reduce risks,
9 very different options may achieve similar risk reduction goals. For example,
10 education about use of emergency medical services or alternatives may achieve as
11 much risk reduction as education about proper use of specific protection devices (e.g.,
12 with application of pesticides). Banning a given product may achieve the same risk
13 reduction as providing incentives for development of alternative products. In such
14 cases, however, the options' other attributes may differ considerably (e.g., flexibility, or
15 cost).
16
17 If those involved in refining options find at this point that they are having trouble
18 determining whether a change improves an option, it may be worthwhile for the
19 interested and affected parties to return to a discussion of the values at stake
20 (Keeney, Value Focused Thinking) and consider redefining the criteria used to
21 evaluate options. For this reason, it is likely to be important that interested and
22 affected parties are represented in the option refinement stage as well.
23
24 Expertise and creativity are key components of structuring good solutions to
25 difficult policy problems. Expertise is only acquired through many years of experience
26 and training, and is likely to change the way people think about the task at hand
27 (Anderson; Van Lehn refs). Such differences often create communication problems
28 between experts and non-experts involved, which should be addressed explicitly.
29 However, systematic consideration of a variety of ways of improving options, along
30 with good communications between decision-makers with expertise in different areas
31 (including interested and affected parties other than EPA), should foster creativity.
32
33 Options can be refined by redefinition in at least the following ways:
6-56
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 Aggregation - combining options,
2 a) Increase the scope of a proposed action (e.g. the actor, time-frame,
3 resources allocated, kinds of tools used, or intervention points in the hazardous
4 process)
5 b) Combine complementary options to create option sets - with due
6 consideration to what exactly is to be done, as well as how much of each action
7
8 Dis-aggregation - changing or eliminating parts of option sets, and
9 a) Reduce the scope of a proposed action (e.g. the actor, time-frame,
10 resources allocated, kinds of tools used, or intervention points in the hazardous
11 process)
12 b) Eliminate parts of option sets
13
14 Some combination of aggregation and dis-aggregation
15
•»" Each of these approaches raises further questions, such as what the optimal
scope of a proposed action is, or what mix of actions and in what proportions is likely
18 to be optimal or best on the most criteria. In any proposed change, risk tradeoffs
19 should be reconsidered explicitly, as they are likely to change as well.
20
21 As mentioned above, how criteria are considered is also likely to influence
22 option refinement. Such influences are not likely to be obvious, but may be serious.
23 For example, criteria that are easier to measure may be weighted more heavily
24 implicitly (i.e., receive more attention, be considered more thoroughly) simply because
25 they are easier to evaluate and discuss. Similarly, if a criterion is usually considered
26 under a given set of constraints that set of constraints may be imposed implicitly.
27
28 Other kinds of cognitive traps are discussed at length in the judgment and
29 decision-making literature (Kahneman and Tversky refs; Pious; Arkes and Hammond).
30 Some examples:
31
6-57
-------
1
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
a) Context effects generally (e.g., framing effects) - an option described in
2 terms of losses may be evaluated differently than the same option described in
3 terms of gains;
4
5 b) Use of heuristics such as availability (a risk that comes more easily to mind
6 may be considered riskier than a comparable risk that is not as "available");
7
8 c) Prominence - people may prioritize or weight an option implicitly based on
9 the dimension that is most important to them; and
10
11 d) Uncertainty - people may weigh less ambiguous attributes of an option more
12 heavily, in implicit tradeoffs, because of an inherent preference for avoiding
13 ambiguity (Hsee ref).
14
15 Knowing when to stop refining an option set is as important as at least
16 attempting to refine the option set originally considered. Several guidelines for
17 stopping rules are proposed, both in terms of the refinement process, as well as in
18 terms of its original goals.
19
20 Process rules could include stopping once each option has been examined at
21 least once; considered input from participants with differing perspectives (all interested
22 and affected parties); and/or examined historical precedents, approaches used in
23 other countries, and approaches used for other risks (by analogy).
24
25 Achievement-oriented stopping rules will be focused on whether the resulting
26 options comprise a manageable or workable set. This might be determined by
27 considering resource limitations, by undertaking back-of-the-envelope value of
28 information analyses, or by achieving some set that can easily be screened to a
29 predetermined size.
30
31 This step in the risk reduction option ranking procedure is a quality check.
32 While elements of the refinement process discussed above are likely to have been
33 incorporated in earlier phases of the procedure, including this step explicitly helps
6-58
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 ensure that those creating the options and ranking them will not become locked into
2 an unnecessarily restricted way of thinking about options. The goal of this step is to
3 improve the options being considered, by explicitly considering whether they could be
4 redefined to better meet the criteria. To keep this step manageable, specific stopping
5 rules are recommended.
6
7 6.9 Select an Option
8
9 The decision-maker is now faced with selecting the "best" risk reduction
10 option(s) from a set of options or group of options that have been screened and
11 optimized to address the defined environmental problem. Thus, the goal of this step
12 in the process is to actually select the risk reduction option for the defined problem.
13 As with other components of the process, this selection process should be as
14 transparent as possible to those outside the selection process. In many ways, it is
15 even more important for this selection process to be made transparent since this
16 activity will ultimately result in the risk reduction option that will be carried forth. The
decision-maker will ultimately have to make the decision based upon the current
18 policies of the Agency but it is a basic tenant of this effort that any decision be fully
19 delineated and documented, including any factors taken into account and how they
20 are weighted relative to one another. This section discusses decision analysis
21 methods that are appropriate for use in the selection of a risk reduction option by
22 being rigorous enough to provide the necessary transparency to the selection process,
23 by being robust enough to deal with the uncertainties of complex environmental risk
24 reduction option selection, and generally being compatible with environmental risk
25 reduction option selection.
26
27 In previous steps, there has been a "down selection" of risk reduction options
28 and potentially some changes or aggregation were made to the original options in
29 order to ensure that the best mix of options or groups of options are being considered.
30 Now the optimized, aggregated, and screened list of risk reduction options will be
31 further evaluated in an attempt to select the best option or set of options. This
32 selection among the various risk reduction options will be based upon a comparative
33 evaluation of the options versus a set of criteria. Using the evaluation of the alternative
34 options relative to these criteria as input, one or more decision analysis methods,
6-59
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 described below, will be used to make the final selection of the risk reduction option or
2 set of options. As previously noted, the evaluation criteria span the range of issues
3 appropriate for consideration when making an environmental decision from those
4 dealing with technical performance and cost factors to those dealing with social
5 factors. As was the case in the screening steps, the goal is to quantify the score or
6 value for each evaluation criteria for each risk reduction option. The degree to which
7 the scores of each evaluation criteria can be quantified and with what degree of
8 certainty, will influence the type of decision analysis method used, as discussed
9 below.
10
11 There are numerous decision analysis methods that can be used for assessing
12 the alternative risk reduction options and several good literature sources that deal with
13 the multi-attribute decision analysis. The selection of the appropriate decision
14 analysis procedures will depend on many factors including the degree of quantitation
15 that can be brought to each of the evaluation criteria and the need for a transparent,
16 well delineated, selection process. The general types of decision analysis methods
17 that could be used for this analysis range from those that are most qualitative to the
18 those that are most quantitative:
19
20 a) Holistic,
21
22 b) Cost Benefit,
23
24 c) Matrix qualitative (High to Low) Ranking,
25
26 d) Decision Factor Analysis (numerical rating factors with weighting on the
27 factors), and
28
29 e) Optimization (multi-attribute decision procedures).
30
31 In the end, the decision-maker will likely choose a hybrid of these processes in
32 making the decision on the best mix of risk reduction options in order to take
33 advantage of the special features of the different decision analysis tools for the
6-60
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 particular environmental problem. In the following paragraphs we discuss the basis of
2 each of these approaches and the potential utility of the decision approach for the
3 types of selection encountered relative environmental risk reduction options.
4
5 The holistic decision process refers to the completely (or almost completely)
6 qualitative procedure of making the decision among the various risk reduction options
7 based upon intuition, professional judgment and certain known facts about the various
8 evaluation criteria. This procedure can be used for consideration of the other criteria
9 when one dominant criteria such as environmental performance has been quantified
10 more rigorously. This approach suffers from the fact that it does not adequately allow
11 delineation of the decision process due to the use of personal judgment.
12
13 Cost benefit analysis is a decision analysis procedure that can be used to
14 evaluate the relative costs and benefits of the risk reduction options. This procedure,
15 for example, would allow a comparison of the costs for each pound of pollutant
16 removed by the various risk reduction options. If each of the evaluation criteria could
be cast as a monetary cost or benefit to society then a comprehensive quantitative
18 evaluation of costs and benefits could be made. For complex environmental problems
19 with numerous risk reduction options and a large number of criteria, this approach is
20 limited by the large informational requirement.
21
22 Matrix comparisons involve the development of a matrix of the assessment of
23 each of the evaluation criteria for each risk reduction option. A quantitative (numerical
24 value) or a qualitative (high, medium, low) ranking is developed for each criteria for
25 each risk reduction option. The decision-maker then must choose the risk reduction
26 option to implement that provides the best balance in simultaneously meeting all of the
27 criteria. It is of paramount importance that the decision-maker delineate the tradeoffs
28 that were made in the final decision and provide the rationale for the final selection
29 given the different criteria considered to be important. Some potential decision rules
30 that are useful in the final selection are provided in the following section.
31
32 Decision factor analysis is similar to the matrix approach but differs in that each
33 of the evaluation criteria is weighted and scored with a numerical value during the
34 evaluation of each risk reduction option. In this approach, each option is evaluated
6-61
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 either qualitatively or quantitatively relative to each evaluation criteria and given a
2 quantitative numerical score. The weight placed on each criterion reflects the relative
3 importance of this criteria to the decision-maker for this particular environmental
4 problem. The selected risk reduction option is thus the option with the highest
5 cumulative score given the numerical score and weighting for each evaluation criteria.
6 The advantage of this approach is that it clearly delineates the factors used in the
7 evaluation and the weighting applied to the criteria. The disadvantage of this
8 approach is the need to develop quantitative scores that discriminate between the
9 options that may be difficult to quantify for complex environmental problems and
10 complex criteria such as social equity. Furthermore, not all preferences are consistent
11 with this approach (e.g. lexicographic preferences).
12
13 Optimization refers to the decision-making process that attempts to quantify the
14 decision into a formulated relationship between an objective function that is to be
15 maximized (or minimized) relative to a series of constraints (or criteria). The objective
16 function can be a function of cost, benefits, and detriments and can account for
17 tradeoffs among these by incorporating relative weighting factors and functional
18 relationships. There are a number of methods for determining the optimum set of
19 decision variables and commercial software is available to carry out the procedures.
20 While this procedure is the most quantitative approach it is often not suitable for
21 complex environmental problems due to the difficulty in establishing functional
22 relationships between decision variables especially with qualitative factors and due to
23 the nonlinear relationships that often exists among the various factors.
24
25 We conducted a number of example evaluations involving the actual selection
26 of risk reduction options for a series of hypothetical environmental problems. The
27 environmental problems were oriented around different contextual viewpoints,
28 specifically environmental stressor (e.g., VOC emissions as related to urban ozone),
29 the media (water, air, solid waste) context and location (place based) context. In
30 these examples, the selection criteria included both factors that were quantifiable and
31 those that were unlikely to be quantifiable for the set of different risk reduction options.
32 The qualitative nature of some of the inputs to the decision, tends to move the
33 selection to be of a subjective nature. However, a purely holistic approach does not
34 provide the appropriate degree of delineation and therefore transparency.
6-62
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1
2 In order to accomplish this selection for an environmental problem with a wide
3 array of potential solutions, it may be necessary, due to time and resource constraints,
4 to further narrow the field with the qualitative matrix approach in order to generate a
5 manageable set of highly ranked options and then conduct a more rigorous analysis of
6 the criteria for each of the highly ranked options. The combination of techniques
7 represents a hybrid of the selection processes discussed in the previous sections but
8 may improve the overall selection process efficiency and ultimate result.
9
10 The final selection of the optimal risk reduction option or set of options can be
11 made by applying different decision rules. The use of decision rules is particularly
12 important when the matrix approach is used for the selection process due to the
13 complexity of the particular environmental problem under analysis. For example in the
14 special case called "dominance" where one single risk reduction option is highest
15 ranked relative to every criteria it would clearly be selected. Other possible decision
16 rules include using a lexicographic rule (just looking at a single criterion), selecting the
option(s) with the highest ranking on a specific criterion, or excluding options with the
18 lowest ranking on any one criterion. The policy decision maker could also select the
19 option(s) whose lowest ranking on any criterion is higher than the lowest rankings for
20 other options, or use some model to combine rankings on individual criteria into one or
21 more scores (as discussed above). Again, the exact application of the decision rule(s)
22 can not be a priori specified. The policy decision maker must, however, document any
23 decision rule employed and provide the rationale for that decision rule in the context of
24 the particular environmental risk reduction options considered. It must be possible to
25 follow the thought process through the decision-making to the final selection of the risk
26 reduction option or set of options.
27
28 One of the greatest challenges in selecting a risk reduction option, or set of
29 options, in addressing a complex environmental problem is dealing with the
30 uncertainties in the performance of the options relative to the criteria. In order for the
31 selection process to be transparent to those on the outside of the process, the
32 process must document the uncertainties in the evaluation of each criteria and deal
33 with the uncertainties relative to the final selection in an open manner. As discussed
34 above, the presence of uncertainties for more complex problems and options will tend
6-63
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 to push the decision analysis process to more qualitative methods such as matrix
2 approaches with low, medium, and high rankings as opposed to decision factor or
3 optimization approaches with quantitative measurement. In any case, it is important to
4 deal with the uncertainties within the selection process framework.
5
6 There are several potential methods of dealing with the impacts of uncertainties
7 in the scores of the evaluation criteria. One method is to have a separate criteria for
8 each risk reduction option that scores the overall uncertainty of the performance of the
9 option relative to all of the other evaluation criteria. A risk reduction option with the
10 same overall score relative to the other criteria but a high degree of uncertainty would
11 be less favorable relative to those with higher degrees of uncertainty but equivalent
12 scores on other criteria. It is important not to reject a better option solely on the basis
13 of the certainty of the evaluation criteria. Nonetheless, uncertainty could be used as
14 one more criterion in the evaluation of the overall acceptability of an option versus
15 other options.
16
17 A second approach is to attempt to quantify the uncertainty range relative to
18 each criteria. Given this range of criteria scores, a sensitivity analysis of uncertainties
19 for the highest ranked risk reduction options can be undertaken to determine if the
20 rank order of the risk reduction options changes over the range of uncertainty
21 assumptions. This is essentially a form of sensitivity analysis on the options. The
22 most comprehensive method is to conduct the process in an iterative manner by
23 collecting new data for those parameters that are so sensitive as to change the
24 selection of risk reduction options. This last method can also have a significant impact
25 on time and costs due to the resources required to collect the additional data.
26
27 Ideally the selection process will lead to a single option or single set of options
28 that are considered to be "best" by the decision-maker and the decision process,
29 including criteria and criteria weighting, will be clearly delineated for those outside the
30 process. Under certain situations there may no clear winner, but rather a set of
31 options may be very close with no clear discriminators among the evaluation criteria.
32 In this case, it may be necessary to analyze other criteria or further quantify special
33 criteria to help delineate the better case.
34
6-64
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 6.10 Document the Process
2
3 The final decision should be transparent, i.e., the motivation and criteria used in
4 reaching the decision should be obvious. Therefore, it is important to document not
5 just the final decision, but also the process leading to the decision.
6 Both the process and the documentation need to contain sufficient detail to
7 substantiate decisions, but should not be so detailed as to be unusable. Therefore,
8 we recommend simple documentation, with tables emphasizing the breadth of options
9 considered, and the values and judgments used in evaluating those options, and
!0 ultimately in selecting the best option. The process of making judgments is inherently
11 complicated, especially in a public forum, because of the need to elicit different values
12 and input from a broad spectrum of people. Thus, the need to document, and be
13 consciously aware of biases and decisions when making evaluations.
14
15 6.11 Quantify Option Effectiveness
1P
} Measurement is a critical component in risk reduction evaluation. The bottom
18 line in implementation of risk reduction options is to determine whether or not the goal
19 is achieved. Once implemented, it is critical that the effectiveness of the risk reduction
20 option or set of options be quantified. If we cannot measure the success of our
21 performance, than we probably cannot determine whether or not we are achieving our
22 goal. It is critical that both what we measure and how we measure are built in before
23 we begin implementing selected risk reduction options.
24
25 The following are the four basic approaches to measurement:
26
27 Direct - Measures that directly quantify the current status against a set goal or
28 goals for improvement (also referred to as outcome measures).
29
30 Indirect - Measures a component that has a direct impact on the specific area
31 for improvement, but does not specifically provide quantitative measurement of
32 the specific improvement goal (including administrative process or stressor
33 measures).
6-65
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 Leading Indicator - Measurement of a specific parameter that provides an
2 indication of the probable success to achieve a goal.
3
4 Lagging Indicator - Measures after the fact a specific parameter that may
5 directly or indirectly measure the performance against the improvement goal or
6 goals.
7
8 To better understand these four approaches, the discussion below describes
9 measurement approaches using two examples. The first example discusses a typical
10 set of risk reduction options to achieve a reduction in cases of childhood asthma (the
11 goal). The second example describes a set of risk reduction options to improve the
12 biodiversity in a receiving stream. We selected these examples to describe
13 measurement approaches in both public health and ecosystem environmental
14 problems.
15
16 For this example, the goal of a specific set of risk reduction options is to reduce
17 the number of childhood asthma cases. Therefore, measurement of the number of
18 childhood asthma cases on an annual basis is a direct measurement of the goal. This
19 is also a lagging indicator because the cases have already occurred and, therefore,
20 the measurement system using this indicator tracks the result from the output of the
21 system as it currently exists.
22
23 For purposes of this example, we will assume that there is a correlation
24 between air quality and childhood asthma cases. Therefore, air quality can be
25 monitored within the study area that provides an indirect measurement of the goal.
26 The assumption is that improvements in air quality will result in reductions in the
27 number of childhood asthma cases. Therefore, measurements of air quality indirectly
28 predict improvements in the number of childhood asthma cases and also represent a
29 leading indicator. Expanding the community's rideshare program is one of the
30 mechanisms that could be implemented to assist in improving air quality. Therefore,
31 the measurement of the rideshare program (decreases in single vehicle travel miles)
32 represents an indirect indictor on air quality and is also a leading indicator for air
33 quality.
6-66
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 As can be seen from this example, a number of parameters can be
2 quantitatively measured to help define the system and predict whether the risk
3 reduction option(s) will achieve the overall goal. It is important to understand the
4 interrelationship of what is being measured and how it relates to the specific
5 improvement goal.
6
7 A typical goal for an aquatic ecosystem is to improve both the quantity and
8 diversity of the biota in a stream, river, or lake. Measurement of biota diversity is a
9 direct measurement and a lagging indicator. A number of techniques have been
10 described in the literature to perform biodiversity measurements. Clearly, the water
11 quality has direct impact on the biota diversity. Therefore, measuring dissolved
12 oxygen, temperature, salinity, and other water quality parameters are critical in
13 defining the water environment in which the biota live. These measurements of water
14 quality parameters are indirect measurements of biota diversity and are leading
15 indicators. Improvements in the water quality parameters typically lead to
16 improvements in biota. A secondary indirect measurement can be reduction in
combined sewer overflow to the receiving stream. Typically, combined sewer overflow
16 adds pollutants to the receiving stream and negatively impacts water quality, and
19 subsequently negatively impacts biodiversity. Projects can be implemented to reduce
20 combined sewer overflow and the reductions in the quantity of sewer overflow as a
21 function of time can be measured, which represent a leading indicator of water quality.
22
23
24 Reduction in the quantity of combined sewer overflow is just one example of a
25 risk reduction option that could be implemented to improve biodiversity. Clearly, water
26 quality is one of the key controlling conditions within the ecosystem impacting
27 biodiversity. Therefore, other risk reduction options that improve water quality also
28 need to be monitored. Those risk reduction options that have the greatest impact on
29 controlling the key water quality parameters must be monitored to determine their
30 impact on the ability to achieve the overall goal.
31
32 In addition to water quality, the physical conditions (including water velocity, low
33 flow water volume, and bottom substrate) within the aquatic ecosystem also play a
34 major role in determining biodiversity. The purpose of this discussion is to alert the
6-67
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 reader to the fact that ecosystems are complex and that there are a number of
2 parameters and conditions that impact biodiversity. It is, therefore, critical to
3 determine the key controlling parameters that impact the specific goal to be achieved
4 and to quantitatively monitor the risk reduction options which are implemented to
5 improve those conditions in the eco-system parameters which lead to achieving the
6 goal.
7
8 In order to evaluate the effectiveness of the risk reduction options implemented,
9 it is necessary to establish a monitoring system/program that provides sufficient data
10 to predict performance and effectiveness. Measurement program plans should
11 describe:
12
13 a) Location and types of samples to be collected;
14
15 b) Frequency of sampling;
16
17 c) Analytical methods;
18
19 d) Quality control; and
20
21 e) Management review.
22
23 The location, type, and frequency of sampling is critical to ensure that sufficient
24 data are collected to predict the performance of a specific risk reduction option or of
25 the system as a whole. In the previous example of the childhood asthma cases, it
26 may be beneficial to collect the number of cases on a monthly basis to see if there is
27 any variation in frequency with time of year. Since there is frequently a direct
28 correlation with air quality and there are seasonal variations in air quality, one would
29 expect seasonal variations in the number of childhood asthma cases. If the number of
30 cases is only collected on an annual basis, the evidence of seasonal variability would
31 not be available. Collecting the data on a daily basis would be too frequent and
32 provide little added benefit over collection on a monthly basis. In the case of the
33 aquatic ecosystem, it is critical to perform the biodiversity analysis at the same time
6-68
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 each year to eliminate seasonal variations. For example, in a river ecosystem, the
2 biodiversity is very different in the summer from the winter. For biodiversity, it is
3 important to compare data from the same time of year. However, water quality data
4 should be collected year-round to understand whether there are seasonal variation
5 which may have an impact upon the biota. For example, spring runoff events may
6 transport pollutants into the receiving body of water that may subsequently impact the
7 biodiversity.
8
9 Data collection is extremely resource-intensive. It is, therefore, critical to select
10 the number of locations, appropriate monitoring parameters, and sampling frequency
11 to optimize the cost impact to the project. Sampling needs to be frequent enough to
12 be representative of the study program. Where appropriate, the frequency should be
13 sufficient to perform proper statistical analysis.
14
15 Most samples that are collected of physical systems, such as water quality
16 samples from a receiving stream, will require laboratory analysis. The analytical
methods employed should be consistent with the nature of the samples and with the
18 most recently approved method of analysis. For example, the analytical method for
19 preparing stream water samples for measuring metals is different from the analytical
20 procedures used to prepare samples for analysis of stream sediment materials. There
21 are a number of existing references that specify current and approved procedures,
22 including Standard Methods for the Examination of Water and Wastewater and
23 procedures published by American Standards for Testing and Materials (ASTM).
24-
25 In all cases, sampling programs require proper quality control. This includes
26 the collection of field blanks, trip blanks and laboratory blanks and spikes to ensure
27 that there has been no unknown contaminant added as a result of the sample
28 collection methodology, transportation or the analytical procedure. Depending on the
29 nature of the samples and the type of analysis being performed, there are general
30 recommended procedures on quality control. It is recommended that before a
31 sampling program be implemented, a formal quality control program be written and
32 approved.
33
6-69
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 Management review of the data is critical to ensure that sampling and analytical
2 methodologies are properly followed, quality control procedures fully implemented,
3 and performance tracked. As required, the sampling program should be modified to
4 ensure efficient and cost effective sampling and analysis are performed. Frequently,
5 sampling programs are modified over time to focus on specific areas of interest and
6 reduce the sampling frequency and analysis for parameters that show little variation.
7
8 To improve the overall implementation of risk reduction options, it is critical that
9 the performance be communicated to the public. The monitoring data should be
10 assembled in a simple to understand format and presented periodically at public
11 hearings/meetings. This is one of the best methods to ensure acceptance of the
12 program being implemented and to receive feedback from interested parties on the
13 performance of the program.
14
15
6-70
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 6.12 References Cited
2
3 C and EN. 1996. Criminal enforcement efforts by EPA continue to grow apace.
4 Chemical and Engineering News. July 29, pp. 33.
5
6 CHEM ENG. 1996. EPA enforcement continues at full throttle. Chemical Engineering,
7 July, pp. 47.
8
9 Davis, M.L. and Chou, G.P. 1992. RCRA Land disposal restrictions for contaminated
10 debris. Hazardous Materials Control, March/April, pp. 30-33.
11
12 Frantzis, 1.1993. Methodology for municipal landfill sites selection. Waste
13 Management and Research, Vol. U, pp. 441-451.
14
15 Keeney, R.L. 1992. Value-Focused Thinking. Harvard University Press.
• ^
. / Knight, M.J. 1985. Public acceptance and siting of hazardous waste disposal facilities.
18 Bulletin of the International Association of Engineering Geology, No. 32, pp. 83-
19 89.
20
21 Levin, M.ll. 1990. Implementing pollution prevention: incentives and irrationalities.
22 Journal of Air and Waste Management Association, Vol. 40, No. 9. pp. 1227-
23 1231.
24
25 McGarity, T. 1994. Radical Technology Forcing in Environmental Regulation. Loyola
26 of Los Angeles Law Review, v.27, p. 943.
27
28 Norman, M.E. and J.D. Keenan. 1996. Market incentives to reduce non-point
29 agricultural nutrient pollution: a theoretical and implementational discussion.
30 Journal of Environmental Systems, Vol. 24, No. 2, pp. 13167.
31
32 OTA 1993. Environmental policy tools: a user's guide. Office of Technology
33 Assessment, United States Congress, Washington, DC, 212 pp.
6-71
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
Quote
1 Schwartz, S.F., R.A. McBride, and R.L Powell. 1989. Models for aiding hazardous
2 waste facility siting decisions. Journal of Environmental Systems, vol. 18, No.
3 2., pp. 97-122.
4
5 Smith, P.M. 1996. Environmental evaluation: fuzzy impact aggregation. Journal of
6 Environmental Systems, vol. 24, No. 2, pp. 191-204.
7
8 UN 1995. Work programme on indicators of sustainable development of the
9 Commission on Sustainable Development. Division for Sustainable
10 Development, Department for Policy Coordination and Sustainable
11 Development, United Nations, New York.
12
13 U.S. EPA. 1990. Reducing Risk: The Report of the Strategic Options Subcommittee,
14 Relative Risk Reduction Project, Appendix C. EPA SAB-EC-xxx, Science
15 Advisory Board, U.S. Environmental Protection Agency, Washington, DC 140p.
16
17 Wright, E.G., H.I. Inyang, and V.B. Myers. 1993. Risk reduction through regulatory
18 control of waste disposal facility siting. Journal of Environmental Systems, Vol.
19 22, No. 1, pp. 27-33.
20
21 Zeiss, C. and B. Paddon. 1992. Management principles for negotiating waste facility
22 siting agreements. Journal of the Air and Waste Management Association, Vol.
23 42, No. 10, pp. 1296-1304.
24'
6-72
-------
PART V IMPLEMENTATION AND PERFORMANCE EVALUATION
-------
PREFACE
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
CHAPTER 7. PERFORMANCE EVALUATION: THE USE OF
ENVIRONMENTAL DECISION REPORT CARDS
TABLE OF CONTENTS
1
2
3 7.1 Introduction 7-1
4
5 7.2 Types of Performance Measures 7-3
6
7 7.3 Improved Report Cards 7-8
8 7.3.1 Employing a Broader Range of Measures 7-8
9 7.3.2 Reporting on Environmental Outcomes 7-10
10 7.3.2.1 Environmental Health 7-10
11 7.3.2.2 Ecological Health 7-13
12 7.3.3 Additional Design Considerations 7-15
13
1 7.4 Implications for Existing Monitoring Systems 7-17
15
16 7.5 Summary and Recommendations 7-19
17
18 7.6 References Cited 7-22
19
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 CHAPTER 7. PERFORMANCE EVALUATION: THE DESIGN AND USE
2 OF ENVIRONMENTAL REPORT CARDS
3
4
5 7.1 Introduction
6
7 The lack of a widely accepted and commonly used method for evaluating the
8 effectiveness of governmental (including environmental) programs represents one of
9 the great challenges facing agencies such as the EPA as the millennium draws to a
10 close. With the passage of the Government Performance and Results Act (GPRA) in
11 1993, the Congress has mandated that an agency must do much more to answer the
12 bottom-line question "How are we doing?" if it expects continued, let alone expanded,
13 support. In the context of EPA, the answer to this question is even more important now
14 than in the past because of the growing number of complex environmental issues, the
15 evolution of institutional roles from narrowly focused mandates to broader public
16 demand for high quality protection of human health and ecological integrity, and the
" "* intensified scrutiny of the cost of environmental programs.
>
19 In the same way that GPRA demands accountability for performance at a
20 government agency-wide level, the IED framework seeks accountability for
21 performance at the level of individual integrated environmental decisions. In the former
22 case, the question is "What is the state of the environment, and what is the impact of
23 Agency actions on that state and the trends in the environment?" In the latter case, the
24 question is "What is the state of that portion of the environment that a specific
25 management decision was designed to affect, and what information is available that
26 might suggest adaptive management options that could obtain more desirable results
27 more efficiently?"
28
29 The term "environmental report card" has been used to refer to the evaluations
30 that are applicable to both the more global GPRA-like and the more specific lED-like
31 situations. For example, recently several groups have recommended that a "report
32 card" be developed for national environmental programs (e.g., Enterprise for the
33 Environment, 1997; Vice President Al Gore, in his direction to the National Science and
34 Technology Council, year??). In many respects the Annual Reports from the Council
35 on Environmental Quality served as prototypes for a national environmental report card.
7-1
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 In addition, several regional initiatives have generated report cards, including reports of
2 the U.S. Forest Service and Bureau of Land Management in conjunction with the
3 Northwest Forest Plan, the Chesapeake Bay Program's annual State-of-the-Bay
4 publication, the State of the Great Lakes report (Environment Canada and EPA, 1997),
5 the State of Boston Harbor report, and others (Young, Harwell to provide cites). Each
6 of these reports provides a periodic update of governmental agency activities and, to a
7 limited extent, related environmental improvements.
8
9 In contrast to these environmental report cards that assess the environment
10 more broadly, the IED uses the term "environmental decision report card (EDRC)" to
11 refer to an evaluation of specific decisions.
12
13 As depicted in the framework diagram (Figure 1-3), the EDRC is the formal
14 mechanism for feedback and course correction. Such feedback on the extent and
15 distribution of environmental effects is helpful in determining whether the relative
16 seriousness of risks was accurately characterized in the first place and whether specific
17 risks have changed as a result of an implemented risk reduction program. Report cards
18 should also provide the basis for evaluating the performance of specific risk reduction
19 programs or decisions, as judged against decision criteria such as efficacy at reducing
20 aggregate risk, cost, equity, and time required to achieve risk reduction goals.
21 Therefore, environmental report cards should provide the information needed to a)
22 identify opportunities for course correction and adaptive management, i.e., modification
23 of risk reduction approaches in light of performance information or new information on
24 risks; and b) assign specific accountability-to individuals, programs, or organizations-
25 for environmental results. A reporting system that focuses on environmental outcomes
26 associated with individual decisions will provide greater focus and discipline by explicitly
27 expressing the relationship between an environmental investment (e.g., money,
28 information, skills, and time) and the results achieved.
29
30 The enhanced accountability provided by a well-designed report card
31 mechanism is particularly important in the context of the greater flexibility espoused in
32 the IED framework for both government agencies and private interests to approach
33 problems with innovative solutions. As the system of integrated environmental
34 decision-making evolves, with its focus on reducing aggregate risk through flexible
35 approaches to risk reduction, reporting on the rate of achievement of particular single-
7-2
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 stressor standards (or other individual characteristics affecting environmental health)
2 will fail to tell the whole story. In its stead, there needs to be a mechanism to gauge the
3 collective impact of sets of risk management approaches on actual outcomes.
4
5 A well-designed and appropriately implemented outcomes reporting system also
6 can engender long-term public support for environmental programs. By providing
7 periodic objective reports that link decisions to outcomes on the health of ecosystems,
8 human health, and quality of life — in other words, information about the things people
9 really care about — rather than specific chemical risks whose impacts may not be
10 obviously relevant to most people (e.g., mercury or 1,3-butadiene), the reporting system
11 has the capacity to generate a strong sense of ownership and investment on the part of
12 the public. (For a more detailed discussion of the benefits from performance-based
13 reporting systems, see Enterprise for the Environment, 1997.)
14
15 While interest in performance evaluation is growing, EDRCs and the broader
16 environmental report cards will be only as useful as the information contained in them.
17 In the following sections, we discuss various types of performance measures and some
issues associated with their use in environmental reporting systems.
19
20 7.2 Types of Performance Measures
21
22 With the growing interest in performance evaluation, a number of closely related
23 terms have emerged to refer to different performance measures; i.e., those measures
24 that are used in constructing the report card. For the sake of clarity, the Committee
25* defines the following types of performance measures:
26
27 a) Process Measures, i.e., measures of administrative effort or program actions
28 that are presumed to result in environmental or health improvements (e.g.,
29 numbers of permits issued, number of enforcement cases pursued, number of
30 contaminated sites cleaned up to standards).
31
32 b) Stressor Measures (levels): measures (levels) of stressors in the
33 environment that are used to determine attainment or non-attainment of desired
34 reductions in stressor levels (e.g., total emissions of a pollutant, levels of
35 dissolved oxygen or turbidity levels in a stream, and density of roads in a
7-3
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 watershed).
2
3 c) Exposure Measures: measures of the co-occurrence or contact between an
4 individual or population and environmental stressor(s) over a defined time period.
5 The term exposure is traditionally associated with chemical stressors (e.g.,
6 stressor levels in food, concentrations of stressors in tissues, time-activity
7 measures, and total exposure to a contaminant via all routes), whereas the term
8 co-occurrence is often used as a broader term applicable to chemical, physical,
9 and biological stressors. Exposure measures are more directly related to effects
10 than ambient levels of stressors in environmental media because they address
11 direct contact with the stressor. These measures can also be more readily linked
12 to risk management decisions than effects measures since causes of adverse
13 effects are often multi-factorial.
14
15 d) Effects Measures: measures of human and/or ecological effects, the
16 changes in which can be used to assess the impact of an environmental risk
17 reduction program (e.g., asthma rates, cancer rates, acres of wetlands gained or
18 lost, local extinctions of important species). Another application of effects
19 measures is for condition assessment, in which a suite of effects measures are
20 evaluated and reported in combination to characterize the health or condition of
21 an entire population or ecosystem. Condition assessment provides a baseline
22 against which to evaluate the success of specific environmental decisions or
23 multiple decisions impacting a population or geographic region. In addition,
24 reporting on condition provides feedback on whether the problem formulation
25 has correctly identified the set of stressors associated with the environmental or
26 health consequences of concern.
27
28 The Committee considers measures of effects or condition to be environmental
29 outcome measures. We note that this definition differs from the Agency's definition of
30 "environmental outcomes," which also includes measures of stressor levels (EPA
31 Strategic Plan, 1997). We recommend, however, that stressor measures be kept
32 distinct from environmental outcome measures because a) changes (increases or
33 decreases) in stressor levels do not necessarily translate into changes (increases or
34 decreases) in risk, and b) the public's environmental goals are typically in terms of
35 desired states of health or condition, rather than desired stressor levels.
7-4
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 The different types of measures can be thought of in terms of a spectrum of
2 measures ranging from least directly to most directly related to actual environmental
3 outcomes, which are of primary interest to the public. In many cases, the spectrum of
4 measures will also correspond to a spectrum of response times, ranging from shorter
5 term to longer term measures. For example, it is relatively quick and easy to document
6 the number of permits issued in a year, but it is more difficult and time-consuming to
7 determine the actual impact of that process measure on the condition of the
8 environment. Figure 7-1 illustrates this spectrum and relates the Committee's
9 terminology to other commonly used terms for environmental performance measures.
10
Figure 7-1. Spectrum of Performance Measures
Environmental Outcomes
Process Stressor Exposure Effects/Condition
Measures Measures Measures Measures
activity measures' pressure indicators3 co-occurrence4 adverse effects5
output measures^ release measures health outcomes9
response indicators' emission measures state indicators3
'beans' ecological indicators7
<_____ >
Least directly related Most directly related
to Environmental Outcomes to Environmental Outcomes
'EPA Strategic Plan; 2GPRA; 3OECD Framework; 'Chapter 2; 5Ecorisk Guidelines; 'Understanding Risk; 7EMAP
11 Examples of the various types of performance measures and their use in concert
12 to evaluate the success of an IED project is illustrated in Figure 7-2 using a watershed
13 restoration example.
14
15
7-5
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
Figure 7-2. Evaluating the Success of an Integrated Environmental Decision: A
Watershed Example
To illustrate the use of each of the types of performance measures to report on a specific
Integrated Environmental Decision, consider a watershed management or restoration program. In
this example, the ultimate goal of the program is to maintain or re-establish an ecosystem that
supports a diversity of habitat types along with their resident communities of plants and animals,
supports essential ecological functions, and is self-sustaining. Here, the Integrated Environmental
Decision will consist of a set of interrelated actions, many of which will be designed to address
multiple stressors in order to achieve a reduction in the aggregate nsk from those stressors. Other
IED program actions will focus on restoring damage from past stressors (such as the restoration of
riparian zones damaged by livestock operations in order to decrease sedimentation downstream,
provide shade and cooler temperatures for aquatic species, and provide additional nutrients to the
system from dropping leaves).
The evaluation criteria used to judge overall IED program results will include measures of
habitat quality (such as length of intact corridors of natural riparian vegetation), water quality and
temperature, hydrology that mimics natural variations, the extent of connectivity between floodplains
and the river, nutrient balance, presence of sustainable native populations, and the like. Taken
together, these effects measures effectively describe the condition of the watershed, i.e., whether
the watershed, in fact, can sustainably support native populations and their habitats and maintain
ecological functions - and they therefore report on the aggregate results of the IED program.
Each of these effects measures can also be used as to evaluate the success of individual
actions within the IED program. Following the example above, measures of length of intact riparian
comdor, water temperature, and nutrient balance would be used to assess the success of specific
actions taken to restore riparian zones.
The time frame required to see changes in the effects measures will vary. Some, such as
population levels of short-lived species of interest, may be detected after only one year of a
management regime that alters pollutant inputs and water releases to the system. Other
environmental responses may take more than ten years to be detectable.
In addition to the effects measures and condition assessment, direct reporting on the
decreased pressure from various stressors will be useful. In the example above, such stressor
measures might include the decrease in the rate of new riparian damage and increases in the
release of cool water from dams in the summer. Stressor measures relating to other actions that are
also part of the IED program might include reductions in ambient pollution levels, decreases in the
number of unscreened pumps that injure fish, and the restoration of periodic flood flows.
Finally, process measures will provide insight into the level of effort expended and provide a
shorter-term indicator that the program is proceeding as planned. Examples of process measures
might include conservation easement contracts signed for the management of riparian corridors,
changes in the regulations governing water releases from dams, and numbers of water pollution
permits updated with new effluent limits.
7-6
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Although the focus of this chapter is on evaluating decisions on the basis of
2 environmental outcomes, integrated environmental decision-making may seek to
3 achieve other goals such as environmental justice, quality of life, inter-generational
4 equity, or public "right to know." In such cases, it will be important to define
5 performance measures to evaluate these other categories of outcomes, referred to in a
6 recent NRC report (1996) as social, economic, and ethical outcomes.
7
8
7-7
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 7.3 Improved Report Cards
2
3 7.3.1 Employing a Broader Range of Measures
4
5 In order to provide a meaningful assessment of environmental program
6 performance, it is critical that EDRCs provide accurate information not only about
7 process measures (such as how many industrial units now have permits and are
8 meeting air pollution requirements, or how much money has been spent controlling
9 water pollution), but also on environmental and health outcomes. Moreover, the
10 specific measures selected should include those that are related to desired states of
11 human health or ecological condition (such as the proportion of at-risk children that are
12 free from pollutant-related respiratory diseases, or the maintenance of native
13 biodiversity in a stream).
14
15 In cases where outcome measures are not available, or not measurable in the
16 short term, the selection of the most appropriate measures of impact of an
17 environmental decision should be influenced by the strength of the association between
18 the measure and the desired environmental outcome. If, for example, exposure to a
19 chemical is identified as the causative agent for a particular disease, and exposure is
20 closely correlated with ambient concentrations of a stressor, then a stressor measure
21 may be an adequate predictor of resultant health effects. In other cases (e.g.,
22 persistent bioaccumulative toxics), environmental concentrations in a particular medium
23 may not be related clearly to human dose or exposure. In such cases, direct measures
24 of exposure (e.g., fish tissue contaminant levels) or of the incidence rate of adverse
25 effects in populations at risk (e.g., subsistence fishers or fish-eating wildlife) may be
26 needed to evaluate progress toward the desired environmental outcome.
27
28 The SAB urges the Agency to place increased emphasis on outcome measures,
29 developing new ones where required, because the ultimate goal of integrated
30 environmental decision-making is the actual reduction of risks and adverse effects. We
31 recognize that environmental outcomes, whether measures of health, ecological
32 condition, or quality of life, may not be observable over short time frames, so that
33 process, stressor, and exposure measures will continue to be important in
34 environmental decision-making and management. Outcome measures, however, are
35 an important means of verifying the theoretical basis for the control or abatement
7-8
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 options chosen. In other words, if the postulated relationship between stressors and
2 effects is inaccurate, then stressor or exposure reductions might not produce the
3 expected improvements in adverse effects or condition. A robust reporting system will
4 include a mix of process, stressor, exposure, and outcome measures.
5
6 Even the Agency's traditional list of process measures will need to be expanded
7 in order to evaluate the performance of some of the new approaches to environmental
8 management. In recent years, the Agency has added more tools to the environmental
9 toolbox, bringing to bear such new approaches as economic incentives, negotiated
10 agreements, and the like (see Chapter 6). Therefore, the collection of data to evaluate
11 these new approaches should evolve as well. There are several important inputs
12 currently missing from most reporting systems that would provide valuable information
13 for EDRCs that assess how well environmental programs have worked and what
14 changes or adjustment might be made. In the case of a marketable permit system, for
15 example, systematic reporting about the number of transactions, the average price per
16 unit of the traded commodity, the number of market participants, the net reduction in
17 pollution output, and average marginal cost per unit of pollution reduced should be
compared with initial predictions. To the extent that a risk management program
19 involves more than one tool, the entire panoply of actions should be reviewed for
20 efficacy, cost, and other measures of performance. Separate information should also
21 be provided on the "non-quantitative" inputs to decision-making, including the effects of
22 the risk reduction program on sustainabilrty and equity.
23
24 While all of these parameters are often reported somewhere, seldom are they
25 reported in the same place, in terms understandable to the public, and in a format that
26 provides useful input into the next iteration of decision-making on a particular issue.
27 This feedback loop is essential to the functioning of the proposed IED framework.
28
29 In summary, over the years the Agency has developed a traditional set of
30 systems to track administrative process measures and stressor measures. However, in
31 addition to these measures, the IED framework emphasizes the need to strengthen
32 reporting by a) expanding the range of measures used to evaluate the performance of
33 new (non-traditional) environmental management approaches, and b) shifting from the
34 traditional focus on process and stressor measures to a more balanced suite of
35 measures, including measures of exposure and environmental outcomes.
7-9
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
7.3.2 Reporting on Environmental Outcomes
7.3.2.1 Environmental Health
Setting up an outcomes-based reporting
system requires an understanding of the chain of
events linking human activities to stressor levels to
effects of concern (whether health, ecological, or
quality of life); hence the need for process, stressor,
exposure and especially effects measures in an
EDRC. The expected response time for various
measures along this chain of events will vary. Thus,
the time frame for monitoring and reporting on suites
of measures needs to be established based on the
time expected for changes to occur.
In looking at current reporting systems, it is
clear that the nation has much more fully developed
systems for gathering and analyzing information on
outcome measures for human health than it does for
ecological condition. The Department of Health and
Human Services (HHS) -the second largest
Department in the federal government - oversees
an extensive effort to fund and coordinate
information collection and analysis at local, state,
and national levels, which provides input on
measures of effects and condition for human health.
The Department has introduced a number of useful
approaches in this regard. For example, the HHS
effort to establish national human health goals—Healthy People 2000—and to track
progress towards these goals (HHS, 1990; 1995) using outcome measures has served
as a model for the Agency's own effort to articulate environmental goals (EPA, 1996).
In addition to HHS, there are many other groups - at various scales - that also
periodically release studies of health effects/trends in different geographic areas; e.g,
World Resources Institute, World Watch, and the World Health Organization.
Reporting on Quality of Life
Human welfare and quality of life
(QOL), defined in terms of non-
health related impacts from
environmental exposures or
ecological change, is inextricably
linked with both ecological condition
and human health issues. Progress
toward achieving QOL goals can
and should be measured in an
integrated performance evaluation
system. In some cases, QOL
measures will relate to exposures to
chemical agents (e.g., nuisance
odors), and thus might be included
in an overall report on health. Other
indicators of QOL will denve from
the consequences of a change in
ecological condition (e.g., changes
in biological diversity, availability of
recreationally important species),
and might be reported in concert
with measures or indicators of
ecological effects. Yet other QOL
measures may relate to economic
consequences of environmental
exposures or ecological changes
(e.g., changes in property values,
job impacts of changes in the
availability of harvestable natural
resources).
7-10
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 However, current systems often provide fragmented health information, such as
2 data on only the most serious outcomes; e.g., deaths from chronic obstructive
3 pulmonary disease, rather than total incidence of respiratory problems. Many important
4 health outcomes, such as reproductive health or neurologic health of children, are not
5 well captured by these systems. In many cases, environmental data are collected
6 without consideration of how they might be used to understand the role of the
7 environment in reported health outcomes or how the data might be used to attribute risk
8 from exposures to environmental stressors. As a result, these data are of limited utility
9 in assessing the relative contribution of environmentally mediated factors relative to
10 other major drivers, including nutrition, education/lifestyle, economic status, and
11 genetics.
12
13 We recommend, therefore, that the Agency work with HHS and other
14 governmental and non-governmental organizations to develop an assessment that
15 examines the impact of exposures to environmental stressors on human health, with a
16 particular - but not exclusive - focus on those stressors that fall within the Agency's
17 sphere of influence. Emphasis should be placed on measuring total or aggregate
exposures to various stressors, i.e., by all routes of exposure — inhalation, ingestion,
19 and dermal contact. Such an "environmental health" report card could serve as a
20 component of a larger "human health report card" released by some of these larger
21 organizations which could eventually attribute health outcomes to specific
22 environmental risks or combinations of such risks. Any system that is developed should
23 be systematically scanned for changes that may indicate problems. Reviews at present
24 are episodic and so may not recognize early changes in health indicators.
25
26 Efforts to report through EDRCs on the effectiveness of environmental protection
27 programs in reducing human health risks would benefit from the development of a
28 conceptual framework describing the relationship between goals, interventions, and
29 desired outcomes. Evaluating the effectiveness of an environmental intervention on
30 human systems will require consideration of the same factors that are used to
31 determine risk; namely the measure of a health effect (outcome), the measure of
32 exposure to identify potential stressors of concern, and the establishment of human
33 dose. The health outcome would be the endpoint of choice for evaluation. However,
34 when an effect takes many years to occur (e.g., cancer), measures of early steps along
35 the path to the outcome or even exposure measures should be used for evaluating the
7-11
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 effectiveness of interventions.
2
3 At the same time, we need to be aware that implementation of an EDRC
4 reporting system for human health risk reduction will be hampered, in many cases, by a
5 lack of data in key areas; e.g., exposures, differential susceptibilities among
6 subpopulations, and incidence of non-fatal diseases. It is important to recognize that
7 there are critical gaps in currently available data (discussed further in section 7.4). At
8 the same time, these gaps should not limit efforts to collect the data that do exist to
9 evaluate specific programs today or constrain the consideration of desirable
10 performance measures for which data might be collected in the future.
11
12 Even where data on environmental health are available, the challenge of relating
13 information on exposures and effects remains. Current data often are not sufficient to
14 link exposures in populations to health outcomes. Occupational/workplace exposures
15 are more easily dealt with in this regard than residential exposures. Geographic
16 mapping systems are a potential tool for linking residences to exposures. However, the
17 usefulness of mapping approaches is still dependent on the precision possible for
18 assessing exposures by exact location. Obtaining such precision is frequently a difficult
19 task in rural and suburban areas. However, efforts are underway to improve the
20 recording of location for both exposures and, in some cases, outcome data, which will
21 improve the ability to match and model exposure and outcome data in the future.
22
23 To further complicate matters, the susceptibility of populations is known to vary,
24 and science is moving rapidly to determine what constitutes these susceptibility
26 characteristics. At present, we can judge susceptibility by general characteristics such
26 as age, race, and sex, but even these have not been used extensively to judge the
27 differences in risks. An important challenge lying ahead is to develop methods and
28 reporting systems that use traditional exposure information, population characteristics,
29 and rapidly expanding data on biomarkers of susceptibility and exposure to expand our
30 ability to be more specific in identifying subsets of the general population that are at risk
31 of disease.
32
33
7-12
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 7.3.2.2 Ecological Health
2
3 The focus of the I ED framework is on improving the environment by making
4 individual - yet integrated - environmental decisions and tracking their impact on
5 human and ecological health through appropriate measures consolidated and reported
6 in an EDRC. In this regard, discussions of EDRCs for ecological health and human
7 health are quite similar. However, in the case of ecological health, the particular
8 outcomes of concern are less obvious, certainly to the public. Whereas a human health
9 outcome of reduced childhood asthma is easily grasped and enjoys a broad consensus
10 as a desired goal, some ecological outcomes (e.g., increased biodiversity) may be more
11 esoteric. Thus, the Agency, working with the scientific community, should provide
12 leadership in defining specific ecological outcome measures that can be used to
13 quantify progress toward the widely held goal of protecting ecological integrity.
14
15 This challenge of clearly articulating ecological outcomes should be confronted
16 early, in the Problem Formulation Phase of the IED. Through open and frank
17 discussion among assessors, managers, stakeholders, and the public, common
agreement should be reached on the desired outcomes and the risk reduction options
19 of choice. In addition, the selection of ecological outcome measures should be guided
20 by a conceptual model, also developed in the Problem Formulation Phase, that relates
21 expected changes in specific environmental outcomes to anticipated changes in
22 stressors resulting from specific management approaches or interventions. The
23 Problem Formulation Phase should also settle on the measures or indicators of those
24 changes that will be monitored and reported. Generalized conceptual models of
25 various ecosystem types and landscape units may be used to provide a scientific
26 foundation for the conceptual framework.
27
28 While easy to give, in the past this advice has been difficult to follow. But there
29 is reason to believe that much more is possible today than was possible just a few
30 years ago. For instance, the idea of a conceptual model is working its way into the
31 mainstream (Ref: EPA Eco RA GLs). The notion of appropriate monitoring endpoints
32 is being integrated into ecological risk management plans (Ref: South Florida). The
33 concept of adaptive management to address ecological problems is becoming more
34 commonplace (Ref: SETAC on adaptive management). The practice of stakeholder
35 involvement has gained wide acceptance (Ref: Comm on RA/RM)
7-13
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Therefore, the Agency should vigorously pursue the development of ecological
2 EDRCs to evaluate the performance of its integrated environmental decisions. At the
3 same time, the Agency should be aware of and support - by contribution to and
4 utilization of the results of - efforts to generate report cards to assess the overall
5 condition of the nation's environment. These efforts can incorporate the results of
6 individual EPA risk reduction actions in the process of addressing the broader questions
7 of ecological health; e.g., "Are our landscapes and ecosystems functionally and
8 structurally intact, supporting a broad range of native animal and plant communities, or
9 not?" The National Science and Technology Council recently concluded that
10 comprehensive assessment of the nation's natural resources will require an integrated
11 national strategy for environmental monitoring and research (CENR, 1997). An ongoing
12 project by the Heinz Center, funded by the White House's Office of Science and
13 Technology Policy, is intended to produce a national environmental report card,
14 integrating environmental program results across the federal government, by the year
15 2001 (Heinz Center, 1998).
16
17 The choice of appropriate ecological outcome measures should be guided by a
18 conceptual model that relates specific environmental goals and sub-goals to
19 management approaches or interventions. The conceptual model should, in turn, link
20 expected changes in stressor levels to expected changes in human or ecological
21 condition associated with specific risk reduction actions. In order to assess ecological
22 condition, it will also be important to select measures of ecological attributes for which
23 cause-and-effect relationships with stressors (at least those being managed) have not
24 been postulated. The model should also indicate the measures or indicators of that
25 change that will be monitored and reported. In addition, it is important to evaluate
26 comprehensively the essential characteristics of landscapes and ecosystems across
27 levels of hierarchy and time scales in a way that will allow aggregation and
28 disaggregation of information for a variety of geographical boundaries. In order to
29 provide a scientific foundation for the conceptual framework, it would be useful to
30 employ generalized conceptual models of various ecosystem types and landscape
31 units.
32
33
34
35
7-14
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 7.3.3 Additional Design Considerations
2
3 Environmental Decision Report Cards (EDRCs) should exhibit the following
4 general characteristics:
5
6 a) Transparent: First and foremost, each EDRC should be derived in a clear
7 and methodical fashion from objective measurements. Whenever expert judgment is
8 required to aggregate components of the report card (such as aggregating disease
9 information on separate subpopulations or aggregating information regarding the quality
10 of different habitats within an ecosystem), the method of aggregation should be laid out
11 explicitly. If algorithms are used to combine different types of measurements, the
12 values assigned to the elements in the algorithm should be documented and justified.
13 Ideally, EDRCs should include the results of sensitivity analysis and uncertainty
14 analysis. Clear documentation of the algorithms used also helps to assure continuity
15 over time, so that report card results from one decade can be fairly compared to the
16 next since for some outcomes, such comparisons will need to be made over many
17 years, even decades or generations.
19 b) Relate closely to environmental goals: Measures included in the report
20 card should be directly related to the specific goals of the risk reduction program. In
21 other words, the design of the report card should allow one to understand how changes
22 in performance measures relate to the attainment of environmental goals and vice
23 versa. Although different versions of the EDRC (with differing amounts of technical
24 detail and aggregation) will be needed by the various audiences and users of
25 performance information, the underlying scientific framework must be consistent.
26 Ideally, the logic upon which the EDRC is based should be both systematic and
27 hierarchical so that the information can be a) aggregated- at least qualitatively~to a
28 national or international scale, and b) disaggregated to meet regional and local needs in
29 other report card exercises.
30
31 c) Build a continuing historical record: The EDRC should set forth a frame of
32 reference for each important ecological, human health, or quality of life goal, indicating
33 its past and current status, desired long-term condition, and shorter term milestones.
34 To be truly effective, the EDRCs must be maintained and compared from time to time in
35 order to (1) determine the trend, (2) become aware of new factors that might be
7-15
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 impacting on the goal, and (3) learn the lessons that adaptive management has to
2 teach.
3
4 d) Provide an integrated picture: The report should be organized to convey a
5 more integrated picture of health (e.g., a healthy terrestrial ecosystem), rather than just
6 progress toward stressor/pollutant-specific goals; e.g., a target effluent level. In
7 commenting on the draft EPA document, Environmental Goals for America (EPA,
8 1996), the SAB noted that the structure of the draft report was not ideal for this purpose
9 because it was "centered on the current program office- or media-specific issues" (SAB,
10 1997).
11
12 When the IED framework is applied to multiple environmental risks, the
13 performance measures selected during Phase II and monitored during Phase III may be
14 tied to specific risk reduction tools. Indeed, designers and implementers of risk
15 reduction programs require information on the relative effectiveness of selected risk
16 reduction approaches in meeting risk reduction targets, as assessed against a number
17 of decision criteria (e.g., timeliness, cost-effectiveness, and equity). Such
18 reassessment requires that risk reduction be attributed to specific risk reduction tools, to
19 the extent possible. At some point, however, it will also be important that performance
20 measures be reported in combination in an integrated report card so that overall
21 progress toward achieving a goal can be assessed.
22
23 e) Credible: An additional consideration in the design of report cards is the
24 question of who will generate them. It is valuable, from the standpoint of learning, for
25 those implementing the risk reduction measures (whether in the public or private sector)
26 to generate such reports. It is possible, however, that information will be gathered and
27 reported by some group not directly involved in the implementation. Such a scheme
28 has the advantage of independent reporting but reduces the opportunity for learning by
29 those implementing the plans. A third possibility is for the report cards to be prepared
30 by those implementing the plan (self-reporting), combined with an outside audit
31 program by auditors not involved in the implementation. This latter combination has the
32 advantage that those implementing the plans leam as they produce their reports; they
33 also keep their eyes on the yardsticks set up for measurement as they implement
34 plans; at the same time, the problems that can frequently arise with self-reporting
35 schemes are counteracted by the periodic audits.
7-16
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2 7.4 Implications for Existing Monitoring Systems
3
4 Decisions on the design of a report card system have implications for information
5 collection systems, both those that track administrative processes and those that collect
6 environmental data (e.g., ambient monitoring programs). In selecting performance
7 measures, there will often be a gap between what it would be desirable to know and
8 what it may be feasible to know. This gap may exist as a result of several factors,
9 including the limited state of knowledge about the relationship between stressors and
10 effects, the costs involved in obtaining certain types of information, and the willingness
11 of affected people to provide information. Nonetheless, it is important that careful
12 consideration of the most desirable performance measures, including those based on
13 an established chain from process and stressor measures to exposure and outcome
14 measures, influence the types of information that are actually collected.
15
16 Existing monitoring systems, both those geared toward environmental condition
17 and those intended to assess human health and welfare, provide invaluable
information, particularly regarding the implementation of current legal mandates.
19 Compliance monitoring and ambient air and water quality monitoring networks, for
20 example, provide significant process as well as some outcome data. The EDRCs and
21 overall report cards designed to answer broader questions, however, will likely require
22 changes to existing monitoring systems.
23
24 Ecological Health Outcomes: Assessments of ecological integrity, for
26 example, require information not only on water quality, air quality, soil quality, water
26 flow, and populations of certain species, which are commonly monitored today, but also
27 measurements of biological community structure, the presence of successional states,
28 diversity of habitat types across the landscape, connectivity, altered topography,
29 hydrodynamic patterns, and so forth. This latter set of parameters is not typically
30 monitored by EPA, but has received greater attention by other federal agencies (USGS,
31 USFS). This fact emphasizes the importance of strengthening and maintaining the
32 collaboration among the many agencies that conduct monitoring, an effort begun
33 several years ago under the auspices of the Committee on Environment and Natural
34 Resources (CENR, 1997). Linking monitoring efforts into a coherent framework will
35 improve the quality and robustness of the EDRCs.
7-17
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2 Human Health Outcomes: As noted above, much of the data, both on
3 environmental exposures and on health outcomes, needed to evaluate the
4 effectiveness of a risk reduction program are not currently being collected. The
5 measurement of disease for the total population exists only for diseases recorded on
6 death certificates. Mortality data do not provide information on the incidence or the
7 occurrence of disease, but only death from a specific disease. The incidence of cancer
8 is now being recorded in many states (ref CDC Cancer Registry Program), but the
9 cases cannot be tied to exposures because of the long latency of the disease. Still
10 more important, many health outcomes today are associated with environmental
11 exposures but records are not routinely kept on these diseases for the general
12 population. For example, asthma episodes, behavioral changes in children, and minor
13 neurologic deficits are known to be important effects of various exposures, but there is
14 no complete database regarding these outcomes that could serve to link these effects
15 to specific exposures. Testing for respiratory function, intelligence, hearing, and vision
16 is done on many segments of the population, but the information resources are
17 fragmented and have not been used to evaluate the effects of environmental change.
18 The available data linking environmental exposures to health outcomes have come
19 from special studies on limited and often selected populations. Further, these
20 outcomes are usually evaluated using surveys designed to detect trends, rather than to
21 look for potential causes. For many problems, only certain subpopulations are at risk,
22 either because of their susceptibility or special lifestyles that may increase exposures.
23 Little information is available specifically on these subsets to allow for evaluation.
24
25 Exposure: The databases on human exposure have often been limited by
26 infrequent national sampling sources, e.g., the Center for Disease Control's National
27 Health and Human Nutrition Examination Survey (NHANES), and the Total Exposure
28 Assessment Methodology (TEAM) study (EPA, 1987). States have far more data
29 available but these data often are not in electronic format and available for easy use.
30 Air and water monitoring data are only indirect measures of exposure and reflect only a
31 limited number of monitoring stations. In some situations, only non-compliance data
32 are recorded. This practice limits the usefulness of the information for modeling
33 exposures of individuals over space and time. Research to develop a better
34 understanding of the relationships between centrally located monitoring sites and the
35 actual human exposures could make some of these data more useful. Ultimately,
7-18
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 however, more direct monitoring of human exposures will be required to demonstrate
2 the effectiveness of various risk management strategies.
3
4 Measurements of pollutants in environmental media are less intrusive than
5 human exposure measurements. However, the knowledge needed to link exposure of
6 individuals to measures of the pollutant in various environmental media has not been
7 well developed. For example, we have monitored outdoor air for many years.
8 However, to use such data for exposure assessment, we must have information on
9 time-activity patterns and the penetration of outdoor air pollutants into buildings since
10 most of the population spends about 90% of the time indoors. In addition, exposures
11 to indoor pollutant sources (e.g., environmental tobacco smoke) will add to those from
12 outdoor air, and will also contribute to adverse health effects. Similarly, measurements
13 of a pollutant in food supplies may not capture changes, both losses and gains, that
14 occur during food preparation in homes and restaurants. Ultimately, considerably more
15 measurement of exposures will be required, as well as measurement of changes in
16 exposure over time, in order to meet the challenges of performance evaluation.
17
In short, the concept of report cards will challenge assessors to make the most
19 out of the data that are available and, perhaps, to re-define and re-design the systems
20 in place to gather environmental data. The goal is to answer questions that people
21 want to have answered.
22
23 7.5 Summary and Recommendations
24
25 The extent of environmental data collected (e.g., on contaminant levels in
26 various environmental media, on tissue burdens, on human exposures, and disease) is
27 expansive and holds the promise of improved measures to generate effective
28 Environmental Decision Report Cards (EDRCs). At the same time, in order to achieve
29 this promise, the current information reporting systems would benefit from additional
30 development and, in some cases, restructuring. An improved performance evaluation
31 system will be characterized by clear articulation of objectives during the Problem
32 Formulation (Phase I), careful selection of performance measures in Analysis and
33 Decision-Making (Phase II), and refinement, integration, and possible expansion of
34 existing data collection efforts in Implementation and Performance Evaluation (Phase
35 III).
7-19
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 In closing, the SAB would like to direct the Agency's attention to the promise and
2 challenge of environmental reporting as discussed in this report. That is, reporting in
3 the context of the IED framework would move the Agency further along the spectrum of
4 measures illustrated in Figure 7-1, increasing the attention to observable environmental
5 outcomes from specific management programs. In addition, the challenge of combining
6 information from a number of integrated decisions into an overall environmental report
7 card remains. With respect to this next generation of performance evaluation, the SAB
8 offers some initial observations:
9
10 a) Implementation of a more integrated form of environmental decision-making
11 will require an improved environmental report card system that provides reliable and
12 accurate information about the issues people care about: how healthy are the nation's
13 citizens? is the integrity of our ecological systems intact? is the quality of life getting
14 better or worse? In order to answer these questions effectively, government agencies
15 and the public need to develop a reporting system that focuses on the desired states of
16 human health and ecological condition, rather than program-specific objectives.
17
18 b) Such a system should build on the report cards devised to evaluate specific
19 integrated decisions, along with information from other sources, to report on the overall
20 human health and ecological condition.
21
22 c) In this regard, the Agency, in concert with others in the National Science and
23 Technology Council, can play a leading role in designing the improved reporting system
24 by developing a conceptual framework that: (1) defines ecological integrity and desired
25, human health results in terms of measurements that can be used at the local, state and
26 national levels, and (2) provides a systematic and transparent methodology for
27 aggregating these measurements into an overall picture of health and integrity.
28
29 d) Research on selected report card components, such as methods of accurately
30 measuring certain ecological characteristics at the landscape scale, methods of
31 identifying susceptible human subpopulations at risk for disease, and methods for
32 combining these into more aggregated measures of performance will be required and
33 should proceed concurrently with the development of an overall report card framework.
34
35 e) Specifically, current data collection and analysis systems should be upgraded
7-20
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 so that information can be assembled in a format that relates directly to the conceptual
2 framework, as well as the EDRC, and that lends itself to aggregation along a variety of
3 geographical boundaries.
4
5 f) The Agency should provide a results-based environmental report card regularly
6 to the public and to policymakers, and use such a report card — along with specific
7 EDRCs — as a tool to support re-evaluation of program implementation within the
8 Agency.
9
7-21
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 7.6 References Cited
2
3 Committee on Environment and Natural Resources. 1997. Integrating the Nation's
4 Environmental Monitoring and Research Networks and Programs: A Proposed
5 Framework. National Science and Technology Council, Executive Office of the
6 President, Washington, DC.
7
8 Enterprise for the Environment. 1997. The Environmental Protection System in
9 Transition: Toward a More Desirable Future.
10
11 Environment Canada and U.S. Environmental Protection Agency. 1997. State of the
12 Great Lakes: 1997-The Year of the Nearshore. Great Lakes National Program
13 Office, Chicago, IL (EPA 905-R-97-013).
14
15 National Research Council. 1996. Understanding Risk: Informing Decisions in a
16 Democratic Society. National Academy Press, Washington, DC.
17
18 Science Advisory Board. 1997. Review of the Environmental Goals for America: With
19 Milestones for 2005. (EPA-SAB-EC-97-007)
20
21 U.S. Department of Health and Human Services. 1990. Healthy People 2000. Office
22 of Disease Prevention and Health Promotion, Washington, DC.
23
24 U.S. Department of Health and Human Services. 1995. Healthy People 2000
25 Midcourse Review and 1995 Revisions. Office of Disease Prevention and Health
26 Promotion, Washington, DC.
27
28 U.S. Environmental Protection Agency. 1987. The Total Exposure Assessment
29 Methodology (TEAM) Study: Summary and Analysis: Volume 1. U.S. EPA Office
30 of Research and Development, Washington, DC, EPA/600/6-87/002a, June
31 1987.
32
33 U.S. Environmental Protection Agency. 1996. Environmental Goals for America: With
34 Milestones for 2005 (December 1996 draft).
35
7-22
-------
PART VI CONCLUSIONS AND RECOMMENDATIONS
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
i CHAPTER 8: CONCLUSIONS AND RECOMMENDATIONS
2
3
4 Over the course of the Integrated Risk Project, the Steering Committee and the
5 project subcommittees came to a number of conclusions about the need for and
6 benefits of improved methods of assessing and managing aggregate risks, broadening
7 benefit/cost analyses to include difficult-to-monetize values, explicitly considering the
8 role of values in decision-making, and so forth. The preceding chapters include
9 suggestions and recommendations for specific improvements in the methodologies and
10 approaches that underlie the IED framework. In addition, however, the Steering
11 Committee came to a number of broad conclusions as a result of the discussions and
12 thinking that led to the proposed IED framework itself. These overall conclusions and
13 recommendations, which are the focus of a companion document, Integrated
14 Environmental Decision-making in the 21st Century: Summary Recommendations, are
15 summarized below:
16
17 Recommendation 1: EPA should accelerate the transition to integrated,
outcomes-based environmental protection and apply the an integrated
i* environmental decision-making framework in selected cases, while
20 maintaining the safeguards afforded by the current system.
21
22 Previous environmental management approaches, both regulatory and non-
23 regulatory, have resulted in substantial improvements in human health and ecological
24 condition, particularly in regard to chemical risks. While maintaining the current
25 regulatory and management structure, the IED framework offers an approach for going
26 beyond current protections by taking a truly integrated look at risks, opportunities for
27 risk reduction, and the consequences of addressing (or failing to address) those risks.
28 The focus of this next generation of environmental decision-making should be on
29 environmental results; that is, on demonstrable outcomes (improvements) in the
30 environment resulting from integrated action. In order to further incorporate integrated
31 environmental decision-making into EPA programs, the Agency should select a few test
32 cases where the IED framework can be applied explicitly. Such test cases offer the
33 best means for demonstrating the potential of the IED framework, identifying needed
34 improvements in its underlying methodologies, and realizing its benefits with real-world
35 problems.
8-1
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1
2 Recommendation 2: Because science plays a critical role in protecting the
3 environment, EPA should commit the resources necessary to expand the
4 scientific foundation for integrated decision-making and outcomes-based
5 environmental management.
6
7 An important theme of the Integrated Risk Project is the need to integrate
8 scientific and technical information on risks and risk reduction opportunities with
9 information on societal values, aspirations, and goals. The call for greater inclusion of
10 multiple disciplines and points of view, however, must not obscure the fact that science
11 and scientific methods are critical to sound environmental decision-making. Science
12 has a unique and critical role to play in protection of the environment. It is through
13 scientific investigations that many environmental problems are first discovered (e.g., the
14 depletion of stratospheric ozone, hypotheses about environmental endocrine
15 disrupters). Science also is instrumental in developing, testing, and evaluating risk
16 reduction options, and social sciences offer techniques for assessing societal
17 preferences and wants. Implementation of the 1ED framework, particularly its
18 application to multiple-stressor problems, will in many cases require new science, both
19 empirical and theoretical. In order to gain the greatest benefit from the IED approach,
20 the Agency will need to invest in research designed to support integrated risk
21 assessment and management.
22
23 Recommendation 3: EPA should apply and encourage the broader use of
24 risk comparison methodologies, such as those described in this document,
25 that clearly identify how scientific information and judgment are
26 incorporated into risk comparisons.
27
28 Scientific information on risk, such as quantified risk assessments and
29 scientifically demonstrated relationships between stressors and effects, provides the
30 essential basis for making objective risk comparisons. However, desired scientific
31 information often is incomplete or absent, and scientists have to use their best
32 professional judgment to bridge important gaps in the data. In other cases, scientific
33 information and analysis, by themselves, are not sufficient for comparing risks; e.g.,
34 when comparing human healths risks vs. ecosystem risks or cancer risk in adults vs.
35 neurologic risk to children. In these instances, it is public values that come into play in
36 making the comparisons. Thus, methodologies used to compare environmental risks-
8-2
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
whether human health risks, ecosystem risks, or both-should incorporate not only the
2 most up-to-date scientific information, but also should identify explicitly where
3 professional judgment and values have influenced the results. The SAB has developed
4 two prototype methodologies for risk comparisons that meet these criteria. These
5 approaches should be points of departure for the Agency as it seeks to develop and
6 use science-based methodologies for comparing relative environmental risks.
7
8 Recommendation 4: EPA should use a broader range of risk reduction
9 options in combination to manage environmental risks.
10
11 In 1990, the SAB recommended in Reducing fl/sfr that the nation make greater
12 use of all the tools, including market forces, information, and product specifications,
13 available to reduce risk. Now, almost a decade later, many of those tools are being
14 used to a greater extent than ever before. The challenge for EPA is not only to expand
15 the use of those various tools, but also to use them in creative, coordinated ways to
16 reduce multiple risks to multiple receptors in communities and ecosystems across the
17 country. The SAB has developed a risk reduction options methodology that can be
1 ° applied fruitfully to single stressors. However, the methodology should be especially
useful for identifying multi-dimensional strategies to control complex environmental
20 problems involving many sources, stressors, and receptors.
21
22 Recommendation 5: When evaluating risk reduction options, EPA should
23 weigh the full range of advantages and disadvantages, both those
24 measured in dollars as costs and benefits and those for which there may
25 not be a comprehensive dollar measure, such as sustainability and equity.
26
27 The valuation framework that underlies environmental decisions is the simple
28 formulation of whether the gains that accrue from protective actions are worth what is
29 given up to attain them. There is a subsidiary question: would other possible actions be
30 preferable? This question highlights the importance of taking all effects, including long-
31 term effects, into account. It also raises the sometimes hidden issue of how people
32 value different aspects of the environment. Some of society's environmental values
33 can be measured directly in monetizable terms, and others can be inferred and
34 translated into monetizable terms with some confidence. But other things that people
35 value, such as sustainability and equity, often may be expressed only qualitatively, yet
36 are of no less importance for that reason. As a part of this project, the SAB has
8-3
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 described the applicability and limitations of the benefit/cost framework, and suggested
2 areas where new approaches to characterizing values are essential. The SAB also has
3 attempted to define better the full range of relevant questions that must be considered
4 in ecological valuation.
5
6 Recommendation 6: EPA should make fuller use of the scientific methods
7 available to characterize public values and incorporate those values into
8 goal-setting and decision-making.
9
10 Community and national values have been and will continue to be a primary
11 driver of community-level and national efforts to protect the environment. However,
12 values usually are not weighed transparently in the decision-making process. Rather,
13 they are usually implicit in the judgments made by decision-makers. Thus, they
14 influence decisions in ways that are not clear to, or renewable by, the public.
15
16 Because public values undoubtedly help shape environmental decisions, it is
17 important to understand and document their role in and influence on decision-making.
18 It is also important to elicit public values systematically, differentiate values from
19 technical information as a part of decision-making, and include their effects on
20 decisions as part of the public record. Individual and community values should be
21 solicited systematically by social scientists and other appropriately trained individuals.
22 The deliberative processes that are used in arriving at decisions should involve
23 professionals trained in fields like consensus-building and dispute resolution. EPA
24 should make more extensive use of existing expertise in the areas of behavioral
25 science and decision logic so that a more complete representation of community values
26 is incorporated into the Agency's decisions.
27
28 Recommendation 7: EPA should identify, collect, and disseminate
29 scientifically-based environmental metrics organized in new ways to
30 support a more integrated approach to managing environmental risk.
31
32 The transition to and effectiveness of integrated, outcomes-based environmental
33 protection will depend to a large extent on the availability and utilization of appropriate
34 information in the area of exposure, human health, ecological health, and quality of life.
35 Current information collection efforts, however, often are insufficient, inadequately
36 organized, or focused on inappropriate endpoints. In the area of human health, for
8-4
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
example, data on non-fatal outcomes of environmental exposures, such as asthma, are
2 not being collected except in a rudimentary manner or at certain sites. Only mortality
3 information is collected systematically at all locations and can be related to some
4 limited information on selected exposures. With regard to ecological data, more
5 comprehensive information is needed on the current status of ecosystems such as
6 wetlands, lakes, forests, and grasslands. This information should include the extent to
7 which each ecosystem is exposed to and affected by non-chemical stressors such as
8 habitat conversion, habitat fragmentation, and invasions of exotic species.
9
10 The challenge facing EPA and the nation is not only one of collecting the right
11 environmental data, but of finding new ways to manage and use that data. Working
12 with other federal and states agencies, EPA should take a leadership role in:
13 identifying critical environmental data gaps, including data on exposures and health and
14 ecological outcomes; integrating the largely fragmented data collection efforts already
15 underway; and disseminating integrated environmental information to decision-makers
16 and the public.
17
i g Recommendation 8: EPA should develop a system of "report cards" to
organize and disseminate information on the status of ecological and
20 human health and the quality of life in order to assess the effectiveness of
2i its environmental decisions and to guide future environmental
22 management.
23
24 One of the most valuable uses of environmental data is to measure the results of
25 the actions society takes to reduce environmental risk. However, even if such data did
26" exist, through a vigorous implementation of Recommendation 7, there would still be a
27 need to develop widely accepted and commonly used methods for evaluating a) the
28 state of our environment and b) the success of national environmental protection efforts
29 .. where success is measured in terms of demonstrable outcomes in the health of
30 humans and ecosystems. Regarding the state of the environment, EPA should work
31 with federal and state entities to develop indicators of ecological and human health
32 conditions in our country. Such a environmental report card would be readily
33 comprehensible to the public and policy makers alike, serving to galvanize them on
34 action toward specific, measurable goals. The concept of a report card should also be
35 applied to evaluating the impact and effectiveness of individual decisions made under
36 the IED methodology; i.e., an environmental decision report card (EDRC).
8-5
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
1 Reporting on measures of progress towards ecological, human health, and
2 quality of life goals will provide a number of benefits. First, the exercise of defining
3 performance measures and reporting on them will bring more focus and discipline to
4 the Agency by expressing the relationship between investments (measured in time,
5 money, or information) and environmental results. Second, shorter term measures of
6 progress (including process measures or measures of stressor reductions) will be
7 useful for accountability and course correction. Third, longer term measures of
8 progress (such as improvements in overall human and ecosystem health and quality of
9 life) will be most helpful in determining whether goals are being met, whether further
10 actions are needed to control well-recognized stressors, or whether new actions are
11 needed to control new stressors.
12
13 To strengthen their credibility and utility, environmental report cards should
14 contain information derived from objective measurements, be transparent and clearly
15 documented, and provide integrated information on progress towards multiple inter-
16 related environmental goals.
17
18 Recommendation 9: EPA should expand and develop new collaborative
19 working relationships with other federal and non-federal governmental
20 agencies and others who also will be involved in integrated environmental
21 decision-making.
22
23 Inherent in the I ED framework is the idea that problem formulation and decision-
24 making must match in scale and location. In some cases, decisions may be most
25 effective when local or state governments play the primary role. In others, coordinated
25 action across several levels of government or among a number of state and/or local
27 governments will be required. In still others, the current emphasis on centralized
28 decision-making may be the preferred approach. In any event, integrated thinking
29 about environmental problems will tend to drive decision-making to the agency or level
30 of government where decisions are most appropriately made. As a consequence,
31 EPA's role will evolve to one in which the depth of control gives way to broader
32 involvement in partnership with other agencies. EPA will continue to be responsible for
33 implementing and enforcing federal environmental laws and statutes, conducting
34 environmental research and development, and conducting stressor-specific risk
35 assessments. At the same time, the Agency can exert national leadership to bring
8-6
-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
i together appropriate agencies and stakeholders to explore integrated approaches to
2 environmental problems.
3
4 Recommendation 10: EPA should aggressively explore options for
5 reducing risks from significant stressors that currently are addressed
6 inadequately by the nation's environmental institutions.
7
8 Over the course of the Integrated Risk Project, it became clear to the SAB that a
9 number of important human health and ecological risks are not being addressed
10 adequately by the nation's environmental institutions. In many cases this is because
11 risk management responsibility is not clearly assigned to any one government entity, or
12 is scattered over many agencies and/or levels or government. This fragmented
13 approach results in uncoordinated and incomplete efforts to identify cause-and-effect
14 linkages and to manage those risks. With regard to ecological risks, the SAB has
15 concluded that many of the highest ranking risks (e.g., hydrologic alterations, harvesting
16 of living marine resources, habitat conversion, climate change, and introduction of
17 exotic species) are associated with physical and biological, rather than chemical,
1? stressors, which do not fall clearly within the purview of any single federal agency.
1 Important human health risks that remain unaddressed include those for which the
20 environmental exposure link is suspected but not certain (e.g., asthma, brain cancer,
21 and non-Hodgkins lymphoma) and risks associated with susceptible and/or
22 compromised human populations.
23
24 To control many of these inadequately addressed risks will require the kind of
25 integrated decision-making envisioned in this report. It will also required a new kind of
26 integrated institutional leadership. When EPA determines that serious risks are not
27 being addressed effectively by existing environmental institutions or decision-making
28 systems, the Agency has a responsibility to inform the public about those risks and
29 bring together the appropriate federal, state, and local agencies to address them.
8-7
------- |