Willard W. Smith. Jr.
Office of Policy and Management
U.S. Environmental Protection Agency
Rodney Lorang
Sobotka and Company, Inc.
Washington, DC
DRAFT RfPOR[
February, 1981
-------
TABLE OF CONTENTS
PAGE
Executive Summary S-l
Legislative Implications . . . S-2
Analytic Conclusions ....... S-3
I, Introduction 1-1
A. Background 1-2
A.1 The Need For Government Regulation 1-2
A.2 Strategies For Achieving Environmental Goals .... 1-2
Ambient-Based Strategies ... .... 1-4
Technology-Based Strategies ............. 1-5
Mixed Strategies 1-7
A.3 Selecting Approaches For Executing Strategies .... 1-8
Economic Efficiency 1-8
Cost-Effectiveness 1-9
Effectiveness 1-10
Feasibility 1-11
Equity 1-12
Other Considerations ..... 1-13
B. The Current Approach to Air Pollution Regulation . 1-13
B.l Regulation of Stationary Sources .... 1-14
Existing Plants .................... 1-14
New Plants and Modifications . 1-15
B.2 Regulation of Mobile Sources ............ 1-16
C. Attributes of the Current Approach 1-18
C Stationary Sources 1-18
Ambient 'Standards 1-18
Technology-Based Standards 1-22
C.2 Mobile Sources 1-26
1 Introduction to Economic Approaches II-1
A. Charge-Based Approaches II-2
A. 1 Stand-Alone Charges . 11-2
A.2 Charges as a Supplement When Standards Are Specified. 11-12
-------
PAGE
B. Trading Approaches 11-15
B,1 Overview of Trading Approaches . ..... 11-15
Marketable Permits . 11-16
Controlled Trading . 11-17
Averaging 11-18
B»2 Operation of Trading Markets 11-18
Market Transactions 11-18
Banking 11-21
3.3 Attributes of Trading Approaches 11-21
ill Designing and Selecting Specific Economic Approaches . . . III-l
A. Choosing an Economic Approach III-l
A.l Charges Set Equal to Marginal Social Costs ...... III-l
A,2 Charges and Trading to Pursue Targets III-2
B. Specific Charge-Based Approaches ........... XII-4
B.l Stand-Alone Charges III-4
Rate-Setting Principle 111—4
Application of Rate Ill-5
Sources and Emissions Affected ............ 111-6
B.2 Supplemental Charges III-6
Supplemental Charges that Induce Compliance ...... III-7
Supplemental Charges that Provide a "Safety Valve" . . III-7
Combining Trading with Supplemental Charges ...... II1-8
B.3 Use of Fee Proceeds 111-9
C. Specific Trading Approaches .............. 111-10
C.l Review of Controlled Trading and Marketable Permits . . 111-10
C.2 Variations in Controlled Trading and Permit
Approaches 111—12
Market Involvement by the Control Agency ....... 111-14
Establishing the Equivalence of Emissions ....... 111-16
Restrictions on Trading ................ III-l?
Restrictions on Tradeable Reductions 111-17
Restrictions on the Use of Credits ........ Ill-IS
Allocation of Entitlements and Assessments of
Additional Control Obligations ........... 111-20
Discretionary Allocation of Entitlements ..... 111-20
- il -
-------
PAGE
Allocation of antitleaents by Pre-established
Rules 111-21
Allocation of Entitlements by Auction ...... 111-24
Allocation of Additional Future Control 111-25
Obligations .
1 EPA On-Going Experiences and Study Capsule Summaries . , , , IV-1
A. Brief Summaries of Study Results IV-1
A. 1 Stationary Source Studies . . . IV-1
A,2 Mobile Source Studies IV-5
B» EPA On-Going Experiences With Incentives IV-6
B.i Stationary Source Controlled Trading IV-7
Status of Programs . IV-7
Current Efforts Regarding Implemented Controlled
Trading Activities IV-11
Controlled Trading Plans Under Current Legislation , » IV-12
B.2 Mobile Source Averaging IV-13
B.3 Supplemental Charges IV-14
Noncompliance Penalties IV-14
Nonconformance Penalties IV-15
Economically-Based Civil Fines ...... IV-17
B.4 Other On-Going Stationary Source Incentive Efforts . . . IV-17
Reducing Chlorofluorocarbon Emissions . IV-I8
N'SPS Offsets and New Source Bubbles IV-18
PSD Increment Allocation IV-20
Emissions Density Zoning IV-21
Nonattainment Strategies ......... IV-22
Mobile Source Strategies for the 1980"s XY-24
V. Analytic Conclusions and Legislative Proposals . V-l
A. Analytic Conclusions . . • V-l
B, Legislative Implications ¥-9
Appendix - Specific Study Descriptions A-l
1. The Rand Study of Chlorofluorocarbon Emissions , , » , , A-l
2. The Nichols Study of Benzene Emissions A-2
3. The Mathtech/EPA N0X Study ............... A-5
-------
PAGE
4» The Meta Systems Hydrocarbon Study A-9
5. The Putnam, Hayes and Bartlett Marketable
Permits Study . . . . A-l 1
8. The Repetto PSD Study A-l L
7. The 1CF Study of NSPS Offsets A-14
3. The Temple, Barker and Sloarte Study of Noncompliance A-15
Penalties .
9. The TCS Mobile Source Study A-16
iu The Sobotka Study of ISP Averaging for Diesel
Vehicles 4-18
1i The Policy Planning and Evaluation N0X Studies » , . A-20
Bibliography
- iv -
-------
EXECUTIVE SUMMARY
Traditional coramatid-and-conttol approaches to air pollution regulation
typically require companies to use specified pollution control technologies
or require classes of emissions sources to meet performance standards. Such
approaches ration each source's use of the air as a receiver for pollutants
in order to meet National Ambient Air Quality Standards (N'AAQS) or other air
quality goals.
These approaches provide little flexibility to companies in determining
the degree to or the way in which pollution control requirements are met. They
provide no economic incentive to companies to install controls quickly or to
go beyond regulatory requirements. They provide little incentive to innovation
which could decrease long term pollution control costs or make more effective
control possible. These characteristics probably result in a regulatory program
that costs more and in the long run may do less to protect the environment than
might otherwise be possible.
Economic incentives can overcome some of the limitations of traditional
regulatory approaches by providing the flexibility that is missing with comma rtd-
and-control approaches and giving sources incentives to use that flexibility.
Because polluters have greater knowledge about their cm control costs, plans
for development and possibilities for technological advance than government
officials, they can use this knowledge to respond to pollution control require-
ments in a cost-effective way.
This report responds to the requirement in Section 405(g) of the Clean
Air Act (CAA), as Amended, for a general report to Congress on economic incen-
tive approaches to air pollution control. Section 405 of the CAA directs EPA
and the Council of Economic Advisors (CEA) to identify, study and assess eco-
nomic measures which could*.
o strengthen the effectiveness of existing regulatory requirements by
encouraging compliance;
o provide incentives to abate pollution to a greater degree than is
required by existing regulations; and
-------
S-2
o serve as a primary means to control air pollution problems that are
not addressed by existing regulations,
A. report to Congress analyzing the use of economic incentives to control sta-
tionary source MQX emissions was completed in early 1981 in accord with Sec-
tion, 405(f) of the Glean Air Act Amendments.
This Executive Summary discusses the implications of any desirable changes
to existing legislation, and then presents the major analytic conclusions
reached by EPA and CEA,
Legislative Implications
-------
5-3
Analytic Goaelttsiotis1
o REPLACEMENT OF TOE IN-PLACE REGULATORY SYSTEM WITH STAND-ALONE EMISSION'S
CHARGES OR MARKETABLE PERMITS IS SOT NECESSARY. ALL SIGNIFICANT BENEFITS
OF ECONOMIC INCENTIVES CAN BE ATTAINED THROUGH CAREFUL SELECTION AND
DESIGN OF CHARGE OR TRADING APPROACHES WHICH SUPPLEMENT THE CURRENT
REGULATORY SYSTEM. The current regulatory system has accomplished a great
deal, and it would be disruptive and confusing to replace this system.
If the full potential of supplemental charges and "controlled trading"
programs (the buying and selling of obligations to reduce emissions or of
credits for abatement which was not required) could be achieved, no sig-
nificant incremental benefits would be available from use of stand-alone
charges or a "pure" system of marketable permits. However, stand-alone
charges and permits would offer a fresh start that may be important to
achieving the potential benefits of economic approaches in certain unre-
gulated areas.
0 WITH ITS OFFSETt BU3BLE, AND EMISSIONS BANKING POLICIES» EPA HAS MADE
SIGNIFICANT STRIDES TOWARDS INCORPORATING MARKET INCENTIVES IN OS AIR
PROGRAM. In each of these policies, polluters have incentives to seek
out sources of low-cost emission reductions. The result will be a more
cost-effective pollution abatement program. However, these policies have
not yet been fully implemented by the states. One reason has been the
cumbersome administrative procedures and certain restrictions associated
with these policies. EPA has taken steps to streamline the process and
to remove restrictions. EPA is continuing to explore additional improve-
ments.
o NONCOMPLIANCE PENALTIES SIGNIFICANTLY STRENGTHEN THE EFFECTIVENESS OF
CURRENT REGULATORY REQUIREMENTS BY ENCOURAGING COMPLIANCE. Noncompliance
penalties are a special form of economic incentive intended to neutralize
the economic benefits which would otherwise accrue to stationary sources
that delay compliance with regulatory requirements. EPA has implemented
these penalties, and EPA and CEA expect them to be highly effective.
1 Only the most significant analytic conclusions are reported here, A
more complete set of conclusions is given in Chapter V.
-------
S-4
IT IS DIFFICULT TO DIRECTLY PURSUE ECONOMICALLY EFFICIENCI OUTCOMES. Eco-
nomic efficiency requires thac damages from emissions and control costs be
equalized at the margin. In the absence of an ability to determine damages
accurately, efficient outcomes cannot be identified. Thus the operational
test of efficiency is that of cost-effectiveness, i.e., achieving an
environmental objective at least cost. This is the standard of efficiency
pursued by EPA in its consideration of incentive approaches. Any approach
thac is economically efficient will also be cost-effective.
BOTH SUPPLEMENTAL CHARGES AND "CONTROLLED TRADING" CAN PROVIDE INCENTIVES
TO ABATE POLLUTION TO A GREATER DEGREE THAN IS REQUIRED BY CURRENT STAN-
DARDS. Supplemental charges imposed on some or all residual emissions
can be as effective as stand-alone charges in providing a continuous
incentive to sources to go beyond current standards and to find innova-
tive ways to reduce control costs. Controlled trading provides identical
incentives for each source and can lead to a reduction in total emissions
as emission reduction credits are created for later use or sale.
STAND-ALONE CHARGES AM? UNRESTRICTED TRADING OF MARKETABLE PERMITS CAM
BEST SERVE AS A PRIMARY MEANS TO CONTROL THOSE AIR POLLUTION PROBLEMS
WHICH ARE SOT ADDRESSED OR DEALT WITH SUCCESSFULLY BY CURRENT REGULATIONS.
Air pollution problems which are not being addressed by the current regu-
lations should receive careful scrutiny as candidates for implementing
emission charges or marketable permits. In these cases, economic incen-
tives cannot disrupt an existing program, and may provide a better ap-
proach to as yet unsolved problems.
II GENERAL, A MARKETABLE PERMIT SYSTEM IS PREFERABLE TO A CHARGE SYSTEM
FOR ATTAINING AM) MAINTAINING AN AMBIENT STANDARD. We base this con-
clusion on the following findings of our comparative analysis of these
two economic policies:
— Under a charge system the quantity of pollutants emitted depends upon
the response of sources to the costs imposed by a charge. Thus,
under a charge system, administering agency has greater certainty in
the short run that standards will be met than under a permit systems,
unless charges are initially set very high.
-------
— To implement an efficient charge system to attain or maintain an am-
bient standard, the adtainistracing agency must acquire information
about sources* control costs. This is a difficult and expensive
undertaking if costs are to be determined accurately. Under a permit
system,.the quantity of emissions is -fixed by the quantity of permits
issued, so the agency does not need detailed source-by-source cost
data. Cost data would still useful to design permit systems that
operate smoothly. These data will be revealed "by sources as they buy
and sell permits from one another.
— A marketable permit system self-adjusts to inflation and growth. A
charge system requires that the agency make periodic adjustments to
these factors, adjustments which depend upon uncertain and perhaps
expensive data. Furthermore, frequent changing of charge rates may
undercut the credibility of a charge system.
— A marketable permit system is administratively and legally similar
to permit programs now operated under regulatory control programs.
This means that it could be administered alongside of existing regu-
latory programs more easily than could a charge system, and would
probably encounter less opposition from vested interests. A market-
able permit system is also similar to the Offset and Bubble Policies
currently in force.
0 CHARGE-BASED APPROACHES ARE MOST ATTRACTIVE WHERE TRADITIONAL APPROACHES
WILL HAVE A DIFFICULT TIME MEETING AIR QUALITY STANDARDS ON A SCHEDULE,
WHERE THERE IS A FAIRLY LONG LAG BETWEEN EMISSIONS AND SUBSEQUENT ENVI-
RONMENTAL EFFECTS, OR WHERE MAXIMUM FEASIBLE CONTROL EFFORT IS SOUGHT.
Where attainment of air quality goals on schedule is in doubt, or there
is no grave concern over short-term damages during the period of necessary
adjustments in charge rates, charges can provide a means to induce cost-
-effective control steps with very little information on hand. Where maxi-
mum feasible efforts are sought, charges can induce control efforts that
would not have been feasible to identify In advance and efforts that
could not have been induced through specific enforceable control obliga-
tions.
-------
8-6
THE USEFULNESS AND FEASIBILITY OF A MARKETABLE PERMIT APPROACH DEPENDS
UPON WHETHER AN ALTERNATIVE CONTROLLED TRADING APPROACH COULD APPROXIMATE
THE CHARACTERISTICS OF PERMITS, Marketable peraits offer little incre-
mental benefit, if controlled trading works well. This requires that
restrictions on controlled trading be no more extensive than restriction
on the sale and use of marketable permits, and that states have the autho-
rity and will to allocate control obligations in a manner that will assure
attainment.
-------
1
INTRODUCTION
Traditional "command-and-control" approaches to air pollution regulation
typically require sources of emissions (plants, automobiles, etc.) to use pol-
lution control technologies specified in the regulations, or require classes
of emissions sources to meet performance standards. These approaches provide
little flexibility to sources in determining the degree or the way in which
pollution control requirements are met. While there may exist an incentive to
minimize the cost of a given level of control, they provide no economic incen-
tive to install controls quickly or to go beyond regulatory requirements.
They provide little incentive to innovation which could decrease long-term
pollution control costs or make more effective control possible. These charac-
teristics result in a regulatory program that costs more, and does less to
protect the environment, than might otherwise be possible.
Economic incentives can overcome some of the limitations of traditional
regulatory approaches by providing the flexibility that is missing with command-
and-control approaches and by letting the market reward companies which use
that flexibility intelligently. EPA has already implemented some economic
approaches and has studied or is considering several aore. This report examines
economic approaches that can improve the efficiency and effectiveness of the
existing environmental program. In addition, existing provisions of the Glean
Air Act (CAA) or of current EPA regulations and policies that reduce economic
efficiency are noted.
This chapter briefly describes the existing environmental program for
controlling air pollution. Chapter II introduces the economic approaches that
ate discussed in this report. Chapter ill provides a discussion of the key
considerations that are important for selecting and designing economic ap-
proaches. Chapter IV describes the economic approaches that EPA is now
using or considering, and summarizes the results of studies of the feasi-
bility of using economic approaches in other cases. The Appendix describes the
studies in greater detail. Chapter V discusses EPA's conclusions regarding
the use of economic approaches, and identifies the steps that are prohibited
or inhibited by the current provisions of the Clean Air Act.
-------
A, BACKGROUND
A i THE NEED FOE GOVERNMENT REGULATION
The need for environmental regulation arises because companies do not pay
for che damages that result from the pollution they emit. Since in the ab-
sence of regulation the ability to pollute is "free," there is no incentive
for companies to reduce emissions. Those companies that pay to reduce emis-
sions are at a competitive disadvantage conpared to companies that do not.
Yet pollution causes damages to society in the form of health problems, damage
to property, damage to the environment and reduced esthetic values. Society
would be better off if polluters controlled their emissions, but in the absence
of some form of regulation, few if any will do so.
Environmental regulations reduce the damages incurred by society by re-
quiring polluters to reduce emissions. The costs of control are then reflected
in the prices that emitters charge for their products, so that the cor.suoers
that benefit from the products indirectly pay for the controls.
A. 2 STRATEGIES FOR ACHIEVING ENVIRONMENTAL GOALS
Ideally, decisions about the stringency of environmental regulations for
each source should be based on the relationship of control costs to the value
of the damages avoided by the additional control effort. This holds true
regardless of whether the air quality goal is to achieve acceptable levels of
health impacts, to provide for a margin of growth or to control degradation of
air quality. Unfortunately, control strategies cannot be directly based on
cost-benefit considerations because damages are usually very difficult to mea-
sure in physical terms, much less In economic terms, and are difficult to assign
to particular sources.1 As an illustration, Figure 1-1 is a schematic of the
steps (and variables) that must be taken Into account in translating emissions
into estimates of physical damages. Given the uncertainty associated with
I For a discussion of the difficulties in measuring damages and valuing
benefits see, for example, Environmental Quality ~ 1979, pp, 648-6 "> 2, and
Environmental Quality - 1975,"ppT~333-347 and"~pp. 496-519. These reports are,
respectively, the Tenth and Fifth Annual Reports of the Council on Environ-
mental Quality.
-------
1-3
Figure 1-1
Schematic of Steps in Translating
Emissions into Damages
policy variables
SO2 emissions
I AIR QUALITY MODELLING
terrain
catalysts
in atmosphere
sunlight
humidity
~-Dispersion ¦
Transformation
to SO^,
Transport
aetaorological
conditions
[Outdoor ambient
1 levels
ESTIMATION OF PHYSICAL DAMAGES
-Visibility impacts
-Acid rain impacts
-Crop and other
vegetation damage
-Property value
impacts
-Corrosion and
soiling
-Recreation effects
I Exposure
impact from
other sources
Indoor ambient
levels
population
distribution
organism
dose "to
I
physiological
response
~ •
I MONETIZE DAMAGES
[Health effects, |
[whole population!*
!Health effects,i
"Imost sensitive I
I individuals I
! Attach money values to 1
[health effects and other!
[ impacts 1
ESTABLISH REGULATIONS
I Relate monetized damages
I(benefits) to control costs.
I Set policy variables
i at cost-effective levels.
1 Set ambient standards 1
I to protect health. !
jSet policy variables
I to achieve standards.!
-------
1-4
benefits analysis, it is understandable that the Clean Air Act has not based
protection of the environment and health on direct cost-benefit evaluations.
Regulatory strategies have been based on considerations that can be more
readily identified and acted upon than cost-benefit considerations. The stra-
tegies that have been used to date to achieve air quality goals are based on
achieving specified ambient air quality levels or reducing emissions to levels
determined by technological considerations. This does not mean chat these sur-
rogate strategies can not be used to control pollution to socially desirable
levels —¦ the levels that would have been required using cost-benefit analysis,1
Rather, it means that Congress has made the judgment that these strategies were
the most desirable at the time for achieving the levels of control that were
deemed socially desirable. EPA's responsibility has been to implement these
legislatively-mandated strategies In a manner that is consistent with legisla-
tive requirements and {where permitted) at lowest cost.
These ambient- and technology-based approaches are described briefly below,
Antbleat-Based Strategies
Specification of target ambient concentrations can be used as a surrogate
for a damages-related (cost-benefit) assessment of each source as a basis for
regulation. There are a number of difficulties in using ambient-based strate-
gies, but as illustrated in Figure 1-1, such strategies involve fewer uncer-
tainties than strategies based on source-by-source cost-benefit considerations.
This is primarily because ambient—based strategies can be implemented without
explictly monetizing the damages from emissions or assigning them to sources.
The pursuit of air quality goals through use of ambient standards is a
difficult undertaking where modeling projections of ambient concentration
levels are used as a basis for establishing control requirements. Accurate
1 For an excellent review of environmental benefits studies, see A. Myrick.
Freeman III, "The Benefits of Air and Water Pollution Control: A Review and
Synthesis of Recent Estimates," prepared for the Council on Environmental
Quality, December 1979, A summary of Freeman's study is contained on pp. 654-
655 of CEQ1s 1979 Annual Report.
-------
prediction of the impact of individual plants on ambient concentrations re-
quires complex modeling and substantial data. Moreover, air quality models
currently available require substantial judgment because they are not yet well
specified, they tend to oversimplify reality and, in most cases, the available
data are not ideal, la addition, political problems may arise if later verifi-
cation efforts by means of monitoring show controls were either too little or
too much to attain the ambient targets.
Typically, ambient targets are set at levels that are considered to be
"thresholds" that protect human health and welfare. Establishing thresholds
is difficult because the applicable data are limited, inconclusive or ambiguous.
Furthermore, for most regulated pollutants, a sharp line does not always exist
that separates benign from adverse conditions; differences in health impacts
may exist for different segments of the population, for short-term versus long-
term exposure, and for different combinations of pollutants. Because of these
difficulties and the implementation uncertainties discussed above, air quality
standards are targeted at the most vulnerable segments of the population and
are designed with a margin of safety.
Ambient concentrations cannot be used as a basis for regulation when the
¦target levels cannot be accurately determined or where the social goals are not
reflected in aabient targets. Target levels may not be accurately determined
where available information, indicates a health problem but is inadequate to
identify a "safe" level, if it exists, or where available modeling is inadequate
to link emissions to ambient concentrations, such as in the case of long-range
transport problems like acid rain. Target levels cannot be used to represent
some air quality goals such as slowing the degradation of air quality in areas
where ambient "standards" are being met. In these cases, technology-based
strategies can be used.
Technology-Based Strategies
In the absence of operative ambient concentration targets, technology is
used to provide the basis for determining how far companies should go in
reducing emissions.
-------
1-6
The direct relationship of technology-based standards to primary environ-
mental goals can be difficult to define or to quantify because technology-based
strategies are generally used precisely in those cases where environmental
goals are especially difficult to quantify. Nonetheless, just as ambient tar-
gets have an underlying rationale (i.e., health), technology-based standards
also have underlying rationales. For example;
o As the economy expands and the population grows, it is inevitable that
the potential for emissions from mobile and stationary sources will
increase, la the face of an increasingly difficult pollution problem,
it might be less costly and more effective in the long run to devise
and follow decision rules that require the installation of "best con-
trols" into the stock of capital, particularly where "best" pollution
controls can he incorporated during initial construction. As older
plants and mobile sources are retired and replaced, the overall pollu-
tion potential for a given level of economic activity will drop and
the long run pollution problem could be less severe than it would be
if new plants and mobile sources had not faced stringent emission re-
duction obligations. In fact, it is sometimes argued that the absence
of requirements of pollution control technology on new sources could
preclude the efficient installation of such equipment in the long run
if new environmental evidence later indicates the need for more strin-
gent controls.
o We have very uncertain knowledge about future growth patterns and
about long-run damages to the environment, some o£ which may not be
reversible. With our lack of knowledge, some overcontrol and added
expense may be more advisable than accepting risks which cannot be
reliably assessed. While more stringent ambient concentrations could
provide a "hedge" or margin of safety for protecting the nation's
health and welfare, this could involve unacceptable economic disruption,
A technology-based approach can tie the amount of "hedge" acquired to
the ability of new sources to incorporate stringent controls into their
production planning process.
o In areas chat are meeting ambient standards, the quality of the air
will degrade as economic and population growth results in new sources
-------
1-7
of emissions. The degradation of air quality can be slowed or coo-
trolled by requiring new sources to meet some level of control even
though it is not needed to attain ambient targets. In addition, such
requirements can "be used to ration, the air resource "between new sources
over a longer time period — otherwise there is a possibility that the
first new sources will use up the margin for growth. In this manner,
the ability of the air resource to accomodate growth will be prolonged,
and the need to obtain more controls from existing plants to accomodate
growth will be deferred.
Certainly congressional sympathy at the time the Clean Mr Act was passed
appeared to be with such arguments, as reflected in -the technology-based require-
ments for new sources. Moreover, less information and- studies regarding the
use of economic incentive approaches were available for consideration at the
time.
To determine whether technology-based requirements are "reasonable," a
combination of criteria have been weighed by regulatory agencies, depending
on the problem that is being addressed and statutory requirements. These
criteria include technical feasibility, cost, energy impacts and the perceived
seriousness of the pollution problem. Establishing technology-based standards
can be difficult and time-consuming since there are many classes of sources
and even within a class production facilities may be unique.
Mixed Strategies
Some strategies are based on a combination of ambient and technology
considerations.
Mixed strategies axe useful when ambient standards cannot be attained and
another method for determining and limiting the control obligations of sources
must be used. For example, as discussed later in this chapter, some areas have
failed to attain ambient standards for some pollutants, and may not be able to
attain the standards in the near tern. How far should existing sources be
required to go in further reducing their emissions? In this case, the Clean
Air Act requires that existing sources be required to reduce emissions to the
-------
1-8
extent determined by available technology — Reasonably Available Control
Technology.
Mixed strategies are also useful when ambient standards cannot be attained
with existing technology and the ambient targets are used as a 'basis to "force"
the development of adequate technology. For example, as discussed later in
this chapter, certain standards for mobile sources have been related to techno-
logy which is expected to be available and, in this manner, a clear target for
industrial efforts to develop the needed technology is provided.
A.3 SELECTING APPROACHES FOR EXECUTING STRATEGIES
Ambient- and technology-based strategies or any other environmental strat-
egy can be Implemented by using eomaani-amd-eoatrol (i.e., typical regulatory)
approaches, economic incentive approaches, or both. Selecting the approach
to be used requires balancing the key attributes (or operating characteristics)
of the different approaches. the key attributes are introduced below and
discussed in general terms. Comoand-and-control approaches currently in use
are discussed ia terms of these attributes at the end of this Chapter and
economic approaches are evaluated throughout the remainder of this report.
Economic Efficiency
Approaches are efficient if the cost of reducing emissions on the margin^
is equal to the damages that would result if the emissions were less inten-
sively controlled. Thus, it only makes sense to require that an additional ton
of emissions be abated up to the point where the damage caused by that ton is
greater than the cost to control that ton. An equivalent definition of effi-
ciency is that point at which the difference between total benefits and total
costs (B-C) is greatest.
1 "The margin" is the point where decisions are made about the last incre-
ment of control effort, and it is here that costs and benefits must be compared
to make determinations about economic efficiency. This concept is explained
more fully in Chapter II,
-------
1-9
In most cases, damages fro® emissions cannot be measured or precisely cal-
culated. However, even if damages are riot explicitly known, every regulatory
action involves an implicit assessment that the cost of the regulation does not
exceed its social value. To the extent that social costs are known, some ap-
proaches may be better able to assure efficiency than others. In the absence
of an ability to determine damages accurately, the operational economic goal is
that of cost-effectiveness, i.e., achieving an environmental objective at least
cost. Any approach chat is economically efficient will also be cost-effective.
Cost-Effactiveness
Environmental approaches are cost-effective if they achieve environmental
objectives at the least cost. That is, it is sore cost-effective to achieve a
redaction of 100 tons of emissions at a cost of SO.10 per pound than at a cost
of $0-50 per pound. Similarly, approaches are cost-effective if, for a given
cost, they are more effective in achieving environmental objectives. Thus,
over a given time frame it is more cost-effective to pay $20,000 to achieve a
reduction of 100 tons of emissions than to achieve a reduction of 20 tons of
emissions.
Different approaches to environmental control may be cost-effective at
different levels, for example, consider a plant that has five emission points,
each of which is emitting ten tons of pollution pet year, for a total of fifty
tons. Some control approaches may be cost-effective for the individual emis-
sion points within the plant — the five enisaion points may each be best con-
trolled to two tons per year at a cost of $200,000 each. The result will he
a total cost of one million dollars and total remaining emissions of ten tons
per year.. While this may be the least cost for controlling individual emis-
sions points, it may not be the least cost approach for reducing the overall
level of emissions for the entire plant to ten tons per year. Other approaches
may be cost-effective on a plant-wide basis, so that the total cost for a plant
to achieve a particular environmental goal is minimized. Thus, for example, it
aay be possible to control four of the emission points to one ton per year at
a cost of $225,000 each and the remaining emission point to six tons per year
at a cost of $50,000; the remaining emissions will still be ten tons per year,
but Che total cost will be $950,000. But even this plant-wide approach neglects
-------
I-10
opportunities to reduce total cost because it fails to include other plants in
the universe of opportunities to reduce costs.
Thus, to be "truly" cost-effective, approaches must be evaluated for all
sources and the overall environmental goal must be achieved at the least total
cost. In. the case of achieving target emission levels {or ambient targets),
this will occur when the marginal cost of controlling the last increments of
pollution (or ambient impacts) is the same for every regulated emission point
-- one would never pay $0,50 per pound to control emissions at one emissions
point when emissions could be further controlled at $0,10 per pound at another
emissions point.
Effectiveness
Effectiveness refers to the ability of an approach to achieve the desired
objectives. Approaches may differ in the extent to which objectives are
attained, in the certainty (or reliability) of attaining objectives, and in
timing. Environmental objectives include protecting health and welfare, pro-
tecting air quality by controlling degradation, encouraging the development
of new or less expensive ways to reduce emissions, obtaining more emissions
reductions than required by regulation, and reducing the states* ability to
use less stringent environmental regulations as a means o£ attracting new
industry.
Environmental objectives are not always apparant. They are often repre-
sented by the attainment of ambient standards or the installation of specified
pollution control technology. However, the objectives that underlie these
measures should be kept in mind when evaluating the effectiveness (and the
cost-effectiveness) of different approaches — the measure should not be mis-
taken for the objective itself. For example, a technology-based standard may
be used simply because it is not possible to specify appropriate ambient levels*
The underlying objective in this case is to achieve a specified essissions reduc-
tion, not to develop or install a particular technology. Thus, while one
approach might be superior in achieving the development of technology to reduce
emissions at a particular emission point, it might not be superior in achieving
the underlying objective, i.e., achieving a given level of emission reduction.
*
-------
I-l 1
The manner in which the environmental strategy is described can facilitate
or hinder the evaluation of effectiveness. Environmental strategies can be
correctly described in terms of either obligations to control or "entitlements"1
to emit. Both of these descriptions will be correct in every case, but one will
generally provide a clearer perspective. Consider a regulation that requires an
emission, reduction of 10 tons from a plant that is emitting 100 tons. One
viewpoint is that the regulation creates an obligation to reduce emissions by
10 tons and that the control authority has allocated the requirement to control
emissions. Another viewpoint is that the regulation creates an entitlement to
emit 90 tons, and that the control authority has rationed the "emissions-
absorbing capacity" of the atmosphere. The desirability of using one or the
other perspective depends on the environmental objectives.
Feasibility
Approaches are administratively feasible to the degree their information
and resource requirements can be met at acceptable levels of effort and cost.
In some cases, approaches may not be feasible because the needed information
is not available. la other cases, the cost of design, implementation and
enforcement may exceed available resources.
Approaches must also be politically feasible. They must be capable of
attracting sufficient support to secure passage of the legislation. Of course,
even with political support, an approach must possess legislative feasibility.
In broadest terms, this means it must be workable within the context of existing,
functioning legislation.
Trade-offs between feasibility and other attributes (such as effectiveness
and cost-effectiveness) are especially likely. For example, some approaches
may be more effective for the same information and resources than other
1 "Entitlement" is used throughout this report in the economist's sense
of permission to emit certain levels of pollutants rather than in the specific
legal sense of rights conferred on polluters. Because the government is able
to reclaim entitlements under current practice, entitlements are "leased" by
sources rather than owned.
-------
1-12
approaches. In other cases, some approaches will require less information and
resources but require a longer period of time to achieve the desired environ-
mental objectives or impose higher control costs on sources.
Equity
Equity considerations would "be straightforward if the regulations for each
emitter could be based on the direct social cost of the damages that result
from its emissions. In these cases, the responsibility for control and the
cost of control would be apportioned in direct causal relationship to the
social damages. But because damage cannot be precisely attributed to and
quantified for each emitter in an air quality control region, this approach
is unlikely to be feasible for determining equity.
When the regulations are not based on the damages that result from emis-
sions, regulators must make additional decisions about who pays, how much they
pay and when they pay to control emissions, Ml approaches that are not based
on damages will raise the same equity issues, although some approaches deal
with them more explicitly than do others.
Equity or fairness involves both (1) allocation or "income distribution"
issues and (2) procedural neutrality or "equal treatment of equals." Assignment
or alteration of control obligations always raises income distribution issues,
but these are most recognizable when a new program is first established. "Dif-
ferent approaches are possible, and they may lead to different conclusions.
Common approaches are based on principles related to "equal treatment of equals,"
"benefits" and "ability to pay." "Equal treatment of equals" is a difficult
principle to apply In practice, since there are always many inconsistent dimen-
sions against which to judge equality of both sources and requirements. The
principle is sometimes seen, particularly in the political arena, as satisfied
if uniform requirements are imposed on all sources or all major sources. This
shifts attention to the assumptions used for defining uniform requirements
rather than on assembling groups of "equals," Under a "benefit principle"
those who benefit from pollution control would pay the costs of control in
proportion to benefits received; under a similarly founded but converse "damages
principle" those who cause damage by polluting would pay in proportion to
damages caused. Under an "ability to pay principle" costs of control would be
-------
1-13
borne in proportion to ability to pay. This criterion is frequently used in
tax policy, but is lass applicable when dealing with production decisions by
economically motivated companies rather than with taxes on individuals. Com-
panies will not have "abilities to pay" as much as economic reactions Co changes
in production cost. For example, "rich" companies may shut down heavily-
regulated sources as uneconomic, or control costs aay be passed through to con-
sumers with low "ability to pay" but with no teal option other than purchase
of the product.
Other Considerations
Other considerations can be important when evaluating the relative merits
of alternative approaches. These include adaptability in accomodating chang-
ing environmental problems, economic circumstances, and growth; and the crea-
tion of incentives to innovate and to control beyond regulatory requirements.
3. THE CURRENT APPROACH TO AIR POLLUTION REGULATION
The pollutants that are regulated under the Clean Air Act fall into one
of three categories: criteria pollutants, hazardous air pollutants, or desig-
nated pollutants. Criteria pollutants are regulated by National Ambient Air
Quality Standards (N'AAQS) that are established by EPA at levels intended to
protect the national health and welfare. Pollutants that are not regulated
by KMQS but that may cause an increase in mortality or serious illness are
regulated by National Emission Standards for Hazardous Air Pollution (NESHAP).
Pollutants that are not controlled under either NAAQS or NT.SHAP may be "desig-
nated" pollutants and regulated under Section lil(d) of the Clean Air Act.
The current approach to regulating these pollutants depends on the type
of pollutant that is being controlled, on whether the regulated emitter is a
new or existing source, and on the severity of the air pollution problea in
the area affected by the regulated source. The following discussion briefly
describes the approaches that are now used by EPA and the states1 to regulate
stationary and mobile sources.
1 In general, states have chosen to use comniand-ana-control approaches to
meet air quality goals. However, the Clean Air Act does not specifically
require the states to use a command-and-control approach to state regulation
of existing sources.
-------
1-14
B»1 REGULATION OF STATIONARY SOURCES
Existing Plants
The Clean Air Act directs EPA to establish National Ambient Air Quality
Standards (NAAQS) at levels that protect the public health and welfare. States
are required to prepare State Implementation Plans (SIPs) for attaining the
NAAQS within the time frames specified in the Act. Typically, states meet these
standards by imposing specific requirements on emissions points in existing
plants to reduce their emissions. The states adopt these requirements as regula-
tions that are the basis of the State Implementation Plans.
As long as the SIP demonstrates attainment by the statutory deadlines, the
states exercise great discretion in the regulation of existing sources. The
SIPs must demonstrate attainment for total suspended particulates, nitrogen
oxide, and sulfur dioxide by 1982; and fox ozone and carbon monoxide by 1987,
However, if states cannot attain the ambient standards for these pollutants
by the statutory deadlines, the Glean Air Act directs EPA to disapprove the
SIP, impose a ban on construction of new sources and require the states to
impose technology-based standards on sources in existing plants.1 The standards
for existing plants are called Reasonably Available Control technology (RACT).
The current approach also includes the "bubble policy" which is explained
in some detail in Chapter I?. Once the control requirements for a plant have
been established 'by a state, the bubble policy encourages companies to propose
alternative control requirements that achieve the same environmental goal for
their plants (whether it is an ambient level or an amount of emissions reduc-
tion). In this manner, plant managers have the opportunity to achieve environ-
mental requirements at the lowest possible cost — thus achieving cost-effec-
tiveness "within" plants. In addition, plant managers may propose alternative
control requirements that lessen controls for their plants but that involve
additional compensating controls at other plants ¦— thus achieving cost-effec-
tiveness "among" plants.
1 For ozone and carton monoxide, despite the 1987 attainment date, if the
state cannot demonstrate attainment of standards by 1982, the state is required
to impose RACT-level technology on existing sources.
-------
1-15
Existing plants may be required to meet NESHA?, which can be either
emissions standards or, if emissions standards are not practicable, design
standards. In addition, if Section 111(d) requirements are established for a
category of new plants under NSPS (as discussed in the next section), states
must establish requirements for the designated pollutant for existing plants
in the same category. The states have substantial discretion in establishing
Section 111(d) requirements.
lew Plants and Modifications
Emissions points in new sources oust always meet technology-based stan-
dards where these have been established regardless of the air quality in the
area in which the new plant is sited. Each technology-based standard and its
rationale is described below.
At a minimum, the Clean Air Act requires most types of emission, points in
new plants to meet emissions levels based on Hew Source Performance Standards
(NSPS).1 NSPS reduces the pollution problem in the future — avoiding pollu-
tion problems can be easier than dealing with problems after they have already
developed. This is accomplished by providing a means for the continued reduc-
tion of emissions in the loag run; as older plants are retired, the remaining
plants will be well controlled and the pollution problem will be less severe
in the future than otherwise. The pollutants for which the NSPS are applicable
are criteria pollutants and designated pollutants. The NSPS emissions levels
are established nationally for different categories of sources (emission units)
and are based upon consideration of the performance of adequately demonstrated
technology, cost, energy requirements and non-air quality impacts.
To prevent the significant deterioration of air quality in chose areas
that are now meeting the N'AAQS (PSD areas), sources in new plants may have to
install control equipment based on Best Available Control Technology (BACT),
which must be at least as stringent as NSPS. BACT provides a way to manage
the deterioration of air quality, to ration the air resource and to provide
a case-by-case mechanism for establishing control levels for new sources where
1 NSPS have not been established for all source categories.
-------
1-18
MSPS has not been defined. BACT is required for every pollutant that is regu-
lated under the Clean Air Act and emission levels are determined on a case-by-
case basis using the same considerations as for NSPS. USPS and 1ES1AFS repre-
sent the minimum levels allowable for BACT.
In areas that have not attained the SAAQS ("nonattainment areas"), new
plants must Install technology based upon the Lowest Achievable Emissions Rate
(LAER) for the pollutants that exceed the NAAQS. LAER minimizes the immediate
impact of growth on the non-attainment problem as well as the long run impact
of growth on the ability of areas to eventually come into or to maintain at-
tainment. LAER is the most stringent requirement for new sources, though it
coincides in many cases with the definition of NSPS or BACT. LAER is deter-
mined on a case-by-case basis for each affected sources, and must be either
the lowest emission rate achieved in practice or the raost stringent SIP require-
ment in the country for sources of the kind in question. In addition, unless
there is a "growth margin" in an area due to stringent control of existing
sources, new plants that locate in nonattainment areas must obtain offsetting
reductions in emissions frota existing sources in the area. These "offsets"
compensate for the remaining increase in emissions that occur after the appli-
cation. of LAER.
New sources are also required to meet any applicable National Emissions
Standards for Hazardous Air Pollutants (NESHAP).
B.2 REGULATION OF MOBILE SOURCES
New vehicles must meet emissions standards that are established separately
for classes of vehicles (e.g., passenger cars, light-duty trucks, heavy-duty
trucks). These standards are based on the use of technologies that are expected
to be available when the new vehicles are manufactured. The standards are
designed to ensure that all vehicles in a class will be able to meet the emis-
sions standards based on the expected performance of the anticipated technology.
Thus, the performance of heavier vehicles in a class will often limit the
stringency of a standard, even though lighter vehicles nay be expected to have
better emissions performance. The standards are periodically reviewed to ensure
that they are based upon the best technology that Is expected to be available.
-------
1-17
New passenger cars that do not meet the standards cannot be sold. As
specified in Section 206(g)(3) of the Clean Air Act Amendments of 1977, a noncon-
formance penalty structure is to be established that permits the sale of trucks
that cannot, for technological reasons., meet emission standards. This is
described in Chapter 1? and the Appendix.
The standards for new vehicle emissions performance must be met after
expected deterioration in performance over 50,000 miles of operation. An
extensive program of performance and durability testing is used to establish
new vehicle performance levels and to estimate deterioration factors. Manu-
facturer-recommended maintenance is performed during durability testing.*
For 1981 and later model years, manufacturers are required to warrant to
vehicle owners that emissions performance will be within the standards for
50',000 miles with normal maintenance* If a particular warranted -vehicle does
not perform as required when in use for reasons other than owner neglect or
abuse,^ manufacturers are required to pay for repairs (and for the diagnosis to
identify the problem). Where failures for a model of vehicle are due to design,
manufacturers may have to recall the model. Emission recalls usually will not
involve major problems, since manufacturer efforts and EPA testing will identify
major problems before a model is placed on the market.
1 The tn-use emissions performance of vehicles depends on initial design
and, for pre-taodel year 1981 vehicles, on routine maintenance. The use of leaded
fuels in vehicles equipped with catalytic control devices can also degrade eais-
s ions performance. Owners often have little incentive to beat the expense of
voluntary testing or of any maintenance directed specifically at emissions con-
trol for pre-1981 model year vehicles, because there is usually no significant
vehicle performance or fuel economy benefit from such maintenance. For new
vehicles with computerized engine controls routine maintenance is not as impor-
tant to emission performance, although periodic testing can detect systems
which have stopped functioning. (GM systems are designed with a dashboard
failure indicator,) Vehicles with nonfunctioning computer controls have higher
emissions and lower fuel economy, but performance is not affected.
2 "Neglect or abuse" need not involve deliberate tampering with the control
system. Failure to replace sparkplugs on schedule or to detect loose wires or
electrical short circuits can constitute neglect, and nistuning by a home
mechanic can constitute abuse. Diagnostic costs are relatively low with pre-
modal year 1981 vehicles, but may be high for computer-controlled vehicles.
-------
1-18
States which have not demonstrated attainment for HC and CO are required,
as part of a good faith effort to reach attainment, to take steps to establish
a mandatory program for periodic vehicle emission inspection and maintenance
I&M. 1 I&M programs may expose manufacturers to warranty claims in the future,
but have no effect on EPA recall activities because no mechanism is in place
to use inspection data to identify defective vehicles. The failure rate can
be varied depending on what is needed to reach attainment. Mandatory l&M
programs will typically fail 20 to 30 percent of tested vehicles: for pre-1981
model year vehicles, 20 to 30 percent of vehicles account for 75 percent of
excess emissions, although higher percentages fail to meet standards by small
amounts; for computer-controlled cars, 5 percent of all vehicles are expected
to account for 75 percent of excess emissions.
C- ATTRIBUTES OF THE CITOHST AfPROACg
The current approach is to implement ambient- and technology-based stra-
tegies using command-and-control methods. These methods generally specify the
manner in which sources can meet environmental goals. Thus, for example,
specific emission points within a plant are typically required to meet certain
levels of performance. This section briefly evaluates the command-and-control
aspects of the current approach in order to provide a context for considering
economic approaches. The economic approaches that are already in use are covered
in mote detail later in this report.
C.l STATIONARY SOURCES
Ambient Standards
o Economic Efficiency
The relative economic efficiency of ambient standards depends on whether
they have been established at economically efficient levels and on whether con-
trol occurs in a cost-effective manner.
1 Recently, California became the first state to be penalized through
loss of some federal highway and sewer grant funds for failing to institute a
mandatory I&M program in air quality control regions which had not demonstrated
attainment.
-------
-------
1-19
Information regarding the benefits and costs of ambient air quality stan-
dards are limited, but generally indicate that their benefits exceed the costs.
For example, Freeman'si study for CEO concluded that the total national bene-
fits realized from reductions in air pollution since 1970 lie in the range of
roughly S5 billion to $51 billion per year, with the "most reasonable estimate"
of annual benefits of air quality improvement enjoyed in 1978 being $21.4
billion. Freeman views this estimate as quite conservative. This can be com-
pared to the estimate of costs to the nation for air pollution control for 1978
contained in CEQ's 1979 Annual Report^ of $16,6 billi on. This does not mean
that the standards are actually optimal — optimality occurs where the dif-
ference between benefits and costs is maximized, not where benefits simply
exceed costs. Moreover, it is unlikely that the national standards could be
economically efficient everywhere, since the ambient targets are uniform
nationally, while both the marginal damages from air pollution and the costs
of control at the margin to attain the standards are likely to vary locally.
Nevertheless, in the absence of the capability to conduct cost-benefit analysis
that is adequate for the purpose of standard-setting, uniform ambient concen-
trations remain a practical, health-related substitute.
o Effectiveness
So ambient standards have as yet been met at all times in all areas of the
country and deadlines for achieving standards have been repeatedly missed. Thus,
the current approach of comma,nd-aftd-control regulation has not been fully effec-
tive in attaining ambient targets, and has not lived up to the theoretical capa-
bility of traditional regulation to control the timing of emission reductions.
Nevertheless, substantial improvement has been made. From 1972-1978 ambient
levels of particulates (smoke and dust) were reduced 10 percent, sulfur dioxide
by 17 percent, carbon monoxide by 35 percent and lead by 26 percent. Ozone
levels remained essentially stable over this period, with 1979 showing a 3
percent decrease from 1978 levels. Violations of ambient standards in 50 of
the aost polluted countries have remained constant or declined from 1974 to
1 Op. cit., p. vi.
2 Op. cit., p. 686.
-------
1-20
1977. Over the same period, the number of unhealthtul and very unhealthy days
in 25 major metropolitan areas declined by 15 and 32 percent, respectively,
Where additional controls are needed to attain standards, the command-and-
control approach may not succeed in identifying sufficient reductions to achieve
ambient standards because of the difficulty in identifying yet additional ways
to control emissions. As discussed below, the feasibility of obtaining addi-
tional reductions through comraand-and-control alone is low. Moreover, coamand-
and-control approaches do not provide any on-going incentive for corapanies to
find new ways to continue to reduce emissions. Thus, commard-and-control
approaches are likely to be less effective in the future than they have been to
date.
o Cost-Effectiveness
Since regulators have limited knowledge about the specifics of individual
plants, the SIP requirements are not necessarily the least-cost ways of achieving
the desired results. Moreover, even if the SIP requirements are cost-effective
for individual plants, they would not result in the lowest overall cost unless
regulators were able to optimize among all sources of emissions. Of course,
implementation of the bubble policy has increased the cost-effectiveness of SIP
requirements by allowing plant managers to optimize controls within and among
plants, To the extent that trading markets develop as a result of the bubble
policy and to the extent that trading takes place the overall cost of attaining
and maintaining ambient standards to existing sources can be minimized.I
o feasibility
Coaanand-and-contro L regulations are typically easier to develop for the
initial stages of control because initial control requirements are generally
based on the application of well known control technologies for basic sources of
emissions. As command-and-control regulations become increasingly stringent,
regulators must know more of the specific characteristics of individual sources
1 As discussed later in this chapter, trading under the bubble policy
is limited to existing plants.
-------
1-21
of emissions and more about sophisticated control methods. In gome cases, the
needed control methods may not yet exist or be readily apparant to regulators,
but could be readily developed or employed by industry if adequate incentives
were available. Thus, where ambient standards have not been net and more strin-
gent control requirements are needed, command-and-control approaches will be
less feasible to use than before.
o Equity
A number of allocation decisions have been made by states and by Congress.
To meet ambient-based standards, states have allocated control obligations (or
entitlements to eait) anong existing sources of emissions. The basis for the
allocations have included technical feasibility, cost and economic impact. The
cost for additional controls will be higher than in the past and it will be
increasingly difficult for states to determine which sources should be obligated
to control additional emissions (or conversely which sources should be entitled
to emit store pollution than others). Control opportunities will increasingly
be special cases, making "equal treatment of equals" ran even more difficult
principle to apply than in the past.
Allocation decisions have been made that affect transfers of wealth between
new and existing sources. In some cases (such as nonattainaent areas) existing
sources own all of the entitlements; new sources must purchase emission reduc-
tions (offsets) from existing sources. In other cases, states have created a
margin for growth in their SIPs; in these cases, states have essentially-
required existing sources to generate emission reductions (offsets) which the
state provides at no cost to new sources.
o Other Considerations
The current approach does not provide much incentive to innovate or to
control emissions in excess of requirements. However, to the extent that the
bubble policy is used by industry and as long as offsets are needed, some incen-
tives will exist to develop new ways to meet SIP requirements.
-------
1-22
Technology-based Standards
Technology-based standards have been primarily used Co establish control
levels for new mobile sources (discussed later in this section) and for new
and modified stationary sources. The technology-based standards far stationary
sources achieve several objectives: SSPS helps reduce the long-run pollution
problem when the capital stock turns over; BACT rations the capacity of the
air in PSD areas and is a backup to SSPS; and LAER allows growth in nonattain-
ment areas while minimizing its environmental impact, Additional objectives
of these requirements may be to "force" technology, to provide a hedge against
uncertainty about achieving auabient standards, and about the level at which
ambient standards should be set.
The attributes of these requirements is discussed below.
o Economic Efficiency
LAER typically represents the highest marginal costs of control in non-
attainment areas. As in the case of existing stationary sources, It is not
known how the cost of LAER compares to the marginal damages associated with
current ambient levels- However, since LAER typically represents the highest
marginal costs, it is the most likely requirement to be inefficient, if inef-
ficiency occurs. There is some evidence that LAER is inefficient in at least
some areas. Some air quality regions have come into attainment without paying
the marginal costs associated with LAER. If the ambient standards in these
areas are efficient, then the cost of LAER must he too high; conversely, if the
cost of LAER is not too high, then areas that have attained ambient standards
without paying the cost of LAER are undercontrolling, and the ambient standards
would need to be made more stringent for these areas if economic efficiency is
to be achieved. However, we do not know whether LAER is too stringent in some
cases or whether some local ambient standards are too lax from the standpoint
of economic efficiency. This result is not unexpected since the ambient stan-
dards are national instead of local — it may well be that the ambient standards
are as economically efficient as it Is possible to be for a national standard.
The efficiency of NSPS and BACT is more difficult to assess because they
seem, to be intended to achieve several objectives that are not directly related
-------
1-23
to ambient standards. However, ic is not likely that NSPS and BACT can be
economically efficient. The primary reason is that there is a wide disparity
in the marginal costs of control for these standards — for example, the marginal
cost of N'SPS control of a pollutant can wary widely from one industry to the
next. Thus, regardless of the objectives, since the marginal costs vary widely,
it is not likely that the objectives are being attained in a cost-effective
manner (unless the benefits to be achieved are related more to industries than to
pollutants). Since cost-effectiveness is a prerequisite of efficiency, it is
also not likely that the requirements are economically efficient.
However, even if the marginal costs for all NSPS and for all BACT standards
are equalized, these requirements may still not be efficient. NSPS is largely
oriented toward achieving future environmental benefits and BACT Is largely
oriented toward managing growth, Both either tend to delay the time when ambient
concentrations in attainment areas begin to approach the ambient standards or
tend to lower the cost of obtaining emission reductions in the future. If
this is largely true, most of the benefits of N'SPS and BACT are related to
deferring the damages associated with nonattainment or to lowering the cost of
achieving attainment. If the marginal costs currently experienced to achieve
attainment are economically efficient, the value of deferring or avoiding non-
attainment should be no greater than the costs incurred to achieve attainment
now. Typically, the cost of NSPS and BACT is higher than the cost of SIP
requirements where attainment has been achieved, but lower than the cost of
LAER (where attainment has not been achieved). Whether or not NfSPS and BACT
are economically efficient depends on whether LAJ5R is considered to be over-
control or whether ambient standards in attainment areas are considered to be
lax. It may well be that NSPS and BACT (and LAER) represent economically
efficient levels of control effort, while some ambient concentrations are low
because of political or practical constraints or because of the use of national
instead of local ambient targets.
As indicated above, there are a number of uncertainties about the effi-
ciency of technology—based standards. These uncertainties do not mean that
the technology-based standards are necessarily inefficient, They do strongly
indicate that it would be desirable to minimize the cost associated with
achieving the reductions represented by these requirements.
-------
1-24
o Effectiveness
These technology-based standards have In most cases led to the installa-
tion of control measures as required. la some cases, the standards have led
to the use of technologies that were not previously in wide commercial use.
However, when standards that are meant to be technology-forcing cannot be aet
(or when the consequences to industry of not meeting the standard are so severe
that they lose credibility as enforcement tools), technology-forcing comnand-
and-cootrol approaches alone may not lead to compliance.
o Cost-Effectiveness
The cost-effectiveness of technology-based standards depends on the ob-
jectives. As discussed earlier, Che cost-effectiveness calculation for a
given approach would differ, for example, depending 011 whether the intent was
to reduce emissions by a particular amount, to reduce emissions at particular
plants, or to reduce emissions at particular plants at specified emission
points using innovative or advanced technologies. Different approaches could
achieve these objectives for different costs.
The Clean Air Act, as it is currently written, is interpreted to require
the control of emission reductions at new sources, regardless of which of the
underlying objectives is most important. Thus, trading against technology
requirements to achieve the same amount of emissions reductions at other plants
is not permitted. As discussed in later chapters, such trading could be used
to achieve the same emission reductions at lower total cost while still
achieving objectives such as reducing emissions as the capital stock turns
over, rationing the capacity of the air resource, or achieving innovation (the
innovation would occur in existing plants to a greater extent). If all of
the objectives of technology-based standards can still be met if trading is
allowed, then the existing approach for regulating new plants is not cost-
effective.
o Feasibility
"S?S, 3aCT and LAER have been developed for a number of classes of plants
and for a number of individual plants. While experience has shown that these
-------
1-25
standards can be developed, they are sometimes difficult to develop and the
subject of controversy. The exact specification of these standards Is important
because, unlike existing plants that can use single or multi-plant bubbles,
there is no flexibility for new sources in meeting the standards. In addition,
the administrative cost of implementing some new source requirements may be
higher than necessary because of the frequent similarity of N'SPS, 3ACT, and LA£R
— some of the of the elaborate case-by-case analysis (and delay) for BACT and
LAER is, in restrospect, duplicative.
o Equity
New sources have been required to control emissions to levels determined
by technological considerations. These requirements essentially ration the
capacity of the atmosphere to absorb pollution. Thus, there has been a
distribution, of wealth between those who locate first and those who locate
later.
There are also wealth distributions between new and existing sources.
Wealth is transferred to existing sources because additional control require-
ments for existing sources may be deferred as a result of the steps taken to
control new sources first (and more stringently). The rationale for this is
that (1) new sources are likely to be more efficient in achieving pollution
control requirements by virtue of their ability to incorporate such controls
into production planning, and (2) existing sources were constructed at a time
when they did not have to meet stringent control standards and it is, in a
sense, unfair to penalize then to the same extent as new sources for require-
ments they could not have foreseen. In addition, to the extent that the
standards are intended to force technology, new sources are obligated to
contribute to the commercialization of technology. However, in some cases
wealth is also transferred to new sources from existing sources where states
have created a margin for growth in their SIPs; in these cases, states have
essentially required existing sources to provide emission reductions (offsets)
at no cost.
-------
1-26
o Other Considerations
Command-and-control approaches can limit total emissions in the face of
economic growth only by becoming increasingly stringent, While technology-based
requireraeo.cs can force use of new technologies that will permit more effective
control, no other incentives to innovation are provided. In addition, where
regulations are based on technology, innovation may be discouraged —¦ there may
be a tendency to "freeze" technology at current levels.
C ? MOBILE SOURCES
. Emission standards for mobile sources are technology-based, and intended
to force the development of new technology and to help attain ambient stan-
dards. Thus, much of the previous discussion of ambient-based and technology-
based strategies for stationary sources applies to mobile sources as well.
Technology-forcing mobile source standards are unlikely to be economically
efficient since the same emissions performance is required of each vehicle in
a class regardless of where it is used or of the costs of meeting standards.
(It would be more efficient to vary control requirements according to the cost
of control and the need for control in a given area.) "Economic efficiency and
coat-effectiveness are also limited by the use of vehicle classes and the lack
of flexibility accorded manufacturers: attention is necessarily focused on
solving the most difficult technical control problems rather than on reducing
emissions at the least cost possible. In the past, these efforts resulted in
some degradation of vehicle performance and fuel economy, and may have placed
heavier domestic vehicles at a competitive disadvantage to lighter foreign
vehicles. These technological tradeoffs have been eliminated as performance
and fuel economy characteristics have been improved. However, control technol-
ogy has now reached a point where increasingly stringent standards may not be
as cost-effective as past mobile source measures, or as other control efforts.
Mobile source standards have been fairly effective in forcing the develop-
ment of cleaner new vehicles, although some delays in compliance have had to
be accepted. The existing program has been less effective in controlling the
emissions of vehicles after they have been sold. Recall is expensive for
manufacturers and must be used with discretion. Owners have little economic
-------
1-27
incentive to maintain controls in the absence of mandatory I&M or to respond
to recall notices. Mandatory I&M is not popular with manufacturers (as it may
lead to recalls and warranty claims), with states or with vehicle owners.
-------
II
INTRODUCTION TO ECONOMIC APPROACHES
Economic approaches to air pollution control can be expected to be more
economically efficient than command-and—control regulations because they can
better exploit and adjust to differences in the costs of controls experienced
by different companies. Economic approaches manage this primarily be placing
greater decision-making power in the hands of companies, which have better ac-
cess to information about relevant costs than do the regulators and which can
act on this information. This is accomplished by creating an economic value
for reducing emissions that allows companies to make economically efficient
decisions about pollution control. This economic value is created when the
government allows companies to trade entitlements to emit Cor obligations to
reduce emissions) or when the government places a charge on undesirable act-
ivities (such as "excess" emissions).
This report groups economic approaches to air pollution control into two
classes: charge-based approaches and trading approaches. Charges induce
emitters to make socially desirable decisions about controlling pollution by
requiring them to pay a fee for emissions in excess of defined levels; these
levels may or may not allow some emissions without charge. Trading opportuni-
ties induce emitters to meet environmental requirements in the most cost-
effective manner by allowing them to buy or sell either entitlements to emit
or requirements to make emission reductions. Thus, trading systems establish
a market for such entitlements or emission reductions.
In many cases, either pricing or trading can be used to achieve a given
goal, although often one approach will be more desirable than the other. "Pric-
ing approaches are a more radical departure from existing regulatory programs.
However, in some cases economic incentives can be used as supplements to en-
hance command-and-control systems.
In this chapter, economic approaches are discussed in terms of the cri-
teria identified in the introduction—effectiveness, economic efficiency, cost-
effectiveness, feasibility, equity and other considerations such as adaptabil-
ity to change and effects on innovation. (The order in which these attributes
are addressed will vary as necessary to aid exposition.)
-------
II-2
A. CHARGE-BASED APPROACHES
Charges can be used as the sole means of achieving environmental objec-
tives or they can be used to supplement approaches that are already in use.
These ara discussed separately as stand—alone charges and supplementary charges.
A.1 STAND-ALONE CHARGES
Charges set a price on use of the air. The basic concept is that a charge
is established for every unit Cor every unit in excess of a specified limit)
of pollution emitted, and sources have the choice of paying the cost of reduc-
ing pollution or paying a charge to emit it. Sources will pay to reduce pol-
lution to the extent that it is cheaper than paying the charge, and can reduce
emissions in any way they wish. At the margin—the last unit of pollution
controlled'—sources will pay up to the amount of the charge to reduce pollu-
tion. As a result, no source will pay more at the margin per unit of pollution
reduction than any other source. In this way, charges lead to cost-effective
controls within companies and among companies.
Conceptually, the charge should be based on the environmental danages
caused by a source's emissions. In this way, no source will pay more for con-
trolling emissions than it is worth to society to do so.
Charges can also be used as an alternative means to meet the same environ-
mental objectives as to comraand-and-control regulations- The charge can be
set at a level that is expected to induce companies to reduce enough emissions
to achieve the goals that would "have been attained by the traditional regula-
tory approach. Because traditional approaches usually do not set requirements
on the basis of source-by-source marginal control costs, charges will result
in lower total costs of emission, controls than traditional regulation. Even
if traditional regulation was sensitive to marginal control costs or charges
were levied in a uniform fashion across categories of polluters, charges would
result in lower costs because they leave technical decisions in the hands of
those with the most information, and provide incentives to develop cheaper
control methods.
The above considerations are discussed in greater detail below.
-------
11-3
Cos t-Eff e etiyene ss
Figure II-I illustrates the choice offered to a firm by stand-alone charges,
and the process by which a firm's behavior will be determined. The curve MC
shows the firm's marginal control costs—that is: its cost, at any given
level of abatement, of a successive increment of abatement. It is assumed that
marginal control costs increase as abatement efforts become more intense. The
cost level labeled C0 on the scale measuring abatement cost shows the charge
imposed on emissions. It is assumed that this charge is the same for each
unit of pollution.
If the firm represented in the graph is faced with an emissions charge of
Cc dollars per unit of emissions, the firm will have to decide whether to pay
the charge or to "purchase" additional abatement by incurring the costs of
additional reduction in emissions. If the firm was purchasing I4 units of
abatement prior to imposition of the charge, the charge the firm now faces is
greater than the cost of further abatement, and continues to be so until point
L0. Therefore, the firm will continue to increase its abatement until its
marginal cost of abatement equals the charge at LQ« The firm will pay the
charge for its emissions that remain unabated at the abatement level LQ be-
cause the charge is lower than the marginal cost of abatement for those units
of emissions. The firm's total control costs at the new emissions level are
represented by the cross-hatched area under the marginal cost curve, MC.
There is an incentive for every emitter that is faced with an emissions
charge to behave like the source in Figure II-l, reducing pollution to the ex-
tent that it is less expensive to install controls rather than pay the estab-
lished charge. Because all sources base their behavior on the charge, no
source will pay more per unit of pollution reduction at the margin than any
other source.
This equalization of marginal costs is a desirable outcome, because it
means that the pattern of control effort which results from a charge will be
fully cost-effective. Those sources with relatively higher abatement costs
will abate less; those with lower, more. Whatever the resulting level of
control, the total overall cost of control will be the minimum possible for
-------
11-4
Figure II-l.
Effects of a Charge on an Emitter
$
Cost of
Abatement
MC
Units of
emissions
Aggregate Abatement by Source
-------
II—5
that level. Similarly, whatever the total expenditure on controls» the level
of control achieved will be the maximum possible for expenditures of that
magnitude. Thus, charges not only lead to cost-effectiveness within plants
but also among plants. In this way, charge-based approaches also contribute
to a better use of all of society's productive resources than is likely to be
achieved with a consnand-and-con.trol approach alone.
Efficiency
Stand-alone charges can. be economically efficient if the charges are based
on the damages caused by a source's emissions. So source should pay more for
controlling emissions at the margin than the amount of damages an additional
unit of pollution would impose on society. The total level of emissions that
would result from such a charge is said to be socially optimal because when the
optimum point is reached all controls with a value in excess of their costs
have been undertaken.
This concept is illustrated in Figure 11-2, which shows how L0, the so-
cially optimal level of emissions, is determined. As la Figure 11-1, marginal
control costs are shown for successive amounts of emissions removed by the
curve MC, but in this case the curve reflects all control opportunities avail-
able to society rather than the opportunities available to a single source.
The curve MSB is a representation of the marginal social benefit of abatement,
that is—the value to society of the damage averted by a marginal unit increase
in pollution abatement- The curve MSB includes health and all other benefits
to society and is expressed in dollar terns. It slopes "downward," indicating
that at low abatement levels (high emissions levels) the social value of a
marginal increase in abatement is high; when abatement is already high, in
contrast, the marginal social benefit from an additional unit of abatement is
low. The optimal charge rate in Figure II-2 is C0, since this is the only
charge which will induce source behavior that will equalize the marginal
costs of abatement to the marginal social benefit from abatement. A higher
charge of G]_ would induce control that is more expensive than it is worth,
as the b% units of abatement that would be purchased by society at a marginal
cost equal to C]_ are valued at the margin at only ¥]_.
-------
11-6
Figure II-2.
Setting the Charge Rate to Yield
A Socially Optimal Level of Emissions
$
Cos c,
Charge
Sate
MSB
MC
C
o
Units of
Emissions
L0 lx
Aggregate Abatement
-------
II—7
Figure 11-2 illustrates the quantity of information that is required to
determine an optimal charge rate. In that illustration^ information about the
quantity-based relationships (the "shape of the curve") of both the social
benefits from abatement and control costs would be needed. Only if marginal
social benefits are equal for all units of emissions is knowledge of marginal
social benefits from abatement adequate to define the optimal rate.
Unfortunately, information about the cost of damages associated with
emissions is limited and difficult to quantify. Such information would include
the environmental dose-response relationships, the population at risk, and as-
sumptions regarding the valuation of physical measures of benefits. Knowledge
of the damages associated with one level of total emissions or with marginal
emissions for one source does not necessarily provide an adequate basis to
predict social costs for other emissions levels or for other sources in other
locations.
The amount of abatement that is "optimal" will depend upon control costs
and the social benefits from abatement. Depending upon the shape of the bene-
fit curve, nearly total abatement of emissions may be justified; the use of
economics does not necessarily imply more pollution than would occur under
other decision rules. Figure 11-3 illustrates this point. In that figure
marginal control costs are (as before) increasing. If the marginal social
benefits of abatement are best represented by the downward sloping curve
MSBi, then the optimal level of abatement is L| and the optimal charge is
the relatively low C». If the marginal benefits of abatement are high and
constant as in MSB2, however, marginal control costs of C2 are justified and
the optimal level of abatement will be L?, very near the total abatement
point L-p,
Feasibility
Setting charge rates at the point where marginal social benefits equal
marginal control costs is only possible where the benefits are quantifiable.
Unfortunately, generally they are not.
-------
11—8
Figure II-3.
Dependence of "Optimal" Abatement on Shape of Benefits Curve
Abatement
Cost,
Charge Rat
MC
MSBi
Aggregate Abatement
-------
11-9
Data difficulties regarding estimation of benefits occur because emis-
sions have overlapping effects and because effects change as emissions change.
Different pollutants occurring together (or pollutants from more than one
source) may both cause the same type of effect, but the benefits of controlling
one source will depend on the controls imposed on other sources. Or there may
be two kinds of effects (e.g.> health, property values) arising from the saree
pollutant which must "be estimated separately, thus creating a risk of double
counting, Another problem is continuing uncertainty about-the measurement bene-
fits even in simple cases: Less than perfect assumptions about unknown para-
meters, relationships and values will be involved, and measurement errors will
compound the problem; incomplete knowledge of the effects of long-term expo-
sure on humans and plants and inappropriate extrapolations of damages based on
assumptions about the shape of the damage function and the size and exposure of
the population at risk will result in damage estimates of uncertain reliabil-
ity. A third data problem arises due to systematic errors in the use of aggre-
gate or average measures of the air pollution variable. Finally, there are
problems regarding the valuation of damages, including the difficulties inher-
ent in deriving estimates at different levels of aggregation (individual regu-
lations versus nationwide). Because the damages caused by emissions may not
be reversible, uncertainty about estimates might call for a conservative ap-
proach to setting charge rates.
Because It is often difficult to set an optimal charge rate based on
marginal social benefits of pollution, charges could be set to achieve more
modest objectives. If control costs can be estimated, for example, charges
could be set at a rate that would lead to a predictable level of emissions.
If the resulting level of control is less than what was sought, the charge
could be raised to induce the desired level of control; or if too much control
has occurred, the charge could be lowered. This use of charges provides an
alternative means of achieving the objectives of traditional command-and-
control regulation, but at a lower total cost of control and with less
information.
Moreover, charges can be useful where control authorities are unable to
use traditional approaches to secure needed progress toward attainment. If
timing (the level of emissions in the short as well as the long run) is impor-
tant, however, more information may be needed to ensure that the charge is
-------
11-10
Initially set at an appropriate level. This is because the responses of pol-
luters to a charge Is difficult to predict in the short run without consider-
able information. In addition, some target ambient levels can only be achieved
by controlling emissions from particular sources; this holds true for "local"
pollutants such as particulates, nitrogen oxides and sulfur oxides, but is
not the case with area pollutants such as hydrocarbons.! Achieving the target
ambient levels for local pollutants requires a nonuniform charge system that
will induce the targeted firms to reduce emissions by Che targeted amounts.
Developing nonuniform charge systems to Induce particular behavior by sources
at acceptable costs can be as difficult as establishing source-by-source
command-and-control regulations.
In any of these applications, charges allow more control over the costs
incurred by sources, while cotnmand-and-control regulations allow more control
over level and timing of emission reductions. If more is known about accept-
able control costs than about damages froa emissions, charges may be the ap-
proach which makes best use of available information to reach a good outcome.
This is especially true if control costs vary a great deal with the level of
control, while marginal damages are relatively constant.
There are legal problems implementing charge systems under current legis-
lation. Congress has been unwilling in the past to relinquish control over any
area that is related to setting taxes or tax rates, and may be precluded by
the courts from delegating the inherently legislative authority to establish
and periodically change charge rates. However> if an agency was left with
little discretion in determining Charge rates according to the instructions
set out by Congress, the courts might find that no delegation had occurred.
Constitutional problems may also be raised by charge rates that differ from
source tc source if they are characterized as discriminatory taxes. But "dis-
criminatory™ rates (rates which vary to reflect ambient impacts) will be opti-
mal for many pollutants.
1 While it is likely the spatial and temporal pattern of emissions and
their chemical composition affect the pattern of ozone concentration, we do not
have operational models to deal with the detailed photochemical processes.
Thus, for some pollutants such as hydrocarbons it is assumed that emissions are
an appropriate, proportional measure of ambient air quality impacts.
-------
11-11
Effectiveness
As indicated in Figure II-1, it will be difficult to predict the effects
of a charge in advance, unless all sources' marginal cost curves MC are known.
Charge-based approaches do not guarantee that total emissions will be reduced
to a particular level or that a particular ambient level will be attained,
unless the system operates with exact knowledge of each firm's abatement costs.
Therefore, the effects of charges on total emissions usually cannot be accu-
rately predicted in advance. Initially, if the charge is set too low, indus-
try's response will fall short of the goal; or the charge may be set too high,
and industry may control more emissions than desired. Charges could be varied
over time, but this could lead to a loss of credibility with the public.
Charges will reduce emissions at the lowest cost if they are stable bat the
unpredictability of source responses to charges limits their effectiveness in
meeting air quality goals where speed is important. In contrast, cottmand-and-
control approaches allow prediction of the level of emissions, but costs of
achieving these goals are not controlled. Charges are more likely to be effec-
tive in practice where another system is not already in place, with the asso-
ciated vested interests.
Where, information or resources are limited, charges can be more effective
than command-and-control regulations. Charges will eventually induce cost-
effective controls regardless of the agency's ability to identify then in
advance. In addition, charges can induce innovation because firms will have a
continuing incentive to develop new ways to reduce emissions. Charges are
also very credible tools for technology forcing because firms know with cer-
tainty that they will pay a charge if they fail to reduce emissions; charges
hold no hope for polluters to avoid costs by delay or postponement of dead-
lines. Charges eliminate the incentive to delay control efforts that exists
with command-and-control systems.
Equity
Stand-alone charges affect all covered sources in the same way; each is
required to pay an equal amount per unit of regulated emissions which are not
eliminated. Typically, charges will not result in Che same pattern of control
effort or the same costs to sources as traditional approaches, la particular,
-------
11-12
if companies prefer to pay some charges rather than reduce emissions to zero,
charges will result In money transfers from industry to the government that
will not occur with other approaches. To reduce these transfers, charges
might be imposed only on emissions in excess of some threshold established
for each source, (Thresholds might also be established to confine charges to a
supplemental role. This use of charges is discussed In detail after the next
subsection).
Other Considerations
The effects of charges will depend upon the level of the charge relative
to control costs and on the number of sources that wish to emit. As inflation
drives up control costs and growth drives up demand for use of the air as a
receiver of emissions, charges need to be adjusted if they are to remain
equally effective at controlling total emissions,! However, assuming statu-
tory authority, charges can be increased at least as easily as more stringent
command-and-control regulations can be instituted. In addition, charges pro-
vide a continuous incentive to sources to find new and less expensive ways to
reduce emissions in order to avoid paying them. This may also create a need
for adjustments in charge rates to avoid overcontrol.
A,2 CHARGES AS A SUPPLEMENT WHEN STANDARDS ARE SPECIFIED
Charges may be used as supplements to the current connnand-and-control
approach to regulation, in order to combine some of the characteristics of
each approach that have been identified in the previous chapter and the pre-
vious section. When used as an addition to established standards, charges can
(1) provide added incentives to comply with requirements, (2) provide an in-
centive to exceed requirements, or (3) simultaneously provide a credible
alternative to compliance and an incentive to develop new technology.
Charges can Increase the effectiveness of command-and-control requirements
by eliminating economic incentives to delay installation of controls. In this
1 If the environmental goal is to match control costs to damages rather
than to limit total emissions and if damages per unit of emissions are rather
constant, charge rates may not need to be adjusted.
-------
11-13
context, such charges are similar in purpose to "noncompliance penalties,"
which are discussed in more detail in Chapter IV and are a feature of the
current provisions of the Glean Air Act. These penalties are set separately
for each noncomplying source at a level equal to the costs savings obtained by
the source's failure to comply with regulatory requirements on schedule.
Supplemental charges can also be assessed for emissions remaining after
command-and-control requirements are met. Such charges provide a continuing
incentive to sources to find ways to reduce emissions further at lower cost-
Finally, sources could be allowed to pay a "nonconformance" fee as a
temporary alternative to compliance. A program of this kind, by payment of a
penalty charge, allows continued operation of sources faced with technically
infeasible or high cost requirements, yet maintains incentives to come into
compliance eventually. Supplemental fees of this kind can Increase the effec-
tiveness of technology-forcing standards by providing a credible enforcement
alternative that induces eventual compliance. Congress, in the Clean Air Act
Amendments of 1977, authorized a program of this kind in the mobile source
area, where technological laggards may find it difficult to comply with heavy-
duty truck standards on schedule.
Effectiveness and Feasibility
Supplemental charges used to provide incentives to comply with regulatory
requirements can greatly increase the effectiveness of command-and-control
requirements, by eliminating the economic benefits of delaying compliance-
Charges set to provide a temporary alternative Co compliance can also increase
the effectiveness and feasibility of command-and-control systems by providing
a credible alternative to immediate compliance. This alternative eliminates
any prospect for regulatory tolerance of delays based on questions of technical
feasibility or economic impact.
Supplemental charges used to provide an incentive for additional emissions
reduction will be as effective as stand-alone charges of the same level, but
involve fewer institutional risks. If supplemental charges were subjected to
successful legal attack as an unconstitutional discriminatory tax or as based
-------
11-14
on ait unconstitutional delegation of legislative authority, existing regula-
tory requirements would still be in place. Stand-alone charges involve no
such backup.
Such supplementary charges are easier to set than stand-alone charges
because primary reliance can be placed on the existing regulatory requirements
to meet minimum standards. In addition, the availability of a workable "safety
valve" in the event of unacceptable cost or technological infeasibility can
allow control authorities to set technology-based standards with less informa-
tion than would be required if these standards had to be enforced by themselves.
These charges can be set In anticipation of particular expenditures by particu-
ar firms, and can be adjusted retroactively to reflect actual cost experiences.
Therefore, charges of this kind are Eeasible.
Efficiency
Supplemental charges which provide an alternative to compliance with
regulatory requirements, or which are intended to force technology, may (where
sufficient information is available) be set at a level which reflects the
marginal social costs of emissions. Such charges will contribute to efficiency
in the allocation of resources, since they prevent purchase of overly expensive
increments of abatement.
Equity
Those supplemental charges that simply increase the effectiveness of
existing regulatory requirements do not raise new equity issues. Moreover,
the nonconformance and noncompliance type fees eliminate the "inequity" of
soae sources failing to comply while others do. However, supplemental charges
will typically involve money transfers froa industry to government, and charges
which provide an alternative to compliance and set a ceiling on marginal con-
trol costs can affect the total costs Incurred by some sources. To this ex-
tent charges will have a different, though not necessarily better or worse,
equity impact than do coomand-and—control requirements.
-------
11-15
Co s t-Effect iveneas
Supplemental charges which set a ceiling on marginal control costs should
increase the cost-effectiveness of command-and-control requirements because
they systematically eliminate the requirement to comply with control obliga-
tions that are more expensive than is considered justifiable. Marginal control
costs will be equalized for sources with the highest costs, and total costs
incurred by some individual companies will decrease even after money transfers
to the government. Charges which force or encourage the development of new
technology may also contribute to cost—effectiveness by lowering the overall
cost of control in the long run.
Other Considerations
Like stand-alone charges, supplemental charges may have to be adjusted to
reflect inflation technological change and levels of emissions. If imposed on
emissions which are allowable under the established standards, they provide
the same incentive to innovation as stand-alone charges. Section 111 contains
an assessment of the factors that are Important in choosing the appropriate
economic approach.
B, TRADING APPROACHES
Trading can be used either as a primary means of achieving specified
environmental goals in a cost-effective manner or as a means of enhancing the
effectiveness, feasibility and cost-effectiveness of traditional regulation,
This section describes the general features of trading approaches, discusses
the operation of trading markets and then summarizes the attributes of trading
approaches. The next chapter will discuss more detailed design considerations
of trading programs and Chapter IV will describe already existing trading pro-
grams.
B.l OVERVIEW OF TRADING APPROACHES
This section describes marketable permits, stationary source controlled
trading and mobile source averaging. Marketable permits Include a continuum of
trading approaches that regulate the quantity of pollution that can be emitted.
-------
11-18
In its "pure" form, marketable permits refer to a system that contains a air.imuia
of government intervention and provides entitlements to emit prescribed amounts
of pollution. The absence of restrictions and wide coverage of a "pure" system
would encourage an active permit market. Stationary source controlled trading
and mobile source averaging are variants of marketable permits which have been
developed as supplementary approaches to traditional regulation. Trading under
these more restrictive program is likely to be much less active.
Marketable Permits
Marketable permits are entitlements to emit specified amounts of pollu-
tants in a period of time at a specified location, or area. As in the case of
command-and-control regulations, marketable permits "ration™ the amount of
pollution that control authorities are willing to allow. Companies may only
emit the amounts specified in the permits they own. Permits may be bought
where a seller is willing and may be sold by their owners. Companies will be
willing to pay for permits an amount up to their cost of reducing their emis-
sions. As with charge-based approaches, the resulting overall cost to com-
panies will be minimized. The marketable permit approach is best applied to
pollutants that occur within a definable geographical area. Under certain
supply conditions entitlements will trade at a price equal to the rate that
would be used under a charge-based approach. In this case, a marketable permit
policy will result in control costs equivalent to those under charge-based
approaches, with the added advantage that the level and timing of emission
reduction is known with a certainty limited only by enforcement capabilities.
A key issue in designing a marketable permit system is the manner in
which permits are to be allocated initially. Allocation options include
auctions, distribution based on historical use, and distribution based on
government discretion.1 There are apparent, important competitive and wealth
transfer implications for any marketable permit approach. Of course, the
imposition of any regulatory approach has the same types of implications—they
are simply more obvious for marketable permits.
1 The Government has a long history of awarding valuable rights, e.g., the
right to plant crops, the right to the use of water, the right to import or to
buy low cost domestic crude oil, landing "slots" at airports, etc.
-------
11-17
"Controlled trading" approaches have been developed and iapleraented as
regulatory supplements to reduce the costs of complying with requirements under
the existing regulatory framework. Trading can reduce the cost of command-
and-control regulations by allowing companies to meet their control obligations
in less expensive ways. This can be accomplished by allowing companies to use
alternative control strategies for emissions sources within their plants as
long as they achieve the same environmental goal (total emissions or ambient
impacts), and are just as reliable and as enforceable as the original source-
specific control requirements. EPA* s "bubble™ policy is already in use and
allows such flexibility for existing plants to aeet SIP requirements. The
bubble policy also encourages multi-plant bubbles; plants can meet their con-
trol obligations by allocating emission reductions among all of the emissions
sources in the plants. Thus, trading establishes a market for emissions
reductions that allows companies to minimize their cost of control. It also
creates incentives for the equalization of the marginal costs of control and
can lead to cost-effectiveness within plants and among plants.
Trading nay provide a means for new sources to locate in nonattainraent
areas while actually contributing toward progress in reducing emissions. A
new source can purchase emission reductions or "offsets" from existing sources
to compensate for the emissions it adds to the area. Alternatively, a govern-
ment seeking to attract new industry could provide offsets by taking actions to
generate emissions reductions itself or inducing other sources to do so with or
without compensation. Such government-mediated transactions have been more
common in the past than arms-length bargaining between firms: these transac-
tions do not necessarily have the same optimality properties as trades between
firms.
In principle, controlled trading could be expanded until a traditional
approach with controlled trading would be more like a marketable permit system
than a coramand-and-contro1 approach. The practical aspects of a comparison
between these approaches are examined more fully in Chapter 111.
-------
11-18
Averaging
Averaging for mobile sources offers benefits similar to that of the bub-
ble policy for stationary sources. Under this approach, emissions standards
for individual vehicles would "be replaced with average standards applicable
to a manufacturer's entire fleet or to broad classes of vehicles made by the
manufacturer. Emissions of vehicles that are lower than, the standards can be
used to offset or compensate for emissions of vehicles that exceed the stan-
dard. This would allow manufacturers to control more emissions where costs
are low or where sales volume is high, and to reduce controls elsewhere. If
properly designed, overall emissions should be no higher than under the current
approach, but the overall cost of control should be less.
B.2 OPERATION OF TRADING MARKETS
Market Transactions
Once entitlements or obligations to control have been allocated to sources,
the resulting markets for entitlements or obligations operate in essentially
the same way as any other market. The discussion in this section is cast in
terms of entitlements, but is equally applicable if obligations to reduce
emissions are traded.
Unless entitlements are auctioned, it is likely that sources will face
different marginal costs in attempting to reduce emissions to levels consistent
with the entitlements in hand. Most sources will, of course, use some of the
entitlements they possess themselves. Some of these sources will be potential
buyers of additional entitlements and some others will be potential sellers.
Sot all entitlements initially allocated will find their way into the market.
After permits are initially allocated, they may be bought and sold. An
emitter will be willing to pay up to its marginal cost of reducing emissions
for permits to pollute so as to avoid the cost of emissions reductions. An
emitter will be willing to sell permits if it can do so far a price equal to
or higher than its marginal cost of abatement. The price which "clears the
market" by bringing supply and demand into balance is determined by both buyers
-------
IT- 19
and sellers. At higher prices there will he nore willingness to supply en-
titlements, since more expensive control measures can be justified in order to
free an entitlement for sale. However, there will be less interest in higher
priced entitlements on the part of buyers- Conversely, at lower prices for
entitlements c.here will be fewer entitlements offered for sale, since sources
will use low value entitlements to cover their own emissions rather than to
undertake more expensive control efforts, and there will be more interested
buyers. At prices that will be determined by the market, these tendencies
will balance each other. When this occurs, all sources incur the same costs
for abatement at the margin. Abatement which could occur at a lower cost will
have been undertaken to make entitlements available for sale, and abatement
that is more expensive will be avoided through purchase of entitlements.
This process is set out in Figure 11-4, which combines an illustration
of the determination of a market-clearing price in 4(a) with an illustration
of the effects of that price on the decision of one source in 4(b),
In 4(a), the market demand curve for entitlements D is the summation over
all sources of the marginal costs of abatement; that is, the sua, of the emis-
sions reductions that would be undertaken by each firm for any given marginal
cost. The supply of entitlements, E0 on the diagram, is fixed "by the govern-
ment. Thus, for any given total, number of entitlements issued, the price at
which they would be traded on the market is the marginal cost of attaining the
abatement level corresponding to the entitlements issued, shown by CQ, the
price at which supply and demand for entitlements are in balance. Although
the amount of entitlements actually bought and sold in the market will depend
upon individual sources' marginal cost curves and their original allocation of
entitlements, total abatement undertaken will not. This will be determined by
the quantity of entitlements originally allocated.
In 4(b), the role of the individual source in this market is illustrated.
One source with marginal costs illustrated by the curve MC responds to the
market-clearing entitlements price by reducing emissions by the amount k0.
This source may then become either a buyer or a seller of entitlements, depend-
ing on whether its allocation of entitlements is less than or greater than the
emissions level remaining when the abatement is L0. If more abatement is
required, the source will purchase entitlements in lieu of undertaking control
-------
11-20
Figure II-4.
The Market for Entitlements
4(a)
Determination of Market Price
4(b)
Determination of Source Abatement
Cost
Cost
Emission
Units
Emission
L0 Units
Entitlements Issued
(All Sources)
Abatement by Individual Source
-------
11-21
at costs in excess of C0. The crucial factor is that marginal control costs
remain at C0, for this source and for all sources.
Banking
t* g.
Trading Is most effective when there Is an active market for emissions
reductions. Because it can be difficult to identify potential sellers of emis-
sion reductions, a number of sellers must sometimes be contacted to obtain the
desired amount of reduction. Thus, there can be uncertainty as to the valid-
ity of the reductions available and actual reductions can take time to obtain.
These factors can affect siting decisions of new sources that require offsets
or existing sources that are seeking alternatives to meeting their require-
ments. The market can be facilitated by the establishment of emission "banks,"
which provide for the storing of emission reductions for ti.se at a later time.
Such banks are convenient for sources which reduce emissions, and provide a
way for purchasers to quickly obtain the emission reduction they need.
B.3 ATTRIBUTES OF TRADING APPROACHES
Effectiveness
Trading approaches, whether used in place of or to supplement traditional
regulation can increase or exceed the effectiveness of coannand-and-control
approaches in two ways. By reducing costs, trading reduces Incentives to de-
lay compliance. In addition, when emissions banking is used trading can lead
to extra or earlier emission reductions for as long as reductions are "stored"
in the bank.
Cost-Effectiveness
Since trading approaches can lead to the same emission reductions as
cotiMiand-and-controi approaches at the lower costs associated with charges,
trading approaches can be as cost-effective as charges and considerably
more cost-effective than command-and-control approaches alone. Even if total
emissions are identical so that there is an equality of marginal social bene-
fits between approaches with and without trading, the total costs of control
can be expected to be less with trading.
-------
11-22
Economic Efficiency
in addition to being more cost-effective than command~and-corxtrol ap-
proaches, trading approaches will tend to be more economically efficient.
Where-marginal control costs exceeded marginal social benefits prior to trad-
ing, trading should increase economic efficiency by reducing expenditures for
control. If trading approaches work well they will be as efficient as charge-
based approaches for the same level of total control, because all sources will
have the same marginal cost of control when the aarket for entitlements clears.
This is an important consideration, because trading can be used regardless of
the total quantity of entitlements initially allocated, or the distribution of
those entitlements among sources, and still lead to cost-effective control ef-
forts. The conditions under which trading is likely to work well, are examined
in greater detail in Chapter 111.
#
Trading approaches us«d in lieu of conuaand-and-control regulation can be
designed and implemented with the sane information as the system they replace.
Supplementary trading can be added to a coramand-and-control system with no new
information, although, as discussed in the next chapter, attention must be
given to ensuring that trading will not interfere with attaining the goals of
the coM»and-and-control regulations. Trading can provide information about
control costs that might not be available otherwise.
Equity
The equity implications of trading approaches that replace traditional
approaches are no different than those faced in command-and-control systems;
new regulatory requirements mean that decisions must be made about who pays
for controls and how much they pay. Marketable permits can be allocated in
the same way as the requirements of traditional regulations are allocated.
As discussed in the next chapter, entitlements can also be allocated in
market-oriented ways, such as through auctions. Although equity and allocation
issues are the same for command-and-control regulations, trading can give them
added visibility because money will change hands as entitlements are traded.
-------
-------
Supplementary trading should not affect the equity decisions that were
made when developing the conssand-and-tontrol regulations that form the basis
for trading. The initial allocation of entitlements will not be affected, and
all sources in similar situations can he given similar opportunities to trade.
Other Considerations
Trading does not require the government to establish a price for entitle-
ments or to adjust that price to reflect inflation or to offset the effects of
economic growth. Prices will be set by supply and demand for entitlements in
the market, will adjust automatically to reflect inflation and increased demand
as a result of economic growth, and will also adjust as a result of technolog-
ical improvements in production as well as control technologies. Consequently,
the government need not, and indeed should not, directly intervene to affect
market prices of entitlements.
Trading stimulates technological change because a market for entitlements
gives all sources an incentive to find ways to reduce emissions further to
make entitlements available for sale.
-------
HI
DESIGNING AND SELECTING
SPECIFIC ECONOMIC APPROACHES
The general discussion of charges and trading in the Introduction and in
the previous chapter provides a starting point for matching broad categories
of economic approaches to goals. This chapter begin by drawing this material
together to suggest in general terms when particular approaches might be appro-
priate, The chapter then identifies variations in the design (major variables
or parameters) of charge-based and trading approaches, Identifies the informa-
tion required to use each variation, and discusses in detail the implications
of particular design variations. No attempt is made here to characterize par-
ticular approaches or potential design variations as desirable or to identify
what can and cannot be done under the current statute: these conclusions are
contained In Chapter V.
This section attempts to identify the most important factors in choosing
an economic approach. Stand-alone charges set equal to marginal social costs
(the economist's classic case) is briefly discussed first. The use of charges
to pursue air quality or emission targets is then compared to the use of trad-
ing approaches to pursue the same goals.
Choices among economic approaches need not be completely constrained by
the exlsteace of a previous regulatory program, because some adaptation is
possible. Controlled trading or supplemental charges can be used if a regu-
latory system already exists and is not to be replaced; marketable permits or
stand-alone charges can "be used where new control systems are needed. There-
fore, some form of charge, either supplemental or stand-alone charges, can be
used under all circumstances. Expansions of controlled trading and marketable
permits also act as complements to one another, so one of these approaches can
also be used in all situations.
4.1 CHARGES SET EQUAL TO MARGINAL SOCIAL COSTS
Where the marginal social benefits of emissions are. known or are being
xaplicitly estimated, no economic approach, and of course no other control
-------
111-2
mechanism, is more economically efficient than stand-alone charges, because
this approach leads to optimal emissions at minimum total cost. Allocation
and equity issues are. avoided; each source is required to pay for its damages
at the margin. Money transfers to the government make the hidden costs of
pollution apparent to businesses and contribute to a better allocation of all
of society's productive resources. Crude estimates of marginal social bene-
rits are Implicit in some current air pollution control programs, and sore
frequent and better estimates will become important as the stringency of con-
trols and therefore control costs increase. The use of charges may therefore
become more practicable in the future.
If marginal social benefits are not known, stand-alone charges can be used
to pursue a target level of air quality or total emissions,
A.2 CHARGES AND TRADING TO PURSUE TARGETS
Either stand-alone charges or trading can be used when a target level
of emissions or air quality is sought. The major difference between these ap-
proaches is that charges cannot directly control total emissions while trading
approaches can. If this attribute of charges is unacceptable, trading should
be preferred over charges. If it is acceptable, other considerations are
important when comparing charges and trading.
Stand-alone charp.es are better able to reduce control costs than trading
where markets for trading entitlements are very thin. However, charges are
less appropriate than trading where ambient concentrations or emission re-
duction goals must be met because charges cannot directly and certainly con-
trol total emissions In the short run. The desired responses of polluters
can, of coarse, be achieved in time as rates are adjusted up and down depending
on associated emissions or air quality.
Charges are more likely to be seen as fair, and therefore should be more
politically acceptable, where they can be set at uniform rates for all sources.
This means that charges are aost likely to be suited to pollutants for which
emissions are assumed to be representative of ambient impacts and to pollutants
the impacts of which are relatively insensitive to source location and other
source characteristics. Seasonality or time-of-day characteristics must be
-------
III-3
carefully analyzed la considering particular pollutants as candidates for
charges. Finally, charges are not well suited to situations where marginal
control costs and emissions levels after controls are In, place will be high.
In this situation, the rates to induce appropriate responses are likely to "be
/
high, and transfers to the government for emissions which cannot be reduced
will be large,
Trading can have important advantages over charges in certain circum-
stances, particularly where Information is limited and costs are highly vari-
able between sources. Trading approaches do not require information about
control costs, because entitlements prices are set in the market. Trading
approaches require less adjustment over time than charge-based approaches
because they adjust automatically to changes in the supply of and demand for
entitlements, to inflation, and to changes in control costs. Trading ap-
proaches are, therefore, superior where control technology is evolving or
regional economic growth or decline is relatively rapid and air quality goals
are stable.
The major limitation of trading approaches is that they can function well
only where significant trade-off opportunities exist between sources and where
the market for trades between firms is not too thin. Several factors can
contribute to an active and broad-based market. More trading will occur if
major continuous point sources with different control costs account for a
large portion of emissions and if firms are in the process of making control
decisions when the trading option is established. Such decisions are more
likely if new sources are entering an area or environmental requirements are
changing than if industrial activity is stable and adequate controls are in
place.
Trading will be easier to implement if a large portion of emissions in an
area are reasonably interchangeable, because under such circumstances Impacts
are not highly dependent on source characteristics and location. However,
even if air quality impacts are highly variable and information is poor, trad-
ing approaches can still have an advantage over charge-based and command-and-
control approaches. Trading allows and encourages sources to seek out cost-
reducing opportunities and allows the agency to look at relative impacts for
one pair of traders at a time. While charges also encourage cost reductions,
-------
I1I-4
neitner charges nor commar).d-and-con£rol approaches can easily make use of such
partial information on relative air quality impacts.
B. SPECIFIC CHARGE-BASED APPROACHES
' '' " ' """rni, rrr nmr
Stand—alone charges, supplemental charges, and the use of proceeds from
such charges, are discussed in turn below.
B.l STAND-ALONE CHARGES
Stand-alone charges can differ (a) in the principle used to set: rates;
(b) as to how that rate is applied; and (c) as to which sources are affected.
late-Setting Principle
Stand-alone charges may be used to bring marginal control costs and
damages into balance, to pursue ambient air quality goals or to induce use of
particular control technology.
If the information were available, rates for stand-alone charges could
(and ideally would) be set equal to the marginal social damages from emissions.
The goal of this approach is to reach an equilibrium level of pollution abate-
ment in which the charge rate, marginal social benefits frotn abatement and
marginal control costs are equal. Any control approach (charges, command-and-
control, or trading) can reach this goal if social costs at an equilibrium
level of control can be estimated and marginal control costs are known.
Charges are the simplest way to achieve this goal when information ls avail-
able. Other approaches will usually pursue less ambitious goals.
Estimates of marginal social costs may involve impacts on air quality
and the relationship of air quality to damages to structures, crops and indi-
viduals, or quantification of effects on aesthetic values. This information
is more likely to exist, where the marginal social costs of emissions are con-
stant over a wide range or where a sharp change in marginal social costs occurs
at some level of total emissions. Where marginal social costs vary gradually
with the level of total emissions, information on control costs will also be
-------
111-5
needed to predict the equilibrium level of emissions which will result from a
given charge.
To reach a specified emissions or air quality target with a charge system
the control agency must know how particular sources will respond to different-
levels of charges. If the timing of control efforts is important, charge ap-
proaches will require at least as much information about sources as coramanel-
and-control approaches, and will require more unless the coamand-and-control
approach was sensitive to control costs. This information is likely to be
more expensive and difficult to collect where charges are imposed In a new
area than where pre-existing requirements ate being replaced. Where informa-
tion on anticipated responses is not sufficient, more information may be
accumulated through trial and error by varying charge rates, experimentally
or according to an announced schedule. Experimental variations would take
more time, and the uncertainty would make business planning difficult. Control
cost curves would change after imposition of a regime of controls, making the
determination of the necessary variation in charge rates more difficult. Such
charges would also probably encounter considerable opposition la the political
arena. Uncertainty can be reduced if control authorities find acceptable any-
total emissions outcome within the likely scope of a specified charge rate or
schedule. Use of announced schedules of rate escalations is possible where
it is acceptable to delay reaching environmental goals. Schedules provide a
better basis for business planning, but complete predictability is not possible.
Even if charges could be increased on an announced schedule until targets were
attainedt deviations from the schedule would then he needed to avoid over-con-
trol and to offset both the effects of inflation on costs and the effects of
economic growth on total, emissions.
Application of tote
Rates under a stand-alone charge may be applied to emissions, to ambient
impacts, to the performance or non performance of maintenance procedures which
affect emissions, or to any other factor (such as production levels, raw mate-
rial inputs, or source characteristics or locations) which is closely corre-
lated with emissions or air quality impacts. Where charges are related to
emissions, emissions rates or levels can be established through direct monitor-
ing or by reference to design and operating parameters which affect emissions.
-------
T.II-6
Where charges are keyed to ambient impacts, relative impacts on air qualitv
could be predicted by air quality modelling and measured by monitors. If
charges are used to control local pollutants, the rates per unit of emissions
will typically vary from source to source.
Sources and Emissions Affected
Charges may he assessed against all emission sources of a pollutant or
only against selected sources. Selective application could be based on the
need fcr control by a particular source (e.g., for pollutants with local im-
pacts), or on the costs of control by particular classes of sources (if envi-
ronmental goals can be met by controlling only some sources). Use of any of
these bases requires substantial information, although only more detailed
cost information would be necessary in addition to the data that should be
collected to design traditional approaches.
All emissions by a source may be subjected to charges, or charges may be
confined to emissions in excess of some base level for each source. The use
of a threshold before charges apply can be appropriate where the environmental
goal is to encourage a known amount of further emissions reduction. It will
be as difficult and costly to set thresholds for charges as it would be to set
traditional regulatory requirements for each source. The necessary information
about sources is more likely to be available where thresholds can be "based on
preexisting regulatory requirements which will typically already have catego-
rized sources in a Meaningful way.
Selective application of charges across source categories or continuing
charges to emissions in excess of a base level can also reduce the money trans-
fers from sources to the government.
B.2 SUPPLEMENTAL CHARGES
Supplemental charges that encourage controls beyond required levels have
design considerations similar to stand-alone charges. Supplemental charges
are likely to be useful whenever control relies primarily on source-specific
standards either to induce compliance with regulatory requirements, to place
a cap on marginal control costs or to provide credibility for technology-forcing
-------
III-7
standards. Each case is discussed below. In addition, the Implication of com-
bining supplemental charges with trading is discussed.
Suppleaental Charges that Induce Compliance
Compliance-inducing charges can be "highly effective solutions to the dif-
ficult problem of making it in the economic interest of sources to reduce
emissions to a target level. Charges which are intended to induce compliance
tan be set at a uniform level, or be set separately for each source, and can
he based on different economic criteria. While payments are not based on a
measurement of emissions, EPA's existing noncompliance penalties for stationary
sources induce compliance by assessing the source-specific savings from de-
laying compliance with regulatory requirements. Planned nonconformance penal-
ties for heavy-duty trucks will use a uniform charge rate related to the mar-
ginal costs incurred by competitors to come Into compliance with regulatory
standards; the initial cost-based rate will escalate over time to provide
stronger incentives to comply. The total amount of the penalty will of course
also be greater when emissions exceed standards by greater degrees. Culpabil-
ity and damages need not be assessed to set these economically-based fees, but
control costs must be known. This is a comparatively straightforward process
because fees can be adjusted after actual control costs have actually been in-
curred to come into compliance.
Supplemental Charges that Provide a "Safety Valve"
Supplemental charges could also be used to set a ceiling on marginal con-
trol costs incurred, by providing sources the option of paying a fee as a
short- or long-run alternative to compliance. (While a regulatory variance
procedure accomplishes the same purpose, a charge can be tailored to achieve a
more efficient results.) A fee set to serve these functions would typically
be uniform for all excess emissions. A supplemental charge should be set low
enough so that at least a few sources would pay the fee for a short period of
time. A fee set at this level would act as a "safety valve" in case control
requirements prove to be infeasible or too expensive for some sources. This
can be particularly important if. control obligations are so costly or difficult
to meet that some sources would otherwise unexpectedly shut down. Alterna-
tively, the fee could be set at a lower level to ensure that environmental
-------
II1-8
benefits and economic costs were "balanced for most sources, by capping marginal
control costs.
Supplemental fees set at a moderate level to provide a safety valve, and a
stand-alone charge for emissions in excess of som« baseline levels are in some
respects equivalent approaches. Both seek to balance environmental objectives
and costs by relating control effort to marginal control costs. However, the
supplemental fees are merely an overlay to a set of regulatory requirements
and trading opportunities that may be sufficient to attain air quality goals
even with the safety valve in place, while the stand-alone charges provide no
assurance that minimum environmental standards will be met at any particular
time.
Combining Trading with Supplemental Charges
The existence or nonexistence of trading should be considered in using a
supplemental fee. If control Led trading is not worki ng well enough, some
sources may have no practical ability to come into compliance when their own
control costs are very high or control is not feasible. If trading is working
well and emissions reduction credits are available for sale, the opportunity
to purchase entitlements or "credits"! can serve as a safety valve for cost
and feasibility pressures. If trading is permitted, the implications of a
safety valve fee will depend on the level of the charge relative to the price
of any credits available for sale. In the absence of a fee, credits will
change hands at a price which reflects—and balances—supply and demand. If a
safety valve charge is set at a price in excess of the market price for credits
the safety valve becomes operationally irrelevant; sources will either control
their emissions or purchase credits in the private market rather than pay the
charge. (The charge may still serve to convince sources that the agency in-
tends to stand by the requirements that it imposes.) If the charge is set be-
low the market price for credits, it will provide added flexibility to sources.
Less control will be undertaken and there will be substantial effects on
1 Credits are salable entitlements to emit, which are created by reducing
emissions below the levels required by applicable regulations.
-------
111—9
trading. Less trading will occur than otherwise, as some sources pay Che
charge rather than purchase emission reduction credits. A new and lower market
price for credits will be established to reflect this reduced demand for cred-
its.
B.3 USE OF FEE PROCEEDS
The revenue generated by emissions charges could be returned to the U.S.
Treasury, or be used to finance agency operations, to purchase emission reduc-
tions from sources, or to compensate those who are damaged b> pollution.
In general, revenues from charges will decline as industry adjusts to the
charge by reducing emissions. If revenues are to he made available to fund
the operation or programs of the control agency (this would require statutory
authority), there may be an incentive to prevent this decline, creating a con-
flict of interest, to avoid this, proceeds can be returned to the Treasury.
This would be consistent with common fiscal policy practices, which typically
separate spending decisions from revenue collection.
Government purchase of emission reductions using fee proceeds duplicates
a trade in which one source pays another to generate an emissions reduction,
But governmental willingness to trade purchase reductions (whatever the source
of revenues) can disrupt private trade. If sources have responded to a charge
by reducing emissions or if trading has resulted in establishment of a market-
clearing price for credits, the government will not be able to secure any addi
tional reductions at a price per unit equal to that iaplied by the fee or the
market equilibrium price for credits. If the government is expected to offer
higher prices for reductions than are available in the market, the private
market for trades will inevitably be disrupted as potential sellers of credits
withdraw to await government intervention.
Compensation payments attempt to replace a private transaction in which
polluters pay victims to accept the consequences of pollution. If such trans-
actions could in fact be arranged and were preferred by all parties, payment
of compensation in lieu of further control would be more economically effi-
cient. Use of fee proceeds to compensate those who are injured, or to mitigate
the effects of pollution, could make higher levels of emissions politically
-------
Ill-10
tolerable. A program to compensate those harmed by pollution is currently
operable in Japan.
As shown above, the use of tee proceeds to compensate those who are
damaged by pollution, or to purchase emission reductions puts the control agency
in the position of attempting to replace private transactions. This is not a
role the government necessarily should play, or is likely to play well. For
example, compensation transactions are desirable if they occur as a result of
private exchanges, because there are no forced sales of anyone's rights and
because preferences are respected even if they are unusual, An attempt by
government to duplicate such transactions would at best result la "fair™ com-
pensation, not in an outcome where all participants felt better off. Similarly,
private trades of entitlements result in equal marginal control costs and cost-
effective control. Bat if the government becomes involved, some sources would
fee required to pay the government a fee or reduce emissions, while others would
be paid a fee by the government to reduce emissions. Moreover, the government
would in sons cases pay more for reductions that it was willing to see sources
pay themselves. The governaent discouraged high cost control by providing the
alternative of paying a fee. Some subsidy might well be in order where ex-
tremely high cost control was socially justified, bat the use of a fee to
generate revenues would be a clumsy and disruptive procedure,
c. specific. mmae j®?ioacbss
c.i REVIEW OF CONTROLLED TMMMG AND MARKETABLE PERMITS
In this report, trading approaches have been formally divided into "con-
trolled trading" and "marketable permit" groupings, but formal distinctions are
probably less important than practical differences and potential similarities.
In practice, trading approaches fall on a continuum where the importance of
trading increases as restrictions on trading decrease. Either controlled trad-
ing or marketable permits could in principle be implemented at either end of
this continuum. Formally, the approaches arc distinguished by the manner in
which the asset that is traded is created. With marketable permits, a central
agency creates and distributes entitlements or "penalts" to emit quantities of
a pollutant. With controlled trading, sources create entitlements of "credits"
by reducing emissions below levels set in baseline regulatory requirements.
-------
111-11
Of course, this baseline can be specified in the same way that permit alloca-
tions were specified.
Because controlled trading assets are defined and measured by reference to
a regulatory baseline, it is the baseline program rather than trading which
will be central to the goal of environmental protection. Controlled trading is
firmly rooted in a set of regulatory specifications which could stand without
(and usually will have pre-dated) trading. 4s a practical matter this context
usually means that controlled trading programs involve many restrictions. This
may be less true with permits, but no history exists and theory suggests that
restrictions would play identical roles under either approach.
The most familiar "controlled trading" approaches are EPA's existing
offset trading program and bubble policy. These particular controlled trading
approachesj which are still evolving, are discussed more fully in Chapter IV.
As described earlier, an offset program allows new facilities to be constructed
in nonattainment areas, provided emissions which reraain after a new source has
installed specified controls are offset by reductions in emissions from exist-
ing sources. The bubble policy allows existing plants in attainment areas to
increase emissions at some sources to levels In excess of baseline 3IP require-
ments, in return for reduced emissions at other sources.
In their "pure" form, marketable permits are envisioned as replacing
command-and—control systems where the latter exist. Emissions covered by
permits are allowed, and all emissions which exist must be so covered. Thus,
marketable permits are usually viewed as a complete and distinct approach to
the control of eaissions rather than a means of reducing the costs incurred by
sources in meeting a baseline set of cotnmand-and-control requirements. The
pure form of marketable permits is most appropriate where no control program
has been established or if it appears feasible to replace the existing control
program.
When a conmand-and-control regulatory program is already in existence (or
can be created) the usefulness of a marketable permit approach will depend in
practice upon the extent to which an alternative controlled trading approach
could effectively approximate the characteristics of permits. If states have
the authority .and will to allocate control obligations in a manner that will
-------
IIT-12
assure attainment, and restrictions on controlled trading are minimized, per-
mits are not needed to control emissions or to allow additional control cost
reductions. But operational differences between controlled trading and "pure"
marketable permits or political ability to use particular characteristics in
different contexts can affect the way in which these trading approaches are
actually implemented. In some cases, permits say be the more attractive alter-
native simply because they offer the possibility of a clean slate when begin-
ning the regulatory process or introducing trading opportunities* Finally, the
need to explicitly allocate permits suggests that "pure" marketable permit
approaches are better suited to situations where a major change in regulatory
requirements or control technology is taking place than to stable control
situations. A changing situation provides an occasion for a new initial
allocation of entitlements and avoids the disruptive effects of imposing a
new formal structure on a stable control situation,
Co 2 VARIATIONS IN CONTROLLED TRADING AND PERMIT APPROACHES
Controlled trading and permit approaches have several design parameters.
This section discusses five areas in which choices of characteristics are
possible, and the implications of particular choices in each of these areas.
In most cases choices of characteristics in each area can be made indepen-
dently of choices_ in the other areas; thus a very large number of potential
trading approaches exists.
The five areas discussed are (a) the existence and scope of a banking
program; (b) agency involvement in market operations; (c) the determination of
equivalent emissions; (d) restrictions on trading—which sources are allowed
to and choose to trade, which emissions reductions are creditable, and how
emissions reduction credits may be used; and (e) how eaissions entitlements
and assessments requiring additional emissions reductions are allocated. The
implications of different characteristics in each of these areas will change
as a trading approach evolves, so that designs which are appropriate when
trading plays a limited role can become inappropriate as trading becomes more
common.
-------
111-13
Banking
"Banking" gives sources a right to sell or use* emission reduction credits
or entitlements at a time most convenient to the source. In the absence of
banking, trades must be contemporaneous, which presents a serious barrier to
trading transactions, regardless of the other characteristics of a control
system that included trading. Banking "rules" may set out the conditions
under which a control agency can restrict the future use of banked credits.
The issues involved in such actions by a control agency are discussed in a
later subsection on the allocation of additional control obligations. This
subsection discusses the useful consequences of banking that should be recog-
nized when designing trading approaches. Chapter IV contains a more detailed
description of EPA* s existing hanking activity.
If emission reductions must occur at the same time as the emissions in-
creases which they offset, trading can be very difficult for both buyers and
sellers* Without bankings would-be buyers of credits must locate sellers and
may find the delays this could involve unacceptable from a planning or finan-
cial standpoint. Without banking, sellers must wait to generate credits at a
time that is convenient for buyers rather than appropriate to their own opera-
tion and investment plans, or generate credits without complete information on
how the agency will treat the credits in the future. Because it is risky to
make financial commitments when other actors are not bound, credits will typi-
cally not be deliberately created In jurisdictions without banking in anticipa-
tion of sale or use, unless a buyer or an immediate use for the credits has
been identified and the control agency has been sounded out. Even emission
reductions which would be economically justified might be deferred until a
simultaneous exchange can be worked out to make use of valuable reduction
credits.
Banking 'eliminates this unnecessary timing restriction and reduces uncer-
tainty by allowing the size, ownership and type of emission reductions to be
certified when a redaction is undertaken, and allowing the credits to be held
1 Under the current banking proposals, use must be approved on a case-by-
case basis.
-------
111-14
for use or sale at a later time. This aak.es it possible for the trading system
to induce creation of surplus redactions, and avoids three practical proglems
with trading. First, investments ia controls tend to be lumpy in size and
constrained as to timing; banking allows firms to create surplus reductions
when convenient (e.g., when some new equipment must be Installed) for sale
when possible.^ Second, banking nay be used as an occasion to focus market
activity, by channeling reductions through a central institution; this can
reduce the transactions costs involved in finding and negotiating with poten-
tial suppliers of emissions reduction credits. Finally, with banking, emis-
sion reduction credits may be readily available for sale, or a firm might be
able to buy a future reduction or an option on a future redaction now. This
reduces uncertainty about the costs of meeting environmental requirements
and thereby facilitates business planning.
Market Involvement by the Control Agency
Some involvement by the air pollution control agency in trading trans-
actions will usually be unavoidable, but the scope of Involvement can vary
widely and perhaps should vary depending on the maturity of the market. Ini-
tially, an agency might see a need to help establish a market or to help fi-
nance emission reductions. However, financing reductions will be less of a
problem when the market value of credits is clear from experience, and mature
markets will be able to sustain themselves with less agency involvement than
new markets. Since agency involvement can hinder operation of a mature market,
agencies- should do no more than is clearly necessary.
The least amount of involvement that Is likely to be feasible would be to
do nothing more than certify entitlements for sale. The least centralized
1 This coin lias two sides. By eliminating timing restrictions, banking
can permit the sale of credits which would otherwise have lapsed under applica-
ble rules upon shutdown of an uneconomic or obsolete facility; in such cases
flexibility prevents a net emissions decrease. But these cases may be rare.
The timing of source shutdowns can be delayed to permit sale of the credits
even, in the absence of banking. By eliminating the incentive to delay shutdown
in this more typical case, banking can lead to reductions ia total emissions
from earlier closures.
-------
Ill—15
form of market operation would then allow free trading of certified entitle-
ments between sources, perhaps with the assistance of private brokers. Because
brokers could profit from arranging trades, they could provide valuable and
specialized services that would "help establish the controlled trading approach,
and could assume the financial risks involved in purchasing an inventory of
emissions reductions tor future sale. The agency would still operate any bank-
ing system, and might set any necessary exchange ratios for trades between
sources to protect air quality.
The smallest (and a likely effective) step an agency might take toward
centralizing market functions would be to act as a clearinghouse for informa-
tion about potential sellers and users of emission reduction credits. The
next step would be to become directly involved in the initiation of private
trades by encouraging sources to reduce emissions in order to generate en-
titlements for banking and by helping potential buyers and sellers locate one
another.
in a fully centralized market an agency would establish itself or some
other entity as the sole buyer of entitlements generated by emissions redactions
and as the sole seller of entitlements to offset increased emissions. Numerous
variations on this approach exist. Once acquired, credits or permits could be
sold at a prespecified price, at a negotiated price in particular transactions,
or at periodic auctions. Use of a centralized purchase approach could provide
a means to guarantee a market to those that generate emissions reductions and
help to establish a controlled trading approach quickly. However, attempts to
encourage the market in this way would expose the control agency to culpability
for financial risks which are not usually accepted by government agencies.
Attempts to limit these risks could lead to purchase prices that were lower
and sales prices that were higher than a market-clearing price. Some useful
trades would not occur under these circumstances. In extreme cases, an agency
might take and give credits rather then buying and selling them, either by im-
posing control requirements in excess of those needed for attainment or by con-
fiscating credits when a source reduces emissions.
-------
ra-16
Establishing the Equivalcnce of Emi. s s i on s
Trades of equal quantities of emissions from sources in different loca-
tions or with different charact'erist ics can result in adverse effects on air
quality, Where ambient concentrations represent the environmental goal an
agency must ensure that emissions which are traded have equivalent air quality
implications. In the simplest trades presented to an agency for approval the
sources of emissions will he very similar (e.g., emissions from stacks of near-
ly similar height) and the overall ambient impacts of emissions from the trad-
ing sources will be the same as long as the overall emissions level does not
increase. A unit for unit exchange of emissions of the same pollutant between
such sources would be acceptable. More complicated cases make it more diffi-
cult for an agency to assure that ai r quality after a trade is equivalent to
air quality prior to that trade. Mr quality modelling is usually used to
measure equivalence when sources have different impacts on air quality.
Equivalence questions can also arise because the relationship of control
efforts to emissions is uncertain or because the relationship of emissions to
ambient impacts Is not well understood (i.e., available models work imper-
fectly)- Differences in the timing of control efforts or in the reliability
and enforceability of proposed control measures can also raise questions.
Even if these questions were addressed adequately for baseline emission levels
and control approaches, they must he addressed again when trades which will
change these factors are proposed. In other cases, companies may propose
emission reductions from previously unregulated sources of emissions that are
not well understood or may propose use of technology that does not have an
established record of control efficiency.
Most questions about equivalence can be addressed by requiring demon-
strations of proposed approaches before trades are permitted or before a final
exchange ratio Is specified, or by building aa air quality cushion into the
trade by setting an exchange ratio that reduces total emissions.
Uncertainty regarding the government's judgments about equivalence can
impede trading because sources will not be able to confidently assess the
value of the commodity to be exchanged. flume sources may be discouraged
from attempting trades if there is any need to reveal the existence of a
-------
Ill—17
control opportunity to the control agency before a trade has been worked out,
To the extent that control agencies can reduce uncertainty and give source
tools which can be used without agency assistance, trading will be facilitated.
Restrictions on Trading
Greater cost savings will be achieved with a trading approach if trading
is maximized. Ideally every emissions reduction that would not otherwise occur
should generate a tradeafale credit, and reduction credits or permits could be
used (once an equivalence ratio was established) to offset any emission in-
crease. Restrictions (such as conditions or bans on the use of credits by new
sources or by sources in nonattainaent areas) nay be used in some trading
approaches, however, either as a result of historical development or as an
indirect way to keep trading from interferring with regulatory strategies.
The subsections below discuss restrictions on trades of reductions generated
by particular sources and restrictions on use of redaction credits by partic-
ular sources. In general, alternatives to restrictions exist in the form of
clearer specification of baseline obligations, or direct action to offset the
consequences of trading,
o Restrictions on tradeable reductions
Existing regulatory programs contain significant restrictions on the use
of emission reductions generated by certain sources; in some cases these re-
strictions are cast In terms of requirements that reductions be "real" and
"permanent," In practice this often means that reductions which were relied
upon in demonstrating attainment under the state SIP (usually from the shut-
down of an existing source or from plans to Impose more stringent controls on
existing sources) could not be used to offset longer-lived emission increases.
In other cases restrictions prevent trades of credits generated by emissions
reductions which might be quickly replaced by emissions from a new entity or
new facility which replaces the seller. For example, a company which owned a
source with uncontrolled emissions might negate the effectiveness of its agree-
ment to reduce emissions at that source through a sale of assets to a new
company. 'Or gains from reductions achieved by shutdown of a facility for
which there is undiminished market demand could be neutralized if an equivalent
facility could be established to meet demand without the purchase of offsets.
-------
111-18
Restrictions based on the concern that reductions be real and permanent
have their base in uncertainty about the ownership of entitlements when a source
shuts down. More generally, the issue involves definition and measurement of
the entitlements implied fay pretrading control obligations in a dynamic situa-
tion. If permissible emissions were well defined both as to quantity and
time for ail existing sources, and all new sources were required to secure
entitlements to offset their emissions, restrictions on the sale of credits
generated by nonobligated emission reductions could be avoided.
The ability of a control agency to eliminate restrictions on trades of
certain credits is directly related to its ability to specify ownership of
entitlements in likely future situations- This proved to bs possible when the
alternative was blocking economic growth In nonattainraent areas. Although new
obligations may need to be Imposed on existing sources in these areas to dea-
onstrate attainment, existing sources are able to sell credits as offsets to
new sources when they reduce emissions below currently required levels.
Son® restrictions on trades of particular emission reductions are simply
based on applicable statutory provisions. For example, reductions to meet
control obligations must now be achieved by using continuous rather than in-
termittent controls; this restriction has been extended to the use of intermit-
tent controls to generate credits for sale. This contributes nothing to air
quality if intermittent controls would result in real, permanent and enforce-
able emission reductions.
o Restrictions on the Use of Credits
Under Lhe current statute, credits cannot be used to exceed technology-
based requirements such as NSPS, LAER. or BA.CT. Because these are stringent
and high-cost control requirements, this restriction is a major barrier to
realizing the full cost-reducing potential of controlled trading. The re-
striction has no effect on emissions in the short-run, since trading does not
increase or decrease allowable emissions. Stringent statutory standards are
instead apparently imposed on new sources in an attempt to preserve or slow
the use of available air increments for economic growth and to assure con-
tinuing air quality improvement--regardless of ambient standards—as the capi-
tal stock turns over. These goals were described ^ore fully in Chapter T.
-------
111-19
However, trading by new sources need not be prohibited to ensure that these
regulatory goals are achieved. Some restriction on offsetting actions will he
needed in some situations f but these can he kept to a minimum.
Restrictions on the use of credits by -new sources can better "preserve"
air quality for future growth only if they make imposition of added require-
ments that may be desired at a later time more feasible. This is a doubtful
proposition: at some cost, existing sources would be able to reduce emissions
in the future, providing room for economic growth, and the new source which
purchased entitlements could also install more controls later. It is some-
times argued requiring physical controls on the new source today will be easier
because it would be more expensive to retrofit controls later*, however, any
potential savings is offset by the certain expense of denying sources an op-
portunity to install the most cost-effective controls today.
If future economic growth can be accommodated, there is less need to be
concerned with emission reduction as the capital stock turns over. In any
event, restricting use of the air resource by new sources will contribute to
emission reductions as the capital stock turns over only if credits sold by
existing sources and credits generated by shutdown of existing sources are
treated inconsistently. If credits may be sold even when sources shut down it
makes no difference when in a source's lifetime credits are sold. Similarly,
if existing sources are only able to trade explicitly temporary entitlements
to new sources and cannot sell credits when they shutdown, restrictions on the
use of such temporary credits by new sources play no role In reducing emissions
over tine. In the first case, credits live on after the existing source re-
gardless of the timing of any trades. In the second case, credits cease to
exist after the source shuts down, regardless of any trades. It is only whan
credits which are sold become permanent and credits which are not sold are
forfeited upon shutdown that any rationale for a restriction exists, tod even
in this case other means exist to meet targets for air quality improvements
over tine.
-------
111-20
Allocation of Entitlements and Assessments of Additional Control
Obligations
The allocation of entitlements and assessments imposes costs or confers
benefits on particular segments of the population; thus the primary issues
involved in the allocation are the distribution of wealth and income. If
trading works well, allocation decisions will not adversely affect environ-
mental goals, economic efficiency or cost-effectiveness. Trading can also
increase the political acceptability of different approaches to the alloca-
tion of entitlements and additional control obligations, since trading allows
sources to adapt to agency allocation decisions through trades before install-
ing controls. The adaptive potential of trading means that the way in which
entitlements are allocated ts not as important as assigning all entitlements
to some source and allowing trading to assure that all entitlements are pro-
perly valued. There are several approaches which can be used to allocate
entitlements and to assess additional control obligations: initial allocation
can be taken from the distribution of emissions as reflected in the baseline
regulatory program if one exists, or it can be managed by relying on agency
discretion, on rules, or on an auction. Discretionary, rule-based and auction
allocation of entitlements are each discussed separately in the subsections
Immediately below. The fourth subsection discusses the use of discretionary
and rule-based approaches to allocate additional control obligations ("assess-
ments") subsequent to the initial allocation.
o Discretionary Allocation of Entitlements
Initial entitlements could be allocated by an air quality control agency
in any manner it sees fit, either without charge or upon payment of a fixed
price, and without conditions or with ancillary restrictions attached. An
agency choosing discretionary allocation could adopt as its guiding principle
equity or fairness, extrapolation of existing emissions based on a regulatory
baseline, cost minimization, a matching of control obligations with the
ability to pay for controls, or local or agency preferences for particular
kinds of economic activity. These same concerns can play a role in discre-
tionary allocation of additional control obligations. If allocation is based
oa any criteria other than cost minimization, trading will be extensive as
-------
Ill— 21
sources adjust to control costs and trading will therefore play a major role
in reducing total control costs. To the extent allocation is sensitive to
relative control costs, the need for subsequent trades to minimize costs will
be reduced. Direct discretionary allocation avoids the problem of large pay-
ments by sources to the government.
Discretionary allocation, can encourage a focus on technical consider-
ations or on the ability of a source to afford controls on its own facilities.
These concerns are important where trading does not exist but should become
less important when"trading works well. Trading allows a source which cannot
reduce Its own emissions further, given current technology, to pay another
source to generate reductions. Similarly, a source which cannot afford to
install additional controls on its own facility may he able to afford lower
cost controls elsewhere.
Discretionary allocation require information about particular sources to
provide some basis for allocation decisions; thus, these decisions will be
expensive to make and will reflect any biases or flaws in the data used. On
the other hand, a discretionary approach can make use of all relevant and
available information; if information is good, decisions can provide for fair
allocations that will meet air quality goals with a minimum of economic dis-
ruption and at a low total control cost.
o Allocation of Entitlements by Pre-established Rules
Rule-based allocation of initial entitlements will typically condition
receipt of an entitlement on some action by a source or potential source. The
most common current rule is first-come, first-served allocation or allocation
on demand, which simply requires a source to get in line to obtain permission
for constructing a facility that meets certain conditions. This approach is
used in PSD areas and is implicit in mobile source regulations. First-come,
first-served allocation is simple, familiar and easily understood. It is
perceived by some to be "fair", based on the notions that either everybody has
an equal chance to be first in line or that those who want the resource most
will make the effort to be first, and thus deserve the first chance. On the
other hand, the first-come, first-served rule as now used in PSD areas involves
Inefficiencies that might be avoided by other rule-based approaches. In part
-------
111-22
this is because sources must simply Install some controls and be first in line
to receive entitlements and in part because assigned increments are not readily
tradeable. Thus, there is no opportunity for other potential new sources to
indicate that they place a higher value on available entitlements until after
the source which is granted entitlements without charge is constructed. Only
at that point can the source which holds the entitlements resell them, as an
"existing" scarce. Finally, the first-come, first-served rule generally ap-
plies only to those sources which need but lack entitlements. Some sources
may be exempted or may already have sufficient entitlements, so that incen-
tives to reduce emissions in order to preserve air increments will not be
everywhere the same.
Pre-established rules which could act as alternatives to first-cone, first-
served roles include rules which require other positive steps by sources to
reduce emissions. For example, the offset policy now in force in nonattainment
areas could be adapted to the needs of clean air areas by requiring new source
purchase of partial offsets prior to siting In clean air areas. This would in
effect gradually increase the price of the air as it became increasingly scarce.
Designing such an approach to achieve the greatest possible benefits would be
a complex task, and the approach would increase the costs to those who receive
entitlenemts in the near term. It oust also be pointed out that the markets
for offset trades in PSD areas may be thin.
Rule-based allocation of assessments can also rely on a formula to guide
allocation rather than on actions by sources. For example, the simplest form-
ula would apportion equal percentage reduction requirements to all existing
sources. However, a formula could be adjusted to reflect past control efforts
or be used to impose obligations only oa certain categories of sources (e.g.,
those that can afford additional controls).
formulaic assessaents in their simplest form (i.e., those based on equal
percentage reduction or "rollbacks" by ail sources) can provide an administra-
tively low—cost and less controversial way to allocate additional control
obligations; because no discretion is exercised, little information is re-
quired and there is no occasion for second-guessing agency decisions. The
opportunity to use a fixed simple formula also aak.es it unnecessary to be con-
cerned with the effects of trading on the ability of the agency to impose
-------
111-23
controls on particular sources. Simple formulaic assessments encourage trading
and are most feasible where trading works well, because sources can then take
technical capabilities and differences in control costs into account through
trading before controls are actually installed. However, if trading does not
work well, formulaic assessments will result in patterns of control effort that
are not cost-effective, since no account is taken of polluters' marginal control
costs. Cost-effectiveness will also be poor where emissions from different
sources have different ambient impacts, since unnecessary control will be under-
taken.
A formulaic approach may be difficult to impose in its simple form, espe-
cially on top of preexisting regulatory requirements. It may not be fair to
disregard previous control efforts nor wise to rely on trading to overcome
problems of technical feasibility nor politically acceptable to ignore ability
to pay. Ideas about equity that focus on outcomes will clash with ideas which
focus on changes. Where these concerns must be accommodated, the costs of
allocating assessments will increase and agency decisions will be open to
question. The importance of these concerns will vary. If existing control
requirements are believed to be essentially fair, a uniform additional assess-
ment may also be seen as fair. However, if some categories of sources have
been required to do more than others, a formula will be less acceptable.
Once the limits of technological capability are reached, a formulaic
approach places more reliance on trading. If trading exists only on paper,
defending technically impossible requirements by invoking formulas and trading
will be difficult. However, if trading is frequently used, sources will
recognize that assessments are essentially assignments of financial liability.
(This will be especially clear if a supplemental charge is usee.) If sources
which cannot afford additional control obligations are free to sell emissions
reduction credits upon shutting down, less importance might be attached to
financial feasibility;1 but if credits from shutting down revert to the
control agency (or if society is unwilling to face the prospect of reducing
economic activity to reduce emissions), financial feasibility aust be given
1 It will be important, however, to recognize the associated economic
impacts from shutdown.
-------
111-24
more attention. Therefore, it is important to consider characteristics which
promote the success of trading If a formulaic allocation approach is used.
o Allocation of Entitlements by Auction
Allocation by auction can take many forms. Auction may involve separate
and sequential sales of lots with the "high" bidder purchasing the lot being
offered on the block,I This is a traditional "English" auction. In the less
familiar "Dutch" auction, quantity orders are taken at successively lower prices
until the market clears, with all units sold at the clearing price. This system
has some attractive properties, which are explained below. Variations on and
alternatives to these classic "English" and "Dutch" systems abound. Different
approaches are likely to result in different costs to sources, in different
allocations of entitlements, and in the generation of different amounts of
information about control costs.
A slight variation of the Dutch-type auction can produce excellent infor-
mation about control costs* In this variation, sealed bids specifying a price
and a maximum quantity of entitlements would be submitted, The air quality
agency would fill as many bids as it could, all at the price of the last bid
it was able to fill. , So bidder would "bid more than entitlements were worth to
him because control would be cheaper than permits, and none would bid less than
the maximum amount entitlements are truly worth to them because the price paid
would be set by the last bid filled; a bid which was higher than necessary to
win would imply no excess costs to the bidder.
Because auctions bring market mechanisms to bear in determining the initial
allocation of entitlements, they will result in an initial allocation which is
closer to the minimum control cost allocation than other approaches. By bring-
ing traders together at a single place and tine, auctions can also help over-
come problems with thin markets. "Thin" markets could also be strengthened by
allowing potential new sources to bid for entitlements for future use- This
1 Because auctions determine the initial allocation of marketable permits,
the differential impacts of emissions from different sources may need to be
considered in comparing bids.
-------
II1-25
would increase the number of bidders, especially tf these entitlements could
he banked with some payments deferred until use.
Barring collusion, it must be expected that any auction will result in
substantial money transfers to the government. Transfers will he larger with
English-type auctions than Dutch-type "because the latter provide for bids to
be filled at a single, and lower, market clearing price. Because these trans-
fers occur industry may prefer allocation mechanisms other than auctions.
This could encourage criticism of auctions on the basis of problems which are
solvable.
Collusion or other strategic bidding can occur with auctions. Sources
may attempt to keep their costs low, or they may attempt to deny needed en-
titlements to competitors. The possibility of noncompetitive behavior de-
creases as the number of participants in the auction increases, Noncompetitive
outcomes can he discouraged by using price/quantity information obtained in
legally-binding sealed bids or by restricting the amount or percentage of en-
titlements that can be obtained by each "bidder. Of course, to the extent that
firms themselves face uncertainty about the future, they may miscalculate in
their demand for permits. A market in "futures" and "options" should minimize
any inefficiencies arising as a result of this situation.
o Allocation of Additional (Future) Control Obligations
After control obligations are initially assigned, trading approaches can
differ in the manner in which future assessments (additional control obliga-
tions) are imposed on existing sources, on trades and on banked credits. Such
assessments may be necessary to secure additional emissions reductions because
an agency has learned more about the ambient impacts of existing emissions or
because uncontrolled emissions have increased more than was expected.
The ability of an agency to impose additional control obligations can
either contribute to or jeopardize the success of a trading approach. If an
agency is secure in its ability to impose additional obligations when neces-
sary, there will be no need to rely on turnover of capital stock to reduce
omissions, and there will be less incentive to impose restrictions on how and
when emissions reduction credits may be created and used. Trading can be made
-------
II1-26
more widely available, and greater cost savings can be achieved. On the other
hand, unpredictable assignment of new control obligations can create enough
uncertainty in the private sector about "baseline" control requirements to
inhibit trading 2nd jeopardize the cost savings trading is intended to permit.
Trading approaches will work well only where sources are reasonably secure in
their expectations about the continuing value of entitlements. Trading can be
inhibited if an agency imposes assessments directly and specifically on already
banked credits or trading transactions, pressures sources to donate entitle-
ments to new sources to promote economic growth, or refuses to certify (confis-
cates) reductions from shutdowns. Imposition of assessments on trades and on
banked credits may occur where an agency atteapts to exploit opportunities to
improve air quality without increasing total control costs.
From the agency's point of view assessments on banked credits and on trad-
ing transactions which bypass banking can be attractive opportunities to improve
air quality. Reductions are clearly technically feasible, and the gains from
a trade may be great enough that a "tax" on the emission reductions generated
may not prevent some potential trades from remaining financially attractive.i
So loag as the source selling emissions reduction credits continues to profit
fro® the transaction, it may appear that no serious equity or efficiency pro-
blems are presented by this kind of practice.
In reality, direct assessments against trades or tanked credits imply that
the control agency has not yet succeeded in balancing air quality goals and
costs as it wished and intends to move toward a better balance by relying on
cancellation or discounting of banked credits, allowance of excessively small
1 Banked emissions reduction credits clearly reflect technically feasible
controls, which a source could afford at the time to undertake In expectation
of a profit without endangering the viability of its business. (Viability
could be endangered if credits cannot be sold.) In most cases the reductions
reflect relatively low cost controls which have already "been installed. Total
control costs and economic disruption may be lower if these credits are never
used or are used only partially due to an assessment than if other sources are
required to undertake additional controls to generate the same improvement in
air quality. Finally, the very existence of the credits is based largely on a
concern with reducing control costs, a goal which an agency may see as of
secondary importance, when air quality Is thought to be inadequate-
-------
II1-27
credits for emissions reduction, and "taxation™ of trades (through use of exces-
sively low offset ratios). These approaches are also discriminatory; only
sources which propose to trade have their stock of entitlements reduced. As a
result, sources may trade or batik less. Because banked and traded credits may
be based on innovative approaches Co pollution control, measures directed at
these credits nay have an adverse effect on innovation as well as on trading.!
Of course, there may be cases where, if trading itself is not discouraged, a
new source that oast obtain a 2:1 ratio of offsets reduction instead of a 1:1
ratio may have a greater incentive to economize on its use of the air resource
— and may innovate to do so.
1 An exception to these concerns exist where assessments against sources
affect banked credits or trades; changes in requirements imposed on sources
can affect the measure of "excess" reductions, and this may have to be traced
through to banked credits, reductions proposed for use in trades, or even
to sources which purchased credits with fair warning.
-------
IV
Ef A ON-GOING EXPERIENCES
AND STUDY CAPSULE SUMMARIES
This section briefly describes and assesses (1) EPA*s analyses, on-going
experiences with economic incentives, and plans to proceed with implementing
such approaches under the current statute and (2) EPA-sponsored studies of
economic approaches that have not as yet resulted in implementation efforts.
(These studies have also been used in preparing earlier sections of this report.)
A- BRIEF SUMMARIES OF STUDY RESULTS
EPA studies have examined charge—based and trading approaches to air
quality control, in some cases comparing both approaches to the same specific
problem. Exhibit IV.1 and If.2 identify these studies, and major conclusions
are summarized briefly below. A more detailed discussion of the methodologies
and implications of each study Is provided in the Appendix.
A.l STATIONARY SOURCE STUDIES
Eight recent studies which contain fairly detailed examinations of charge-
based or trading approaches to stationary source emissions control are briefly
summarized in this subsection. Seven of these studies were done for EPA and
one for the Council on Rnvlronaental Quality. Three studies examined both
stand-alone charges and trading approaches, two examined only one of these
approaches, and one study examined the economic impacts of noncompliance penal-
ties. These studies covered a wide range of control problems and economic
approaches, and they generally confirm the results which can be predicted on
the basis of theory; all of the stationary source studies (except the narrowly
focused study of noncompliance penalties) found that economic approaches offered
substantial benefits over command and control approaches in terns of the effec-
tiveness or the costs of control, or both.
The Rand Corporation compared mandatory controls and economic incentives
for red .icing nationwide chlorofluorocarbon (CFC) emissions (a nonc.riteria
pollutant) from nonaerosol applications. Sand found that using economic ap-
proaches in place of regulation would reduce the costs of control from $185 to
-------
Exhibit IV.1
Summary of Stationary Source Studies
Approaches Pollutant Source Selected Study
Study Examined__ Character!sties Description Conclusions
Rand
1
1
1
1
i
0
0
1
stand-alone charges: [
uniform rate |
marketable permits |
!
o
0
o
noncriteria
nonhazardous
area impacts
o
o
unusually varied
nationwide
o
o
Cost savings of 40% at same emissions.
Greater emissions reduction feasible
than with traditional approach.
Nichols
1
1
1
1
stand-alone charges: j
uniform rate [
impact-based |
!
1
o
0
hazardous
local health
impacts
o
Q
all plants of
same subject-type
widely dispersed
o
o
Charges much more cost-effective than
stringent uniform standards.
Charges somewhat more cost-effective
than well designed or less stringent
traditional approaches.
Mathtech
i
1
1
i
I
!
i
o
0
!
stand-alone charges: |
uniform rate and j
impact-based j
marketable permits |
1
I
0
0
criteria
local short-term
impacts
inventoried
(sources3 Chicago
0
o
Economic approaches can reduce control
costs by up to 90%.
uniform charges on all sources lead to
higher total control costs than tradi-
tional approaches.
Meta
Sys terns
1
1
1
1
1
stand alone charges: |
uniform rate |
1
0
o
criteria
area impacts
typical hydrocar-
bon emitters,
South Coast
Air Basin of
California
0
o
Marginal costs vary widely among sources.
A uniform charge could reduce control
costs by 10% while reducing emissions
by 25%.
Putnam,
Hayes &
Bartlett
I
1
1
1
marketable permits )
1
1
all pollutants
considered
none
Marketable permits are feasible, but
better for area impact pollutants than
for local impact pollutants.
Repetto
1
1
1
[
1
allocation approaches!
for PSD increments j
1
1
N/A
N/A
0
o
First-come, first-served approach has
serious deficiencies.
Depending on conditions, partial offset
or auction approaches are superior.
ICF
1
1
NSPS offsets |
1
N/A
utility boilers
Cost savings of 1 to 11%.
Temple,
Barker &
Sloane
1
1
i
noncompliance |
penalties 1
1
N/A
o
0
iron and steel
industry
electric utility I
industry f
o
o
Very slight macro and industry effects.
Some firms noticeably affected.
-------
IV-3
Exhibit IV.2
S«tfy of Mobile Source Studies
Study
Approaches Examined
Selected Study Conclusions
ICS Management Group
"
charges on new
¦vehicle emissions
o More cost effective than
current approach, but would
increase vehicle prices.
charges on in-use
vehicle emissions
0 Analysis inconclusive.
averaging
o Large cost advantages, with-
out the political drawbacks
of charges on new vehicle
emissions.
Sobotka & Company
averaging (diesel
TSP only)
o Averaging would be feasible
and fair.
o Cost savings of about 10Z.
Policy Planning and
Evaluation
charges on new vehicle
emissions (NO* only)
N/A - Study focused on design
issues.
averaging and market-
able permits (N0X only)
o Feasible and could offer
cost benefits.
-------
IV—4
$108 million, although substantial dollar transfers from polluting sourc.es to
government would result if emissions charges were used. Rand is now examining
the transfer problem in greater detail. Rand also found that economic ap-
proaches would allow greater reduction in total emissions than would be feas-
ible with, a traditional regulatory approach.
Albert L* Nichols of Harvard University compared various mandatory control
and charge-based approaches for benzene emissions (a hazardous pollutant) from
maleic anhydride plants. He found charges to be much more cost-effective than
simple regulatory approaches at reducing health impacts, but these differences
decreased if regulation became more sophisticated or if stringencies typical
of hazardous pollutants were demanded. The situation with the plants Nichols
examined has changed greatly since he conducted his study.
Mathtech, Inc., compared both charge-based and trading approaches to a com-
plex and difficult control problem: meeting a hypothetical short-term anbient
N0X standard in the Chicago air basin, The impacts of N0X emissions are source
specific and localized- This study showed that economic approaches could be
more costly than sophisticated regulation if an overly simple approach was
taken. Well designed economic approaches—approaches sensitive to source im-
pacts and control costs—offered control cost savings of up to 90 percent.
Mathtech concluded that economic approaches were feasible for this complex
problem, but would require sophisticated administration. Marketable permits
were favored over charges because of their smaller demands for agency ini-
tiative in information collection and because of the greater -certainty in
achieving environmenta1 standards.
Meta Systems, Inc., examined emissions charges to control hydrocarbon emis-
sions in the South Coast Air Basin of California. Hydrocarbon emissions are a
relatively simple control problem, since impacts are assumed to he a function
of total emissions. Trading approaches were not examined. Marginal control
costs for the sources and control alternatives examined were found to be highly
variable but to divide easily into high and low cost groups. \ uniform emis-
sion charge which made use of this division reduced total control costs by about
10 percent from a baseline using mandatory standards, while reducing total emis-
sions by about 25 percent below the baseline level. This study was not con-
cerned with designing an overall charge system but with case study calculation
-------
XV
of the charge rates necessary to achieve emissions levels equal to those under
existing requirements,
Putnam, Hayes and Bartlett, Inc., examined the design details of marketable
permits approaches. Some contributions made by this study have been absorbed
into earlier sections of this report, Putnam, Hayes and Bartlett concluded
that pennies were a workable approach but would be more difficult to apply to NGX
than to other pollutants.
Rober:. Re pet to of Harvard University examined alternative approaches to
the allocation of PSD increments. He concluded that the current "first cone,
first served" approach has serious deficiencies. Auctions were suggested as an
alternative approach in areas where new source growth is relatively large;
partial offset requirements were suggested for use in other areas. This study
also argued that the use of ambient air quality standards can interfere with a
desirable balancing of the benefits of cleaner air against control costs. EPA
has examined PSD increment allocation in internal studies as well.
ICF, Inc., examined the effects of extending offsets to NSPS requirements
for hypothetical utilities with similar plants in various parts of the country.
Cost savings of one to eleven percent were shown, with no effect on air quality.
Finally, Temple, Barker and Sloane, Inc., analyzed the economic impacts of
noncompliance penalties on the iron and steel and electric utility industries.
Average impacts were extremely snail, but some flms would be noticeably af-
fected by these penalties.
A.2 MOBILE SOURCE STUDIES
Economic approaches to mobile source emissions were examined in four
studies, one of which compared both charge-based and trading approaches.
Much of the analysis in these studies consists of informed speculation about
the feasibility and desirability of economic approaches, with special atten-
tion to emissions averaging {the mobile source equivalent of a bubble). Anal-
ysis of this kind has been absorbed into other sections of this report.
-------
IV-6
TCS Management Group, lac., (TCS) examined both charges and a form of
trading to reduce the emissions of motor vehicles while in use rather than just
at the tise of sale. Alternatives studied included emissions charges on manu-
facturers based on both tests at the time of sale and tests (as part of the state
inspection and maintenance_programs) of vehicles in use, emissions averaging,
and nodifications and supplements to more traditional regulatory approaches.
TCS found that emissions averaging and charges on new car emissions would per-
mit more cost-effective control. Averaging would reduce new car prices and he
administratively and politically feasible. Charges would raise new car prices
significantly and probably encounter considerable opposition.
Policy Planning & Evaluation, Inc., examined two approaches: one directed
at mobile source N0X emissions charges in one study and the other directed at
trading approaches in a companion study. These studies served primarily to
identify potential problems in the implementation of either of these approaches.
Testing and certification programs were seen as one area in which changes might
be needed to support any economic approach.
Finally, Sohotka & Company, Inc. (Sobotka), examined averaging for light
duty diesel vehicle particulate emissions, as an alternative to the recently
promulgated standards for model year 1985. This study found cost savings of
about 10 percent with averaging, at a slightly reduced level of total emissions.
Averaging was found to be both feasible and fair.
B EPA QM-C01SG mSRlElCES WITH 1SCEHTIVES
EPA analyses, experiences, and plans regarding four on-going areas of
implementation of incentive approaches are discussed in this section. First,
EPA's current stationary source controlled trading programs are discussed.
EPA is well along in its efforts to support the states' implementation of off-
set banking and trading and the bubble reform and has had significant exper-
ience with those parts of the controlled trading programs which have existed
for the longest time. However, banking and the bubble reform are relatively
new programs and are still evolving fairly quickly. The second area discussed
is mobile source averaging proposals (a form of controlled trading). EPA has
not yet implemented an approach of this kind, hut has taken some steps in this
direction. Third, penalty systems with some similarity to supplemental charges
-------
IV-7
are discussed. EPA has begun to use such penalty systems as the basis for
calculation of civil fines for the phasedown of lead in gasoline, in noncom-
pliance pecalties for stationary sources, and in nonconformance penalties
ror heavy duty trucks. EPA is willing to consider marketable permits or
stand-alone charges where found appropriate, but has not yet proceeded to pro-
mulgate regulations. Statutory change to facilitate use of these approaches
is discussed in Section ¥.
•N
Programs which EPA is implementing are described in more detail here
than in the previous sections. For each of these EPA programs, this section
sets out the program's status (including implementation history and problems
being experienced now), and describes EPA's current and planned activities.
B 1 STATIONARY SOURCE CONTROLLED TRADING
Status of Programs
o Offsets Banking and Trading
When first adopted in 1970, the CAA prevented major new sources of air
pollution from locating in nonattainment areas. In 1976 EPA adopted a policy
which, allowed such sources to enter nonattainment -areas, provided emissions
that remained after stringent controls were installed, were offset by reduc-
tions in the emissions of existing sources.! The offsets policy is now well
established for use in connection with new source siting in nonattainment
areas«
1 This policy was incorporated in the 197? amendments to the CM, and EPA
issued an interpretation of the new provision in January 1979* The general
applicability of this federal policy ended as of July 1 1979, when the states
were required to submit revised SIPs, States which wished to allow industrial
growth in nonattainment areas were required to provide a mechanism for con-
trolling ambient air impacts in these SIPs. The options available to the
states were to set SIP requirements for existing sources at a stringent enough
level to generate a margin for growth or to establish an offsets policy. All
but eight states submitted draft SIPs which provided for an offsets approach.
Those which did not do so have no pressing nonattainment problems. Many of
the states established their offset programs by incorporating EPA's 1979 in-
terpretation of the CAA revisions into their SIPs; some added specifications
requiring new sources to more than offset their emissions.
-------
IV-8
Under an offsets program, companies that wish to site new facilities in
nonattainment areas may either reduce emissions at existing facilities which
they operate in these areas (internal offsets), or purchase or otherwise se-
cure offsets from non company-owned facilities (external offsets) that, exist
In these areas. As of August 1979 about 500 permits for new sources had been
approved in the various states tinder applicable offsets policies, and no re
were pending. Ninety-five percent of these cases involved internal offsets.
In all but two or three of the 15 or so external offset cases, offsets had
been donated to the new source without charge. In some of these donation
cases the air quality agency encouraged the donation and in some cases even
encouraged existing sources to take the steps needed to generate the offset.
In other cases the control or state agency themselves implicitly asserted
control over offsets growing out of source shutdowns in the area and assigned
these to the new source without charge. Finally, in sons cases state agencies
took steps to generate offsets, which were donated to new private sources of
pollution. In short, commercial offset transactions have been very rare.
More recent offset cases probably involve a somewhat higher proportion of
external offset trades, but the pattern of heavy reliance on internal and
agency mediated offsets continues.
The limited role now being played by commercial offset trades, and the
lack of an effective and pervasive set of state-run Institutions for offsets
tradings are matters of concern to EPA and CEA. Occasional ad hoc external
offset arrangements cannot provide as significant reductions in total control
costs as is otherwise possible or significantly increase incentives to innovate.,
Because so few offset proposals involve commercial transactions, existing
programs have provided little incentive for existing sources to reduce their
emissions in anticipation of an offset sale- Mew sources have not been able
to predict whether offsets will be available in an area or what they are likely
to cost. Siting decisions for new facilities aay be strongly affected by
these uncertainties and by the ultimate availability and price of external
offsets.
A major barrier to commercial offset transactions in the past has been
the lack a of mechanism by which credits for emissions reductions could be saved
until a purchaser was found or a company chose to expaad Its operations ia a
-------
IV- 9
nonattainroent area. With no opportunity to savn or "bank" credits for reduc-
tions, existing sources had no incentive to generate in advance a pool of
credits which they might use for growth or which new sources wight purchase.
Instead, existing sources had an incentive not to reduce emissions until a use
for the credits was identified. Where a future need for credits was antici-
pated, uneconomic production processes might he kept in operation to avoid
losing an offset opportunity. Similarly, the size of a pollution control
system to be installed might be chosen to just oeet current standards, even if
replacement with a larger control system (which could economically achieve
more stringent control of emissions than required by current standards) would
otherwise be selected.
New sources also faced problems with this system, since it could he a for-
midable task to locate and come to terms with existing sources which could
reduce emissions but had not yet done so. Given the difficulty of the search
and negotiation problems faced by potential new sources, the leverage avail-
able to an air quality agency, and the interests of localities in facilitating
industrial growth, it is not surprising that so few of the hundreds of offsets
which have been proposed to date have involved commercial transactions between
firms.
la order to address this problem, EPA integrated banking of emissions re-
duction credits into its offset policy in January 1979. So far, very few
states have taken the steps necessary to establish an operating banking system,
but EPA believes that banking will eventually play a major role in making ex-
ternal offset trades more corason in nonattainment areas. Banking will also
help to facilitate bubbles (discussed below), and help new sources either meet
their offset obligations under PSD programs or avoid triggering thresholds for
PSD review,
o The Bobble Policy
The bubble policy encourages industry to develop less costly ways to meet
pollution control requirements for existing plants than the measures specified
in SIPs. It is premised on the belief that plant managers have far more in-
formation about their plants and, fat* greater incentive than the government to
reduce the costs of meeting air pollution control requirements. Subject to
Hew Source Review procedures, the policy is as follows: as long as the area
-------
IV-10
affected by the plant's emissions continues to attain and maintain NAAQS, the
policy permits a plant to relax controls at emission points where control costs
are high in exchange for extra controls at emission points with, lower control
costs. Bubbles which involve more than one plant may also be proposed, whether
the plants are under the control of a single company or not. In order to
assure that new requirements which result from a bubble will be enforceable,
state SIPs must either impose aggregate emissions limits on groups of sources
which bubble, or be revised to reflect the changes involved in each bubble.
SIP revisions involve public hearings at both the state and federal levels,
EPA promulgated its bubble policy in December 1979 as the culmination of
several years of effort to integrate market incentives into the existing regu-
latory structure. Because of the division of authority which exists under the
Clear Air Act, the policy is an option which states may (but need not) include
in their SIPs. Some states have embraced the policy enthusiastically; others
have chosen not to implement the policy at this time. EPA is continuing to
evaluate and refine the bubble policy and its procedures for reviewing and
approving bubble applications in order to improve the policy's usefulness.
Policy issues receiving attention now include the requirements for demonstra-
ting equivalent impacts, the potential for parallel state and federal proces-
sing of applications to expedite reviews, one-time EM approval of states*
own criteria for approving bubbles, and the broadening of exceptions that
permit bubble trades in nonattainraent areas. In addition, implementation of
the policy has served as a catalyst for efforts to reform the SIP revision
process.
Although the bubble policy is a new program, early indications are that
the policy will be successful where implemented. Industry has shown a high
degree of interest in using bubbles despite the short period of time available
for devising new approaches to meet 1982 SIP requirements. As of early
November 1980, about 20 formal bubble proposals had been received by the
states and a few had reached EPA for review. Industries proposing bubbles
include steel and other metals, paper and plastic manufacturing, paper and can
coating and utilities. Pollutants include VOC, SO2 and TSP. Control approach-
es include process changes, fuel switching and new control hardware. The
majority of current bubble proposals involve changes in hydrocarbon controls
for product coating lines.
-------
IV-11
The first proposals submitted estimate significant savings in annual
operating costs and in capital expenditures. The Naragansett. Electric Com-
pany's kmItiplant bubble for SOt emissions was approved and is expected to
reduce fuel costs by $3 million per year. ARMCO Steel estimates reduction of
$.14-16 million in capital costs by replacing controls on process TSP emissions
with open dust controls. The majority of; the bubbles being developed are
expected to improve air quality. A few proposals involve innovative control
approaches. For example, 3M Company is planning to use waterbase solvents and
a solventless coating process to reduce VOC emissions at a tape coating plant.
Two bubbles have received preliminary EPA approval. Actual approval of
several bubbles is expected to stimulate further interest in using bubbles, as
will the need for companies to retire and replace older pollution control
equipment,
Current Efforts Regarding Implemented Controlled Trading Activities
The nature of EPA* s activities in the controlled trading area has been
determined largely by the fact that implementation of banking and trading and
bubble programs is primarily a state responsibility under the existing statute.
EPA is therefore initially confined to interpreting the statute and defining
policies which may be used by the state, encouraging the states to establish
programs and assisting them should they choose to do so, and encouraging
industry to take advantage of the opportunities presented by these programs.
Because offset trades and bubbles will involve SIP revisions (which EPA must
approve or disapprove), EPA can also effectively establish minimum criteria for
the acceptability of these proposals.
Banking and trading and the bubble policy all involve similar kinds of
activities. EPA's banking and trading project has published several manuals
describing particular aspects of the establishment and operation of different
banking and trading systems, as well as papers detailing the benefits these
systems can provide. EPA has also published a manual for industry on use of
the bubble policy; this manual emphasizes the advantages of working closely
with the states and EPA during the development of bubble proposals. A second
booklet on multiplant bubbles is being prepared. The banking and trading
project has developed models and analytical tools for use by the states and
-------
IV-12
provides the states with other direct technical assistance. Members of the
project are available to make presentations, participate in workshops, help
resolve issues, and perform research ia selected areas. EPA staff involved in
the bubble reform also participate in numerous industry briefings, seminars
and conferences, and have worked closely with particular comparies that are
considering bubble proposals-
Public information activities in connection with controlled trading pro-
grams are extensive. Newsletters have been established for both banking and
trading and the bubble. The banking and trading project publishes an anno-
tated bibliography which covers both of these programs and the wider context
of economic approaches to regulation. The bibliography is kept up to date and
includes useful material published by EPA and others- Mew banking and trad-
ing and bubbles proposals are identified by the regions and tracked by head-
quarters. Headquarters helps measure progress and spot problea.s, helps the
regions coordinate these issues, and facilitates diffusion of experience.
"Progress reports" are included in the program newsletter.
EPA is now examining the effects of the bubble policy on industrial inno-
vation. As part of its assessment of strategies for attaining ambient air
quality standards in nonattainment areas, EPA has also attempted to assess the
potential effects -of offset trading In particular areas. A drsft report by
Energy and Environmental Analysis, Inc. in April 1979 proposed analytic proce-
dures for this task. Key cost assumptions underlying the proposed model are
still being evaluated. No comprehensive quantitative analyses of the effects
or potential effects of offset banking and trading, or of the babble policy,
exist at this time. EPA has focused its attention on identifying and solving
problems in the implementation of these prograas and on identifying useful
future steps.
Controlled Trading Plans Under Current Legislation
EPA's future activities in the controlled trading area will focus, in
particular, on emissions banking and integrating the banking and trading and
bubble programs. The establishment of active hanking prograas should encourage
and greatly facilitate offsets trading and bubbles, since operation of a trad-
ing program should be a relatively simple matter once a banking system is in
-------
IV-13
place. Bubbles involving more than one plant or firm can also be more easily
arranged if one party's emission reductions do not have to occur simultaneously
with the emission increases of another party.
In addition, EPA will increasingly focus its attention on the administra-
tive and procedural complexities involved in the banking and trading and bubble
programs. EPA is already examining changes in the SIP revision process to
speed the review of trading and bubble proposals. EPA will continue to look
for ways to streamline procedures where this can be done under the CAAS while
providing adequate assurance that air quality is not jeopardized.
In the long ran, controlled trading programs should he developed into a
more flexible and useable system that will integrate trading in emission
entitlements and obligations and will consider futures and options in these
assets as appropriate.
B.2 MOBILE SOURCE AVERAGING
EPA Is continuing to examine mobile source averaging approaches, and has
issued ar. Advance Not ice of Proposed Rulemaking to implement such a program
for N0X emissions from heavy duty engines and light duty trucks. A public
workshop took place la January 1981 in connection with this proposal, and a
proposed rule is planned for Spring 1981. This activity reflects EPA's and
CEA's conclusion that, allowing tradeoffs of emissions from different sources
is as promising for mobile as for stationary sources. Averaging can reduce
costs for manufacturers while maintaining emissions levels, and could lead to
improved fuel economy. Manufacturers could use the flexibility provided by
averaging in several ways. For example, manufacturers might test new models
in the market before investing large amounts in pollution control or they
might design particular models to conform to standards in export markets.
SPA'« initial analysis of averaging approaches (in connection with the
recent rulemaking on TSP standards for light duty diesels) preceeded comple-
tion of the contractor reports on mobile source averaging discussed elsewhere
in this section and in the Appendix. While the contractor reports demonstrated
the large potential benefits of averaging, these studies and the early EPA
-------
IV-14
study also made it clear that averaging could raise difficult problems involv-
ing equity, program administration and competition between nanufacturers.
Some of the potential problems with averaging can be minimized b> careful pro-
gram design. For example, EPA has determined that averaging probably should
be restricted to within the same vehicle class (cars, light duty trucks, heavy
duty trucks), and should be applicable only at the level of individual engine
families (i.e., families within the same class could be averaged but not vehi-
cles within a family, or vehicles or families ia different classes).
B. 3 SUmEMENTAL CHABGBS
Supplemental charges can be used to induce compliance with regulatory
requirements, or to limit marginal coattol costs by providing sources with an
alternative to control. 1PA has implemented a system with a similarity of pur-
pose to supplemental charges of the first kind under SecLion 120 of the CAA, for
stationary sources (nonconpliance penalties) and has also begun to implement
such charges for mobile sources (nonconformance penalties). Unlike stand-alone
emissions charges and supplemental "safety valve" charges, noncompliance penal-
ties are not assessed on the degree emissions exceed a standard but solely on
the cost of compliance technology. Moreover, noncompliance and nonconformance
penalties are not intended to offer the sources the continuing choice of reduc-
ing emissions or paying a fee: they are intended to encourage compliance with
regulatory requirements. Nonconformance penalties do permit sjch a choice
in the short tun, but will quickly escalate beyond the point at which paying
the charge penalty is a viable option.
Noncompliance Penalties
At the direction of Congress EPA has implemented a program of noncompli-
ance penalties for stationary sources. These penalties are not subject to any
ceiling and will be assessed through administrative mechanisms, avoiding the
delays that often accompany legal action to impose civil fines. Objective
economic considerations (costs of equipment, tax laws, prevailing interest
rates, etc.) are used to set penalty levels which offset the economic benefits
to polluters of delaying expenditures on pollution control measures. Noncom-
pliance penalties ate a mew program at the federal level, but were proven to
-------
I?-15
be effective and workable at the state level before Congress required EPA to
take similar steps.
IPA promulgated its final rule on noncompliance penalties in July 1980*,
however, EPA delayed general implementation of the regulations until January I,
1981 in order to train agency personnel and to better coordinate noncompliance
penalty activities with. IPA enforcement actions.
Penalties apply to all major stationary sources which are in violation of
any SIP requirement, NSPS, or NESHAP, or which have failed to comply with in-
terim requirements established in a court order or consent decree. Because
available resources do not permit immediate action against all 2000 or so
sources in these categories (without jeopardizing other agency missions) EPA.
has established priorities for enforcement. These focus first on those sources
which, have never come into compliance with applicable requirements and which
are not complying with enforceable requirements in court orders or consent de-
crees. It is expected that many sources in these groups will negotiate con-
sent decrees containing interim requirements before the regulations are fully
implemented- EPA does not intend to impose noncompliance penalties on sources
which are economically unable to simultaneously pay penalties and make the
investments necessary to come into compliance. The Temple, Barker and Sloane
report discussed earlier suggests that this will not be a widespread problem,
Nonconformance Penalties
EPA is also implementing nonconformance penalties for mobile sources, as
specified by Section 206(g)(3) of the Clean Mr Act, The statute permits EPA
to allow the production and sale of heavy duty engines and vehicles which can
not meet emissions standards, provided emissions are not in excess of estab-
lished upper limits and provided a charge (the nonconformance penalty) is
paid. Nonconformance penalties allow EPA to set standards that are expected
to be within the technical capacity of most firms in the industry at the time
the standards become effective, without the risk that some firms will be unable
to meet these standards and will be forced out of the market. This approach
allows firms to pursue the development of reasonable control technologies
without the fear that vehicle or engine sales must cease if a specific deadline
is aissed. It allows manufacturers who purchase engines, or control technologies
-------
IV-16
to be protected against temporary shortages, and enhances ccmpeti.ti.on by
keeping more suppliers and more vehicle manufacturers in the narket. At the
same time the approach can enhance the credibilty of EPA standards by making
it less likely that standards will need to be changed or delay-id. This can
result in more effective protection of the environment. The requirement that
emissions be below established ceilings prevents the sale of grossly polluting
engines and vehicles.
In addition to providing that penalties should be set so as to protect the
competitive positions of manufacturers who do comply with emissions standards,
the CM specifies penalties are to be larger the further the emissions are from
meeting standards and they are to increase over time. This last characteristic
assures that the penalties will provide a continuing incentive to manufacturers
to bring their engines and vehicles into compliance.
EPA intends to promulgate initial nonconformance penalties for 1984 model
year heavy duty engines in March 1981, and thereafter to provide for nonconfor-
mance penalties where there is evidence that compliance with a standard cannot
be achieved without substantial development and/or substantial design work.
Thus subsequent promulgations will be directed to penalties for 1985 (and later)
heavy duty engines and for light duty truck emission standards. Penalties are
(by statute) aimed at the technological laggard. Because Manufacturers are
currently meeting those emissions standards required through model year 1983,
1984 is the first year nonconformance penalties are applicable.
After considerable analysis and discussion of alternative bases for calcu-
lating nonconformance penalty levels, EPA has decided to propose a fee-setting
approach based on a uniform penalty rate for each unit of emissions in excess
of the standard. Where the data are available, the penalty rate Jill be based
on the marginal costs of control for the same pollutant in similar engines or
vehicles produced by other manufacturers. Penalties will be adjusted upward
over time, in conformance with the statute. This approach may often result in
total penalty payments which exceed the costs conforming manufacturers incur
to come into compliance. However, this approach is expected to protect these
conforming manufacturers from adverse competitive impacts and to provide incen-
tives tor nonconforming manufacturers to come into compliance, a;; is required
-------
IV-1?
by statute. In addition, the technical difficulties involved in alternative
approaches to setting penalty levels are severe.
Economically-Based Ely11 Fines
EPA has applied the economic penalty Idea to the calculation of civil
fines on petroleum refiners .for violation of the schedule for phasedown of
lead content in gasoline. Lead is used as an additive to increase octane
levels in gasoline; octane levels may also be increased in other ways but
these are usually more expensive than adding lead. Thus, in the absence of
penalties equal to cost savings, firms which do not comply with EPA's regula-
tions in this area would experience a direct and readily measareable economic
benefit. Unfortunately, the implementation of an economic penalty based on
cost savings has had to be compromised in this program, because under the
Clean Mr Act the maximum civil penalty that can be assessed for these viola-
tions is $10,000 per day. The economic benefits to refiners larger than
roughly .125,000 barrels per day of production from a violation of these pro-
visions may exceed §10,000 per day; these refiners will find that it is still
economically advantageous to violate the statute.
B.4 QTHBt 0S-e01i© STATIONARY SOURCE INCESTIVE EFFORTS
In addition to the economic incentive approaches which EPA has begun to
implement, EPA is examining other economic incentives that could contribute to
the solution of difficult problems under the current regulatory system. This
subsection briefly describes the approaches EPA is examining. These include;
o the use of permits or charges to reduce chlorofluorocarbon emissions;
o allowing the use of offsets to meet NSPS requirements (this Is a form
of "new source" bubble);
o a wide range of approaches to the allocation of PSD increments;
o the use of emissions density zoning; and
o the use of tradeable "rollback" obligations and supplemental emission
charges to address chronic nonattainiaent problems.
-------
IV—18
Reducing Chlorofluorocarbon Emissions
SPA is planning to propose regulations in the spring of 1.981 for control
of chlorofluorocarbons (CFC) emissions to the atmosphere from non-aerosol uses.
A marketable permit approach applicable either to producers or first purchasers
of CFC's is anticipated, CFC's are particularly appealing for the application
of economic incentives because: (1) they represent an area previously unregu-
lated; (2) there is a direct relationship between emissions and ambient im-
pacts; (3) at current and anticipated emission levels, there do not appear to
be significant exposure levels beyond which especially serious environmental
effects occur; (4) CFC emissions do not result in localized or "hot spot" ef-
fects; (5) there is a fairly long lag between emissions and subsequent effects
oil the atmosphere (cumulative effects are the most significant); and (6) be-
cause emissions are directly related to the amounts purchased and used in pro-
duction, no serious measurement problems exist.
Because of the need for more detailed information to support the option of
establishing a cap on CFC emissions at 1980 levels and to explore the relative
merits of an incentive system applicable to purchasers as well as producers,
the land Corporation is performing additional analyses for EPA regarding the
implied control technologies„ costs, and resulting economic impacts. Problems
arising as a result of potentially large transfer payments generated by the
operation of an economic incentive system and the effects of CFC regulation on
innovation in the U.S. and abroad are included in the analysis. Associated
studies by Resources for the Future (to design charge systems for auctioning
CFC permits) and a professor at the University of California (Berkeley) (to
design a system for auctioning CFC's) will also contribute to the development
of an approach to CFC regulation.
KSFS Offsets and Hew Source Bubbles
Currently, new sources cannot trade because technology-based standards
must be met. The implications of the restriction were discussed in Chapters
I and III, which cast doubt on the need for such restrictions. Moreover, the
1CF analysis of NSPS offsets demonstrated that allowing trades between new and
existing sources can substantially reduce control costs, even if trades are
-------
IV-19
confined to internal offsets for electric utilities. Even larger cost reduc-
tions nay be available if electric utilities could purchase "external" offsets
from other industries (e.g., from smelting operations), and trades between
other new and existing sources are also likely to offer significant cost sav-
ings. These savings aay be sufficient in particular cases or identifiable
classes of cases to make such transactions financially attractive even if high
"offset ratios" are required of the new source.
EPA. is continuing its analysis of USPS offsets. The analysis of NSPS
offsets now in progress is focusing on the utility situation, and examining
potential cost savings (under improved assumptions and in different situa-
tions), differential regional impacts, energy impacts and possible effects on
coal use.
New source trades which are not constrained by technology-based standards
(such as NSPS, BACT and LAER) represent an expansion of the bubble policy to
include new sources In some situations. Even though technology-based standards
could still determine emission reduction obligations, neither NSPS offsets nor
new source bubbles more generally would be penalssable under the current sta-
tute (which has been interpreted to require meeting technology-based standards
at the plant site). EPA may reconnmend legislative changes to permit new source
trading in some situations, depending upon the results of additional analysis
now in progress. Initial conclusions about legislative changes are contained
in Chapter V.
Analysis of new source bubbles as a general approach is focusing on de-
termining the implications for long-term emissions.
Hew source bubbles relax the requirement that new sources directly meet
technology-based standards, requiring instead that an equivalent emission re-
duction he secured somewhere. While emissions will initially be equal with or
without new source bubbles (and may even be less with the bubble, if the offset
ratio is greater than uaity), trading can always reallocate emissions between
sources with different remaining lifes and thereby affect long-term emissions.
When this occurs, different steps may need to be taken to create room for future
economic growth or to assure improvements in air quality as the capital stock
-------
IV-20
turns over. As discussed in Chapter III, these effects will be present when-
ever (and only if) a trade increases the expected lifetime of an entitlement.
This can of course only occur as a result of trading rules if ent itleoents are
forfeited to the government when a source shuts down, and it
-------
IV-21
Among the alternative approaches EPA has examined are emissions density
zoning (discussed in the next subsection), partial offset requirements like
those examined by Repetto, generating a growth margin by imposing stringent
requirements on existing sources (this approach is now in use in some states),
rigorous 3A.CT review of all new sources, "administrative reservation of the
increment" (explained below), and allocation according to local preferences
favoring the presence of certain kinds of sources.
Partial offset requirements (another allocation rule introduced in Chapter
III) would increase costs to industry to site in PSD areas, since new obliga-
tions would be created. While partial offsets could in theory provide for
more efficient allocation of the increment, it would be difficult to set offset
ratios which achieved a fully efficient outcome. Ideally, the ratio should
balance added cost now (relative to expected control costs in the future)
against the greater value of emissions reduction earlier in time. The benefits
of partial offsets are limited, because it is only feasible where offsets are
available, and if offsets are available there should be no absolute exclusion
of sources which place a high value on use of the air.
"Administrative reservation of the increment" involves directly restrict-
ing (and specifying in advance) the percentage of the remaining increment that
a single source could use, or the percentage that could be used by all sources
in a given time period* This approach does not increase efficiency and would
probably encounter considerable opposition from industry. Moreover, it ad-
dresses a problem that is not important in practice, since sources which con-
sume large percentages of the increment usually do so only in a small geo-
graphic area*
Rigorous BACT review is likely to be sufficient to permit siting in most
areas in the forseeable future. Where this is not feasible, local assessment
of the costs and benefits of allowing one source to site rather than another
is an option that should be considered.
Emissions Penalty Zoning
Emissions density zoning is essentially a way to simplify and express
the results of air quality modelling. Once the impacts of different levels of
-------
IV-22
emissions in different areas have been assessed by modelling, it is possible
to specify an acceptable scenario tor apportioning total emissions of a par-
ticular pollutant in each small subdivision, or "zone" within a large geographic
area. This specification may need to be somewhat conservative, because source
characteristics can affect impacts and because groupings of sources near zone
boundries could lead to a violation of air quality standards.
"Zoning densities" can be used to allocate PSD increments or to make it
easier for sources to plan and execute trades under the offset and bubble
policies. PSD review might be simplified for sources within density limits,
or entitlements to emit from within zones might be auctioned. As the limits
established by zoning densities were approached, a return to more conventional
and cumbersome case-by-case review would always be possible.
Nonattainment Strategies
As discussed earlier, continuing nonattainment problems can resist solu-
tion by conventional means because of the difficulties that exist in identify-
ing additional emission reduction opportunities and in allocating them to
sources. This is likely to occur because sources in these areas have typically
already undertaken standard control measures and agency information about new
control opportunties is likely to be limited. Moreover, political considera-
tions may make it difficult to impose yet additional control obligations on
particular sources of emissions. As a result> EPA has been examining several
potential economic approaches to the problem of reducing emissions in areas
that continue to have nonattainment problems. The approaches being examined
can be used separately or in combination.
The information problem can be reduced by relying on economic approaches
that lead to cost-effective results regardless of the initial allocation of
control requirements, Thus, for example, states could identify the overall
control obligations that will lead to attainment, and allocate them through
a uniform percentage rollback in allowable emissions. If these obligations
(or "emission reduction assessments") are tradable as under the bubble policy,
then trading should lead to cost-effective controls overall. In addition,
the consequences of allocating transferable emission reduction assessments
-------
IV-2 3
(TERAs) to sources that are technically incapable of fully reducing emissions
is minimized because such sources can purchase reductions in the tradirg mar-
ket..
As aa alternative, an emissions charge could be applied to all regaining
emissions or to the amounts allocated as emission reduction assessments. If
charges were applied to all retraining emissions, sources would undertake con-
trol efforts until the marginal cost of control equals the charge and would
have a continuing incentive to find ways to reduce emissions at lower cost.
If charges apply only to emission reduction assessments, sources would elect
to pay the charge if it were less than the cost of reducing emissions. These
sources would have ongoing incentive to find ways to reduce emissions at a
cost lower than the charge. Of course, the essential difference "between these
trading and charge-based approaches is that trading approaches allow better
control over the timing of, reductions while charges allow better control over
the cost of reductions.
These approaches can be combined to achieve the benefits of both trading
and charges. To accomplish this, charges would serve to supplement trading.
Thus, the primary control approach would be to allocate emission reduction
assessments aad allow trading to occur, and supplementary charges could be
designed to serve any of several purposes. The charge could be designed to
provide s safety valve for sources which are unable to meet their new control
obligations directly or through purchase of entitlements, or for sources which
preferred to delay compliance for a short period of time. Alternatively, the
emission reduction assessments could be considered to be technology-f orc.-ng and
the charge could be designed as a nonconformance penalty that is low enough to
provide a credible alternative to noncompliance but high enough to encourage the
development of new control approaches- Finally, a supplementary charge could
be assessed for emissions remaining after emission reduction assessments to
provide i continuing incentive to reduce emissions beyond required levels.
WhiLe these economic approaches are likely to be more effective, cost-
effective and feasible than traditional approaches, they require careful evalu-
ation. is discussed in earlier chapters, there are a wide variety of important
-------
IV—24
design parameters that range from setting the level of the charge to the in-
volvement of the control agency in trading markets, EPA's analysis of poten-
tial alternative incentive approaches to nonattainment problems is still at a
relatively early stage. Approaches such as those discussed ?,bove have been
identified and thought has been given to some of their attributes and operating
characteristics. Further analysis is planned,
Mobile Source Strategies for the 1980's
EPA has begun to implement mobile source nonconformance penalties, and
begun a rulemaking to consider use of mobile source averaging. EPA is also
considering expansions and combinations of these approaches, and is examining
enforcements and alternatives to mandatory inspection ana maintenance (-I&M
programs.
Emissions averaging could be the cornerstone of a more flexible and cost-
effective mobile source emissions control program. EPA is currently studying
averaging tor possible implementation to control N(\r emissions from heavy duty
trucks. One approach would be to administer averaging at the time of sale in
the same way that corporate, average fuel economy (CAFE) requirements are ad-
ministered, by relying primarily on manufacturer certifications and vehicle
counts. Alternatively or in acdition, inuse emissions (of vehicles which had
not been abused) could be averaged, A relatively small number of yearly in-
spections of each model year's vehicles (a few thousand annual inspections per
model year for two or three years) would provide a high level of statistical
confidence about the range of actual in-use emissions, and reduce the need for
emissions and durability testing prior to sale. Once average in-use emissions
were known, manufacturers could be assessed a charge if their average fleet-
wide emissions were not in compliance, and could be given a credit against
future charges if average fleet-wide performance was better than necessary.
The use of charges and credits would permit averaging over time, and could be
ieplenented to provide both flexibility and the same strong incentive toward
compliance as nonconformance penalties will provide. A manufacturer could
build cleaner new vehicles rather than recall old ones, bat would not inten-
tionally incur the expense of being out of compliance.
-------
IV-25
This system or something similar would probably allow store effective
control at less cost than the current approach. However, it would not solve
the problems of tampering and neglected maintenance of pre-model year 1981
vehicles, or of nonfunctioning computer control systems on newer vehicles-
Mandatory l&M solves this problem directly and completely, but even if man-
datory l&M is confined to areas which cannot demonstrate attainment it Is
often opposed by vehicle owners and the states. EPA will therefore consider
economic Incentives to aid in encouraging the implementation of mandatory
l&M.
The ~--2S study summarized earlier in this chapter examined two economic
"enhancements" to mandatory inspection and maintenance: direct financial re-
wards to vehicle owners who respond to recall notices, and indirect rewards
(in che fora of waiver of inspection requirements in the following year) to
those who pass a mandatory inspection. The idea of penalties and incentives to
vehicle owners could be expanded, in the context of mandatory or voluntary
inspections. Rewards could be paid for undertaking repairs to pass a voluntary
inspection (records would be needed to prevent "double dipping"), or an emission
fee could be assessed on a fixed schedule unless scheduled na intenance was
performed or voluntary inspections were passed.
Some of these programs could be difficult to administer, and any fee which
varied with predicted rather than measured emissions would probably encounter
substantial opposition* Moreover, designing a fee which was fair could be so
difficult as to make program implementation infeasible unless a simple program
tied to the age of registered vehicles was acceptable, To provide sufficient
incentive, rewards or fees might have to be large. This is because most excess
emissions from in—use vehicles are accounted for by a relatively small number
of grossly polluting vehicles, whose owners might not respond to small rewards
or might pay a small fee rather than take corrective action.
As an alternative to fees or rewards, manufacturers could be required
to perform scheduled maintenance without charge, or even to reward vehicle
owners for bringing vehicles in for maintenance (vehicle owners would bear
the cost of these programs in the form of higher new vehicle prices). Such
programs might encounter as much political opposition as mandatory X&Mt and
could lead to the added expense of unnecessary scheduled maintenance.
-------
IV-2 6
In the long run it would appear that the su.cr.ess of I&M programs wiJ I
require emission control systems that last for the lifetime of motor vehicle
(e.g.„ 100,000 miles). To provide manufacturers the incentive to design such
systems, consideration should be given to (1) longer lived but more limited
warranties (restricted to primary emission control equipment), (2) innovative
technology waivers (whereby development of 100,000-mile emission controls
might permit a delay in meeting certain standards), and (3) grants to pollu-
tion control equipment manufacturers. While meeting environmental goals will
likely necessitate mandatory I&M programs in nonattainment areas, opposition
from the public can be significantly reduced if periodic inspections merely
determined whether or not a long-lived emission control system is in place and
operable. Only where evidence of damage (covered by warranty) or tampering
(requiring out-of-pocket owner expenses to repair) was found would an emissions
test be necessary. Failing the test would then require repairing the system
and reiesting.
-------
- analytic conclusions
AMB LEGISLATIVE PROPOSALS
*> analytic cogg^mig "
A.1 CURRENT PROGRAMS
o REPLACEHENT OF THE IN-PLACE REGULATORY SYSTEM WITH STAND-ALOME EMISSIONS
CHARGES OR MARKETABLE PERMITS IS HOT NECESSARY. AIL SIGNIFICANT BENEFITS
OF ECONOMIC INCENTIVES CAB BE ATTAINED THROUGH CAREFUL SELECTION AND DESIGN
OF CHARGE OR TRADING APPROACHES WHICH SUPPLEMENT THE CURRENT REGULATORY
SYSTEM. If the full potential of supplemental charges and "controlled
trading" programs could be achieved, no significant incremental benefits
would be available from use of stand-alone charges or a "pure" system
of marketable permits. However, stand-alone charges and permits would
offer a fresh start that may be important to achieving the potential
benefits of economic approaches in certain unregulated areas.
o WITH ITS OFFSET, BUBBLE, AND EMISSIONS BANKING POLICIES, EPA HAS MADE
SIGNIFICANT STRIDES TOWARDS INCORPORATING MARKET INCENTIVES IN ITS AIR
PROGRAM. In each of these policies, polluters have incentives to seek
out sources of low-cost emission reductions. The result will be a more
cost-effective pollution abatement program. However, these policies have
not yet been fully implemented by the states. One reason has been the
embersome administrative procedures associated with these policies. EPA
Is currently working on streamlining this process.
\ 2 ECONOMIC EFFICIENCY
o CHARGE-BASED APPROACHES WILL CONTRIBUTE TO EFFICIENCY IN TOE ALLOCATION
OF SOCIETY'S PRODUCTIVE RESOURCES BY FORCING SOURCES TO TAKE INTO CON-
SIDERATION THE DAMAGE WHICH THEY CAUSE TO THE ENVIRONMENT. Traditional
coomand-and-controL approaches are no better than trading approaches in
duplicating this desirable feature of charge-based approaches. Where
the marginal social costs of emissions are known, no approach is more
economically efficient and cost-effective than stand-alone charges.
-------
¥-2
o IT IS DIFFICULT TO DIRECTLY PORSOE ECONOMICALLY EFFICIENT OUTCOMES~ Eco-
nomic efficiency requires that damages from emissions and control costs be
equalized at the margin. In the absence of an ability to determine damages
accurately, efficient outcomes cannot be identified. Thus, the operational
test of efficiency is that of cost-effectiveness, i.-e., achieving an envi-
ronenvironmental objective at least cost. This Is the standard of effic-
iency pursued by EPA in its consideration of incentive approaches. Any
approach that is economically efficient will also be cost-effective.
A,3 EFFECTIVENESS
o NONCOMPLIANCE PENALTIES SIGNIFICANTLY STRENGTHEN THE EFFECTIVENESS OF
CURRENT REGULATORY REQUIREMENTS BY ENCOURAGING COMPLIANCE. Noncompliance
penalties are a special form of economic incentive intended to neutralize
the economic benefits which would otherwise accrue to stationary sources
that delay compliance with regulatory requirements. EPA has implemented
these penalties, and EPA and CEA expect them to be highly effective,
o BOTH SUPPLEMENTAL CHARGES AMD "CONTROLLED TRADING" CAM PROVIDE INCENTIVES
TO ABATE POLLUTION TO A GBEATER DEGREE THAN IS REQUIRED BY CURRENT STAN-
DARDS. Supplemental charges imposed on some or all residual emissions
can be as effective as stand-alone charges in providing a continuous
incentive to sources to go beyond current standards and to find innova-
tive ways to reduce control costs. "Controlled trading" (the buying and
selling of obligations to reduce emissions and of credits for abatement
which was not required) provides identical incentives for each source and
can lead to a reduction in total emissions as emission reduction credits
are created for later use or sale.
o STAND-ALONE CHARGES AND THE "PURE" FORM OF MARKETABLE PERMITS CAN BEST
SERVE AS A PRIMARY MEANS TO CONTROL THOSE AIR POLLUTION PROBLEMS WHICH
ARE NOT ADDRESSED OR DEALT WITH SUCCESSFULLY BY CURRENT REGULATIONS. Air
pollution problems which are not being addressed by the current approach
should receive careful scrutiny as candidates for implementing emission
charge or marketable permit approaches. Ir. these cases, economic incen-
tives cannot disrupt an existing program, and may provide a tetter approach
to as yet unsolved problems.
-------
V-3
o IN GENERAL, TRADING APPROACHES PROVIDE GREATER SHORT RUM CERTAINTY ABOUT
ACHIEVING SPECIFIED LEVELS OF EMISSIONS AMD IN MOST CASES LESS CERTAINTY
ABOUT COSTS TIM CHARGE-BASED APPROACHES. Trading involves a fixed quan-
tity of entitlements and therefore a strict limit on total emissions.
The control costs that will he incurred to meet this limit are difficult
to predict in advance except where polluters are required to submit cer-
tain "demand for permits" data prior to an auction. Charge-based ap-
proaches will limit marginal control costs to the level of the charge,
but the amount of control that can be undertaken at costs below the charge
will be difficult to predict in advance. The effects of a given charge
rate on emissions will vary with inflation, economic growth and innova-
tion. The price of entitlements under a trading approach will vary with
these factors s but total emissions will not vary.
A.4 COST-EFFECTIVENESS
o SUBSTANTIAL CONTROL, COST SAVINGS AND INCREASED COST-EFFECTIVENESS ARE
LIKELY TO RESULT FROM THE USE OF ECONOMIC APPROACHES. This is true
because economic incentives vest greater decision-making responsibility in
sources, and. allow the market to reward good decisions. (Sources them-
selves will also have better information on which to base (control)
decisions. Studies conducted for EPA produced savings estimates ranging
from about 10 percent to about 90 percent of base level control costs at
equal or greater levels of abatement as a result of the use of economic
approaches. Even larger long-term savings, may be possible, because
incentives should be better able to stimulate technological change than
command-and-control approaches alone.
0 TRADING APPROACHES ARE LIKELY TO BE FEASIBLE WHENEVER DIFFERENT SOURCES
WITH OVERLAPPING ENVIRONMENTAL IMPACTS ARE SUBJECT TO REGULATION. No
information beyond that required to establish Initial regulatory require-
ments is needed to design a trading approach.
o A MARKETABLE PERMITS POLICY AND AN EMISSIONS CHARGE POLICY BOTH. RESULT IN
LOWER CONTROL COSTS THAN DOES A REGULATORY POLICY. Moreover, a marketable
permit policy offers as great certainty about resulting emissions and
-------
V-4
ambient air quality as does a regulatory policy, and both permits and tra-
ditional regulation result in greater environmental certainty than does
an emissions charge policy, at least in the short run,
o EMISSION CONTROL STRATEGIES, WHETHER COMMAMP-AIP-COMTROL OR ECONOMIC
INCENTIVE. THAT TAKE MARE GOOP USE OF DIFFERENCES IN EACH INDIVIDUAL
SOURCE'S COST OF CONTROL AHB AMBIENT IMPACT WILL ACHIEVE AMBIENT GOALS AT
SIGNIFICANTLY LOWER COSTS TBM STRATEGIES FOR WHICH EKISSION LIMITATIONS
ARE DEVELOPED ACROSS CATEGORIES OF SOURCES. A quantitative assessment of
pollution control strategies in the Chicago area showed that the approach
that exploited this relationship on a source-by-source basis to minimize
the cost of meeting ambient concentration targets was roughly thirteen
tines less expensive than the "traditional" approach to air pollution
control. However, because of implementation, problems, it is unlikely
that an air pollution, control agency could ever capture all these poten-
tial cost savings. Yet, these savings are of such magnitude that even
limited implementation of cost-effective abatement programs would result
in substantial cost savings.
A.5 FEASIBILITY
o ALTHOUGH MARKET INCENTIVE SCHEMES SHOULD BE USED WHEREVER FEASIBLE. THEY
WILL BE EASIER TO IMPLEMEST FOR POLLUTANTS FOR WHICH A SOURCE'S EMISSIONS
ARE REPRESENTATIVE OF ITS AMBIENT IMPACT. The reasons for this are:
— simplicity and ease of administration: When emissions are representa-
tive of a source's contribution to an ambient problem, a control
agency does not have to undertake (or supervise) elaborate ambient air
quality modeling to ensure that Source A can trade emissions limitations
with Source B without degrading air quality. It needs only to know
both A's and B's emissions,
— market, size: For some pollutants all emissions are thought (or assumed)
to contribute equally to the ambient air quality within an area —
regardless of where emitted within the airshed. Emitters of these
pollutants could thus trade emission limitations with any source within
the airshed without degrading air quality. The potential market in
-------
V-b
emission limitations is therefore quite large, However, for pollutants
where ambient impact depends not only on the quantity of the pollutant
emitted, but on the stack parameters and location of the source, trading
of emission limitations must be restricted to sources that can demon-
strate equivalent impacts on air quality within the sane location.
This has the effect of decreasing market size,
0 M GENERAL, a marketable permit system is preferable to a charge system
FOR ATTAINING ABB MAIKTAIKING AN AKBIEKT STANDARD. We base this con-
clusion. on the following findings of oar comparative analysis of these
two economic policies:
— Under a charge system the quantity of pollutants emitted depends upon
the response of sources to the costs imposed by a charge. Thus, the
• administering agency has greater certainty in the short run that
standards will be net under a permit system.
— To implement an efficient charge system, the administering agency must
acquire information about sources' control costs- This is a difficult
and expensive undertaking if cos.s are to he determined accurately.
Dnder a permit system, the quantity of emissions is fixed by the quan-
tity of permits issued, so the agency does not need detailed source-by-
-source cost data. Cost data would still be useful to design permit
systems that operate smoothly. These data will be revealed by sources
as they buy and sell permits from, one another.
— A marketable permit systea self-adjusts to inflation and growth. A
charge system requires that the agency make periodic adjustments to
these factors, adjustments which depend upon uncertain and perhaps
expensive data. Furthermore, frequent changing of charge rates nay
undercut the credibility of a charge system.
— A marketable permit system is administratively and legally similar
to permit programs now operated under regulatory control programs.
Consequently, it could be administered alongside existing regulatory
programs more easily than could a charge system, and would probably
-------
V--6
encounter less opposition from vested Interests, A marketable permit
system is also similar to the Offset and Bubble Policies currently
in force.
o CHARGE-BASED APPROACHES ME A FEASIBLE WAY TO MEET AMBIENT All QUALITY
STANDARDS ON A FIXED SCHEDULE ONLY IF SUBSTANTIAL INFORMATION ABOUT CON-
TROL COSTS IS AVAILABLE. Information of this kind is needed to predict
the response of sources to charges. Thus, in the absence of this informa-
tion, achievement of air quality standards on a fixed (particularly short-
term) schedule cannot be guaranteed.
o CHARGE-BASED APPROACHES ARE MOST ATTRACTIVE WHERE TRADITIONAL APPROACHES
WILL HAVE A DIFFICULT TIME MEETING AIR QUALITY STANDARDS ON SCHEDULE, WHERE
• THERE IS A FAIRLY LONG LAG BETWEEN EMISSIONS AND SUBSEQUENT ENVIRONMENTAL
EFFECTS, OR WHERE MAXIMUM FEASIBLE CONTROL EFFORT IS SOUGHT. Where attain-
ment of air quality goals on schedule is .-n doubt or there is no grave
concern over short-term damages during the period of necessary adjustment
la charge rates, charges can provide a means to induce cost-effective
control steps with very lltiile information on hand. Where maximum feasible
efforts are sought, charges can induce control efforts that it would not
have been feasible to identify in advance and efforts that could not
have been induced through enforceable regulatory obligations.
o EITHER STiUTO-ALONE CHARGES OR A TRADING APPROACH CM BE USED WKE8 THE
REGULATORS GOAL IS TO ACHIEVE A TARGET LE¥EL OF TOTAL EMISSIONS OR AIR
QUALITY. Trading approaches will be at a disadvantage when markets
would be very thin, however, and charges will be at a disadvantage when
information about control costs is limited and when environmental goals
must be achieved in a relatively short time frame.
o CURRENT APPROACHES TO THE ALLOCATION OF CONTROL OBLIGATIONS WILL BECOME
LESS FEASIBLE AS CONTROL REQUIREMENTS BECOME MORE STRINGENT. ECONOMIC
APPROACHES CAN FACILITATE THE USE OF ALTERNATIVE ALLOCATION METHODS.
Where additional control is required to meet air quality goals and all
easily identifiable and affordable controls have already been required,
it is difficult to assign more stringent requirements to particular
-------
V-
sources. Because economic approaches allow sources additional flexi-
bility, they can permit assignment of control obligations with less
information.
A. 6 -EQUITY
0 ECONOMIC AMD TRADITIONAL APPROACHES RAISE SIMILAR EQUITY ISSUES AMD
PROVIDE GO¥BRNMEBT WITH THE TOOLS TO ADDRESS THESE ISSUES. ECONOMIC
APPROACHES CAM MAKE THE EXISTENCE OF EQUITY ISSUES MORE OBVIOUS, HOWEVER,
All regulation involves the assignment of rights and obligations that
affsct the economic positions of regulated entities. Because economic
approaches can make these assignments explicit and "because they nay in-
volve transfers of money (either from sources to the government or between
sources), the distributive effects of environmental control are likely to
be more prominent.
A.7 IKHOVATXOtf
TlOim, AffKPACHIS. traditional approaches create incentives for sources
to reduce the costs of required controls and for vendors to develop new
equipment that nay form a basis for future standards. Economic approaches
allow sources greater flexibility in reducing costs and increasing abate-
ment and create incentives for sources to take advantage of this flexi-
bility.
A.8 DESIGN C0M5IBSSATJ0SS
o GAINS IN COST-EFFECTIVENESS WITH ECONOMIC APPROACHES WILL BE GREATER
THE MORE THAT SOURCES ARE PERMITTED TO ¥ART CONTROL LEVELS II RESPONSE TO
CHARGE RATES AHD ENTITLEMENT PRICES. If sources are required to meet
certain control requirements regardless of entitlements purchased or fees
paid, or if some emissions reductions receive no credit under the economic
approach, an economic approach will do less to reduce total control costs
than would otherwise be the case.
-------
v-s
0 ALLOWING "BANKING" OF ENTITLEMENTS CONTRIBUTES TO TOE SUCCESS OF A TRADING
APPROACH. BANKING IAS NO ECONOMIC DISADVANTAGES, If potential suppliers
of entitlements can undertake additional control effort whan it is con-
venient for them and have the ownership and size of resulting entitlements
certified for later use or sale, more sources are likely to undertake
unobligated emission reductions. The existence of a bank of available
entitlements will facilitate business planning by potential users of
entitlements.
o THE USEFULNESS AND FEASIBILITY OF A MARKETABLE PERMIT APPROACH DEPENDS
UPON WHETHER AN ALTERNATIVE CONTROLLED TRADING APPROACH COULD APPROXIMATE
THE CHARACTERISTICS OF PERMITS. Marketable permits offer little Incre-
mental benefit if controlled trading works well. This requires that
restrictions on trading be mlxiiiaizad and that states have the authority
and will to allocate control obligations in a manner that will assure
attainment.
° XM. APPROACH AN AGENCY TAKES TO ORGANIZING ENTITLEMENTS MARKETS AND TO
DETERMINING THE EQUIVALENCE OF EMISSIONS WILL, AFFECT THE SUCCESS OF A
TRADING APPROACH. Agency involvment in a market that is able to sustain
itself will interfere with trading. Predictablility in determining the
equivalance of emissions will facilitate trading,
A. 9 OTHER
o THE PRINCIPLES BEHIND EXISTING BEAVY-DPTY TRUCK NONCONFORMANCE PENALTIES
ARE EQUALLY APPLICABLE TO STATIONARY SOURCES. Nonconformance penalties
are a form of supplemental charge that can improve cost-effectiveness,
avoid economic disruption, and make regulation more feasible by providing
a safety valve in situations where control requirements cannnot be met
by all sources at an acceptable cost. However, such charges can disrupt
trading if set below the market-clearing price for entitlements.
o MOBILE SOURCE "AVERAGING- APPROACHES, WHICH APPLY PRINCIPLES SIMILAR TO
THAT OF THE BUBBLE POLICY„ ARE A PROMISING AREA FOR EXPANSION OF THE USE
OF ECONOMIC APPROACHES, Allowing manufacturers of vehicles to meet
-------
¥-9
emissions standards on an average basis would, like the bubble policy,
contribute to wore cost-effective control efforts.
0 CHARGE-BASED APPROACHES AND AUCTIONS OF ENTITLEMENTS RESULT IN MONEY
TRANSFERS FROM INDUSTRY TO THE GOVERNMENT. These transfers can be reduced
by charging only for emissions in excess of some threshold lev=l or by
distributing some entitlements without charge. Revenues from charges or
auction proceeds could be returned to the Treasury, used to compensate
the victims of pollution, used for the purchase of emissions reductions,
or used to fund agency operations.
B LEGISLATIVE IMPLICATIONS
-------
APPKKDIX
SPECIFIC STUDY DESCRIPTIONS
1- The Rand Stody of Chlorofluorocarbon Bnlssloas
Background
The Rand Corporation study of chlorofluorocarbon emissions examined one of
the simplest cases for use of economic approaches which is likely to be en-
countered. Chlorofluorocarbon (CFC) Is a noncritera pollutant, so the control
objective is simply co reduce euisslons rather Chan to attain a particular air
quality standard. CFC is believed to damage upper atmospheric ozone in direct
proportion to emissions, and damage is independent of source location or other
source characteristics- Therefore, traces can cake place easily. Furthermore,
emissions are easy to measure because ail CFC used in a product is eventually
emitted, and there are only a few manufacturers of CFC. These unusual factors
make application of an economic approach at the manufacturer le^el simple and
effective, and application at the user level feasible.
Approach
The quantitative analysis in the Rand study did not distinguish between
charge-based approaches (in this case, excise taxes) and trading approaches
(quotas or marketable permits). Instead, the study assumed for quantitative
purposes that charges would be set at a level that would result in the same
total emissions as a permit system and that this level could be known with cer-
tainty. Rand compared economic 1 approaches to a benchmark set of mandatory
standards devised by Rand to include all control measures which were enforce-
able, technologically feasible, and able to make a contribution to reduced
emissions by 1990.
Results
*
Total CFC emissions increased over time in Rand's benchmark case, but
declined relative to a "no controls" case. The cumulative costs of this
benchmark program were estimated to have a present discounted value of $185
-------
A-2
million. A charge program to achieve the same total emissions reduction in-
volved a present discounted cost of only $108 million. However, the present
discounted value of transfer payments (payments from sources to the government)
associated with such a charge system was estimated by Rand to be about $1.5
billion. Rand's approach did not provide any quantitative basis fot determin-
ing whether one economic approach was superior to another. Rand did note
that in practice permits allow greater certainty about emissions levels than
do charges; however, permits also require attention to allocation issues and
operation of a permit market.
Economic approaches allowed more effective control than mandatory stan-
dards because they provided incentives to sources to take steps which Rand
believed could not be compelled through mandatory standards. Thus, charges
could be set at a level ^hich would prevent any permanent increase in CFC
emissioas as the economy grows; at this level charges could generate roughly
twice the emissions reduction that could "be achieved with mandatory standards,
at about 150 percent of the cos; of the Less effective mandatory standards.
However, transfer payments at this high rate could exceed $6 billion. These
could be reduced or recycled, hut any step in this direction would reduce the
ability of these payments to internalize the social costs of emissions and
thereby promote more efficient allocation of resources. Rand is now examining
the transfer payment problem in greater depth fot EPA,
2. The Nichols Study of Benzene Emissions
Background
Under aa EPA grant, Albert L. Nichols of Harvard University studied ben-
zene emissions from the eight (of nine) maleic anhydride plants in the nation
which used benzene as a feedstock, at the time of the study. Like CFC, benzene
Is a noncriteria pollutant, but the situation studied differed from CFC in
important respects. Benzene is a hazardous hydrocarbon pollutant which af-
fects health near the source of emissions;, thus the impacts of each plant are
localized, and depend upon population exposure factors. These factors were
fifty times greater at the highest exposure plant than at the lowest. The
eight plants examined are sufficiently separately located that they have no
overlapping environmental effects, thus precluding use of a trading approach.
-------
'Marginal control costs per unit reduction In emissions also varied between, the
plants, because three plants lacked any controls while five already achieved
removal levels of 90 to 99 percent to neet state requirements tor control of
hydrocarbons. In fact, the plant with the highest population exposure factor
had -no controls in place and therefore a relatively low marginal cost of con-
trol, This coincidence made it possible to design either mandatory standards
or charges (based on emissions or population exposures) which induced addi-
tional control primarily at plants with low control costs and relatively large
local health impacts.
AoDroach
Nichol's study examined two mandatory control programs. The first would
have required a uniform percentage reduction in emissions for all sources;
the second would have divided plants into two groups facing different stan-
dards to Improve cost-effectiveness. Nichols also looked at uniform emissions
charges o£ various ratess and at population exposure charges which reflected
the differences in impacts of emissions at different plants. These approaches
were compared tinder different assumptions about the value placed on lives
saved.
Results
Nichols found that a uniform emissions charge was only slightly more cost-
effective than a mandatory control approach that divided sources into two
groups facing different standards. However, the more common mandatory program
for a hazardous pollutant would involve a high and uniform percentage removal
requirement for both groups. In fact, EPA collected data only on 97 and 99
percent removal levels, and Nichols found that the most cost-effective uniform
emissions charge was much more cost-effective than even the less stringent (97
percent) uniform reduction standard; the emissions charge provided 86 percent
of the benefits (0.332 lives saved, versus 0.386) at only 28 percent of the
costs ($600 thousand annually, versus $2 million annually), A charge related
to population exposure offered performance similar to this "optimal" emissions
charge. Nichols also found that exposure charges provided an additional 31.2
to $1,8 million in net benefits per year compared to the uniform 97 percent
-------
emissions reduction standard. The value within this range would depend upon
the value assigned to lives saved.
Nichols' favorable judgment on charges did not rest primarily on the
cost/benefits improvements charges offered in the situation he examined. In
factj in this situation some well designed mandatory programs nearly matched
the performance of economic approaches. Nichols stressed instead that charges
could provide a basis for a consistent balancing of control costs and benefits
for different sources, including sources of benzene other than maleic anhy-
dride plants. Nichols favored this outcome over an approach based on benzene's
status as a hazardous air pollutant, which now requires EPA to set regulations
based on use of the best demonstrated technology considering costs. Of course,
Nichols' approach is severely Halted under existing interpretations of Clean
Air requirements. Moreover, some confusion over the control technology re-
quired nay have caused the regulatory approach to fare worse by comparison
with charges than is warranted.
Nichols' strong expressions of concern about the reliabilty of the data
available to him, and physical changes .'_n the examined plants since the data
was collected, suggest that the general pattern of these results should receive
more attention than specific dollar values. Currently, all but two of the
eight plants no longer jse benzene in their production process. Hence such
developments may change the feasibility of a charge system for this industry
segment.
General Implications for Economic Approaches
The pattern of cost-effectiveness which Nichols found in his study ap-
pears to have been largely determined by the structure of marginal control
costs and impacts for the eight plants rather than by the inherent character-
istics of the approaches which he studied. Factors of this kind will play an
important role in determining whether an economic approach can make a useful
contribution in a particular case. In this case, all approaches give similar
results; all resulted primarily in additional controls being installed on a
plant which by coincidence combined very high impacts with low marginal con-
trol costs.
-------
A-5
The kinds of data chat were available to Nichols cannot be expected to be
available in most cases. Even in this case Nichols questioned the reliability
of the data that he had, and it quickly became obsolete. Careful program
specification based on extensive analysis of the effects of different kinds
and levels of charges will not normally be possible.
3. The Mathtech/EPA HQy Study
Background
The Mathtech, Inc., study of N0X in. Chicago examined a difficult case for
economic approaches—attaining a short-term ambient standard for a reactive,
pollutant with highly local impacts. Control over total emissions in a region
is not adequate to meet a stringent short-term standard for a reactive pollu-
tant; emissions from particular sources must be limited or trade-offs between
sources carefully calibrated ia order to avoid local violations of the stan-
dard.
Since the Mathtech study was completed, EPA has made corrections in the
inventory of stationary sources emitting N0X and in the model used to evaluate
quantitatively the efficiency and effectiveness of alternative approaches. The
results reported here are based on EPA'9 updated quantitative analysis. Com-
plete EPA results are reported in An Analysis of Economic Incentives to Control
Emissions of Nitrogen Oxides From Stationary Sources, a report to Congress
called for under Section 405(f) of the CAA Amendments, completed in January
1981.
Approach
Mathtech compared a mandatory control baseline to 'several economic ap-
proaches. The mandatory control alternative specified in the Mathtech/EPA
study was unusually sophisticated and therefore fairly cost-effective. The
baseline did not require control by all sources or all major sources, and it
took account of differences in ambient impacts and control costs for types of
sources. A uniform technology-specific control requirement was imposed, but
only on those classes of sources which could contribute to attainment at least
-------
A-6
cost. Vhe control requirement specified in the baseline was sufficiently
stringent to attain the NO^ air quality standard.
Mathtech compared this mandatory program to three emissions charge pro-
grams: a uniform charge on all sources; a source-category charge system which
set different charge rates for each of the three categories of sources on
which control requirements were Imposed in the mandatory baseline; and a set
of eharges which were tailored to the control costs and ambient impacts of
each source, to minimize the costs of achieving ambient standards. The study
also discussed the use of marketable penults.
Where an ambient air quality goal must he attained for a pollutant with
highly localized effects, it Is not an tasy task to determine optimal charge
rates for each source- Charge rates must not only reflect effects on air
quality at different locations, hut also the relationships between sources
which have overlapping impacts. It must be known whether control by a specific
source will be sufficient to reach attainment In a given area or whether a
higher charge rate that will also induce control by a second source will be
necessary. The Mathtech stjdy solved this problem with a sophisticated mathe-
matical programming model that considered each source's costs of control and
moved toward attainment in each area In a step by step fashion.
Results
The total annual cost of the controls required in Mathtech* s baseline
program was estimated to be $130 million. A uniform emissions charge on all
sources involved much higher control costs than this regulatory baseline.
This is not surprising, since any uniform approach affecting all sources dis-
cards information about relative ambient Impacts and control costs that was
used in setting up more selective mandatory baseline program. The charge rate
was necessarily set high enough for all sources to assure the specific emis-
sions reductions required to prevent local violations of the standards. At
this rate, the total cost to sources was $719 million, of which $414 million
was emissions fee payments. Total emissions under this charge were sharply
reduced from the baseline because the charge would induce a great deal ot
control that would make little contribution to attaining standards. Thus,
-------
A-7
the uniform charge would be more cost-effective than the relatively sophisti-
cated mandatory baseline in reducing emissions, but would be le-?s cost-effec-
tive than the baseline in attaining the MO,, air quality goal.
The source-category charge system examined by Mathtech reduced total con-
trol costs below the mandatory baseline usee in the Mathtech/EPA study, At
charges sufficient to reach air quality goals at all locations, this approach
resulted 1b total costs to sources of $155 million, of which $68 million were
emissions fee payments. Direct control costs under this program would there-
fore be about half those in the mandatory case.
The "customized"* source-by-source charge rate approach examined by Math-
tech improved substantially on the performance of category-based charges and
standards; costs to sources under this approach were only $13 million annually,
of which $4 million were emissions fee payments. This is a 93 percent reduc-
tion in control costs from the "baseline, and a 90 percent reduction in total
costs incurred by sources. Mathtech's analysis of marketable permits focused
on generic properties and design issues.
Mathtech suggested that trading should occur frequently and extend to
"options" and "futures" based or. permits (as well as on trading in the permits
themselves), so that sources could insure themselves against uncertainty about
their future situations. The cost savings actually achieved with trading
would depend on the extent to which trading was successful in reallocating
control obligations to the least cost sources. Therefore, impacts would be
intermediate between the source-category charge system and the source-by-source
charge system. If trading worked perfectly, the pattern of control effort
would be identical to that with source-by-source emissions charges; if sources
traded in such a way as to reduce but not to minimize costs, savings would be
less.
While some reviewers disagreed, Mathtech concluded that charges and per-
mits were feasible even in this complex control situation. Feasibility would
depend upon use of a computerized system (incorporating air quality and cost
modelling) to set charge rates or control trades; however, Mathzech cone i.uded
that this kind of system would be workable, efficient, and low in cost compared
to the cost savings that an economic approach could generate. A potential
-------
A-8
problem identified by Mathtech Is that sources with different air quality im-
pacts would, in. effect, pay different prices for permits to emit the same
amount of pollutant; this could be seen as unfair if the relationship of prices
to air quality impacts were not well -understood.
General Implications for Economic. Approaches.
Specification of charge rates for the Mathtech study required use of a
complex analytical tool uslag a complete set of high quality data. Data re-
quirements were met for study purposes zhrough the use of expediencies which
aay not be generally available; the study was set In Chicago because unusu-
ally good data on HO^ emissions sources existed, and control costs were esti-
mated in a simplified and idealized fashion because the control technology
required to meet the hypothetical standard has not yet been developed. Greater
difficulties can generally be expected ia gathering aoji working with actual
cost data.
The sophisticated modelling approach used by Mathtech led to results which
nay not be generally applicable. Mathtech's model solved for the lowest cost
pattern of control consistent with attainment, Once this was known, charge,
rates could be set at levels consistent with each sources marginal costs of
control in equilibrium. If control by a source was not optimal in equilibrium,
charge rates for that source could be set at zero. Emissions fees were there-
fore paic only by sources which also undertook additional control, and fee
payments were held to a minimum. This will not be a typical situation with
charges, which usually are expected to induce sources to sort themselves out
by control, costs. Since this Las'< had already been performed by the Mathtech
model, fees served only to induce the behavior the central planners had deter-
mined was most efficient. The Mathtech/E?A figures for emissions fee payments
under this approach may therefore not be typical of what would usually be ex-
pected of charge-based approaches of this kind.
The Mathtech study directly addressed the question of whether a charge
based or a trading approach would be best, and showed that total control costs
with a trading approach would be equal to those with a charge-based approach
if both programs could be implemented in an optimal fashion. Mathtech con-
cluded that trading (in the form of marketable permits) would be superior to
-------
\ 9
charges in practice, and the reasoning used should have some general applic-
ability. Trading offers the same potential for cost reductions as charges,
but achieved the same level of certainty as traditional regulation regarding
achievement of air quality goals. Moreover, trading required less initiative
on the part of the control agency to gather information about control costs
and would adjust automatically to inflation, economic growth and
technical change.
These results provide some indication that substitution of a charge-based
approach for mandatory controls (supplemented by trading) to meet ambient
standards will not automatically result in lower total control costs, at least
not for N0X control. Where ambient impacts depend on source location or other
source characteristics, an aaissions charge must discriminate among sources
about as well as the mandatory baseline does in order to reduce costs. Manda-
tory programs with trading can discriminate sufficiently well to question the
advisability of abandoning on-going programs and replacing them with untested
alternatives..
4, The MetA Systems Hydrocarbon Study
Background
Meta Systems, Inc.» examined charge-based approaches to the control of
hydrocarbon emissions in the South Coast Air Basin in a study funded by EPA but
with project responsibility at the Council on Environmental Quality. In keep-
ing with current Environmental practice, the impacts of hydrocarbon emissions
on ambient air quality within an air basin are assumed independent of source
location or characteristics. This simplifies the use of economic approaches
to pursue ambient standards related to hydrocarbon emissions. Like the Nichols
study on benzene, this study focused on a Halted number of plants. However,
in the Meta Systems study these plants did not constitute the universe of
plants of a particular kind; plants were selected to be representative of a
broad range of hydrocarbon-emitting activities in a single air quality control
region. This study did not address implementation problems associated with
charge systems.
-------
A-10
Approach
Meta Systems began its analysis by calculating the charge rates that would
be required to induce each source to take the same steps as would b«? expected
under mandatory controls. The costs and emissions reductions which would re-
sult from application of various charge rate to each source were then examined.
Results
Cost-based charges to induce equivalent control varied among sources from
sero to 59,560 per ton of hydrocarbon Quitted* Only one of the nine sources
saapled, however, had an associated charge in excess of SI,000 per ton. This
wide range of charges reflects a wide range of marginal control costs under
the requirements in place. Because control cost varied so widely it rela-
tively easy to divide the sources into two groups with a uniform emissions
charge, set so that high cost sources could avoid controls while low cost
sources generated necessary emissions reductions. A wide range of charges
existed which would provide increased control and reduced costs simultaneously,
relative to what would have been expected as a result of alternative comraand-
and-controi approaches. Whan ail sources were faced with a charge of $>00 per
ton, total emissions were 25 percent lower than under the mandatory controls,
and expenditures for controls were reduced about 3 percent, froa $12.9 million
to $11.9 million. After-tax control costs exclusive of fee payments also fell
from $8.5 million to $8.0 ailIion. However, emissions fee payments resulted
in total costs to sources roughly equivalent to those with mandatory controls.
General Implications for Economic Approaches
The Meta Systems study demonstrates the feasibility of using engineering
analysis to gather information about control costs for specific facilities. In
the case examined by Meta Systems, marginal control costs varied so widely that
it was easy to specify a uniform emissions charge rate which simultaneously re-
duced cofts and emissions. If this is a typical situation and if administration
of charges is feasible, charge approaches as an alternative to mandatory con-
trols can achieve cost savings. A major unresolved question is whether inform-
ation from specific facilities can be used in setting charge rates for sources
that have not been as extensively characterized as those in this study,
-------
A-ll
5. The Putnamt Hayes and Bartlett Marketable Permits Study.
Approach
Putnam, Hayes and Bartlett, inc., studied the design and application of
marketable permit systems. This conceptual exercise set out to define the
conditions for and describe the operations of marketable permit systems, and
to evaluate the applicability of this approach to S0X, SO9. HC, ISP, and CFC.
Further work is now being done on CFC permits. The study devised and used a
sinplified model of an air quality control region consisting of a limited
number of sources for which emissions and cost characteristics were specified.
Results
Putnam, Hayes and Bartlett concluded that marketable permits could be
used to implement current regulatory policies, but were less readily applic-
able to the control of M)x than to other pollutants. Specific recommendations
on system design were that permanent permits be allocated without charge to
continuous major point sources in amounts (and in units) consistent with
attainment of NAAQS, and that government play a major role In designing the
system, monitoring market operations, and enforcing compliance.
The model devised for the study demonstrated that if trading ended after
a single round of exchanges, the amount of control cost savings achieved would
depend upon the specific trades which were undertaken. A similar conclusion
about savings from trades was reached in the Mathtech study; however, there is
no reason why sources should not continue to trade and retrade permits so long
as cost reductions are available.
6. The Repetto PSD Study
Background
Under an EPA grant, Robert Repetto of Harvard University examined alterna-
tive approaches to PSD increment allocation. So particular pollutants or
grr.ups of sources were studied.
-------
A-12
Approach
The four alternative approaches to allocation of PSD increments which
Repetto examined were first-come, first-served allocation; allocation via
auction; imposition of a requirement to secure offsets at less than a one-to-
one rat:„o from existing sources; and permitting excess emissions upon payment
of a fee. Internal EPA analysis has also examined the use of better trackings
more stringent BACT review, retrofits, administrative reservation of the in-
crement and allocation to locally preferred sources as potential approaches
to the PSD allocation problem. The first two approaches examined by Repetto
ration the supply of entitlements in a PSD area, and the third makes supply
dependent upon the willingness of existing sources to reduce emissions, The
implicit assumption, of the fourth approach is that there is some cost of con-
trol above which it is not worthwhile to maintain ambient standards. Repetto
explored the relationships between these allocation approaches and market
conditions to assess potential effects on the environment and on industrial
expansion. Four demand and four supply conditions, two offset ratios and two
fee levels, and three different competitive situations were examined, for a
total of 768 numerical cases. Nine criteria were examined in each case. Off-
set ratios that were continuously variable were discussed but not numerically
analyzed. Ideally, offset ratios would vary over time to reflect the antici-
pated cost of and need for future controls, and the time value of money, but
this would involve difficult data collection and administrative problems.
Repetto also examined the available evidence on what market conditions
might be in the future and addressed the broader efficiency issues involved in
the use of SAAQS. Repetto added some relevant evidence on the costs of control
to the question of whether use of NAAQS is efficient in the broadest sense.
Results
Repetto concluded that the present first-come, first-served allocation
system contributed less to efficient control than the alternatives he examined,
since incentives toward cost-effectLveness in meeting air quality targets when
a new source enters an area are provided neither to new or existing, sources by
this approach. However, EPA'8 internal analysis suggests that the gains from
-------
A-13
implementing an auction approach In lieu of other alternatives which are work-
able and could be implemented in the short run would probably be small in most
eases.
Repetto's extensive analysis of hypothetically possible cases did not
identify a single alternative approach that was likely to be preferrable to
all others under several criteria and in most conditions. Under a wide range
of market conditions requiring partial offsets tended to provide the strongest
incentives for abatement, the highest payments to existing sources, and the
lowest net increase in emissions. No other ranking of approaches over the
nine criteria was possible over a wide range of market conditions. Repetto
therefore concluded that the most cost-effective approach to allocating the
entitlements would depend upon local conditions in each case. Repetto made
two generalizations about the effects of local conditions on cost-effective-
iwss:
° If projected new source growth is large relative to the level of exist-
ing emissions, and if BACT cost levels do not greatly exceed RACT levels,
an auction mechanism is mos: likely to promote cost-effectiveness.
0 If the existing emissions inventory is large relative to projected
growth, then a variant of the offset market might be more conducive to
cost-ef fectiveness.
EPA* s analysis demonstrates that PSD auctions would be complicated to
administer for pollutants with local impacts, if potential sources had to
acquire bundles of entitlements to impact air quality at all relevant mea-
surement. points at auction. This problem can be eased if some needed en-
titlements can be obtained in a market for entitlements, or if incomplete sets
of entitlements can be sold in such a market.
The evidence on market conditions assembled by Repetto indicated that
market conditions might lead to domination of local markets at particular
times by one or a few buyers or sellers. Demand would usually involve one
major source seeking entitlements in a given air quality region in any given
year; rarely would more than one new source seek entitlements in the same air
quality region, and even sources within one region might have different areas
-------
A-14
of impact. Similarly, most air quality regions contain few large sources of
emissions which could be expected to sell offsets, limiting the potential for
competition on the supply side.
The importance of these competitive conditions is lessened by the likeli-
hood that the available PSD increment will exceed the total demand for such
entitlements in most areas and for most pollutants. The conditions under
which a good is allocated become important only if that good is in scarce
supply relative to demand. Available evidence indicates that this situation
will be rare for TSP, and rare for SO2 by the year 2000, An SO2 increment
allocation problem will exist in the nea„: terms however.
Repetto's discussion of the efficiency of NAAQS combined theoretical and
a
empirical analysis, The theoretical argument against NAAQS as currently set
rests on the assumption that environmental control should be an exercise in
trading off environmental and economic costs under conditions of uncertainty.
If it is believed that the social costs of emissions are approximately propor-
tional to total emissions, while marginal costs of control increase rapidly
over the relevant range, then Repetto suggests that Lt nakes little sense to
set inflexible limits on emissions. Repetto argues that outside of pristine
air areas where aesthetic damages are ar,. issue, the social costs of emissions
in PSD areas are likely to be proportional, to total emissions. Where aesthetic
damages are an issue, low emissions levels can do great damage, but once this
damage is done additional emissions are not a source of additional costs.
Repetto also concludes, on the basis of earlier studies of abatement costs for
SO2, that marginal control costs do escalate rapidly in the relevant range.
The implication, is that in Class 1 areas emissions should "be very tightly
controlled, to the point of compelling sources to locate elsewhere, but that
in the rest of the country the allowed level of emissions should be sensitive
to control costs.
7. The ICF Study of ISPS Offsets
Approach
ICF, Inc., exaaiaed the Impacts of allowing utilities to use emissions
reductions in existing coal-fired boilers to offset emissions in excess of
-------
A-15
NSPS SO2 requirements for new coal-fired boilers. Five hypothetical utilities
in different regions of the country were examined, and each utility was as-
sumed to offset emissions from one new 500 Mw boiler with reductions in emis-
sions from an existing 500 Mw boiler. Different assumptions about SIP require-
ments were aade in different ragions; operating parameters were also varied
froo case to case. These differences, and differences in the costs and avail-
ability of particular coals, resulted in different control approaches being
used in different scenarios.
Results
In eight of the nine cases involving the recently revised SSPS for utility
boilers, allowing SSPS offsets would make it unnecessary for a utility to
install a scrubber. This results in considerable cost savings. The magnitude
of these savings ranged from 0-7 to 10.2 percent of annualized costs or $1.1
to $18.2 million per year for a 500 Mw unit. Capital cost savings ranged from
6 to 29 percent of total facility costs, or $26 million to $148 million per
500 Mw unit.
8. The Temple, Barker and Sloane Study of Noncompliance Penalties
Background
The Temple, Barker and Sloane, Inc. study of noncompliance penalties
analyzed the potential economic impacts of EPA's program on the iron and steel
industry and the electric utility industry. These two industries account cor
a substantial portion of major sources known to be out of compliance and a
substantial percentage of excess emissions of TSP and SO2.
Approach
The report made macr©economic, industry-wide arid company-specific esti-
mates of effects. These estimates used EPA-estimated total penalty payments
of $165 million for iron and steel arid $181 million for electric utilities,
and assumed rapid progress toward compliance.
-------
A-16
Results
The economic, effects of the penalties in these industries as ,1 whole
would be very slight« However, about naif the individual firms in the iron
and steel industry would be "noticeably" affected by penalties, and some util-
ity companies could be significantly affected. Electricity prices could
increase by up to 10 percent for some electric utilities if the entire amount
of penalties were passed through, although the average increase would be only
0.2 percent. Earnings for two sample utility companies would decline by over
20 percent, although the average decline in earnings for the industry would be
only 0,7 percent. These large variations make it clear that sons utility com-
panies are enjoying substantial economic benefits from delayed compliance
with environmental requirements.
Approach
TCS Management Group, Inc., examined two economic enhancements to manda-
tory inspection and .Maintenance programs, and three other economic approaches.
The economic enhancements to t&M were waiver of subsequent inspection when an
inspection was passed, and payment of a reward for response to an emissions
recall. Both enhancements were reasonably cost-effective.
The three other economic approaches were compared to a mandatory control
baseline. One such approach, use of a corporate (fleet-wide) average standard
for control of N0X and CO CHC data were not adequate for analysis), was ex-
amined as an alternative to currently applicable individual vehicle standards.
As specified by TCS, this approach did aot involve trading among manufacturers
but did allow trade-offs in controls among different model lines of the same
manufacturer. TCS also examined both manufacturer charges based on new car
emissions as a replacement for the current regulatory system, and charges
based on NOx and CO emissions of vehicles tested in state inspect i on and
aaintainar.ee programs as a supplement to the current system.
TCS defined the parameters of Its alternative programs and of the poten-
tial responses of manufacturers to fit the data that were available. Cost
-------
A—17
data were available only for meeting 1980 and 1981 emissions levels; there-
fore, manufacturer behavior under averaging and new car charges was restricted
to controlling model lines either to the levels specified in 1981 standards or
to the levels specified in the 1980 standards. Cost data Limits affected
which programs could be analyzed* The stringency of the average standard had
to be set below the existing standards so that manufacturers would be able to
meet the average standard while controlling Home model lines zt 1980 levels.
Since averaging was not compared to other approaches at the same overall
level of control, cost-effectiveness ratios for the program appeared to
be more favorable than otherwise. Similarly, new car charges were based on
the average costs and average emissions reductions expected from the 1981
standards. With charges at this level, manfacturers with above average costs
will chose to pay the charge rather than control emissions, assuring chat
cost-effectiveness ratios would improve and emissions would increase. These
biases were offset by other factors. Since the control alternative assumed
to be available to manufacturers were so restricted, cost-effectiveness ratios
appeared to be less favorable than otherwise. It is not possible to say which
of these offsetting biases predominates.
In the case of in-use emissions charges (as supplements to the existing
control system), biases were not offsetting. TCS found it necessary to arti-
ficially restrict the range of possible responses by manufacturers to in-use
charges and concluded that this would bias downward the measurement of cost-
effectiveness.
Results
TCS concluded that corporate average emissions standards could offer great
cost advantages over individual vehicle standards, without significant econo-
mic or political drawbacks. New car charges were also seen as much more, cost-
effectlve than the current system, ICS concluded that difficulties ia predict-
ing the effects of an in-use charge made quantitative analysis unreliable.
Both charge-based approaches were seen to involve substantial political and
economic problems resulting primarily from, higher new car prices,
TSC found that emissions averaging would lead to a decrease in average
per-vehicle control costs of $258 (in 1980 dollars), which would show up as a
-------
A-18
3.3 percent decrease in mew car prices. Program administration and vehicle
certification costs increased slightly, by about $1 million and $10 Billion,
per year, respectively, relative to the base case. Emissions increased because
controls were less stringent. Cost-effectiveness ratios indicated a potential
for substantial savings from use of an averaging approach. The 1981 standards
without averaging resulted in CO removal costs of S3,533 per ton and N0X
removal costs of $16,299 per ton. Averaging reduced these costs to S2,720
and $15,000 respectively.
TCS found new car charges to be more cost-effective than the current regu-
latory approach. Under the current approach the controls required to meet
1981 standards would result in average control costs of $423 per vehicle above
those required to laeet the 1980 standards, causing a 5.4 percent increase in
vehicle prices. The new car charges examined would induce less total control,
but would involve a 1380 to 1981 cost increase of only $90 per vehicle. Man-
datory controls were substantially less cost-effective than new car charges,
eliminating CO at a resource cost of $",533 per ton and N0X at a resource
cost of 316,299 per ton, in contrast to $1,337 and $11 ,750 respectively.
However, the combined costs of installing controls and paying charges for
emissions which were not eliminated would force manufacturers to impose average
price increases of $852 per vehicle, a 10.9 percent increase. These higher
prices would contribute to inflation and reduce new car sales. On the other
hand, the higher prices would reflect an internalization of the presumed social
costs of emissions, and lead to a more economically efficient allocation of
resources.
Although in-use charges appeared to be less cost-effective in the TCS
analysis than new car charges or averaging approaches, TCS cautioned against a
conclusion that this would be true if any of these programs were actually put
in place*
10. The Sobotka Study of TSP Averaging for Diesel Vehicles
Approach
Sobotka & Company, Inc., examined averaging for TSP emissions from diesel
vehicles as an alternative to the recently promulgated 1985 standards. Programs
-------
A-19
based on averaging across only a manufacturer's diesel fleet, and across both
the gasoline and diesel fleet (corporate averaging) were discussed, but quan-
titative analysis was confined to diesel-only averaging- The. impacts of trad-
ing among manufacturers in connection with averaging and the impacts of allo-
cating entitlements by auction were also measured.
The technology which is expected to be used to meet the 1985 standards
for light-duty diesel TSP has net yet been fully developed, so data of the
quality used by TCS were not available for this study. Sobotka therefore
extrapolated available data on the costs and performance of early prototypes
of the control hardware into a set of cost and performance relationships that
depended upon specific vehicle characteristics.! Manufacturers were allowed
to use any level of control on any vehicle to meet the average standard,
Because manufacturers could be given this flexibility, Sobotka sat the average
standards at a level that resulted in approximately the saae level of total
emissions as the .1985 standard without averaging, and compared the costs of
reducing emissions to this single level under each of the two approaches.
Results
Sobotka found that averaging approaches could be substantially more
cost-effective than individual vehicle standards. Cumulative 1985 through
1989 costs to meet the 1985 standard were estimated to be SI.7 billion (in
undiscounted 1980 dollars); with diesel-only averaging, total emissions de-
clined by about three percent and total cumulative costs fell by $100 million.
The study found that allowing trading among manufacturers in addition to
averaging would permit an additional $20 million in savings, would not have
adverse economic Impacts, and would probably be perceived as fair and be
feasible. Total payments for entitlements under an auction, assuming that
1 Because the data used by Sobotka are extrapolations from a small.
number of prototype tests, they are probably not very reliable, The apparant
precision of the study's scenarios and results should be assessed accordingly.
Some results in the Sobotka study axe sensitive to data characteristics that
may not hold up when hardware to control diesel TSP has been fully developed.
-------
A-20
bidding was competitive and based on control costs, were estimated at about
$350 million. This greatly exceeds the cost savings available from averaging.
11. The Policy Planning and Evaluation NOT Studies
Approach
Policy Planning and Evaluation, Inc. (PP&E), examined mobile source emis-
sions charges and averaging in separate studies; these studies focused on
program design to a greater degree than the TCS and Sobotka studies. PP&E
also examined a marketable permits system that would impose a ceiling on total
emissions from new vehicles regardless of production levels, rather than con-
trolling emissions or average emissions per vehicle. No quantitative analysis
of charges was undertaken, and quantitative analysis of trading approaches was
limited.
Results
PP&E generally concluded that any of the trading approaches it examined
would be feasible and could offer benefits. PP&E also identified specific
areas in which difficulties in implementation would have tc be overcome. The
only quantitative analysis in the PP&E averaging study, based on data froa the
TCS study, looked at the potential benefits of allowing trading between manu-
facturers in connection with an average standard. A control cost reduction of
about six percent was indicated.
PP&E reached mixed conclusions about new vehicle NO* emission charges.
This study found charges to be theoretically sound, but the authors remained
uncertain of their feasibility and likely Impacts. So quantitative analysis
was undertaken.
-------
BIBLIOGRAPHY
Energy and Environmental Analysis, Inc., Analytical Procedures to Estimate Pollu-
tion Control Costs and to Simulate Markets In Emission and Air Quality Off-
sets in Nonattainment Areas (Draft), Prepared for the Environmental Protec-
tion Agency, Contract No - 68-01-3871, Arlington, Virginia; April 5, 1979.
Environmental Protection Agency, 'Assessment and Collection of Noncompliance
Penalties by EPA and Approval of State Noncompliance Penalty Program"
(Final Rule), 45 Federal Register 50086-50258, July 28, 1980.
Environmental Protection Agency, "EPA Advance Notice of Proposed Rulemaking On
Emissions Averaging of Oxides of Nitrogen" (Advance Notice of Proposed
Rulemaking)s 45 Federal Register 79382-79384, November 28, 1980.
Environmental Protection Agency, An Analysis of Economic Incentives to Control
Emissions of Nitrogen Oxides from Stationary Sources, Report to Congress
called for under Section 405(f) of the CAA, Washington, D.C.; January 1981.
ICF, Incorporated, Concept Paper: Emission Reduction Banking and Trading Systems,
Prepared for the Environmental Protection Agency, OPA 113/0, Washington,
D.C., June, 1980.
ICF, Incorporated, Preliminary Evaluation of Economics of NSPS Offsets for
Utility Powerplants (Draft), Prepared for the Environmental Protection
Agency, Washington, D.C.; June, 1980„
Mathtech, Inc. (Robert J. Anderson, Jr., Robert 0. Reid, and Eugene P. Seskin),
An Analysis of Alternative Policies for Attaining and Maintaining a Short-
Term NO? Standard, Prepared for the Environmental Protection Agency,
Council on Environmental Quality, and Council of Economic Advisors, Prince-
ton, New Jersey; September 17, 1979.
Meta Systems, Inc. (Paul Ginberg and Grant W. Schaumberg, Jr.), Economic Incen-
tive Systems for the Control of Hydrocarbon Emissions from Stationary
Sources, Prepared for the Council on Environmental Quality, Washington,
D.C.; September, 1979.
Nichols, Albert L., "Alternative Strategies for Regulation Air-Berne Benezene,"
Incentive Arrangements for Environmental Protection: A Critical Examination
(Draft), pp. 220-335, Prepared for the Environmental Protection Agency under
Grant R805446010, Harvard University; undated.
Policy Planning & Evaluation, Inc. (James Roberts, and Jasbinder Singh), "Economic
and Administrative Feasibility Analysis of Alternative New—Car NQV Emission
Charge Strategies, Prepared for the Environmental Protection Agency, Contract
No. 68-01-5124, McLean, Virginia; December 15, 1979.
Policy Planning & Evaluation, Inc.., (James Roberts), New Automobile ^0,,. Emis-
sions Regulatory Systems: Offset Credits (Emissions Averaging) and Marketable
Permits Systems, Prepared for the Environmental Protection Agency, Contract
No. 68-01-5124, McLean, Virginia; November 12, 1980.
-------
- 2
Putnam, Hayes & Bartlett, Inc. (Robert: L. Hayes and Stephen A. Basse), Noncon-
formance penalties im Heavy Bat? Engines, Prepared for the Environmental
Protection Agency, Itewton, LsMCtasfttST April 5, 1978. '
Putnam, Hayes & Bartlett, Inc. (Robert L. Hayes and Stephen A. Besse), Discus-
sion Paper II on iancotfemaace PenaUto* for Hjsswy Duty Engines, Prepared
Eor
24, '978.
Protection
Newton, Massachusetts; January
Putnam, Hayes 4 Bartlett, Inc.,
trol oj^LgPolltttion,
idge, Massachusetts; October 9, 1980.
land Corporation (Adele R. Palmer, William 1. Mooz, Timothy H. Quinn, and Kath-
leen A. Wolf), Ec;
Emissions from .!
Pcotection Agency,
>, for the
ica, California; June, 1980.
Repetto, Robert, "The Prevention of Deterioration in the Air Quality and the
Use of Economic Systems in Allocating Emissions Rights," Incentive A
aeats for fwitwjwtol Protection: A Critical ttsgatnation (DraftH PP
336-455, Prepared for the Environmental Protection Agency under Grant
R805-446010, Harvard University; undated.
jbotka & Company, lac.,
Emissions (Draft),
for Regulating
§ust"l, 1980.
Diesel. Particulate
TCS Management Group, Inc., Calculating Heavy Duty Engine Konconformance Penalties,
October 12, 1979.
-------
------- |