October 23, 1996

EPA-SAB-COUNCIL-LTR-97-001

Honorable Carol M. Browner
Administrator
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC  20460

      SUBJECT:   Council Review of the Clean Air Act Section 812 Retrospective
                  Study entitled "The Benefits and Costs of the Clean Air Act, 1970
                  to 1990"

Dear Ms. Browner:

      On June 5-6, 1996, the Advisory Council on Clean Air Compliance Analysis
(the "Council") met to review the May 3, 1996 draft of the EPA document, The Benefits
and Costs of the Clean Air Act, 1970-1990: Report to Congress  (EPA, 1996; Also
referred to as the "study," the "Retrospective Study," and the "draft document"). EPA
intends to submit the final version of this Retrospective Study as a Report to
Congress pursuant to Section 812 of the Clean Air Act Amendments(CAAA) of 1990
(CAA, 1990).

      In general, the Council was favorably impressed with this draft. It is based on
an extraordinarily rich set of data and models, and it reflects an enormous amount of
hard, careful, and effective work by Agency staff and contractors. This draft document
is a significant advance over what we  have seen in the past;  as a result, the final
document as an Agency Report to Congress has the potential to be a landmark study.
That document should make significant contributions to our understanding of the
overall costs and  benefits of the Clean Air Act (CAA, see CAA, 1970 and CAA, 1990).
(We recognize that it was impossible to disaggregate the pre-1990 analysis by
programs and look forward to such a disaggregation in the prospective cost-benefit
study of the CAA  now in preparation.)  The final version of this draft document should
advance the state of the art and science of cost-benefit analysis. In addition,

-------
difficulties encountered during this study have important implications for research
priorities, which can be selected to create a basis for improving future reports.
      The Council's task was, of course, not to praise this draft document but to help
improve the final document as an Agency Report to Congress. To that end, individual
Council members have provided detailed editorial suggestions to involved Agency
staff via the SAB staff.  More importantly, the Council agreed that unless the issues
raised in the remainder of this letter are addressed effectively, the final version of this
study will be open to serious criticism.  We believe the Agency can address them all
effectively with relative ease.

General Issues

      The tone of the current draft document is unnecessarily argumentative and
defensive at points.  As several members observed, by sometimes seeming to cross
the line from analysis to advocacy, the authors of this draft document weaken the
strong case made by their analytical work.  In particular, this draft document seems to
deal with the inevitable imperfections of cost-benefit analysis very reluctantly; the idea
that any estimate in this draft document might be at all uncertain does not appear until
p. xvi  of the executive summary1. (The use of "roughly 20 dollars" in the headline
sentence is misleading in this regard.)  The Council believes that an accurate and
complete treatment  of the uncertainties and gaps in knowledge encountered in this
study should be included  and will, on balance, strengthen the final document as an
Agency Report to Congress.

      Similarly, this draft document is  weakened by almost always using a single
discount rate. Because choice of a discount rate is inevitably controversial, the next
draft document should include more complete analyses of sensitivity to that choice (in
Appendices) as well as clear statements in the text of when that choice has a material
impact on an important category of benefits or costs.

      A second set of problems seems to have arisen primarily because of hasty
writing.  It is sometimes difficult, particularly in the Appendices, for even a diligent
reader to understand exactly how the analysis was performed. It would  be  impossible
for a diligent reader to reproduce the results of some steps in that analysis  based on
the information provided in this draft document, and that is not acceptable in a
technical document.  Similarly, obviously relevant literature is sometimes not cited,
even when it is supportive of the proposition being advanced, and controversial
      1 The reference to page numbers, Chapters, Tables and Appendices refers to text being reviewed by the Council
specifically related to the EPA draft document (EPA, 1996).

-------
issues are often passed over quickly in the text without citing more extensive
treatments in the Appendices. Additionally, the Agency should be sure to cite the
primary references, not just secondary references.

      A third set of problems involves careless wording on some sensitive points.
This is particularly troublesome when it occurs in the executive summary, which is
likely to be by far the most widely-read part of the final document as an Agency
Report to Congress. The misleading description of economic valuation on p. xv is a
case in point. The costs used in this study are called "Clean Air Act compliance
costs" at several points, though in principle at least, they clearly include costs of
complying with environmental regulations that were in place before 1970. Contrary to
the assertion on p. xviii, as a matter of pure logic, total benefits well above total costs
in  1990 does not necessarily imply "that substantial additional cost-beneficial
protection remained to be achieved" because it implies nothing about marginal costs
or marginal benefits.  While it may be the case that further "cost-beneficial protection"
can be achieved for some pollutants, the present analysis does not shed any light on
this question. It is not the case (ch.  1, p. 7) that the direct costs of the CAA are
"readily obtained"; the cost estimates here rest on  a number of judgments regarding
the evolution of technology in the absence of the CAA. The draft document regularly
(e.g., p. xvii) refers to unmonetized benefits,  which are present, and ignores the
possibility of unmonetized costs. Such costs may  arise, for instance, because this
analysis (sensibly) does not attempt to quantify possible adverse impacts of regulation
on the rate of technical change - effects that would continue long after 1990.  A
number of additional issues of this sort are raised in the separate memoranda of
individual Council members.

Costs

      In general, this draft document devotes too  little attention to the significant
uncertainties that affect the estimated control costs. Moreover, it is not possible, as it
should be, for a diligent reader to reconstruct the process by which the cost estimates
were derived. In addition to these general problems,  several specific questions
require attention.

      First, the cost numbers are said to be  based primarily on The Cost of Clean
(EPA, 1990). In some cases, however, those numbers appear to be inconsistent with
(and below) the data underlying the annual articles published by  the Department of
Commerce's (DOC) Environmental Economics Division in DOC's The Survey of
Current Business (U.S. DOC, various issues).  Because it is likely to become a
standard  reference, the Agency's Retrospective Study should explain the

-------
discrepancies.  Moreover, the adjustments performed on The Cost of Clean (EPA,
1990) numbers are not well-explained. In principle, they should be reproducible so
that other researchers can understand how the numbers used in the Retrospective
Study were derived and can test alternative assumptions.

      Second, the most important source of cost uncertainty (and potential
underestimation) relates to  the potential "chilling" effect of environmental regulation
on innovation, which represents some cost to society.  Four significant issues in the
following subparagraphs (a to d) deserve attention; all point to omissions that tend to
underestimate costs:

      a)    The draft document presents annualized total direct compliance costs
            (about $25 billion in 1990; see Table 28).  It also presents the
            Jorgenson-Wilcoxen (J-W) estimates of the reduction in real Gross
            National Product (GNP) in percentage terms (1% in 1990; see Table
            29).  This presentation is consistent with our earlier advice to de-
            emphasize changes in GNP because it is an inferior measure of
            changes in economic welfare (SAB, 1996b). However, as the draft
            document itself acknowledges (p. 57), the direct compliance costs are
            also an imperfect measure because of their failure to capture a variety  of
            indirect effects of regulation on economic welfare.  We recommend that
            the draft document include a presentation of the annual reduction in  real
            GNP (about $55 billion in 1990,  or 1% of the 5,524 billion GNP of that
            year) to give the reader an indication of the potential importance on
            indirect effects as contributions to the total cost of the 1970 CAA.

      b)    The draft document as a Retrospective study used as its
            macroeconomic benchmark the  J-W model.  One of the features of this
            model is its effort to incorporate Endogenous Technological Change
            (ETC). In our June 1996 report  (SAB, 1996b), we  had recommended
            against making ETC the base case of the analysis. This draft document
            presents only  the results without ETC.  We understand that the
            estimated reductions in GNP are substantially larger when the model is
            run with ETC.  We recommend that the results of both model runs be
            presented so that the reader can understand the importance of the
            assumptions about ETC for estimates of total costs.

      c)    The annualization of all stock data adds a further complication to this
            problem.  The J-W estimates are for the "flow" of services from
            technology rather than the stock of technology.  The estimate of the loss

-------
             in output should include not only the annual technological losses but
             also the technological losses that continue past 1990.  This is a
             complicated technical and computational problem, but it is unavoidable if
             we wish to get an accurate estimate of the costs of regulation.  Sample
             calculations by one of the Council members suggest that taking account
             of future  losses due to reductions in innovation and technological
             change could increase estimated costs substantially, perhaps by almost
             a factor of two. This issue should be addressed in the draft document.

       d)     A final question involves the work of Gray and Shadbegian (1993; 1995)
             on the "technological chilling  effect" of regulation. They estimate that
             the indirect costs of regulation are significantly larger than the direct
             costs estimated by the Census Department.  This work was reviewed by
             Industrial Economics,  Inc. (lEc) in a memorandum (Peterson and
             Unsworth, 1995) that is quite misleading and flawed in a number of
             respects.2 We are not suggesting that this study should accept the
             Gray-Shadbegian conclusions without qualification, but it  should
             certainly  include sensitivity analyses that allow for these costs (as with
             the J-W costs), recognizing that important elements of costs are
             uncertain. In addition, the questions of future technology  discussed in
             connection with the  J-W estimates are equally relevant for the chilling
             effect of regulation on productivity.

Emissions and  Air Quality

       Numerous assumptions were required to create the emission inventories used
in this study, and this has resulted in significant uncertainties, especially for the no-
control scenario and for volatile organic compounds (VOCs). Although these
uncertainties are unlikely to have  a significant impact on the outcome of this study,
they need to be  more thoroughly articulated (Table 6 should be discussed) in the
chapter and appendix.
     3This is not the place for a detailed review of the I EC memorandum (Peterson and Unsworth, 1995), but a few points are worth
  making. The lEc memorandum is incorrect in stating that the authors in the 1995 study "correct an important methodological
  problem." Based on a conversation with Dr. Gray, who was not given a chance to review the lEc memorandum, the Gray-
  Shadbegian work in 1995 (Gray and Shadbegian, 1995) includes fixed effects in the regression. Dr. Gray stated (conversation on
  June 12, 1996  with Dr. William D. Nordhaus) that he is somewhat suspicious of the fixed-effects approach because of timing and
  other measurement issues; while this approach provides alternative information, it does not supersede the 1993 estimates (Gray and
  Shadbegian, 1993). In addition, the lEc study argues that the Gray-Shadbegian studies "omit certain factors" such as the impact on
  worker productivity. These impacts are in principle included in the benefits and are not an appropriate reason to omit an analysis of
  the impact on productivity.

-------
      Similarly, numerous assumptions have resulted in significant uncertainties in
the predicted no-control scenario air quality measures.  This was especially true for
those pollutants and measures not linearly related to emissions: ozone, visibility, acid
deposition, and, to a lesser extent, particulate matter (PM). As with emissions, the
Clean Air Scientific Advisory Committee (CASAC) Air Quality Models Subcommittee
(CASAC/AQMS) concluded that these uncertainties were unlikely to affect the study's
outcome significantly, but they need to be explicitly stated.  (In particular, Table 7
should be  discussed in the text, and the more extensive discussions in Appendix C
should be  referenced.) The present chapter conveys a degree of precision that is not
justified. [For the sake of completeness, it should be noted  that after discussing the
CASAC/AQMS letter report (SAB, 1996a), dated May 31, 1996 with that group's
Chairman  (Dr. George Wolff), the Council endorsed the CASAC/AQMS report and
conclusions.]

Physical Effects

      The Council's Physical Effects Review Subcommittee (PERS) met on June 4,
1996. The Council met on June 5 and 6, 1996. The PERS Chair (Dr. Morton
Lippmann) and Vice-Chair (Dr. A. Myrick Freeman) attended the Council meeting, and
their deliberations were reviewed at some length.  The Council is comfortable with
their main  conclusions as we understand them. Their work raises important, but
corrigible,  problems of both substance (choice and use of studies) and presentation
(the "low-mid-high" approach, also discussed below). The  Council has approved the
PERS (EPA, 1996c) letter report and has recommended that  it be submitted to the
Administrator of the Agency. This Council report endorsed the findings and
recommendations of the  PERS report, to which the reader is referred for the complete
presentation of details (SAB, 1996c).

Valuation

      We believe that the values applied to statistical lives saved by the CAA should,
in principle, reflect the health status of those saved and their remaining life
expectancy.  We also believe that the value currently applied to a statistical life, $4.8
million (1990 U.S. dollars), significantly overstates the value most people would attach
to the average number of life years saved (per person) by the CAA.  One must
recognize  that this number is based on a set of labor market wage/risk tradeoff
studies of  working age adults with a lower average age and high life expectancy than
the age groups with the higher risks of death due to air pollution considered in this
study.

-------
      We recommend that the premature mortality avoided by the CAA be stated in
terms of the expected number of life years saved, and that life years be valued
instead  of statistical lives.  At the very least, something must be said about the age of
the statistical lives saved.  This can be done in the case of episodic mortality studies
by applying the age-specific coefficients in those studies to the U.S. population.  If the
$4.8 million (1990 U.S. dollars) figure is retained, it should be made clear that this
was derived as the value of saving the statistical life of a 40-year-old. Another
approach is to value the changes in the mortality risks based on estimates in the
literature of how this value varies with age.  We discussed this approach in some
detail and provided relevant references on page 8 of our June 1996 report (SAB,
1996a).

      An alternative approach to presenting the mortality benefits of the CAA that
avoids having to explicitly value either life years or statistical lives is to compute the
cost per life or per life year saved implied  by the Act. For example, one would divide
the $25.1 billion in compliance costs for 1990 by the estimated 79,000 lives saved to
arrive at a cost per life saved of $318,000. This, of course, ignores the benefits of air
pollution control other than reduced mortality.  It also begs the question "Is $318,000
greater  or less than the value of saving a statistical life?" The computation does,
however, permit a comparison of the cost  per life saved of the CAA with the cost per
life (or life year) saved of other health and safety regulations. We believe that
efficient regulation requires that such comparisons be made.

      The study currently  values heart attacks and strokes avoided by the CAA by
the value of preventing a severe case of chronic bronchitis. In view of obvious
differences among the three illnesses, it is essential that this benefits transfer be
justified. This could be accomplished by comparing the medical costs and foregone
earnings associated with each of the three diseases. To the extent that they are
similar,  one could argue that the three diseases place similar limitations on at least
some aspects of performance and that willingness-to-pay to avoid them should
therefore be similar.

      The estimates of the value of reducing lead emissions contains estimates of
the value of Intelligence Quotient (IQ) points lost. Our most recent report indicated
that including induced effects in this calculation represents an error (see SAB, 1996a,
point 4.9, page 12). This error persists in  the current draft document.

      The choice of a discount rate is not adequately discussed  in the current draft
document.  A 3 percent per year real discount rate is used almost exclusively, though
this is not warranted either by Administration policy (which mandates  a 7 percent real

-------
discount rate), by agency practice (which, the Council was told, is generally to use a 5
percent real discount rate), or by documented reasoning. As we have indicated in
earlier reports, we believe that a range of rates should be employed, though
differences in results should be  presented only when they are substantial.  As a
general matter, it is not clear how much the choice of discount  rate would affect the
conclusions of the Retrospective Study.  One clear point is in the estimates of the
value of IQ points lost because of lead in the no-control  scenario. One Agency official
in attendance at our Council meeting indicated that the IQ benefits (in dollars) in lead
would be reduced by 2/3 if a 7 percent discount rate were used.  If true, this should be
made clear.

       Finally, the Council discussed two important issues of presentation of valuation
analysis. First, more disaggregated information on benefits should be given in the
text and executive summary.  At present, the executive summary contains information
on a variety of estimated physical effects in Table 1 and information on a variety of
estimated values in Table 3, and these are reprinted as Tables 11 and 14,
respectively, in the text.  However, only total benefits (computed by multiplying
physical effects times values and summing) are presented - in Figures 3, 21, and 22.
The Council believes that a table of benefit values, disaggregated at roughly the level
of Tables 1 and 3, should be presented in both the text and the executive summary.
Second, because reduced mortality is such an important source of dollar benefits,
both physical effects (Tables 1 and 11) and dollar benefits, the new Tables just
discussed should show the effects of changes in lead  separately from the effects of
changes in other pollutants.  (The Council recognizes the conceptual difficulty of
further disaggregation.)

Uncertainty

      The  analysis of uncertainty given in Chapter 7 of this draft document focuses
on developing a plausible probability distribution for aggregate benefits.  This
distribution is determined by probability distributions for  individual physical effects and
valuation parameters. Unfortunately, while the latter distributions are presented in
appropriate detail in the Appendices, the reader is given no justifications for the
assumptions on which those distributions rest. Some  distributions used seem to be
anchored on the highest and lowest credible estimates of physical effects within the
literature base. This is inappropriate, if only because  the range between the high and
low can be expected to increase as the number of studies included increases, while
the true uncertainty would be  decreasing with the inclusion of more studies (other
things being equal).   Rather,  the distributions used in the analysis of Chapter 7 should
reflect reasoned judgments regarding the implications of the literature. In turn, the

                                       8

-------
ranges of physical effects presented in the text and executive summary - and the
ranges of values that should be presented in both places - should represent
confidence intervals of the distributions used in the analysis of Chapter 7.

      Monte Carlo analysis should also be used to provide information regarding the
relative importance of the many sources of uncertainty that underlie the aggregate
distributions currently summarized in Chapter 7. The first step in this procedure is to
single out one variable, to set all others equal to the corresponding best estimates,
and to compute the distribution of aggregate benefits implied  by the uncertainty in the
singled-out variable alone.  Applying this procedure to all variables treated as
uncertain in the computations summarized in Chapter 7 yields a set of distributions,
the variances of which can be readily compared. Chapter 7 should include a short
table giving the benefit distribution variances implied by the largest few sources of
uncertainty, compared to (or divided by) the variance of the aggregate distribution.
This one-variable-at-a-time analysis is  important because it shows where reduction in
uncertainty would most improve the precision of aggregate benefit estimation.

Research Issues

      Appendix J represents a unique opportunity to use the lessons from the
analytical process created by this study as a foundation for establishing air pollution
research priorities in the future.  The current version  of Appendix J misses this
opportunity by presenting views which  are largely independent of the analysis rather
than flowing logically from  it.

      During the  research process a number of gaps in available information were
uncovered. Some suspected physical  effects (such as ecological and health effects
associated with air toxics) could not be quantified.  Other quantified physical effects
(such as nonmarketed forest services) could not be reliably valued.  Many others
could be quantified and valued,  but either the corresponding uncertainty could not be
well-described or the literature implied  very wide confidence intervals. Collecting and
characterizing these information gaps would provide  a useful  menu of needs and
opportunities for future research.

      Setting priorities among these gaps would require some comparison of the
value of additional information with its cost.  For those effects where estimates were
possible, but only within  wide confidence intervals, it should be possible to gain some
insight about the value of additional information. The one-variable-at-a-time Monte
Carlo studies (described above) could  be used to identify those specific relationships
where greater precision  (resulting in smaller confidence intervals) could have the

-------
greatest impact on the aggregate estimates. Combining these results with some
information on the cost of filling individual research gaps would allow some inferences
about research priorities to be drawn.

      Our review has identified significant uncertainties and gaps in data and support
information.  Since it appears that cost/benefit analyses will become a continuing
method for evaluating the CAA, a strategic plan for research must be developed in
order to identify, validate, prioritize,  and employ the most appropriate variables and
indicators for these types of analyses.

      Some of the necessary variables already exist, but there are major gaps (e.g.,
ecological metrics) which can only be addressed through a focused, well-designed,
strategic plan of research and development. Such a plan can be used to establish
priorities that will effectively direct resources to reduce exposure and adverse health
or environmental responses by controlling sources of concern.

      Appendix J also discusses the "Prospective Study" of post-1990 CAA costs and
benefits that is in preparation.  Two issues seem to us particularly important in this
context.  The first concerns the treatment of mortality as discussed in the Valuation
section above.  A recurring issue in  the Retrospective Study has been the  life-
lengthening impact of reduced air pollution. The Council inquired on this point on a
number of occasions, and the  question is still unresolved.  The  Prospective Study
could be improved if EPA were to increase the research effort devoted to this
important question.  One approach would be to estimate empirical survival curves as
a function of concentrations. In fact, it would appear to be both prudent and possible
to conduct such an analysis of existing data (e.g., Pope et al., 1995) as a part of the
retrospective study.

      A second issue concerns the utility of this work for improving regulatory design.
One of the shortcomings of the current Retrospective Study is the inability to
disaggregate beneath the totality of all regulations. As a result, we cannot use the
current Retrospective Study to determine in a rigorous way which air pollution controls
are cost-beneficial and which are not, or which ones are highly  cost-beneficial and
which ones are less so.

      A major task for the CAA Section 812 Prospective Study, then,  is to ensure that
the future CAA Section 812  Study design allows such a disaggregation.  In the
discussion of the prospective work,  it would be useful to emphasize that the problem
of deploying the enormous machinery used in the Retrospective Study for actual
                                      10

-------
program and regulatory policies is critical and should be absolutely central as the
CAA Section 812 Study team moves on to the prospective studies.

      Once again, it has been a pleasure for the Council members to work with able
and dedicated Agency staff on this important project. We expect that you and
everyone involved will be proud of the quality of the final product, and we look forward
to your response to our findings and recommendations.

                        Sincerely,
                         )r. Richard Schmalensee, Chair
                        Advisory Council on Clean Air
                          Compliance Analysis
                                     11

-------
                                  NOTICE

      This report has been written as part of the activities of the Science Advisory
Board, a public advisory group providing extramural scientific information and advice
to the Administrator and other officials of the Environmental Protection Agency.  The
Board is structured to provide balanced, expert assessment of scientific matters
related to problems facing the Agency. This report has not been reviewed for
approval by the Agency and,  hence, the contents of this report do not necessarily
represent the views and policies of the Environmental Protection Agency, nor of other
agencies in the Executive Branch of the Federal government, nor does mention of
trade names or commercial products constitute a recommendation for use.

-------
            U.S. ENVIRONMENTAL PROTECTION AGENCY
                 SCIENCE ADVISORY BOARD (SAB)
    ADVISORY COUNCIL ON CLEAN AIR COMPLIANCE ANALYSIS

CHAIR
Dr. Richard Schmalensee, Deputy Dean, Sloan School of Management,
Massachusetts Institute of Technology, Cambridge, MA

MEMBERS
Dr. Maureen Cropper, Principal Economist, The World Bank, Washington, DC

Dr. Ronald G. Cummings, Professor of Economics and Noah Langdale, Jr. Professor
of Environmental Policy, Policy Research Center, Georgia State University, University
Plaza, Atlanta, GA

Dr. Daniel Dudek3, Senior Economist, Environmental Defense Fund, New York City,
NY

Dr. A. Myrick Freeman, Professor, Department of Economics, Bowdoin College,
Brunswick, ME

Dr. Robert Mendelsohn, Edwin Weyerhaeuser Davis Professor, School of Forestry
and Environmental Studies, Yale University, New Haven, CT

Dr. William D. Nordhaus, Professor,  Department of Economics, Yale University, New
Haven, CT

Dr. Wallace E. Oates, Professor, Department of Economics, University of Maryland,
College Park, MD

Dr. Paul R. Portney, President, Resources for the Future, Washington, DC

Dr. Thomas H. Tietenberg, Professor, Department of Economics, Colby College,
Waterville, ME

Dr. W. Kip Viscusi, Professor of Law and Economics, Harvard University Law
School, Cambridge, MA
      3     Although a member of the Council, Dr. Dudek was unavailable to participate in this specific review.

-------
LIAISONS
Dr. William Cooper4, Professor of Zoology, Institute for Environmental Toxicology,
Michigan State University, East Lansing, Ml (Liaison from the SAB's Environmental
Processes and Effects Committee)

Dr. Wayne M. Kachel, Project Manager, Mele Associates, Brooks Air Force Base, TX
(Liaison from the SAB's Environmental Engineering Committee)

Dr. Paul Lioy, Professor, Environment and Community Medicine,  Robert Wood
Johnson Medical School, Piscataway, NJ (Liaison from the SAB's  Integrated Human
Exposure Committee)

Dr. George T. Wolff, Principal Scientist, Environmental and Energy Staff, General
Motors Corporation,  Detroit, Ml (Liaison from the SAB's Clean Air  Scientific Advisory
Committee)

SCIENCE ADVISORY BOARD STAFF
Dr. K. Jack Kooyoomjian, Designated Federal Official, Science Advisory Board
(1400), U.S. Environmental Protection Agency, Washington, DC 20460

Mrs. Diana L. Pozun, Staff Secretary, Science Advisory Board (1400), U.S.
Environmental Protection Agency, Washington, DC 20460
      4      Although a liaison from the EPEC, Dr. Cooper was unavailable to participate in this specific review.

-------
                                ABSTRACT

      The Advisory Council on Clean Air Compliance Analysis (the Council; formerly
known as the Clean Air Compliance Analysis Council, CAACAC) of the Science
Advisory Board (SAB) has reviewed on June 5 and 6, 1996 the Agency's May 3,  1996
draft Report to Congress entitled "The Benefits and Costs of the Clean Air Act, 1970-
1990." While the Council was favorably impressed with this draft and noted that it
reflects an enormous amount of careful and  effective work by the Agency staff, a
number of suggestions were offered to improve the final report.

      Some of the major issues that the Council stressed include the estimation and
valuation of changes in mortality, the treatment of uncertainties and gaps in
knowledge, the clear presentation of cost methodology, the exposition of important
issues and  data, the need to conduct a sensitivity analysis on the discount rate, the
need to  consider the indirect costs of regulation,  and the need to display more
disaggregated information on benefits in the text  and executive summary.

      The Council recognized that this is an unusual opportunity to use the lessons
learned  from this study to establish sound research priorities.  The  Council, drawing
on the work of its subcommittees, identified significant uncertainties and gaps in  data
and support information.  The Council also noted that research priorities should reflect
the continuing important role of cost/benefit analysis in program design and
evaluation.
Key Words: Air Pollutants, Clean Air Act, Cost-Benefit Analysis, Economic Valuation,
Valuation Methodologies

-------
                            REFERENCES CITED

CAA. 1970. Clean Air Act. Public Law 91-604, December 31, 1970

CAA. 1990. Clean Air Act. Public Law 101-549, Section 812, 104 STAT2692, November 15,
      1990

EPA. 1990. "Environmental Investments: The Cost of A Clean Environment," U.S.
      Environmental Protection Agency, Office of Policy Planning and Evaluation (OPPE),
      EPA-230-11-90-083, November, 1990

EPA. 1996. "The Benefits and Costs of the Clean Air Act, 1970 to 1990," Prepared for U.S.
      Congress by U.S. Environmental Protection Agency, draft, May 3, 1996

Gray and Shadbegian. 1993. Bureau of Census Discussion Paper CES 93-6

Gray and Shadbegian. 1995. National Bureau of Economic Research (NBER) Working paper
      4994, 1995

Peterson, T.B and R.E. Unsworth.  1995. "The Effect of Environmental Regulations on
      Productivity,"  Memorandum from T. B. Peterson and R.E. Unsworth of Industrial
      Economics, Inc. (lEc), May 12, 1995.

Pope, C.A. Ill; Thun, M.J.; Namboodiri, M.; Dockery, D.W.; Evans, J.S.; Speizer, F.E., and
      Heath, C.W., Jr. 1995. Particulate Air Pollution is a Predictor of Mortality in a
      Prospective Study of U.S. Adults. Am. J. Respir. Care Med., Vol. 151, March 1995, pp.
      669-674

SAB. 1996a. CASAC Comments on Air Quality Modeling for the Section 812 Retrospective
      Study, U.S. Environmental  Protection Agency, Science Advisory Board (EPA-SAB-
      CASAC-LTR-96-007), May  31, 1996

SAB. 1996b.  ACCACA Review of the Agency's Retrospective Study of Section 812 Clean Air
      Act Benefits and Costs from 1970 through  1990, U.S. Environmental Protection Agency,
      Science Advisory Board (EPA-SAB-ACCACA-96-003), June 3, 1996

 SAB. 1996c. Review of "The  Benefits and Costs  of the Clean Air Act, 1970 to 1990" by the
      Physical  Effects  Review Subcommittee (PERS)  of the Advisory Council on Clean Air
      Compliance Analysis (ACCACA), U.S. Environmental Protection Agency, Science Advisory
      Board (EPA-SAB-ACCACA-LTR-96-010),  September 26, 1996

U.S. DOC. various issues. Survey of Current Business, "Pollution Abatement  and Control
      Expenditures," U.S. Department of Commerce (DOC), Environmental Economics Division,
      published in various issues beginning twice/year and now once/year for the past several
      years

-------
                           DISTRIBUTION LIST

Administrator
Deputy Administrator
Assistant Administrators
Regional Administrators
Office of the Administrator
      Office of Cooperative Environmental Management
Deputy Assistant Administrator for Air and Radiation:
      Director, Office of Policy Analysis and review (OPAR)
      Director, Office of Air Quality Planning and Standards (OAQPS)
Deputy Assistant Administrator for Policy, Planning and Evaluation (OPPE)
      Director, Office of Policy Analysis (OPA)
      Director, Office of Regulatory Management and Evaluation (ORME)
      Director, Office of Strategic Planning and Environmental Data (OSPED)
Deputy Assistant Administrator of Research and Development:
      Deputy Assistant Administrator for Science - ORD
      Deputy Assistant Administrator for Management - ORD
      Director, Office of Science Policy
      Directors, Mega Laboratories and Centers
      Directors, Research  Laboratories
EPA Headquarters Library
EPA Regional libraries
National Technical Information System (NTIS)
Library of Congress

-------