MATHTECH
The Technical Research
and Consulting Division of
Mathematica, Inc.
BENEFIT AND NET BENEFIT ANALYSIS OF
ALTERNATIVE NATIONAL AMBIENT AIR QUALITY
STANDARDS FOR PARTICDLATE MATTER
VOLUME I
-------
BENEFIT AND NET BENEFIT ANALYSIS OF
ALTERNATIVE NATIONAL AMBIENT ATR QUALITY
STANDARDS FOR PARTICOLATE MATTER
VOLUME I
Prepared for:
Benefits Analysis Program
Economic Analysis Branch
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
March 1983
-------
EPA PERSPECTIVE
There has been growing concern with the effectiveness and burden of
regulations imposed by the Federal government. In order to improve the
process by which regulations are developed, Executive Order 12291 was
Issued. The order requires that Federal agencies develop and consider, to
the extent permitted by law, Regulatory Impact Analyses (RIA) for the
proposal and promulgation of regulatory actions which are classified as
major. According to the order, a significant component of the RIA is to be
an economic benefit and benefit-cost analysis of the regulatory alternatives
considered. Under the Clean Air Act, the Administrator of EPA may not
consider economic and technological feasibility in setting National Ambient
Air Quality Standards (NAAQS). Although this precludes consideration of
benefit cost analyses in setting NAAQS, it does not necessarily preclude
consideration of benefit analyses for that purpose.
In full support of the Executive Order, the EPA commissioned Mathtech,
Inc. to accomplish an economic benefit and benefit-cost analysis of some of
the alternatives that were thought likely to be considered in the development
of proposed revisions to the NAAQS for particulate matter (PM). The report,
entitled "Benefit and Net Benefit Analysis of Alternative National Ambient
Air Quality Standards for Particulate Matter," documents the results of the
contractor's study. One of the major objectives of the study was to give a
better understanding of the complex technical issues and the resource
requirements associated with complying with the spirit of the Order for the
NAAQS program. In order to achieve this objective, the contractor was
given a wide range of latitude in the use of data, analytic methods, and
underlying assumptions.
It is important to stress that the benefit analysis portion of the
Mathtech study has not had a role to date in the development of proposed
revisions to the NAAQS for particulate matter. Staff recommendations
currently under consideration are based on the scientific and technical
information contained in two EPA documents. They are the "Air Quality
Criteria for Particulate Matter and Sulfur Oxides" and the "Review of the
National Ambient Air Quality Standards for Particulate Matter: Assessment
of Scientific and Technical Information, OAQPS Staff Paper." These documents
have undergone extensive and rigorous review by the public and the Clean
Air Scientific Advisory Committee in accordance with the Agency's established
scientific review policy. Although the Mathtech study reflects the
"state-of-the-art" in particulate matter benefit analysis, the approach and
results have not been subjected to a comparable extensive peer review
process. In addition, some EPA staff have raised questions regarding the
approach taken in the analysis and the significance of the results for
standard setting purposes under the Act. These circumstances do not
necessarily preclude use of the benefit analysis in some manner after
appropriate peer review and further consideration of the questions that
have been raised.
-------
BENEFIT AND NET BENEFIT ANALYSIS OF ALTERNATIVE
NATIONAL AMBIENT AIR QUALITY STANDARDS FOR
PARTICDLATE MATTER
By:
Ernest H. Manuel, Jr.
Robert L. Horst, Jr.
Kathleen M. Brennan
Jennifer M. Hobart
Carol D. Harvey
Jerome T. Bentley
Marcus C. Duff
Daniel E. Klingler
Judith K. Tapiero
With the Assistance of:
David S. Brookshire
Thomas D. Crocker
Ralph C. d'Arge
A. Myrick Freeman. Ill
William D. Schulze
James H. Ware
MATHTECH, INC.
P.O. Box 2392
Princeton, New Jersey 08540
EPA Contract Number 68-02-3826
Project Officer:
Allen C. Basala
Economic Analysis Branch
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
March 1983
-------
The analysis and conclusions presented in this report are
those of the authors and should not be interpreted as necessarily
reflecting the official policies of the U.S. Environmental
Protection Agency.
-------
PREFACE
This report was prepared for the U.S. Environmental Protection Agency
by Mathtech, Inc. The report is organized into five volumes containing a
total of 11 sections as follows:
Volume I
Section 1:
Section 2:
The Benefit Analysis
The Net Benefit Analysis
Volume II
Section 3:
Section 4:
Appendix:
Health Effects Studies in the Epidemiology Literature
Health Effects Studies in the Economics Literature
Valuation of Health Improvements
Volume III
Section 5:
Section 6:
Section 7:
Section 8:
Residential Property Value Studies
Hedonic Wage Studies
Economic Benefits of Reduced Soiling
Benefits of National Visibility Standards
Volume IV
Section 9:
Section 10:
Air Quality Data and Standards
Selected Methodological Issues
Volume V
Section 11:
Supplementary Tables
IV
-------
ACKNOWLEDGMENTS
While preparing this report, we had the benefit of advice, comments
and other assistance from many individuals. Allen Basala, the EPA Project
Officer, and James Bain, former Chief of the Economic Analysis Branch
(EAB), were especially helpful. They provided both overall guidance on
project direction as well as technical review and comment on the report.
Others in EAB who assisted us included Thomas Walton, George Duggan, and
John O'Connor, the current Chief of EAB.
Others within EPA/OAQPS who reviewed parts of the report and assisted
in various ways included Henry Thomas, Jeff Cohen, John Bachman, John
Haines, Joseph Padgett, and Bruce Jordan.
Several individuals within EPA/OPA also provided comments or assis-
tance at various stages of the project. These included Bart Ostro, Alez
Cristofaro, Ralph Luken, Jon Harford. and Paul Stolpman.
Others outside EPA who reviewed parts of the report and provided
comments included V. Kerry Smith, Paul Portney, Lester Lave, Eugene Seskin,
and William Watson. Other Mathtech staff who assisted us in various ways
were Donald Wise, Gary Labovich, and Robert J. Anderson. We also
appreciate the assistance of Al Smith and Ken Brubaker of Argonne National
Laboratory who conducted the parallel analysis of control costs and air
quality impacts.
Naturally, it was not possible to incorporate all comments and
suggestions. Therefore, the individuals listed above do not necessarily
endorse the analyses or conclusions of the report.
The production of a report this length in several draft versions, each
under a tight time constraint, is a j ob which taxes the pa-tience and sanity
of a secretarial staff. Carol Rossell had this difficult task and managed
ably with the assistance of Deborah Piantoni, Gail Gay, and Sally Webb.
Nadine Vogel and Virginia Wyatt, who share the same burden at EAB, also
assisted us on several occasions.
-------
CONTENTS
VOLUME I
Section Page
1. THE BENEFIT ANALYSIS
Introduct ion 1-1
Scope of the Benefit Analysis 1-2
Air Quality Standards 1-2
Coverage of Effects/Benefit Categories 1-3
Overview of the Study Approach 1-4
Identification and Classification of Existing
Research Studies 1-4
Review and Selection of Best Available Studies 1-6
Benefit Calculation Procedures 1-6
Sources of Uncertainty in the Benefit Calculations ... 1-9
Scenarios Analyzed 1-12
Basel ine Air Qual ity 1-12
Alternative Standards 1-13
Index of Exposure 1-18
Averaging Times and Distribution Parameters 1-19
Particle Measures 1-19
Geographic Areas and Monitor Types 1-21
Plausibility Checks 1-23
Measurement and Economic Valuation of Health Improvements . 1-24
Reduced Mortality Risk 1-25
Reduced Morbidity Risk 1-27
Applicable Exposure Ranges for Concentration-
Response Functions 1-27
VI
-------
CONTENTS (Continued)
Section
Page
1. THE BENEFIT ANALYSIS (Continued)
Summary of Results in Specific Benefit Categories 1-34
Health Effects 1-34
Soil ing and Materials Damage 1-40
Visibility and Aesthetics 1-41
Property Value and Labor Market Studies 1-42
Aggregate Incremental Benefits 1-43
Methods of Aggregation 1-43
Results of the Aggregation Procedures 1-51
Incremental Benefits for Other Standards and
Scenarios 1-54
Geographic Distribution of Benefits 1-61
Findings and Conclusions 1-63
Estimates of Incremental Benefits 1-63
Limitations of the Estimates 1-64
References , 1-65
2. NET BENEFIT ANALYSIS
Introduct ion .............................................. 2-1
Benefit-Cost Analysis: Evaluation Criteria ............... 2-2
Incremental Benefits and Costs ....................... 2-2
Cost Effectiveness .................................... 2-3
Efficiency Criterion ........................ . ........ 2-3
Scope of Analysis ....... . ............................ 2-4
Benefit-Cost Analysis : Methodology ....................... 2-5
Benefit-Cost Analysis: Incremental Benefits and
........................................... 2-6
Distributional Effects ............................... 2-6
Vll
-------
CONTENTS (Continued)
Section Page
2. NET BENEFIT ANALYSIS (Continued)
Measurement of Benefits and Costs 2-8
Measurement of Benefits and Costs: Conceptual
Issues 2-8
Estimates of Benefits 2-10
Estimates of Costs 2-14
Consistency of Benefits and Costs Estimates 2-29
Limitations and Qualifications of Benefit-Cost Analysis ... 2-32
Benefit-Cost Analysis of Alternative PM NAAQS 2-33
Benefit-Cost Analysis: Incremental Net Benefits 2-34
Distribution of Incremental Net Benefits 2-47
Conclusions and Qualifications 2-56
Conclusions 2-57
Qualifications 2-59
References 2-60
vui
-------
FIGURES
VOLUME I
Figure No. Paae
1-1. Basic Steps in Estimating Benefits for an Individual
Study 1-8
1-2. Typical Air Quality Scenario 1-16
1-3. Residual Non-Attainment (Scenario "A") 1-17
2-1. Basic Steps in Estimating Emissions Control Costs 2-16
IX
-------
TABLES
VOLUME I
Table No. Paae
1-1. Effects Categories Potentially Relevant to Alternative
PM Standards 1-3
1-2. Sources of Uncertainty in the Benefit Calculations 1-11
1-3. Alternative Air Quality Standards 1-14
1-4. Summary of Concentration Ranges in EPA/OAQPS Staff Paper .. 1-29
1-5. Incremental Health Benefits for the PM10 (70, 250)
Scenario B Standard 1-35
1-6. Comparison of Incremental Health Benefits for the PH10
(70, 250) Scenario B Standard, With and Without a
Lower Bound Applied 1-39
1-7. Incremental Benefits of Reduced Soiling for the PM10
(70, 250) Scenario B Standard 1-41
1-8. Incremental Health and Welfare Benefits for the PM10
(70, 250) Scenario B Standard as Measured from
Property Value and Hedonic Wage Studies 1-43
1-9. Alternative Aggregation Procedures 1-44
1-10. Incremental Benefits for the PM10 (70, 250) Scenario B
Standard 1-52
1-11. Incremental Benefits for Alternative PM10 and TSP
Standards (B Scenarios) 1-55
1-12. Incremental Benefits for Alternative PM10 and TSP
Standards (A Scenarios) 1-58
1-13. Incremental Benefits for Alternative PM10 and TSP
Standards, With Lower Bound Applied (B Scenarios) 1-59
-------
TABLES (Continued)
Table No.
1-14. Incremental Benefits for Alternative PM10 and TSP
Standards, With Lower Bound Applied (A Scenarios) 1-60
1-15. Incremental Benefits by Region for the PM10 (70, 250)
Scenario B Standard 1-62
2-1. Summary of Incremental Cost Estimates 2-27
2-2. Incremental Net Benefits for Alternative PM10 and
TSP Standards (B Scenarios) 2-36
2-3. Incremental Net Benefits for Alternative PM10 and
TSP Standards (A Scenarios) 2-37
2-4. Incremental Net Benefits for Alternative PM10 and
TSP Standards With Lower Bound Applied (B Scenarios) 2-38
2-5. Incremental Net Benefits for Alternative PM10 and
TSP Standards With Lower Bound Applied (A Scenarios) 2-39
2-6. Domain of Economically Preferred Standards 2-43
2-7. Incremental Net Benefits by Region for the PM10
(70, 250) Scenario A Standard 2-49
2-8. Incremental Net Benefits by Region for the PM10
(70, 250) Scenario A Standard With Lower Bound Applied .... 2-50
2-9. Incremental Net Benefits by Region for the TSP
(70, 260) 9-Year Scenario A Standard 2-52
2-10. Incremental Net Benefits by Region for the TSP
(70, 260) Scenario A Standard With Lower Bound Applied .... 2-53
2-11. Capital and After-Tax Annualized Costs (ATAC) of Control,
By Industry, for the TSP (75, 260) 9-Year Standard
(Scenario A) 2-55
-------
SECTION 1
THE BENEFIT ANALYSIS
-------
SECTION 1
THE BENEFIT ANALYSIS
INTRODUCTION
This report develops estimates of the benefits and net benefits of
alternative national ambient air quality standards for particulate matter
(PM). The report has been prepared to assist the U.S. Environmental
Protection Agency in responding to the requirements of Executive Order
12291 which requires an analysis of the costs, benefits, and net benefits
of major regulatory initiatives (1). This section is a summary of the
benefit analysis results and methods documented in detail in Sections 3
through 11 of the report.
Benefit estimates alone are not sufficient to identify the economi-
cally preferred standard. The costs of attaining and maintaining the
standards must also be considered. The benefits and costs of the alterna-
tive standards are compared in Section 2 of this report and the estimated
net benefits are calculated. The cost estimates were developed by another
contractor using analytical methods described in a separate report (2).
While cost analyses for new regulatory initiatives have been routinely
developed by the U.S. Environmental Protection Agency, benefit analyses
have only recently been undertaken. Benefit analysis in the past has been
hindered primarily by limitations of data and methods available for
measuring health and welfare effects of air pollution.
Recognition of the technical complexities confronting benefit analysis
suggests use of an analysis strategy which has three key elements:
1-1
-------
Insuring that the analysis identifies and makes use of the
best data and information currently available.
Incorporating validation checks and sensitivity analyses
into the analysis at each stage where technical uncertainty
is present.
Using a range of estimates or methods for valuing health
improvements.
These efforts help identify the degree of uncertainty present and provide
greater assurance about the quality of the results.
Section 1 is divided into nine parts including this introductory
section. The other eight parts discuss: the scope of the analysis; the
method of analysis; the air quality scenarios; the indices used to measure
air quality; the procedures for measuring and valuing human health improve-
ments; the benefit estimates by benefit type; the aggregate benefits; and
the conclusions.
SCOPE OF THE BENEFIT ANALYSIS
Air Quality Standards
The benefit analysis examines several alternative ambient PM
standards. The alternative standards involve various combinations of
particle size measures, allowable ambient concentrations, implementation
dates, and averaging times. The standards considered are a subset of those
analyzed in the EPA's cost analysis and air quality modeling effort (2).
While the alternatives do not exhaust the list of possibilities, they do
include th« current primary and secondary standards, and a variety of the
alternative revisions to those standards. By design, the primary standard
is intended to protect human health, while the secondary standard is to
provide protection of the public welfare.
1-2
-------
Coverage of Effects/Benefit Categories
Table 1-1 lists the major health and welfare effects categories poten-
tially relevant in the consideration of alternative PM standards. In the
health category, effects of interest include mortality and various types of
acute and chronic morbidity. The EPA/OAQPS Staff Paper (3) indicates that
the potentially relevant morbidity includes effects on respiratory
mechanics and symptoms, aggravation of existing respiratory and
cardiovascular disease, impairment of the body's defense mechanisms, damage
to lung tissues, and care inogenesis. The benefit analysis is abl.e to
provide partial coverage of these effects, with emphasis on mortality and
acute and chronic respiratory effects. Currently available data and
methods limit the scope of coverage.
In the welfare area, the benefit analysis emphasizes residential
soiling effects and, to a lesser extent, industrial soiling. Available
data and methods preclude consideration of soiling in the commercial.
Table 1-1
EFFECTS CATEGORIES POTENTIALLY RELEVANT TO ALTERNATIVE PM STANDARDS
Health Effects
- Mortality
- Acute Morbidity
- Chronic Morbidity
Soiling & Materials Damage
- Residential Facilities
Commercial & Industrial Facilities
- Governmental & Institutional Facilities
Visibility Effects
Regional Haze
- Plume Blight
Acidic Deposition
Aquatic Life
Crops & Forests
Materials
Climatic Effects
Temperature
Precipitation
NonUser Benefits
Bequest Value
Existence Value
Option Value
1-3
-------
governmental, institutional, and most industrial sectors. Acidic deposi-
tion and regional haze are believed to be primarily associated with fine
particles such as sulfates. It is difficult to predict whether and to what
extent fine particle concentrations will be reduced by the control strate-
gies required to attain the alternative PM standards under consideration.
These effects are omitted from the analysis. Incidents of plume blight may
be reduced also and this is not considered in the analysis. Little or no
information is available on the climatic effects of alternative PM concen-
trations. Non-user benefits such as bequest value (the value placed on
preserving a clean environment for future generations) are also not
estimated.
The coverage of potentially relevant effects categories is thus only
partially complete in terms of numbers of categories and coverage within
categories. In this respect, the benefit estimates in this report may be
conservative estimates of actual benefits.
UVKKV1KV OF THE STUDY APPROACH
The time schedule for the analysis did not allow the undertaking of
new effects studies. In view of this constraint, the basic strategy in the
analysis is to make use of existing research results concerning the health
and welfare effects of ambient particulate matter. That is, the existing
research literature is identified, classified, and reviewed. Next, the
best available studies for purposes of a benefit analysis are selected.
Finally, the quantitative relationships (e.g., concentration-response
functions) in these studies are used to estimate the benefits associated
with the alternative PM standards under consideration. A more detailed
overview of these steps is provided in the subsections that follow.
Identification and Classification of Existing Research Studies
For purposes of this analysis, the existing research literature on
health and welfare effects is classified into the following major
categories and subcategories:
1-4
-------
Health effects studies in the epidemiology literature
including studies of
Mortality due to acute (short-term) exposures.
- Acute morbidity due to acute exposures.
Acute morbidity due to chronic (long-term) exposures.
- Chronic morbidity due to chronic exposures.
Health effects studies in the economics literature
including studies of
Mortality due to chronic exposures.
- Acute morbidity due to chronic exposures.
- Chronic morbidity due to chronic exposures.
Soiling and materials damage studies including studies of
effects in the
Household sector.
Manufacturing sector.
As noted previously, available studies and data limit the focus of this
analysis to health effects and soiling or materials damage effects. The
study areas listed above represent the spectrum of available literature in
these effects areas. Visibility studies are also available and are identi-
fied in the analysis but, as noted previously, may not be applicable to the
PM control strategies under consideration.
Note that in the list above, a distinction is made betireen health
studies in the epidemiology literature and those in the economics litera-
ture. While there are technical similarities betireen these two groups of
health studies, the studies are treated separately in the EPA Criteria
Document (4). The two groups are thus kept separate in this analysis to
allow comparison and separate use of their results.
Two additional literature areas are also identified. These include:
Property value studies.
Labor market (hedonic wage) studies.
1-5
-------
Tie basic benefit estimates in this analysis are developed using the
studies in the health and soiling effects areas listed earlier. Property
value and hedonic wage studies provide an alternative method for estimating
air pollution control benefits. They are thus used in this analysis to
provide independent benefit estimates which can serve as a validation check.
on the basic benefit results.
Review mad Selection of Best Available Studies
In each of the literature areas listed above, a comprehensive review
and critique of the available research is undertaken. The starting points
for this review are the EPA Criteria Document and the EPA/OAQPS Staff Paper
for particulate matter. The objective of the review is to identify the
best available studies for a benefit analysis in each area. Studies are
evaluated in terms of such criteria as the soundness of the technical
approach, the adequacy of the underlying data, the adaptability to esti-
mating benefits, and the adequacy of study documentation.
An important feature of this review is a recognition that the various
studies often differ widely in their findings concerning the effects of
particulate matter. In the absence of a consensus, it may not be desirable
to rely solely on one particular study for estimating the benefits in a
particular category. Thus, in each benefit category, an effort is made to
identify both a best available study and one or more plausible alternate
studies. Separate benefit estimates are then developed for each of the
selected studies in that category. Use of several plausible studies to
calculate benefits helps illustrate both the possible range of benefits as
well as the degree of uncertainty present in the existing research base.
Benefit Calculation Procedures
Benefit calculations in this analysis follow a three part procedure.
In the first part (1), the quantitative relationships found in each
selected study are used to calculate the benefits of a given air quality
improvement. In the second part (2), the benefit results for the
1-6
-------
individual studies are used to estimate the benefits in each benefit
category. Recall that this is required because more than one study is
typically selected in each benefit category. In the third part (3). the
results for each benefit category are used to estimate aggregate benefits.
The third part cannot be accomplished by simply adding up all the benefit
categories because some overlap may be present among the categories. The
next several paragraphs provide further details on part (1), while parts
(2) and (3) are discussed subsequently when the benefit results are
presented.
Part (1) calculations encompass estimating the benefits implied by
each study. These calculations are done according to a four or five step
procedure. As shown in Figure 1-1, the first step is to identify the
magnitude of the ambient air quality improvement that is estimated will
occur in each county and year. This is the improvement achieved after
implementation of a particular ambient standard, relative to a baseline
situation reflecting air pollution controls already in place. The proce-
dures for estimating air quality improvements are discussed in Section 9 of
this report.
The second step involves estimating the health and welfare improve-
ments that are expected to occur as a result of the improvement in ambient
air quality. This step makes use of the research findings extracted from
the literature review discussed previously. These findings include either
linear or nonlinear relationships between health or welfare status and
ambient concentrations of particulate matter. These relationships are
discussed in detail irf Sections 3 through 8 of this report. Note that
estimates are required for each county and year in which there is an air
quality improvement.
The third step is to impute an economic value to the estimated changes
in health and welfare status. This is shown in the figure as Step 3a. For
the health studies, this is perhaps the most difficult step conceptually.
There is limited evidence on how to estimate the economic value of some
changes in health status (e.g., improved lung function). For this reason,
-------
STEP
1
Identify air quality improvement
in county i at year t
STEP
2
Estimate health or welfare
improvement in county i at year t
3a
3b
STEP
3
Estimate economic value of the health
or welfare improvement for i, t
STEP
4
Aggregate over t to obtain
discounted present values
STEP
5
Aggregate over i to obtain
regional and national totals
Figure 1-1. Basic steps in estimating benefits for an individual study
1-8
-------
a variety of methods and estimates of value are employed in order to
evaluate how they influence the results. Note also that in some cases it
is possible to estimate economic values directly from the air quality
improvement. This is shown in the figure as Step 3b. Property value
studies are an example of this approach. These studies analyze variations
in residential property values within a metropolitan area to identify the
variation due to perceived differences in environmental amenities such as
air quality, while statistically controlling for variation due to differ-
ences in housing quality and other locational attributes. Studies of this
type thus allow direct estimates of the perceived economic value of
environmental improvements.
The fourth step is to aggregate results over a specific period of
years to obtain discounted present values. The number of years depends on
the particular standard under consideration. In Step 5, benefits for each
county are also summed to obtain regional and national totals.
Source* of Uncertainty in the Benefit Calculations
Benefit analyses are subject to several sources of uncertainty,
especially when human health effects are of interest. The major categories
of uncertainty in this analysis concern:
The scope and magnitude of air quality improvements likely
under alternative standards.
The evidence of the health and welfare effects of PM.
The economic value of reduced adverse health risk.
The coverage of potentially relevant effects/benefit
categories.
Within each major category of uncertainty there are more specific
problems of incomplete or conflicting data and analytic methods. For
example, the air quality analysis uses county-wide linear rollback
techniques. These techniques introduce a large element of uncertainty when
estimating the air quality impacts of source emissions growth or specific
1-9
-------
control options. Similarly, the estimates of health effects are based on
existing community health studies. All of these studies have limitations
either in study design, data or analytic methods which limit the accuracy
or reliability of the results.
Table 1-2 contains a more detailed list of selected major sources of
uncertainty in this study. Items in the list are defined and discussed
further later in this section or in other sections of the report. The
other sections also contain more detailed lists.
For some of the items in Table 1-2, the degree of uncertainty is
estimated quantitatively. For example, alternative benefit calculations
can be carried out using: each available study, each available functional
form implied by each study, upper- and lower-bound (i.e., + 2a) coefficient
estimates, alternative low-level concentrations, upper- and lower-bound
health values, and alternative methods of aggregation. That is, an effort
is made to identify the range of uncertainty present in the benefit
estimates. The estimates of this range are explicitly reported. Sources
of uncertainty that could not be quantified are described qualitatively in
each report section.
One source of uncertainty that is particularly difficult to address is
the variation in completeness of benefits/effects coverage among studies.
This includes two problems. First, studies may provide overlapping
coverage, in which case double counting of benefits may occur. This
problem is addressed in the aggregation procedures discussed shortly.
Second, studies may provide incomplete coverage, which would lead to under-
counting benefits. As an example, this analysis includes very limited
coverage of PM soiling in the industrial sector and excludes entirely the
commercial, institutional, and governmental sectors.
It should also be noted that all of the studies used for estimating
benefits are community studies as opposed to controlled laboratory experi-
ments. As such, the results of these studies are generally in the form of
identified statistical associations between PM concentrations and various
1-10
-------
Table 1-2
SOURCES OF UNCERTAINTY IN THE BENEFIT CALCULATIONS
Scope and Magnitude of Air Quality Improvements*
Use of county-wide rollback to predict air quality change at
worst-case monitor.
- Use of proportionality to predict change elsewhere in county.
- Failure to account for potential air quality improvements
resulting from emissions controls in adjacent counties.
- Approximations required in converting among different PM
measures (TSP, PM10, BS).
Use of 1977-78 as base year.
Evidence of Health and Welfare Effects
- Conflicting evidence across studies.
- Incomplete control for confounding factors.
- Use of study results from one time and area to estimate
effects in another time and area.
- No information on the size distribution of particles present
during the studies.
- Alternative functional forms possible.
- Sampling variation in coefficient estimates.
- Degree of risk at lower level PM concentrations.
Valuation of Health Improvements
- Inability to value some health changes.
- Large variations in estimates of value assigned to mortality
risk reductions.
Coverage of Effects/Benefit Categories
- Omission of potentially relevant categories.
- Incomplete coverage within evaluated categories.
- Possible overlap among certain categories.
* Note that uncertainty in the air quality estimates will affect both the
cost and benefit calculations.
1-11
-------
indicators of health or welfare status. Use of these statistical associa-
tions for benefit calculations requires the inference that the associations
reflect cause-effect relationships. Such inferences are controversial
because cause-effect relationships cannot be proven by statistical associa-
tions. Conclusions about causality are ultimately a matter of judgment,
typically based on a consideration of such factors as the strength of the
statistical association and its consistency and plausibility in comparison
with other studies and facts. Consequently, the benefit analysis is care-
ful to evaluate each group of studies on this basis. This evaluation is
reflected in the subset of studies selected for benefit calculations and in
the sensitivity analyses conducted on the results obtained.
SCENARIOS ANALYZED
Baseline Air Quality
The baseline for all scenarios is a projection of ambient air quality
in each county for the period 1987 through 1995 (or 1989 to 1995). The
baseline projection reflects the following major components: background
ambient concentrations, area source emissions (e.g., roadway dust),
emissions from current stationary sources and emissions from new stationary
sources coming on-line during the period. Thus, ambient air quality in a
county may improve or deteriorate in the baseline scenario, depending on
the relative growth rates for area sources and new stationary sources and
retirement/replacement rates for existing stationary sources.
In the baseline scenario, some pollution controls are assumed to be in
place. In particular, unretired sources and one-half of any replacement
sources are assumed to be controlled at 1978 control levels. The other
replacement sources and all net new sources are assumed to be controlled at
new source control levels. New source controls include BACT (Best Avail-
able Control Technology) which represents NSPS (New Source Performance
Standards) and other new source control requirements. Further details
concerning the baseline can be found in the cost analysis report (2).
1-12
-------
The baseline scenario reflects some air quality improvement relative
to the no-control situation. The alternative PM standards under
consideration would represent an incremental improvement in air quality
compared to the baseline scenario. For this reason, the benefits
associated with the alternative PM standards are incremental benefits.
These are the benefits estimated in this analysis. The total benefits of
both baseline controls and the alternative PM standards are not estimated.
Alternative Standards
The various ambient standards considered in the benefit analysis are
listed in Table 1-3. The standards are listed approximately in order of
increasing stringency. The standards are distinguished by:
Particle measure
- PM10 (particles with an aerodynamic diameter less than
or equal to a nominal 10 |im) ,
- TSP (total suspended particulates, i.e., particles
measured by the "hi-volume" sampler which collects
particles of aerodynamic diameter of up to 25 to 45
(im.)
Ambient concentration allowed
Averaging time and distribution parameter
- Annual, arithmetic mean
Annual, geometric mean
- 24-hour, expected maximum value
- 24-hour, one exceedance per year (i.e., second highest
value)
Implementation date
1989 (all PM10 standards; some TSP standards)
1987 (all other TSP standards)
Attainment status
- Scenario B (all counties attain the standard)
- Scenario A (some counties remain in nonattainment).
1-13
-------
Table 1-3
ALTERNATIVE AIR QUALITY STANDARDS
(in |ig/m3)
Pollutant
PM10
PM10
PM10
PM10
PM10
PM10
TSP
TSP
TSP
TSP
Annual Standard
Concentration
70
55
55
55
55
48
75
75
Parameter
AAM
AAM
AAM
AAM
AAM
AAM
AGM
AGM
24-Hour Standard
Concentration
250
250
200
150
183
260
150
260
150
Parameter
EMV
EMV
EMV
EMV
EMV
EMV
2nd High
2nd High
2nd High
2nd High
Implementation
Date
1989
1989
1989
1989
1989
1989
1989
1989
1987
1987
Attainment
Status
A, B
A. B
A, B
A, B
A, B
A, B
A. B
A, B
A. B
A, B
AAM = Annual arithmetic mean.
AGM = Annual geometric mean of all 24-hour average values.
EMV = Expected 2nd maximum value of all 24-hour average values.
2nd High = Second highest 24-hour average value observed per year.
A = Some counties experience non-attainment.
B = All counties attain and maintain the standard.
-------
In both the cost and benefit analyses, pollution controls are assumed
to be implemented at the beginning of the year (1987 or 1989). In the
Scenario B cases, all counties are also assumed to achieve attainment of
the standard at the beginning of the implementation year and to maintain
the standard through the end of 1995. In the Scenario A cases, some
counties may not be in attainment during some or all of the period. As
indicated in the table, all standards were evaluated under both partial
attainment (Scenario A) and complete attainment (Scenario B) conditions.
Several of the key concepts defined above can be illustrated
graphically as shown in Figure 1-2. The upper curve in the figure
represents the projection of baseline air quality. As noted previously,
the baseline reflects sources and controls in place in 1978 plus growth and
retirement/replacement of sources after that date. The dashed line
represents the standard. The bottom curve identifies the improved air
quality after implementation of additional controls in this case in 1989.
Benefits are generated by the improvement in air quality indicated by the
shaded area. Note that some improvement below the standard can occur.
This results from approximations required in the cost analysis.
Constraints are imposed to insure that the approximations do not result in
predicted air quality improvements below background air quality levels.
Figure 1-2 illustrates the situation with complete attainment of the
standard (type "B" scenarios). Figure 1-3 illustrates the case of partial
attainment (type "A" scenarios). In this case, sufficient control options
are not available to attain and maintain the standard throughout the
applicable time period. As a result, both costs and benefits are lower
than if complete attainment occurred.
As the previous figures indicate, benefits occur over a period of
several years. It is thus convenient to express benefits in discounted
present value terms. In calculating the discounted present value of
benefits, the following conventions are employed:
1-15
-------
Ambient
Concentration
(pg/m3)
Baseline
Source of
Benefits
Standard
With Incremental Controls
1978
1989
I
1995
Year
Figure 1-2. Typical Air Quality Scenario
-------
Ambient
Concentration
i
H-«
-J
Baseline
Source of
Benefits
ual Non-Attainment
With Incremental Controls
Standard
1978
1989
1995
Year
Figure 1-3. Residual Won-Attainment (Scenario "A")
-------
Time horizon corresponding to the standard (7 or 9 years).
Present value as of January 1, 1982.
Real discount rate of 10 percent.
Estimates stated in 1980 dollars.
Exclusion of benefits occurring after 1995.
INDEX OF EXPOSURE
The studies of the health and welfare effects of PM each employed
specific indices of air pollution exposure. Many of the indices are
different from those specified in the air quality standards under consid-
eration. To use the results of these studies in a benefit analysis, the
air quality estimates associated with the alternative TSP and PM10
standards must be converted to indices consistent with the studies. The
studies used in the benefit analysis employed indices with the following
characteristics:
All studies used either an annual average concentration, a
24-hour average concentration, or a second-highest 24-hour
average concentration per year.
All but two studies used a TSP measure; the other two used
British Smoke (BS) and Suspended Particulate Matter (SPM),
respectively. None used a PM10 measure.
The geographic areas included Census tracts, cities,
counties and SMSAs (Standard Metropolitan Statistical
Areas).
The number of monitors in each area ranged from one to a
dozen or more; in an area with multiple monitors, the
original studies typically used the monitor recording the
highest concentrations, or an average across monitors in
the area.
Use of studies with the above characteristics is further complicated by
limitations in the air quality data available from the cost analysis. The
most important limitation is the lack of air quality data other than an
estimate of the county maximum concentration in the baseline, and post-
1-18
-------
control situations. The benefit analysis requires more complete data and
must make additional approximations to supplement the data available.
These approximations, and those associated with the index conversions
described above, introduce another source of uncertainty in the benefit
analysis. These issues are discussed in more detail in Sections 9 and 10,
and to a lesser extent in the other report sections. The issues are
summarized briefly here.
Averaging T^^es md. Distribution P>f eters
Ambient air quality data incorporating the required averaging times
and distribution parameters are supplied from the cost and air quality
analysis. The required averaging times include annual and 24-hour
averages. The required distribution parameters include the arithmetic
mean, geometric mean, and second-highest 24-hour concentration. The proce-
dures employed to develop these data are described in the cost analysis
report referenced earlier.
Particle Measures
To evaluate standards based on TSP, most of the studies are used
directly since they were originally done in terms of TSP. The study based
on SPH is also used directly based on the approximate equivalence of TSP
and SPH measures. The study based on British smoke requires a specific
transformation procedure from TSP to BS which is described in an appendix
to Section 3.
For standards based on PM10, the cost and air quality analysis pro-
vides estimates of both the PH10 and TSP concentrations that result. The
PM10 information cannot be used directly since none of the health or
welfare studies used a PM10 measure. However, the availability of TSP
concentration data makes it possible to estimate approximate benefits using
the TSP studies (and the study requiring the TSP to BS transformation).
This approach involves estimating the benefits of the TSP reduction that
results when controls are applied to attain the PM10 standards.
1-19
-------
Note that using TSP reductions to estimate the benefits of PM10
standards limits the comparisons possible between PM10 and TSP standards.
Comparisons are possible only in terms of alternative levels of TSP
stringency, not particle size. That is, the benefit analysis is not able
to distinguish between a TSP standard and a PM10 standard that results in
the same TSP reduction.
Evidence cited in the EPA/OAQPS Staff Paper indicates that smaller
particles (1 10 (im in diameter) pose a greater risk to the respiratory
system than larger particles. Thus, the benefits of PM10 standards can be
different than suggested by the benefits calculated from the associated TSP
reduction. This could occur if the PM10 controls produce a proportionately
different reduction in PM10 than in TSP.
The cost and air quality analysis provides estimates of the PM10 and
TSP concentrations before and after implementation of controls. A review
of these estimates for selected standards indicates that, in a majority of
counties. PM10 does not decline proportionately more than TSP. This was
found to be the case for both the PH10 and TSP standards examined.* This
apparently results because control options available for PH10 in many cases
also control larger particles as well. Hence, if emissions sources
selected for control are producing both small and large particles, the
proportionate reductions in PM10 and TSP may not differ appreciably.
* The standards examined were the PM10 (70, 250), the PM10 (55, 150) and
the 7-year TSP (75, 260). All were Scenario A (partial attainment)
cases. All comparisons were before and after controls in 1989. All
comparisons were between the PM10 annual arithmetic mean and the TSP
.annual arithmetic mean because the data available for the 24-hour
averaging time were based on different statistical measures and cannot be
directly compared. For all three standards, about 60 percent of the
counties were predicted to experience a proportionately larger reduction
in TSP; the other 40 percent experienced a proportionately larger reduc-
tion in PM10. About the same percentages were observed among the
counties predicted to receive the largest benefits. In most of these
counties, the predicted absolute change in the PM10 fraction was 0.01 to
0.03 under all three standards.
1-20
-------
As indicated above, small differences are predicted between the PH10
standards and TSP standards analyzed, in terms of the resulting PM10
fractions. This implies that health benefits estimated based on the TSP
change may be of similar accuracy for either type of standard. If the
estimates are of similar accuracy, then it is appropriate in this study to
draw comparisons between the PM10 and TSP standards. However, as noted
previously, these represent comparisons of alternative levels of TSP
stringency (i.e., alternative reductions in TSP). That is, the benefit
analysis cannot distinguish among standards on the basis of particle size.
It is important to note that limitations and approximations required
in the cost and air quality analysis make the estimates of the PM10
fraction in each county uncertain. Predicted baseline PM10 levels for each
county in 1989-95 are based on an assumed PM10/TSP ratio of 0.55 in 1978,
and predicted growth rates and small particle fractions for each emissions
source category in the post-1978 period. Actual conditions may differ from
these assumptions.* Furthermore, other factors, such as differences in
small and large particle dispersion patterns, are not accounted for in the
cost and air quality model, not thus in the benefit analysis. These uncer-
tainties suggest that additional caution be used when comparing the benefit
estimates for the PH10 and TSP standards.
Geographic Areas and Monitor Types
Ambient air quality data available from the cost analysis are
estimates of conditions at the "design value" pollution monitoring site in
each county. The design value monitor in a county typically is the monitor
which recorded the highest ambient concentration (TSP annual mean or 24-
hour second high) in 1977 and 1978. Concentrations at other monitors in
the county are thus generally lower than at the design value monitor. Non-
monitored areas in a county may experience higher or lower concentrations
* For example, more recent data indicates that in late 1981-early 1982 the
PM10 fraction was closer to 0.46 on a national average basis (5). The
PM15 fraction was closer to 0.55 at that time.
1-21
-------
than the design value monitor, but lower concentrations are more likely.
Air quality conditions at these other monitors and locations are not esti-
mated in the cost analysis.* The cost and air quality analysis also omits
the possibility of air quality improvements due to controls applied in an
adjacent county.
Since projections of air quality and economic data are available
primarily for county areas, the county is the basic unit of analysis in
this study. This approach provides a good match with the original studies
that used either county or SMSA data. The air quality change in the county
is then taken to be either the change at the design value monitor, or the
average change across all monitors in the county, depending on the measure
used in the original study. When the average change is required, it is
estimated from the design value monitor change by applying the proportion-
ality factor which existed between these two values for each county in the
1978 base year. The proportionality factor on average was 0.74 for the
annual mean and 0.70 for the 24-hour second-highest concentration.
Note that the above method for estimating the average air quality
improvement across monitors in a county is only approximate. This is
because the improvement at a particular monitor may or may not be propor-
tional to the improvement at the design value monitor. That is, the
dispersion properties of PM attenuate the air quality improvement occurring
at various distances from the emissions sources being incrementally
controlled. The proportionality assumption is an approximation to the
actual attenuation. The degree of attenuation cannot be assessed without
detailed dispersion modeling' of the PM emission controls applied in each
county. Such modeling is very time and resource intensive, and thus not
practical in a nationwide study. As a result, the cost analysis relies on
the more approximate county-wide rollback method for predicting air quality
improvements at the design monitor in each county. County-wide rollback
methods assume that the change in air quality (net of background concentra-
* For a small number of counties (12 out of 501), air quality is estimated
for subcounty areas.
1-22
-------
tions) at any location in a county is proportional to the change in total
county emissions. Thus the use of proportionality in the benefit analysis
to estimate the average air quality improvement across other monitors in a
county is essentially a further application of county-wide rollback.
Lack of dispersion modeling also hinders use of studies which measured
air quality effects at the city or Census tract level. For most appro-
priate use of the results of these studies, it would be desirable to have
dispersion modeling results for subcounty areas. For example, for the same
reasons noted previously, use of the air quality change projected at the
county design value monitor may overstate the change which occurs elsewhere
in the county. To provide a better air quality measure for these studies,
the average air quality change across all monitors in a county is also
used. The average change is more representative than the design monitor
change. The average change is estimated as described previously.
Plausibility Checks
As noted above, the ideal situation would be to identify the change in
PM concentration experienced by the population in each part of the county.
In the absence of dispersion modeling results, the approximations described
above attempt to improve upon the available concentration measure, the
design monitor. Two plausibility checks were performed on these approxima-
tions. The first plausibility check concerned the use of the county
average as an additional measure of the PM concentration. This check
involved comparing the county averages in the 1978 base year with corres-
ponding readings at monitors designated by EPA as "population-oriented"
monitors. This analysis found that readings at population monitors are
typically much closer to the county average than to the design value
monitor readings.* This is not too surprising since 73 percent of all
monitors in 1978 are designated as population-oriented. This suggests that
* The correlation between the average of all population monitors in a
county and the average of all monitors in a county ranged between 0.95
and 0.96, depending on the averaging time. For population monitors and
design values, the correlation ranged between 0.48 and 0.50.
1-23
-------
the county average air quality is likely to be more representative than the
design monitor as an estimate of the actual population exposure.
The second plausibility check concerned the accuracy of estimating the
change in county average air quality based on a proportionality relation-
ship to the change at the county design monitor. An analysis was under-
taken to evaluate the quality of this approximation. The analysis, which
is discussed further in Section 10, centers on the Chicago, Illinois area
(Cook County). This area was chosen because it accounts for the largest
fraction of total national benefits among all counties in the analysis and
also had been the subject of detailed dispersion modeling previously. The
analysis compares benefits estimated using two methods. First, benefits
are calculated based on detailed dispersion modeling results for 31 sub-
areas within Cook County. In the second method, dispersion modeling is
used to predict the change at the design monitor only and proportional
changes are assumed throughout the rest of the county. The second method,
which is comparable to the approach used in the national benefit analysis,
leads to estimates which are a factor of 2 to 4 higher than the first
method. This suggests that the assumption of proportional air quality
changes may introduce an upward bias in the national benefit analysis.
However, there are a number of differences between the Chicago analysis and
the national analysis which preclude an accurate estimate of the magnitude
of the bias or its importance relative to other sources of upward and
downward bias in the national analysis.
MEASUREMENT AND ECONOMIC VALUATION OF HEALTH IMPROVEMENTS
A companion study assesses the cost of pollution controls designed to
attain alternative PM standards. Comparison of costs and benefits in
commensurate terms requires placing an economic value on reductions in the
risks of mortality and morbidity. Alternatives to a formal economic valua-
tion process include choosing policy objectives which reduce all risk at
whatever the cost, or which implicitly and sometimes unknowingly value
reductions in risk. The alternative selected in this analysis is to
explicitly value reductions in both mortality and morbidity risk. This
1-24
-------
allows the costs and benefits of the standards to be compared in commen-
surate terms.
Reduced Mortality Risk
The benefits of a reduction in mortality risk are evaluated based on a
four-part calculation. The calculation is done for each geographic area.
The four parts involve estimates of:
The reduction in ambient PM concentration.
The reduction in mortality risk corresponding to the reduc-
tion in ambient PM.
The number of individuals experiencing the risk reduction
(i.e., experiencing the PM reduction).
The dollar value that one individual places on a unit
reduction in mortality risk.
A hypothetical example will illustrate the four-part calculation.
Suppose the annual mortality rate in an area is 15 deaths per 100,000
people. This may include deaths from a variety of causes as well as from
air pollution. If we let r denote the annual mortality rate, then in this
example r = 0.00015.
Another way to interpret r is as a probability. In this view, there
is a 15 in 100,000 chance that an individual chosen at random will die
during the year. That is, the representative individual faces a 0.00015
risk of mortality during the year.
Now suppose that epidemiology studies indicate that reducing air
pollution by, say, 100 |ig/m annual average would cause r to decline by an
1-25
-------
amount Ar. For example, Ar might be 0.00001.* The question then is how to
value this reduction in mortality risk.
In benefit analyses, the economic value of a small reduction in
mortality risk is estimated by adding together the values that individuals
assign to the redaction. Previous studies and surveys have developed
estimates of individual values. These studies and surveys are reviewed in
the Appendix to Volume II of this report. A typical approach is to observe
the wage premiums associated with occupations involving differing degrees
of mortality risk. The studies and surveys suggest that individuals are
willing to pay about $1.00 per year for each reduction of 1 x 10 in
annual mortality risk. For the previous example of a reduction Ar =
(%
0.00001 = 10 x 10 , the individual willingness to pay for this reduction
is about $10.00 per year. This implies that a group of 100,000 individuals
together would be willing to pay a total of (100,000)($10.00) = $1,000,000
if each experienced a risk reduction of 0.00001.
The actual values used in this benefit analysis for a unit reduction
of 1 x 10 in annual mortality risk are:
Minimum estimate: $0.36
Point estimate: $1.58 (All in 1980 $)
Maximum estimate: $2.80
This range reflects the different values implied by the studies and surveys
reviewed in the Appendix to Volume II.
* The analysis can be made more exact by examining changes in mortality
rates for various subgroups in the population (e.g., the elderly). How-
ever, this is appropriate only if specific subgroups were analyzed in the
underlying research study that is used to estimate Ar.
1-Z6
-------
Reduced Morbidity Risk
The benefits of a reduction in morbidity risk are estimated using a
similar four-part calculation as in the case of mortality risk. In step
two, however, the concern is with the reduction in morbidity risk per unit
reduction in ambient PM; and in step four, the issue is one of valuing
small changes in morbidity risk. The reductions in morbidity risk are
estimated using results from previous epidemiology and economic studies.
Valuation of morbidity risk is done taking into account three factors: the
value of fewer expected work days lost due to illness; the value of fewer
expected nonwork days with reduced activity levels due to illness; and the
savings in expected medical expenditures. The method used to value each of
these effects is identified below. The specific values vary from county to
county.
Effect Method of Valuation
Lost workday Average daily wage
Reduced activity day One-half of the
average daily wage
Change in direct Proportional to change
medical expenditures in disease incidence
As discussed in the Appendix to Volume II, the above approach probably
underestimates total willingness to pay for morbidity risk reductions
because it excludes consideration of residual pain, suffering, and incon-
venience.
Applicable Concentration Ranges for Concentration Response
The results of health. effect s studies can often be expressed in the
form of concentration-response functions. Such functions represent a quan-
titative statement of the relationship between health status and ambient
1-27
-------
concentrations of an environmental pollutant such as particulate matter.*
Concentration-response functions are useful in a benefit analysis because
they provide a mathematical basis for calculating health improvements
associated with air quality changes over different concentration ranges.
This subsection is concerned with the ranges over which the concentration-
response functions should be applied.
Concentration Ranges Suggested by CASAC and the EPA/OAQPS Stmff Pmper
The EPA/OAQPS Staff Paper explicitly addressed the question of concen-
tration ranges over which health effects may occur. The Staff Paper
identified one set of ranges at or above which health effects are likely to
occur among sensitive groups in the population; it identified a second,
lower set of ranges over which effects may be possible, but for which the
evidence and level of risk was less certain. These ranges were derived
from the EPA Criteria Document and reflect consideration of the epidemi-
ology studies judged to provide the most reliable quantitative evidence.
The numerical values of these ranges are identified in Table 1-4.
EPA's Clean Air Scientific Advisory Committee (CASAC) also addressed
the question of concentration ranges over which health effects may occur.
Their conclusions were summarized in the CASAC Chairman's closure letter
(6) on the Staff Paper. CASAC concluded that detectable health effects
occur at the upper bound of the "Effects Possible" ranges identified in the
Staff Paper. These ranges are 150 to 350 jig/m3 for 24-hour PM10 concentra-
tions (150 to 250 |ig/m3 BS) and 55 to 110 |ig/m3 for annual average PM10
concentrations (110 to 180 }ig/m3 TSP). The closure letter also made it
clear that these conclusions were based solely on a review of currently
available quantitative evidence from epidemiology studies. That is, CASAC
did not say that there are no health effects at the lower end or below
* Following the practice of the EPA/OAQPS Staff Paper, this report uses the
term "concentration-response function" when referring to community
studies of ambient air pollution effects. The more traditional term
"dose-response function" implies greater knowledge of actual exposure, as
would be the case in a laboratory study.
-------
Table 1-4
SUMMARY OF CONCENTRATION RANGES IN EPA/OAQPS STAFF PAPER*
"Effects Likely"
"Effects Possible"
Short-Term
BS
250-500
. 150-250
PM10
350-600
150-350
Long-Term
TSP
1 180
110-180
PM10
90-110
55-110
* All entries are in (ig/m . Concentrations are given in both original
study units and approximately equivalent PM10 units.
Source: EPA/OAQPS Staff Paper, op. cit.. Tables 1 and 2.
these concentration ranges; it merely said that there was currently no
quantitative evidence of effects.
Other Considerations in Selecting Concentration Ranges for the Benefit
Analysis
Some of the alternative PM standards under consideration may result in
ambient PM concentrations below the lower bounds of the Effects Possible
ranges identified by the Staff Paper. The lower bounds, 150 jig/m for 24-
hour PM10 (150 ng/m3 BS) and 55 (ig/m3 for annual average PM10 (110 |ig/m3
TSP), will hereafter be referred to as the "Staff Paper lower bounds". As
discussed in Sections 9 and 10, concentations below the Staff Paper lower
bounds can result for several reasons. One example is that controls
required in an area to attain a 24-hour standard may lead to associated
reductions in annual average concentrations below the lower-bound annual
concentration. The reverse can also happen. Sections 9 and 10 give
additional examples.
The existence of air quality improvements below the Staff Paper lower
bounds poses difficult problems for the benefit analysis. On the one hand.
1-29
-------
one can qnestion the appropriateness of attributing benefits to air quality
improvements below the Staff Paper lower bounds. The EPA staff has clearly
given careful consideration to selection of the lower bounds. On the other
hand, constraining the benefit calculations by the Staff Paper lower bounds
also presents problems. The problems include: 1) the possibility that
imposing the lower-bound constraints may introduce statistical bias in the
benefit estimates; 2) the uncertainty concerning the existence of effects
below the Staff Paper lower bounds; 3) data limitations which limit atten-
tion to the annual lower-bound concentration only; and 4) possible differ-
ences between the economic principles of benefit-cost analysis and the
criteria used in selecting the "Effects Possible" range for standard
setting purposes. These issues are discussed briefly below and further in
Section 10.
Possible Statistical Bias Practical constraints limit the benefit
analysis to use of concentration-response functions which, in most cases,
are taken directly from the original research studies. None of the studies
used in the benefit analysis imposed the Staff Paper lower bounds at the
time the concentration-response functions were statistically estimated,
either by the authors or by Hathtech.* As discussed in Section 10, it is
improper statistical procedure to impose a lower-bound constraint when
calculating benefits unless it was also imposed or accounted for when the
concentration-response function was originally developed. The impropriety
most likely leads to downward-biased estimates of the level of benefits
with the constraint imposed. That is, applying the constraint may lower
benefits for two reasons first, because benefits below the lower bound
are truncated, and second, because the estimate of benefits above the lower
bound may be downward biased.
Uncertainty About LowerBound Concentrations As noted in the EPA
Staff Paper (7), the available epidemiological data "do not ... show
evidence of clear population thresholds but suggest a continuum of response
* Some of the studies discussed the issue of no-effects thresholds, but
none accounted for such a possibility when estimating concentration-
response functions.
1-30
-------
with both the risk of effects occurring and the magnitude of any potential
effect decreasing with concentration." This situation complicates any
attempt to choose an "Effects Possible" concentration range. The fact that
health effects exist on a continuum creates a difficulty in selecting the
lower bound of the "Effects Possible" range. And the decreasing magnitude
of any effects as concentrations decrease raises the question of whether
particular effects are of sufficient health significance to influence
choice of the range. Both of these factors combine to make selection of
the lower bound subject to informed judgment as well as quantitative
evidence.
The degree to which judgment must ultimately enter the decision is
illustrated by the selection of the lower-bound annual concentration. The
annual lower bound is based largely on a study of two Connecticut towns by
Bouhuys, Beck and Schoenberg (8). The study found evidence of increased
respiratory symptoms (cough, phlegm, and dyspnea) in the town with dirtier
air (Ansonia), but no differences in prevalence rates for chronic bronchi-
tis or lung function. Concentrations of TSP in the cleaner town averaged
3 3
40 ug/m annual mean. Concentrations in Ansonia averaged 63 ug/m annual
mean during the year of the study, and up to 152 ug/m annual mean during
the previous seven years. The EPA Staff Paper characterizes pollution in
3
Ansonia by a median value of 110 ug/m TSP annual mean (9) and this value
is the primary basis for the lower-bound annual concentration (10).*
Alternative conclusions are also possible from the Bouhuys et al.
3 3
study. For example, the choice of 110 ug/m out of the range 63 ug/m to
152 ug/m is not the only option. Furthermore, there is some possibility
that any value in the 63-152 ug/m range would not fully protect against
increased respiratory symptoms. The presence of symptom differences
between the two towns implies symptoms are produced at concentrations at
least as low as in the dirtier town (63-152 ug/m ) and possibly lower.
Thus, a lower bound below 110 ug/m might be necessary if it were desired
* The EPA Staff Paper also cites studies by Ferris et al. (11-12) and Lunn
et al. (13) as providing supporting evidence for this choice. See
Section 10 for further discussion.
1-31
-------
to protect against symptomatic effects with a margin of safety. This study
thus points out the twin difficulties of selecting a lower bound and
deciding on the health significance of observed effects.
An additional issue is the limited extent of the epidemiological data
base available for the selection of critical concentration ranges. Only a
few studies pass the CASAC tests of acceptability, and all of these studies
have technical limitations of varying degrees. Consequently, as new
studies become available, it is possible that they may point towards
different conclusions. A case in point is Ostro's (14) recent reanalysis
of the London mortality data that were the primary basis for the 24-hour
lower bound identified by the Staff Paper. Ostro uses statistical methods
to test formally for the existence of a threshold at various levels (the
Staff Paper's analysis was informal). He finds very strong association
between PH and mortality at concentrations below 200 (ig/m BS, strong
3 3
assoc iation below 100 (ig/m BS, and some association below 75 |ig/m BS.
Depending on the uncertain relationship between BS and PH10 at lower
concentrations, these observed effects levels may be below the Staff Paper
lower bound of 150 fig/m PM10. However, Ostro's reanalysis was not
available to CASAC or the EPA staff at the time they made their recom-
mendations.
Data Limitations The Staff Paper identified lower bounds for both
24-hour and annual mean concentrations. Application of these lower bounds
requires pro-control and post-control air quality data on both a daily and
annual basis. Air quality data available from the control cost and air
quality analysis does not include daily data other than for the second-
highest day per year. Thus, the benefit analysis can apply only the annual
mean lower bound. The annual mean lower bound is directly applied to those
benefit models using the annual mean as the pollution concentration index.
As explained in Section 10, the annual mean lower bound can also be applied
in the benefit calculations for the studies that originally used daily
data. The 24-hour lower bound, however, cannot be applied.
1-32
-------
Differing Objectives The Criteria Document and Staff Paper provide
a comprehensive discussion of PM-related health effects and the concentra-
tions at which the effects appear. As the Staff Paper's use of the Bouhuys
et al. study and others suggests, health effects of a significant or
permanent nature are stressed when selecting the "Effects Possible" concen-
tration range for the primary standard. Effects in the form of increased
symptoms, By themselves, appear to be less influential in determining that
range. Hence, there may be symptomatic effects occurring at concentrations
below the Staff Paper lower bounds.
In contrast, the benefit analysis is concerned with the benefits of
reducing ambient air pollution. Benefits are determined by individuals'
willingness to pay for pollution reductions. Willingness to pay will be
influenced by both risk assessment and risk valuation. Individuals may
value a reduction in symptoms as well as valuing reductions in more signi-
ficant health problems. Individuals may also have differing assessments of.
the health risks from PM. Either of these factors could lead to the
existence of benefits at concentrations below the lower bounds selected
during Staff Paper development. Thus, application of the Staff Paper lower
bounds in the calculation of benefits may be inappropriate because of the
differing objectives of benefit analysis compared to Staff Paper and
Criteria Document development.
Benefit Calculations with Staff Paper Lower Bounds
In view of the uncertainty about calculating benefits below the Staff
Paper lower bounds, two sets of benefit estimates are provided in this
report. One set is based on application of the concentration-response
functions over nearly the full range of air quality improvements predicted
for each county. Only two constraints are imposed in this first set of
estimates: 1) limits implied by the range of PM concentrations present in
the data of the original research studies; or 2) narrower limits if implied
by the results of an individual study. The second set of benefit estimates
has an additional constraint imposed. The second set excludes all health
benefits predicted by the concentration-response functions at levels below
1-33
-------
the Staff Paper lower bounds. 'The summary sections of the report (Sections
1 and 2) provide highlights of both sets of results. Full details are
provided in Section 11. However, the remaining sections of the report are
concerned primarily with benefit calculations without the lower-bound con-
straints imposed.
SUMMARY OP RESULTS IN SPECIFIC BENEFIT CATEGORIES
Health Effects
This analysis considers six potential categories of health effects
from particulate matter and other forms of air pollution: mortality
effects due to either acute or chronic air pollution exposure, acute
morbidity effects due to either acute or chronic exposure, and chronic
morbidity effects due to either acute or chronic exposure.* The existing
research base in each of these areas varies in scope and quality. In
categories where there is some evidence of the relationships between
particulate matter concentrations and rates of mortality and morbidity,
estimates are made of the changes in mortality and morbidity risks
associated with alternative standards. These risk reductions are then
valued using the procedures described previously. The details of the
procedures and results are described in Sections 3 and 4 of the report. An
overview of the health results is provided below.
Economic Benefits of Health
Table 1-5 summarizes the estimated economic benefits of health
improvements associated with imposition of the PM10 (70, 250) Scenario B
standard.** Estimates based on the epidemiology literature and the
economic literature are separately identified. Note that neither
* The terms acute exposure and chronic exposure are taken to be synonymous
with the terms short-term exposure and long-term exposure, respectively.
** Estimates for the other standards are summarized in a later subsection.
Table entries are displayed to the nearest $10 million. However, not
all entries may be accurate to that many significant digits.
1-34
-------
Table 1-5
INCREMENTAL HEALTH BENEFITS FOR THE PH10 (70. 250) SCENARIO B STANDARD*
Health
Effect
Mortality
Morbidity
(Acute)
Morbidity
(Chronic)
Exposure
Acute
Chronic
Acute
Chronic
Acute
Chronic
Epidemiology Literature
Minimum
0.04
NA
0.15
0
NA
0.12
Point
1.12
NA
1.32
0
NA
0.12
Maximum
14.86
NA
11.91
1.44
NA
0.13
Economic Literature
Minimum
NA
0
NA
0.03
NA
2.61
Point
NA
12.72
NA
10.65
NA
11.40
Maximum
NA
62.10
NA
21.49
NA
20.19
1982 discounted present values in billions of 1980 dollars for a 7-year time horizon
(1989-95) and a 10 percent discount rate.
NA = No estimate available.
-------
literature provides estimates in all six health-effect categories, but that
five categories are covered by the two literatures taken together.
One of the largest benefit categories among the epidemiology-based
estimates is the acute exposure mortality category. The mortality benefits
are based on a study by Mazumdar e_t a_l. (15) of total mortality in London
on a daily basis during 14 winters. Both PM (measured as British smoke)
and SO- measurements were included in the study.
The three morbidity estimates from the epidemiology literature are
based on studies by Samet et_ al. (16), Saric e_t al. (17), and Ferris et, .al.
(18-20), respectively. These studies were concerned with acute or chronic
respiratory disease. Because these studies do not clearly identify the
relative importance of PM and SO-, they may provide upward-biased estimates
of PM related respiratory effects. On the other hand, the three morbidity
estimates are also potentially downward biased due to the exclusion of
nonrespiratory diseases and the conservative method for valuing morbidity
effects.
The largest benefits among the economic literature estimates are in
the chronic exposure mortality category. The estimates in this category
are based on the results of nine separate studies. Most of the studies
control for either S02 or S04 (a subset of PM), as well as PM. Six of the
studies find associations between mortality and PM concentrations, while
two find the SO^ fraction of PM to be more influential.* One study finds
little evidence of effects. This uncertainty is reflected in the minimum
estimate of 0.0 for this category. The point estimate emphasizes the
studies by Lave and Seskin (21) and by Lipfert (22-24), and uses results
that are on the low end of the range of effects they observed.
* Sulfates (S04) are primarily a constituent of the fine particle (approxi-
mately < 2.5 jim diameter) fraction of PM. As noted previously, it is
difficult to predict whether and to what extent fine particle concentra-
tions will be reduced by the control strategies required to attain the
alternative PM standards under consideration.
1-36
-------
Estimates for the chronic exposure acute morbidity category are based
on studies by Ostro (25) and by Crocker et al. (26) which look at labor
productivity effects (lost work days). The Ostro study also analyzes
reduced activity days. These morbidity benefits may be underestimates
because of the valuation method mentioned earlier. The first study
controls for both PM and SO., while the other considers only PM in the
final specification. Benefits in the chronic exposure-chronic morbidity
category are based on a single study of labor productivity effects by
Crocker et al. (27) which includes only PM in the final specification.
Range of Uncertainty
The variation between the minimum and maximum estimates in Table 1-5
reflects several of the types of uncertainty present in the analysis. The
general types of uncertainty were discussed previously. The epidemiology
study of acute exposure mortality effects by Mazumdar et al. is illustra-
tive of several specific examples. The variations in this instance
reflect: 1) use of different functional forms for the dose-response
function, both of which are consistent with the observed data; 2) use of
the upper and lower bounds on the 95 percent confidence interval around the
estimated coefficient for particulate matter in the concentration-response
function; 3) use of coefficients estimated over different subsamples of the
data; and 4) use of different estimates for the value of a marginal reduc-
tion in mortality risk. Note that other types of uncertainty are not
reflected in the range. For example, not included are uncertainty about
the air quality data and uncertainty introduced by using the results of a
London study to evaluate U.S. conditions where particle composition and
weather may differ.
The point estimate of benefits in the Mazumdar et al. example is
obtained as follows. First, point estimates for the estimated coefficient
and for the value of a marginal risk reduction are used to obtain new
minimum and maximum benefit estimates based on the different functional
forms and data subsamples. The new estimates have a smaller range, but one
that still differs by an order of magnitude. The geometric mean of the new
1-37
-------
minimum and maximum values in each county is then calculated and summed
over counties to obtain the U.S. total. The geometric mean is a natural
measure of central tendency for numbers which differ by orders of magni-
tude.
Sensitivity to Staff Paper Lower Bounds
As discussed previously, the EPA Staff Paper and the Clean Air
Scientific Advisory Committee (CASAC) identified PM concentration ranges
below which they concluded there was no quantitative evidence of health
effects. These concentrations are 150 to 350 |ig/m3 for 24-hour PM10 and 55
4 3
to 110 (ig/nr for annual average PM10 (110 to 180 (ig/nr for annual average
TSP). In theory, it is possible to calculate benefits under the assumption
that no health benefits are produced by reducing PM below the lower bounds
of these concentration ranges (the Staff Paper lower bounds). However,
there are several practical problems with doing the calculation as part of
this analysis. As discussed previously, these problems range from probable
statistical bias to lack of daily data on PM concentrations. Subject to
these previously discussed limitations. Table 1-6 contains the benefit
estimates that result when all health benefits associated with reducing PM
concentrations below 110 |ig/m annual average TSP are excluded. For
comparison purposes, the table also includes the corresponding estimates
when the lower-bound constraint is not applied. All estimates are point
estimates for each category.
The estimates exhibit varying degrees of sensitivity to the Staff
Paper lower bounds. The variations are due to differences in: 1) the
index of PM exposure used; 2) the nonlinearity of the concentration-
response function; and 3) other prior cutoffs imposed to provide consis-
tency with the range of PM concentrations or effects found in the original
studies. For the less stringent standards such as PM10 (70, 250), the
chronic exposure mortality and morbidity estimates show little sensitivity
to the lower bounds; they exhibit greater sensitivity with the more
stringent standards.
1-38
-------
Table 1-6
COMPARISON OF INCREMENTAL HEALTH BENEFITS FOR THE PM10 (70, 250) SCENARIO B STANDARD.
WITH AND WITHOUT A LOWER BOUND APPLIED*
Health
Effect
Mortality
Morbidity
(Acute)
Morbidity
(Chronic)
Exposure
Acute
Chronic
Acute
Chronic
Acute
Chronic
Epidemiology Literature
Lower Bound**
0.31
NA
0.18
0
NA
0.12
No Lower Bound
1.12
NA
1.32
0
NA
0.12
Economic Literature
Lower Bound**
NA
11.98
NA
1.43
NA
10.62
No Lower Bound
NA
12.72
NA
10.65
NA
11.40
* 1982 discounted present values in billions of 1980 dollars for a 7-year time horizon (1989-
95) and a 10 percent discount rate. Entries are point estimates for each category.
** Lower bound of 110 ug/m3 TSP AAM applied.
NA = No estimate available.
-------
Estimates of Physical Effects
Implicit in the estimates of economic benefits are estimates of
changes in health status. The changes in health status include reduced
risk of mortality and morbidity. For economic valuation purposes, the
consequences of reduced morbidity risk are further categorized into fewer
work days lost, fewer reduced activity days, and reduced direct expendi-
tures for medical care. For informational purposes, estimates for each of
the physical effects categories are also developed. The estimates for each
standard and scenario can be found in the supplementary tables in Section
11 of the report. The estimates are based on the same methods and data
used in calculating economic benefits except that the final step of
economic valuation is not performed.
Soiling and Materials DM*go
Existing research on the soiling and materials effects of particulates
has focused primarily on household soiling, and to a lesser extent on
particle effects on paint, building stone and certain manufacturing
industries. Benefits in this area are estimated using the results from a
number of existing studies which are reviewed in Section 7. As suggested
above, the existing literature to a large extent constrains the estimates
to the benefits of reduced household soiling.
Table 1-7 presents estimates of the benefits of reduced household
soiling with the PH10 (70,250) Scenario B standard. These estimates are
based on three separate studies by Cummings et al. (28), Watson and Jaksch
(29), and Mathtech (30). Highest weight is given to the Mathtech study,
which had lower estimates than the other two. Also shown in Table 1-7 are
estimated benefits of PM reductions in part of the manufacturing sector.
These estimates are also based on a study by Mathtech (31). Only a few
studies were identified concerning soiling effects in the commercial,
governmental, or institutional sectors. Although these studies found no
effects, the results may be unrepresentative because of limitations in
study design. In view of the limited analysis of these other sectors and
1-40
-------
Table 1-7
INCREMENTAL BENEFITS OF REDUCED SOILING FOR THE PM10 (70,250)
SCENARIO B STANDARD*
Household Sector
Manufacturing Sector
Minimum
0.73
0.73
Point
3.14
1.30
Maximum
13.85
9.45
* 1982 discounted present values in billions of 1980 dollars for a 7-year
time horizon (1989-95) and a 10 percent discount rate.
the partial coverage of the manufacturing sector, the benefits of reduced
soiling may be underestimated in this analysis.
Visibility and Aesthetics
Decreases in ambient concentrations of particulate matter may also
generate benefits through improvements in visual range and the aesthetic
quality of the environment. These benefits may accrue in several ways.
For example, improvements in visiblity may enhance urban quality of life.
They may lead to greater enjoyment of recreational activities. Further-
more, preservation of unique scenic vistas can lead to a type of benefit
known as "existence value". Most visibility and aesthetic benefit studies
rely on contingent valuation techniques which make use of specially
designed surveys.
Visibility and aesthetic benefits associated with achieving alterna-
tive PM10 and TSP standards are not estimated in this analysis. However,
contingent valuation studies have been used to develop point estimates of
benefits associated with specified improvements in visual range (see
Section 8). Use of those estimates in this analysis would require
knowledge of the transformation between changes in PM10 or TSP and changes
1-41
-------
in visual range and plume blight. The transformation is complex with
particle size being one of the more important factors. Fine participates
(i.e., those with diameters less than 2.5 Jim) are more influential in
determining visual range. Whether fine particulate concentrations are
reduced markedly as a result of the alternative PM10 and TSP standards
under consideration is a function of the control strategies employed. The
EPA Staff Paper (32) suggests that the control strategies required may not
appreciably reduce fine particulate concentrations and, hence, may not
improve visual range. Changes in plume blight will also depend on the
control strategies selected.
Property Value and Labor Market Studies
There are a variety of studies which indicate that some of the
observed geographic variation in residential property values can be
explained by site attributes including air quality. Therefore, it has been
suggested that property value differentials attributable to air quality
represent the capitalized market value of cleaner air. A smaller but
comparable literature suggests the same phenomenon can be observed in the
geographic variation in labor wage rates. Representative studies from
these literatures are also used to develop benefits estimates (see Sections
5 and 6 for details). Note that these estimates are likely to reflect
market perceptions of both health and welfare benefits of air quality
improvements, and thus cannot be directly added to the health benefits
described previously. However, they can provide an independent estimate of
perceived health and welfare benefits. The property value and wage studies
are used in this study to provide a cross-check on the health and welfare
benefit estimates.
The benefits of the PM10 (70, 250) Scenario B standard as estimated
from the property value and labor market studies are shown in Table 1-8.
The estimate from the property value literature is based on about eight
studies, most of which control for both PM and SO, or S04. The estimate
from the labor market literature is based on two studies by Smith (33) and
Rosen (34), both of which include only PM in their final specifications.
1-42
-------
Table 1-8
INCREMENTAL HEALTH AND WELFARE BENEFITS FOR THE PM10 (70, 250) SCENARIO B
STANDARD AS MEASURED FROM PROPERTY VALUE AND HEDONIC WAGE STUDIES*
Property Value Studies
Hedonic Wage Studies
Minimum
3.43
9.81
Point
6.85
19.81
Maximum
11.42
37.21
* 1982 discounted present values in billions of 1980 dollars for a 7-year
time horizon (1989-95) and a 10 percent discount rate.
The latter may be one explanation for why the labor studies provide higher
estimates. However, another possibility is that the labor studies provide
a more complete estimate of benefits.
AGGREGATE INCREMENTAL BENEFITS
Methods of Aggregation
The previous tables contain estimates of incremental benefits for a
variety of benefit categories. The remaining task is to combine the
estimates for the individual categories in order to develop an estimate of
aggregate incremental benefits. Aggregate numbers are required in order
for any benefit-cost analysis to proceed. However, aggregation cannot be
accomplished by simply summing all of the categories because this might
result in double counting and other problems.
Some of the issues which must be considered in choosing an aggregation
procedure are:
Does the procedure give due consideration to the relative
strength of the evidence across different benefit
categories?
1-43
-------
Table 1-9
ALTERNATIVE AGGREGATION PROCEDURES
Mortality
Acute Morbidity
Chronic Morbidity
Household Sector
Soiling & Materials
Manufacturing Sector
Soiling & Materials
Other Sectors
Soiling & Materials
A
Mazumdar
et al.
Ferris
et al.
B
Mazumdar
et al.
Samet
et al.
Ferris
et al.
C
Mazumdar
et al.
Ostro
Ferris
et al.
Mathtech
Procedure
D
Lave & Seskin;
Lipfert
Ostro
Ferris
et al.
Mathtech
E
Lave & Seskin;
Lipfert
Ostro
Crocker
et al.
Geom. mean of
Cols. D & F
at county
level
Mathtech
F
Col. C +
Col. D
Col. B +
Col. C
Crocker
et al.
Watson &
Jaksch +
Cummings
et al.
Mathteoh
-------
Does the procedure avoid double counting of benefits?
Does the procedure provide complete coverage of benefits?
Since no one procedure best satisfies all three criteria, it is useful to
consider various alternative approaches. Six possible alternatives are
identified in Table 1-9. They give differing weights to the three
criteria. Each alternative is discussed further below and in Section 10.
Procedure A
Aggregation Procedure A is designed to be consistent with the CASAC's
review of the scientific literature. In the Criteria Document for PH and
SO. CASAC distinguished between studies providing quantitative evidence of
pollution-related effects and those providing less quantitative evidence.
Among the health studies considered adaptable for use in this benefit
analysis, the studies by Hazumdar et al. and Ferris et al. were judged by
CASAC to be in the quantitative category (35). Procedure A includes only
these two health studies. (Other studies were judged by CASAC as providing
quantitative evidence, but they could not be adapted for benefit calcula-
tions. )
Procedure A also excludes all non-health studies. The Criteria
Document acknowledges the probable existence of non-health effects such as
soiling (36), but is noncommittal regarding the availability of quantita-
tive evidence for such effects.
It is unlikely that Procedure A involves any double counting of
benefits. This conclusion follows because of its incomplete coverage of
benefit categories.
Procedure B
This procedure is similar to the previous one with the exception that
the health study by Samet et al. has been included. This modification is
consistent with the conclusions of the EPA Staff Paper (37) concerning
1-45
-------
studies providing reasonable evidence of concentration-response relation-
ships for health effects. The Staff Paper includes the Samet et al. study
in this category, as well as the studies in Procedure A.
As with Procedure A, no non-health studies are included. The Staff
Paper, like the Criteria Document, is noncommittal regarding the availa-
bility of quantitative non-health effects evidence.
Procedures C Through F
The studies included in Procedures A and B are among those judged by
CASAC and the EPA Staff Paper as providing the most reliable quantitative
evidence of air pollution effects. Procedures C, D, E, and F make
increasing degrees of use of other studies. This includes studies such as
the one by Ostro which had not yet been formally published prior to closure
of the Criteria Document. It also includes studies that CASAC judged to be
less reliable but that still have some merit. With these latter studies,
it is of interest to trace out their implications. Even though these
studies were judged less reliable, they provide the only available evidence
of the potential magnitude of benefits in some categories (e.g., chronic
exposure mortality). Including calculations based on these studies
provides greater assurance that potentially important effects of PH have
not been omitted from the benefit analysis. At the same time, the
distinction is maintained between these studies and the CASAC-approved
studies by including them in different aggregation procedures.
Procedures C through F differ in the degree to which the additional
studies are used (C is more selective). The procedures also differ in the
importance assigned to avoidance of double counting versus completeness of
coverage (C stresses avoidance of double counting).
Procedure C
Procedure C brings in the study of acute morbidity by Ostro. This
recent study is not cited in the Criteria Document and, in fact, the
1-46
-------
benefit analysis uses later, supplemental results not available prior to
the Criteria Document final draft.
Among the available health studies not already included in Procedures
A and B, the Ostro study is rated very high for purposes of an economic
benefit analysis. The Ostro study has several features which, taken
together, lead to this conclusion. First, the study uses data on indivi-
duals (microdata) rather than macrodata such as vital statistics. Micro-
data are preferred because they improve the accuracy of the estimates of
the dose-response function parameters and enable more direct application of
proper statistical controls. Second, the study uses data samples from 90
medium-sized U.S. cities. The broad scope of the sample provides greater
assurance that the estimates of health effects can be used as the basis for
a national benefit analysis. Third, in identifying the pollution exposure
for each individual, the study locates individuals by city of residence
rather than by county or SMSA. It also uses an average of the readings
among the population-oriented monitors in each city rather than using a
single monitor or worst-case monitor. Fourth, the study controls for
smoking habits as well as a variety of other economic and demographic
variables. While the study does have some limitations (e.g., no control
for dietary habits, limited control for occupational hazards and exposure),
its other features make it clearly superior for use in a benefit analysis.
With the addition of the Ostro study in Procedure C, the study by
Samet et_ al. is omitted to avoid possible double counting of benefits. The
Samet et al. study is concerned with acute respiratory effects, which may
be a subset of the effects captured in the Ostro study. The Ostro study is
also given preference over the acute morbidity study by Saric et al. The
latter study provides less complete coverage of diseases (respiratory
disease only) and limited concentrationresponse information (e.g., no
separation of PM/SO- effects, no information on PM effects below 200 |ig/m3
annual mean, etc.).
Procedure C also incorporates the Mathtech study of household soiling
effects. The Criteria Document and Staff Paper reference this recent study
1-47
-------
but are noncommittal as to its use for quantitative purposes. The study is
included here in order to provide some coverage of soiling effects, in
recognition of its strong analytical features, and because of the favorable
peer review which the study has received.
Note that Procedure C excludes possible chronic exposure mortality
effects, provides limited coverage of chronic morbidity effects (respira-
tory illness only), and offers limited coverage of soiling effects.
Procedure D
Procedure D addresses some of the possible incompleteness of coverage
present with Procedure C. In particular. Procedure C omits the possibility
of benefits from reduced mortality risk due to long-term exposure. There
are many existing, cross-sectional studies of long-term exposure mortality
effects. The majority conclude that mortality effects exist. However,
these studies are not highly regarded by CASAC, so care is taken in how the
results of these studies are used in the benefit analysis. This includes
carefully assessing the limitations of each study and ultimately using
results which are at the low end of the studies finding effects.
Summarizing briefly, nine chronic exposure mortality studies are
reviewed in detail in Section 4 of the report. As noted previously, six of
those studies found statistically significant associations between
mortality rates and PM (generally measured as TSP). Two others found that
it was the sulfate component of PM, rather than TSP, that had an associa-
tion, The ninth found little evidence of association between mortality and
air pollution.
The three studies finding no PM effects have been faulted for inclu-
sion of an excessive number of pollution variables and/or control
variables. This leads to multicollinearity and reduces the possibility of
finding statistically significant PM effects. At least one of the three
no-effect studies also used relatively more recent data when mean levels of
TSP were reduced compared to earlier years, thus reducing the possibility
1-48
-------
of finding significant effects. These negative results (no effects = no
benefits) are nonetheless retained for use as a lower-bound estimate of
zero for this category.
The six positive-effect studies (and the no-effect studies) have been
faulted on other grounds. Some have used relatively aggregate data. Some
have been criticized for the pollution monitoring used. All of the studies
have incomplete statistical controls, which is also true of the studies
included in Procedures A through C.
Particular attention in the benefit analysis is focused on the
mortality studies by Lave and Seskin and more recent studies which have
attempted to address the criticisms of their work. Two studies by Lipfert
are important in this regard. In one study, Lipfert used SMSA mortality
data and air pollution data, as done earlier by Lave and Seskin. He found
results comparable to those of Lave and Seskin and comparable to those
obtained in his parallel analysis using less aggregate data at the city
level. In a second study, Lipfert did a reanalysis of the Lave and Seskin
work and added an approximate control for smoking. This study also found
results consistent with Lave and Seskin's original results. Based on these
findings, the benefit analysis uses results which are on the low end of the
range observed by Lipfert and Lave and Seskin. These results are also
considerably lower than those found in two of the three other studies
finding positive effects. Thus, the point estimate of chronic exposure
mortality benefits included in Procedure D is based on the low side of the
evidence available among studies finding chronic exposure effects.
With the addition of the chronic exposure studies, the acute exposure
study by Hazumdar et al. is dropped from this procedure. This is because
of the possibility for overlap between these estimates. In particular, the
chronic exposure studies are based on annual mortality rates. Annual
mortality rates will include all deaths during the year, including those
deaths that may be due to acute exposures. Thus, it is possible for the
chronic exposure studies to be capturing the mortality effects of both
acute and chronic exposure. The extent to which this may happen is
1-49
-------
unknown, however. It depends on a variety of factors such as the
functional forms for the acute and chronic dose-response relationships, the
statistical correlation between the measures of acute and chronic exposure,
and so on. To be conservative in the estimate of benefits, the acute
exposure study is eliminated from this procedure.
Procedures E and F
Procedures E and F are most easily considered together. Procedure E
addresses the incomplete coverage of the chronic morbidity category and the
underestimation of soiling effects. In Procedures A through D, chronic
morbidity estimates are based on the study by Ferris et al. which includes
only respiratory diseases. In Procedure E, the Ferris .e_t .§_!. study is
replaced by the Crocker et al. study which includes more chronic illnesses.
In Procedure E, the Hathtech study of soiling and materials damage in
parts of the manufacturing sector is included. The coverage of household
sector soiling is also expanded. The latter is done by taking into
consideration the results of studies by Watson and Jaksch and by Cummings
et al. The sum of these two studies is used in Procedure F, in view of the
possibility that together they may overestimate benefits. In contrast, the
Mathtech household study probably underestimates benefits. Procedure E
thus uses a compromise estimate for household soiling: the geometric mean
of the Mathtech estimate and the sum of the estimates based on Cummings .e_t
.§_!. and Watson and Jaksch. The geometric mean is used as a conservative
measure of the average of the two estimates.
The remainder of the estimates used in Procedure F also seek to
provide more complete coverage of benefits, with the possible risk of some
double counting. In particular, the benefits for the acute and chronic
exposure studies for mortality are added together; the same is also done
with the estimates from the acute and chronic exposure studies of acute
morbidity. As a result, Procedure F provides the most complete estimate of
benefits possible with the available studies. However, it may involve some
1-50
-------
double counting. The possibility of double counting is less likely to
arise with any of the procedures A through E.
Results of the Aggregation Procedures
Applying the procedures outlined in Table 1-9 leads to the results
shown in Table 1-10. The benefit estimates shown in the table are for the
PH10 (70, 250) Scenario B standard. The estimates for Subtotal 2 range
from a low of $1.24 billion for Procedure A to $52.4 billion for Procedure
F. Host of the variation is due to the differing estimates of health
benefits. Benefits of reduced mortality risk are estimated to range from
$1.1 billion to $13.8 billion. For acute morbidity, the range is $0 to
$12.0 billion. For chronic morbidity, the estimates range from $0.12
billion to $11.4 billion.
The range of values between Procedures A and F results from the
alternative ways to combine the various studies and benefit categories. It
does not reflect the uncertainty present in the estimates for the indivi-
dual studies. That is, all entries in the table are based on the point
estimate for each study or benefit category. If the additional uncertainty
about each study were incorporated, as reported previously in Tables 1-5
and 1-7, the full range of variation in Subtotal 2 would be $160 million to
$146 billion. These broader estimates of the range of uncertainty for each
standard are reported in the tables contained in Section 11.
Note that the table makes a distinction between Subtotals 1 and 2.
Subtotal 2 includes both manufacturing sector benefits and benefits
accruing to individuals and households (Subtotal 1). The distinction
between the two subtotals enables comparisons to be made with the estimates
from the property value and hedonic wage studies.
The benefit estimates based on the property value and hedonic wage
studies are shown at the bottom of Table 1-10. These estimates are the
same as reported previously in Table 1-8. They are not included in the
subtotals because they are believed to have substantial overlap with the
1-51
-------
Table 1-10
INCREMENTAL BENEFITS FOR THE PM10 (70, 250) SCENARIO B STANDARD*
Benefit Category
Mortality
Acute Morbidity
Chronic Morbidity
Household Sector
Soiling & Materials
Subtotal 1
Manufacturing Sector
Soiling & Materials
Subtotal 2
Property Value Studies
Hedonic Wage Studies
A B C D E F
1.12 1.12 1.12 12.72 12.72 13.
0.0 1.32 10.65 10.65 10.65 11.
0.12 0.12 0.12 0.12 11.40 11.
0.0 0.0 0.73 0.73 3.14 13.
1.24 2.56 12.63 24.24 37.92 51.
t
0.0 0.0 0.0 0.0 1.30 1.
1.24 2.56 12.63 24.24 39.22 52.
Minimum Midpoint Maximum
3.43 6.85 11.42
9.81 19.81 37.21
84
97
40
85
07
30
36
* 1982 discounted present values in billions of 1980 dollars for a 7-year time
horizon (1989-95) and a 10 percent discount rate. Individual entries may not sum
to subtotals due to independent rounding.
-------
benefit categories included in Subtotal 1 (health, household soiling, and
visibility benefits). Rather, they are reported here to serve as a cross-
check on the benefit numbers estimated from the health and welfare effects
studies. Note that the range of property value estimates falls between the
totals for Procedures 8 and C. The range of wage estimates encompasses
Procedures C and D.
Unfortunately, the exact relationship between benefits included in
Subtotal 1, and benefits estimated from property value and wage (PV&W)
studies is unknown. However, several general points can be made. First,
PV&W studies are behavioral studies. That is, they record how individuals
respond to external factors such as market prices, housing (or job)
characteristics, and amenities such as air quality. This implies that
benefit estimates from PV&W studies will depend on: 1) individuals'
perception of air quality; 2) individuals' perception of how air quality
affects health, household soiling, etc.; and 3) how individuals respond to
these perceptions. This means that PV&W studies can account for indivi-
duals' efforts to avert or mitigate the effects of air pollution.
In contrast to PV&W studies, most of the health studies used in this
report do not account for averting or mitigating behavior. Rather, the
central issue is one of identifying the technical relationship between
health status and air quality. If averting or mitigating behavior has
occurred prior to conducting a health study, then health effects will be
reduced and the study's estimate of effects will be understated.
At the same time, the dependence of PV&W studies on perception and
behavioral adjustments can lead PV&W studies to underestimate the benefits
of air pollution reductions. For example, if there are air pollution
health effects which individuals do not perceive, PV&W studies will not
identify these effects.
The possibility of unperceived health effects suggests that PV&W
studies should be viewed as lower-bound estimates of aggregate benefits.
In this view, the property value estimates suggest that aggregation
1-53
-------
procedures A, B, and possibly C provide underestimates. The wage studies
suggest that A. B, C, and possibly D are too low. The PV&W results are
thus consistent with the earlier observations that Procedures A through D
probably provide incomplete coverage of benefits.
The relationship between the estimates from PV&W studies is also not
well understood. The wage studies used for the benefit estimates are both
specified in terms of the "real" wage. The real wage is calculated by
dividing the nominal (actual) wage by a cost-of-living index, one component
of which is the cost of housing. Thus, benefit estimates based on the real
wage models may reflect adjustments to air pollution in both the labor and
property markets. This suggests that the benefit estimates from real wage
models may be inclusive of benefits estimated from property value models.
It also may explain why the benefit estimates are larger with the wage
models than the property value models ($19.81 billion vs. $6.85 billion).
lenefits for Other
Alternative Standards
The previous discussion has focused entirely on the PM10 (70, 250)
Scenario B standard. Incremental benefits for a variety of alternative
standards considered in this analysis are shown in Table 1-11. The
alternatives include all of the standards defined previously in Table 1-3.
They include six PM10 standards with 1989 implementation dates, two TSP
standards with 1989 implementation dates, and two TSP standards with 1987
implementation dates. All standards extend through 1995. All estimates
are displayed to two significant digits.
As noted previously, health studies available for the benefit analysis
do not incorporate particle size information. Benefits shown in the table
for the PM10 standards are based on the TSP change that results. Compari-
sons across PM10 and TSP standards thus reflect only differences in
relative stringency in terms of the TSP reduction; they do not reflect
differences in particle size. If PM10 standards lead to proportionately
1-54
-------
Table 1-11
INCREMENTAL BENEFITS FOR ALTERNATIVE PM10 AND TSP STANDARDS*
(B Scenarios)
i
ui
Ui
Alternative standard
PM10 70 AAM/250 24-hr. 7 yr.
PM10 55 AAM/- 7 yr.
PM10 55 AAM/250 24-hr. 7 yr.
PM10 55 AAM/200 24-hr. 7 yr.
PM10 55 AAM/150 24-hr. 7 yr.
PM10 48 AAM/183 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 7 yr.
TSP -/150 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 9 yr.
TSP -/150 24-hr. 9 yr.
Aggregation Procedure
A
1.2
1.8
1.8
1.8
2.0
2.1
2.1
2.4
2.9
3.3
B
2.6
4.1
4.2
4.3
5.0
5.2
5.2
6.4
7.2
8.9
C
13
22
22
23
27
29
29
36
41
50
D
24
43
43
44
52
56
57
71
81
100
E
39
74
74
76
90
98
100
120
140
170
F
52
98
98
100
120
130
130
160
190
230
* 1982 discounted present values in billions of 1980 dollars at a 10 percent discount
rate. The 7-year time horizon is 1989-95 and the 9-year horizon is 1987-95.
Comparisons between PM10 and TSP standards are in terms of TSP stringency, not particle
size.
-------
larger reductions in PM10 relative to TSP, benefits for the PM10 standards
may be underestimated. Data from the cost and air quality analysis suggest
that proportionately larger reductions do not generally occur. However,
approximations in that analysis are such that the comparisons should still
be interpreted with caution. This is signified by the line in the table
separating the two groups of standards.
The standards are listed in approximate order of increasing stringency
with the PM10 (70, 250) as the least stringent and the TSP (-, 150) as the
most stringent. The ordering is ,only approximate for two reasons. First,
the ordering is different depending on whether the annual or 24-hour con-
centration is used. Second, the PM10 and TSP standards use different
statistical measures. The PM10 standards are stated in terms of arithmetic
means, expected second maximum values, and particles less than or equal to
10 |im in diameter; the TSP standards use geometric means, observed second
highest values and total suspended particles. This means that relative
stringency will vary from county to county, depending on the particle size
distribution and the temporal distribution of air quality in each county.*
An examination of Table 1-11 reveals several general patterns. First,
incremental benefits increase as the stringency of the standard increases.
This results because more stringent standards lead to larger improvements
in air quality, and also increase the number of counties where air quality
improvements occur. Second, the ordering of standards in terms of incre-
mental benefits is insensitive to the choice of a particular aggregation
procedure. That is, the ordering is the same for all six procedures.
Third, standards with earlier implementation dates produce larger incre-
mental benefits. This is also to be expected since earlier implementation
leads to improvements in air quality which are earlier and of longer
duration.
* As an example, it is possible for air quality in one county to have a
larger annual geometric mean and a smaller annual arithmetic mean than in
another county. In this case, it would be possible for a geometric mean
standard to be binding in the first county but not binding in the second
and vice versa for an arithmetic mean standard. The same is true for
comparisons between different monitors.
1-56
-------
It should be noted that large benefits for a'particular standard do
not necessarily imply that the standard is economically preferable to some
other standard with lower benefits. The costs of attaining the standard
must also be considered. This is done in Section 2 of this report.
A Scenarios
Table 1-12 contains estimates of the same ten alternative standards
for the "A" scenario case. Recall that the A scenarios include some
counties where available control options are insufficient to attain the
standards. In the B scenarios, these counties are forced into attainment.
The degree of air quality improvement is thus lower in the A scenarios
relative to the corresponding standards in the B scenarios. Hence,
benefits (and costs) are also lower with the A scenarios.
The A scenario results exhibit the same general patterns as the B
scenarios. Benefits are larger with more stringent standards and with
earlier implementation dates, and these patterns exist for all six
aggregation procedures. As with the B scenarios, costs of control
associated with each standard must also be considered before drawing
conclusions about the economic attractiveness of a particular standard.
Sensitivity to Staff Paper Lower Bounds
As discussed previously, it is of interest to calculate benefits under
the assumption that no health benefits are produced by reducing PM below
the Staff Paper lower bounds of 55 (ig/m3 annual PM10 (110 fig/m3 annual TSP)
and 150 ug/m for 24-hour PM10. As also discussed previously, there are
several practical problems with doing the calculation, ranging from
probable statistical bias to lack of daily data on PM concentrations.
Subject to these previously discussed limitations, Tables 1-13 and 1-14
contain the benefit estimates that result when all health benefits
associated with reducing PM concentrations below 110 (ig/m annual average
TSP are excluded. Table 1-13 provides results for the B scenarios while
Table 1-14 incorporates the A scenarios.
1-57
-------
Table 1-12
INCREMENTAL BENEFITS FOR ALTERNATIVE PM10 AND TSP STANDARDS*
(A Scenarios)
i
ui
00
Alternative Standard
PM10 70 AAM/250 24-hr. 7 yr.
PM10 55 AAM/- 7 yr.
PM10 55 AAM/250 24-hr. 7 yr.
PM10 55 AAM/200 24-hr. 7 yr.
PM10 55 AAM/150 24-hr. 7 yr.
PM10 48 AAM/183 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 7 yr.
TSP -7150 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 9 yr.
TSP -/150 24-hr. 9 yr.
Aggregation Procedure
A
0.88
1.1
1.1
1.1
1.2
1.2
1.2
1.4
1.7
1.9
B
1.7
2.3
2.3
2.4
2.7
2.8
2.8
3.4
3.9
4.7
C
8.2
11
11
12
14
14
14
18
20
25
D
16
22
22
23
26
27
27
34
38
48
E
24
35
35
36
43
43
44
56
62
78
F
32
47
47
49
58
59
60
76
84
110
* 1982 discounted present values in billions of 1980 dollars at a 10 percent discount
rate. The 7-year time horizon is 1989-95 and the 9-year horizon is 1987-95.
Comparisons between PM10 and TSP standards are in terms of TSP stringency, not particle
size.
-------
Table 1-13
INCREMENTAL BENEFITS FOR ALTERNATIVE PM10 AND TSP STANDARDS, WITH LOWER BOUND APPLIED*
(B Scenarios)
i
Ul
Alternative Standard
PM10 70 AAM/250 24-hr. 7 yr.
PM10 55 AAM/- 7 yr.
PM10 55 AAM/250 24-hr. 7 yr.
PM10 55 AAM/200 24-hr. 7 yr.
PM10 55 AAM/150 24-hr. 7 yr.
PM10 48 AAM/183 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 7 yr.
TSP -/150 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 9 yr.
TSP -/150 24-hr. 9 yr.
Aggregation Procedure
A
0.44
0.46
0.46
0.46
0.46
0.46
0.46
0.46
0.57
0.58
B
0.62
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.82
0.82
C
2.6
3.3
3.3
3.4
3.7
3.8
3.8
4.3
5.0
5.7
D
14
19
19
19
19
19
19
19
26
27
E
28
41
41
41
42
43
43
47
60
64
F
40
61
61
62
69
71
71
84
100
120
With TSPAAM lower bound of 110 ug/in applied to all health studies. 1982 discounted
present values in billions of 1980 dollars. Comparisons between PM10 and TSP standards
are in terms of TSP stringency, not particle size.
-------
Table 1-14
INCREMENTAL BENEFITS FOR ALTERNATIVE PM10 AND TSP STANDARDS. WITH LOWER BOUND APPLIED*
(A Scenarios)
i
o^
o
Alternative Standard
PM10 70 AAM/250 24-hr. 7 yr.
PM10 55 AAM/- 7 yr.
PM10 55 AAM/250 24-hr. 7 yr.
PM10 55 AAM/200 24-hr. 7 yr.
PM10 55 AAM/150 24-hr. 7 yr.
PM10 48 AAM/183 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 7 yr.
TSP -/150 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 9 yr.
TSP -/150 24-hr. 9 yr.
Aggregation Procedure
A
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.44
0.44
B
0.47
0.47
0.47
0.47
0.47
0.47
0.47
0.47
0.60
0.60
C
1.8
2.0
2.0
2.0
2.1
2.1
2.1
2.4
2.9
3.2
D
9.4
11
11
11
11
11
11
11
15
16
E
17
20
20
21
22
21
21
23
30
33
F
24
31
31
31
35
35
35
41
49
58
With TSPAAH lower bound of 110 ug/md applied to all health studies. 1982 discounted
present values in billions of 1980 dollars. Comparisons between PM10 and TSP standards
are in terms of TSP stringency, not particle size.
-------
The effect of imposing the Staff Paper lower bound is to reduce the
benefits associated with all of the standards. It reduces the more
stringent standards proportionately more than the less stringent standards.
This results because the more stringent standards produce air quality
improvements which are increasingly below the lower bound concentration.
Hence, benefits show little increase since no credit is taken for below-
lower bound air quality improvements.
Geographic Distribution of Benefits
Benefits of reduced PH concentrations vary considerably among the
different regions of the country. This is the result of several factors
such as differences in baseline air quality, differences in population, and
differences in population growth rates. As an illustration, regional
benefits for the PH10 (70, 250) Scenario B standard are shown in Table 1-
15. The benefits are in discounted present value terms and the regions
shown are the standard Federal administrative regions. As can be seen,
nine of the ten regions receive benefits (the exception is New England).
Particularly large shares of the benefits arise in Regions V (East North
Central), VI (South Central), and IX (South Pacific). These shares change
depending on the particular aggregation procedure used (A through F). For
example. Procedure A suggests that 54 percent of the benefits would occur
in Region IX. Under Procedure E, the estimate for Region IX falls to 32
percent and is exceeded by the 35 percent share in Region V.
Analysis of the other standards indicates that regional shares also
change depending on the standard under consideration. For example, the
PM10 (70, 250) Scenario B standard is the only one in which Region I (New
England) receives no benefits. For all other B scenario standards, Region
I receives positive benefits under all aggregation procedures, and the same
pattern occurs with the A scenarios (assuming no lower bound constraint is
imposed). A general observation is that benefits are more widely shared
with the more stringent standards. This results in large part because more
counties would be in non-attainment with the more stringent standards and
thus more counties would experience an air quality improvement over
1-61
-------
Table 1-15
INCREMENTAL BENEFITS BY REGION FOR THE PM10 (70, 250) SCENARIO B STANDARD*
Ul> A n _ .» j .._
ErA Keg ion
I New England
II New York-New Jersey
III Middle Atlantic
IV South Atlantic
V East North Central
VI South Central
VII Midwest
VIII Mountain
IX South Pacific
X North Pacific
U.S.
A
0.0
0.0
0.08
0.05
0.23
0.08
0.02
0.04
0.67
0.07-
1.24
B
0.0
0.00
0.14
0.13
0.63
0.24
0.04
0.10
1.16
0.14
2.56
Aggregation
C
0.0
0.02
0.64
0.70
3.69
1.38
0.20
0.45
4.86
0.70
12.63
Procedure
D
0.0
0.04
1.08
1.34
8.36
2.89
0.40
0.87
8.01
1.25
24.24
E
0.0
0.06
1.88
2.20
13.68
4.60
0.64
1.44
12.71
2.00
39.22
F
0.0
0.08
2.51
2.94
17.57
6.15
0.86
1.98
17.55
2.72
52.36
* 1982 discounted present values in billions of 1980 dollars for a 7-year time horizon (1989-95)
and a 10 percent discount rate. Individual entries may not sum to U.S. totals due to indepen-
dent roundoffs.
-------
baseline. For example, the PH10 (70, 250) standard would produce air
quality improvements in approximately 90 to 100 counties; the TSP (-/ISO)
standard would result in benefits for approximately 500 counties.
FINDINGS AND CONCLUSIONS
This study estimates the incremental benefits of alternative ambient
air quality standards for PH. The principal conclusion of the study is
that the benefits are economically significant but that the estimates of
benefits are highly uncertain. That is, the study provides a considerable
amount of new information but care must be exercised in the use of the
results. These conclusions are developed further below.
Limitations in available data, methods and scientific evidence make it
difficult to estimate the benefits of alternative PM NAAQS. Each limita-
tion introduces uncertainty and requires an assumption or an exercise of
judgment to fill the gap. The approach taken in this study is to evaluate
the consequences of alternative assumptions and judgments. This provides
the most complete information possible and reduces the possibility of
biased results.
As an example, under the most extreme assumptions, the discounted
present value of benefits of the PM10 (70, 250) Scenario B standard range
from as low as $160 million to as high as $146 billion. The lower estimate
arises when the most restrictive assumptions are used throughout the
analysis; the higher estimate results from use of much less restrictive
assumptions.*
* The stated range does not reflect two additional sources of uncertainty:
1) the air quality data approximations required in the absence of
detailed dispersion modeling; and 2) the possible existence and magnitude
of benefits below the Staff Paper lower bounds. Incorporating the first
of these could potentially reduce both ends of the range by a factor of 2
to 4. Imposing the Staff Paper lower bound of 110 (ig/m annual average
TSP on the health studies makes the range $140 million to $99 billion.
1-63
-------
Using a somewhat less extreme set of assumptions leads to a narrower
range of estimates. For example, with the standard mentioned above, a more
realistic estimate of the range is $1 billion to $50 billion. This is the
range identified previously in Tables 1-10 and 1-11. The range in this
case is due to the use of alternative aggregation procedures. As noted in
the discussion of the tables and Table 1-9, the lower end of the range is
still likely to be an underestimate because of incomplete coverage of
benefits. The upper end is also possibly an overestimate due to some
double counting of benefits. An alternative method for estimating benefits
(property value and wage studies) is consistent with this assessment and
suggests a still narrower range of about $3 billion to $40 billion. This
range would be most consistent with aggregation procedures B through E. It
is difficult to narrow the range further without making more controversial
judgments.
Another finding of the study is that incremental benefits generally
increase as standards become more stringent and as implementation dates are
advanced. This is true for both A and B scenarios and for all aggregation
procedures. (The few exceptions arise when lower bounds are applied.) For
example, with the B scenarios, benefits increase by a factor of 3 to 4 when
moving from the least restrictive standard to the most restrictive
standard. The increase results from two changes: 1) the larger improve-
ment in air quality in each geographic area, and 2) the addition of more
areas that will experience an improvement. However, this does not neces-
sarily imply that the most restrictive standard is the economically
preferred standard because the offsetting costs of pollution control must
also be taken into account. A comparison of benefits with costs is done in
Section 2.
of the
The sources of uncertainty in the benefit analysis are identified in
the basic report sections. They have also been summarized in this section.
They include such problems as: conflicting evidence among available scien-
tific studies; uncertainty about the concentration ranges over which health
1-64
-------
effects occur; incomplete coverage of benefit categories; alternative
methods for valuing health improvements; alternative methods of aggrega-
tion; and approximations introduced by the methods for estimating air
quality improvements. All of these problems have been addressed in the
study. Where possible, the magnitude of the uncertainty has been estimated
directly. These estimates of uncertainty are reflected in the range of
benefit estimates discussed previously. Where quantitative estimates of
the uncertainty are not possible, conservative assumptions are used
instead, and the sources of uncertainty are described qualitatively.
REFERENCES
1. Executive Order 12291 of February 17, 1981: Federal Regulation.
Federal Register 46:13193-13198. February 19, 1981.
2. Smith, A. E. and K. L. Bmbater. Costs and Air Quality Impacts of
Alternative National Ambient Air Quality Standards for Particulate
Matter. Technical Support Document. Argonne National Laboratory,
Argonne, IL, October 1982.
3. U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards. Review of the National Ambient Air Quality Standards
for Particulate Matter: Assessment of Scientific and Technical
Information. OAQPS Staff Paper (EPA-450/5-82-001), Research Triangle
Park, NC, January 1982.
4. U.S. Environmental Protection Agency, Office of Research and Develop-
ment. Air Quality Criteria for Particulate Matter and Sulfur Oxides:
External Review Draft No. 4. Research Triangle Park, NC, Dec. 1981.
5. Pace, T. G. and N. H. Frank. Procedures for Estimating Probability of
Non-Attainment of a PM10 NAAQS Using Total Suspended Particulate or
Inhalable Particulate Data. Draft Guideline. U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards,
Monitoring and Data Analysis Division, November 1982.
6. EPA/OAQPS Staff Paper, op. cit.. Appendix E, p. E-4. (Appendix E is
the letter from the chairman of CASAC to the EPA Administrator con-
cerning CASAC review and closure of the EPA Staff Paper cited above.)
7. Ibid., p. xiii and p. 88.
8. Bouhuys, A., G. J. Beck and J. B. Schoenberg. Do Present Levels of
Air Pollution Outdoors Affect Respiratory Health? Nature, 276:466-
471, 1978.
9. EPA/OAQPS Staff Paper, o£. cit.. p. 62.
1-65
-------
10. Ibid., p. xvi.
11. Ferris, B. G., Jr., H. Chen, S. Puleo and R. L. H. Murphy, Jr.
Chronic Nonspecific Respiratory Disease in Berlin, New Hampshire,
1967-1973: A Further Follow-Up Study. American Review of Respiratory
Disease, 113:475-485, 1976.
12. Ferris, B. G., Jr., I. Higgins, M. W. Higgins and J. M. Peters.
Chronic Non-Specific Respiratory Disease in Berlin, New Hampshire,
1961-1967: A Follow-Up Study. American Review of Respiratory
Disease, 107:110-122, 1973.
13. Lunn, J., J. Knowelden and A. J. Handyside. Patterns of Respiratory
Illness in Sheffield Infant School Children. British Journal of
Preventive Social Medicine, 21:7-16, 1967.
14. Ostro, Bart D. A Search for a Threshold in the Relationship of Air
Pollution to Mortality in London: A Reanalysis of Martin and Bradley.
Working Paper dated September 21, 1982.
15. Mazumdar, S., H. Schimmel, and I. Higgins. Relation of Air Pollution
to Mortality: An Exploration Using Daily Data for 14 London Winters,
1958-1972. Electric Power Research Institute, Palo Alto, 1980.
16. Samet, J. M., F. E. Speizer, Y. Bishop, J. D. Spengler, and B. 6.
Ferris, Jr. The Relationship Between Air Pollution and Emergency Room
Visits in an Industrial Community. Journal of the Air Pollution
Control Association, 31:236-240, 1981.
17. Saric, M., M. Fugas, and 0. Hrustic. Effects of Urban Air Pollution
on School Age Children. Archives of Environmental Health, 36:101-108,
1981.
18. Ferris, et al.. op. cit.. 1976.
19. Ferris, et. ±1., op. cit.. 1973.
20. Ferris, B. G., Jr. and D. 0. Anderson. Prevalence of Chronic Respira-
tory Disease in a New Hampshire Town. American Review of Respiratory
Diseases, 86:165-177, 1962.
21. Lave, L. B. and E. P. Seskin. Air Pollution and Human Health. Johns
Hopkins University Press, Baltimore, 1977.
22. Lipfert, F. W. Sulfur Oxides, Particulates, and Human Mortality:
Synopsis of Statistical Correlations. Journal of the Air Pollution
Control Association, 30:366-371. April 1980.
23. Lipfert, F. W. On the Evaluation of Air Pollution Control Benefits.
Prepared for the National Commission on Air Quality, November 1979.
1-66
-------
24. Lipfert, F. W. The Association of Air Pollution with Human Mortality:
Multiple Regression Results for 136 Cities, 1969. Paper presented at
70th Annual Meeting of the Air Pollution Control Association, June 20-
24, 1977.
25. Ostro, B. D. Morbidity, Air Pollution and Health Statistics. Paper
presented at the Joint Statistical Meetings of the American Statisti-
cal Association and Biometric Society, Detroit, MI, August 12, 1981.
The benefit analysis also uses later results supplied by the author.
26. Crocker, T. D., e_t .§_!. Methods Development for Assessing Air Pollu-
tion Control Benefits - Vol. 1: Experiments in the Economics of Air
Pollution Epidemiology. Prepared for the U.S. Environmental Protec-
tion Agency, Laramie, University of Wyoming, February 1979.
27- Ibid.
28. Cummings, R., H. Burness and R. Norton. Methods Development for
Environmental Control Benefits Assessment - Volume V: Measuring
Household Soiling Damages from Suspended Air Particulates, A Methodo-
logical Inquiry. Draft Report, January 1981.
29. Watson, W. and J. Jaksch. Air Pollution: Household Soiling and
Consumer Welfare Losses. Journal of Environmental Economics and
Management, 9:248-262. 1982.
30. Manuel, E. H., Jr., R. L. Horst, Jr., K. M. Brennan, W. N. Lanen, M.
C. Duff and J. K. Tapiero. Benefits Analysis of Alternative Secondary
National Ambient Air Quality Standards for Sulfur Dioxide and Total
Suspended Particulates - Volume II. Contract No. 68-02-3392, Final
Report to the U.S. Environmental Protection Agency by Mathtech, Inc.,
Princeton, NJ, August 1982.
31. Ibid.. Volume III.
32. EPA/OAQPS Staff Paper, op., cit.. p. 125.
33. Smith, V. K. The Role of Site and Job Characteristics in Hedonic Wage
Models. Forthcoming in Journal of Urban Economics.
34. Rosen, S. Wage-Based Indices of Urban Quality of Life. In: Current
Issues in Urban Economics, P. Mieszkowski and M. Straszheim (eds.).
Johns Hopkins University Press, Baltimore, 1979.
35. EPA Criteria Document, op. cit.. pp. 14-49 to 14-54.
36. Ibid., p. 10-73.
37. EPA/OAQPS Staff Paper, op_. cit.. pp.48-63.
1-67
-------
SECTION 2
NET BENEFIT ANALYSIS
-------
SECTION 2
NET BENEFIT ANALYSIS
INTRODUCTION
This section of the report provides comparisons of the estimated
benefits and costs of the several alternative National Ambient Air Quality
Standards for particulate matter (PM NAAQS). These comparisons of benefits
and costs are referred to as benefit-cost analyses. They provide a frame-
work for evaluating the economic effects of alternative regulatory policies
and are presented as a response to Executive Order 12291 which requires the
identification of the regulatory alternative which will produce the maximum
net benefits to society.
The implementation of a particular air quality standard, for example,
may lead to favorable health and other welfare effects that represent a
clear improvement in the economic well-being of some members of society.
At the same time, however, costs may be incurred as additional resources
are committed to reduce emissions to permissable levels defined by these
air quality standards. These costs cause a reduction in the economic well-
being of some members of society. Given that these costs are generally
incurred as air quality is improved, an evaluation of the net impact of the
standards on society's economic well-being requires an assessment of both
the benefits and costs associated with each of the alternative PM NAAQS.
Some necessary background concepts are developed immediately below.
In particular, the economic criteria employed to evaluate the alternative
PM NAAQS are described. Proper methods of conducting benefit-cost analyses
are reviewed next. Following this, the estimates of both benefits and
costs are described; comments on the appropriateness of these estimates for
2-1
-------
use in the benefit-cost analysis are provided. Next, the limitations of
benefit-cost analysis are discussed. The results of the benefit-cost
analyses are then presented. Finally, summary remarks, conclusions, and
qualifications are offered.
BENEFIT-COST ANALYSIS: EVALUATION CRITERIA
Air quality regulations affect society's economic well-being by
causing a reallocation of productive resources within the economy. Speci-
fically, resources are allocated towards the production of cleaner air and
away from other goods and services that would otherwise be produced.
Benefit-cost analysis provides a method for assessing the desirability of
the alternative PM NAAQS in terms of their impacts on the allocation of
economic resources. These air quality standards are evaluated in terms of
two established criteria: cost-effectiveness and efficiency.
Incremental Benefits and Costs; An analysis of the incremental
benefits and costs associated with each of the alternative PM NAAQS is
required for an evaluation of both the cost-effectiveness and the relative
efficiency of the alternative ambient air quality standards. These are
defined as follows:
The incremental benefits associated with a given PM NAAQS
are defined as the additional benefits resulting from
improvements in air quality over baseline air quality
levels.
The incremental costs associated with a given PM NAAQS are
defined as the additional costs that are incurred to
achieve and maintain improvements in air quality over the
baseline levels.
Note that both incremental benefits and incremental costs are computed
relative to baseline air quality levels.* The term "incremental net
* Recall that "baseline" air quality levels are intended to reflect those
air quality levels that would prevail in the absence of the implementa-
tion of any of the alternative PM NAAQS (see Section 1 for a more
detailed description).
2-2
-------
benefits" refers to the difference between incremental benefits and incre-
mental costs.
Cost-Effectiveness
An analysis of the cost-effectiveness of PH NAAQS is employed to
reduce the set of alternatives that are evaluated in terms of economic
efficiency. Inferior standards are those which require higher costs to
achieve the same or smaller benefits than dominant alternatives. An
inferior standard is not "cost-effective" since the same or higher incre-
mental benefits can be achieved by adopting a less costly standard. Those
alternative PH NAAQS identified as being inferior need not be further
evaluated in terms of economic efficiency.
Efficiency Criterion
The efficiency criterion is used to evaluate the economic desirability
of the reallocation of resources that occurs as the result of the adoption
of an alternative PM NAAQS. A given air quality standard is efficient in
the economic sense if, as a result of its implementation, at least one
individual's well-being is improved without reducing the well-being of any
other member of society. The allocation of resources associated with an
efficient standard is economically preferred to the allocation that exists
prior to its implementation.
It should be recognized, however, that those individuals enjoying the
benefits of improved air quality are not generally the same as those who
bear the cost of controlling pollution emissions. As a result, the
economic well-being of some members of society may be reduced and the
strict criterion stated above will be violated. This need not be the case,
however, if those individuals benefiting from improved air quality are
required to compensate those who bear the costs of pollution control.
Under such a compensation scheme, a given pollution standard would be
judged efficient if those individuals receiving benefits could potentially
compensate exactly those individuals bearing the costs of the standard, and
2-3
-------
still realize a net gain in economic well-being. As is typically the case
with applied benefit-cost analysis, this is the efficiency criterion that
is adopted in this analysis.
The exact compensation arrangement described above will not be imple-
mented. Tax credits on investments in pollution control equipment will
redistribute some of the costs of emissions control, but these and other
provisions will not completely resolve all equity issues. Consequently,
the distribution of estimated benefits and costs among various sectors of
the economy is described for the alternative PM NAAQS later in this section
of the report.
In addition to determining whether a given air quality standard is
efficient, it is also possible to rank the alternative PM NAAQS in terms of
relative efficiency. The PM NAAQS that is most efficient, relative to the
alternatives considered, is the one which provides the largest incremental
net benefits. An analysis of the relative efficiency of the alternative PM
NAAQS considered is described later in this section.
»
Scope of Analysis
Efficiency is only a necessary but not a sufficient condition for
establishing the economic desirability of an air quality standard. Since
there are generally both benefits and costs associated with achieving and
maintaining baseline air quality levels, it is possible that the total
costs could exceed the total benefits associated with a PM NAAQS, even if
incremental net benefits are positive.* This could occur if the cost of
baseline controls exceeds the benefits associated with baseline air quality
levels. If this were the case, society might be better off if no air
quality standards are adopted, including those already in place to achieve
baseline levels of air quality.
* Total benefits include the benefits associated with baseline controls as
well as the incremental benefits realized as a result of the adoption of
a PM NAAQS. Similarly, total costs include both baseline control costs
and incremental costs.
2-4
-------
A standard for which total benefits exceed total costs is termed
"feasible". As this definition indicates, estimates of both baseline and
incremental benefits and costs are required to assess the feasibility of
the alternative PM NAAQS. Estimates of the benefits associated with base-
line controls, however, are unavailable. The alternative PM NAAQS were
evaluated by a sufficient test for feasibility which compared total costs
with incremental benefits under the assumption that no benefits are
associated with baseline controls. These tests, however, often produced
inconclusive results in that the feasibility of many alternative standards
could not be determined. Consequently, the results of these tests are not
reported.*
Ideally, all feasible and cost-effective PM NAAQS should be evaluated
in order to identify the one which maximizes society's well-being as a
result of improvements in ambient air quality. Within the context of
applied benefit-cost analysis, however, it is usually possible to consider
only a limited set of discrete alternatives. Such is the case with this
analysis. Consequently, the focus is on the relative efficiency of a
limited selection of alternatives and not the identification of the most
efficient of all possible feasible and cost-effective PM NAAQS.
BENEFIT-COST ANALYSIS: METHODOLOGY
Appropriate methods of testing for both the cost-effectiveness and
efficiency conditions are described immediately below. As was stated
previously, the tests for cost-effectiveness and efficiency require
analyses of the incremental benefits and costs associated with each of the
alternative PM NAAQS. These benefit-cost analyses are limited in that they
are not employed to evaluate the distributional impacts of these air
quality standards; consequently, a brief discussion of this issue is also
provided.
* Estimates of baseline control costs, which are necessary to replicate the
sufficient tests for feasibility, are reported later in this section.
2-5
-------
An analysis of the incremental benefits and costs associated with the
alternative PM NAAQS can be conducted in the following manner:
Compute the estimated incremental benefits associated with
each alternative PM NAAQS.
Similarly, compute the estimated incremental costs
associated with each alternative standard.
Compute the estimated net incremental benefits (i.e.,
incremental benefits minus incremental costs) generated by
each alternative standard.
Compare the estimated net incremental benefits for each of
the several alternative standards.
Any alternative PH NAAQS that is associated with higher incremental costs,
but the same or smaller incremental benefits than some other standard, is
inferior or cost-ineffective and need not be further evaluated in terms of
efficiency. Any alternative PM NAAQS that produces positive net incre-
mental benefits will provide a more efficient allocation of resources than
what would occur under the baseline air quality scenario. The PM NAAQS
that produces the largest positive net benefits will produce the most
efficient allocation of resources among the standards considered. When the
net incremental benefits associated with a standard are negative, the
baseline air quality scenario yields a more efficient allocation of
resources.
nal Effects
There are two reasons why the distributional impacts associated with
the alternative PM NAAQS are an important issue. These reasons are:
The benefits and costs associated with the alternative
standards are not likely to be distributed evenly across
various sectors of the economy, thus raising equity issues.
2-6
-------
The distribution of adverse impacts may affect the measure-
ments of costs that are appropriate for use in the benefit-
cost analyses.
Both of these points are discussed below in greater detail.
It has already been noted that those individuals who enjoy the
benefits of improved air quality are not generally the same as those who
bear the cost of controlling emissions. Absent any arrangement whereby
beneficiaries compensate those incurring costs, it is likely that the
implementation of an air quality standard will increase the economic well-
being of some individuals and reduce the well-being of others, thus raising
an equity issue. The distribution of costs among those incurring the
expenses of emissions controls raises an additional potential equity issue.
These distributional or equity effects are not typically evaluated
within the framework of applied benefit-cost analysis. In order to do so,
it would be necessary to obtain estimates of the values that society places
on the distributions of economic well-being associated with each of the
alternative PM NAAQS. Estimates of these values are unavailable. As a
result, a separate analysis of these distributional impacts is provided.
The results of this analysis are summarized later in this section. It
should be stressed, however, that no judgments regarding the desirability
of these distributional effects are offered; instead, these impacts are
only described.
The potential distribution of adverse economic impacts associated with
the alternative PM NAAQS should also be considered in measuring appropriate
costs for use in the benefit-cost analysis. The cost estimates provided by
the emissions control phase of the study, for example, are based on the
assumption that no plants close or reduce production under the burden of
additional emissions control costs. Plant closures are likely to produce
an upward bias in these cost estimates. If a significant number of plants
would, in fact, cease operation when faced with additional control costs,
downward adjustments to these earlier cost estimates may be necessary
2-7
-------
before they can be used appropriately in the benefit-cost analyses.* This
issue is discussed in more detail later in this section.
MEASUREMENT OF BENEFITS AND COSTS
A clear understanding of several conceptual issues is necessary for a
proper interpretation of the estimated benefits and costs that are compared
in the benefit-cost analysis. These conceptual issues are discussed
immediately below. Following this, the estimates of both the incremental
benefits and incremental costs associated with each of the alternative PM
NAAQS are discussed. The scope of benefits and costs included in these
estimates are also described and the estimation techniques employed are
reviewed briefly. Finally, a discussion of the consistency between these
benefit and cost estimates is provided.
Measurement of Benefits and Costs: Conceptual Issues
The beneficial effects of improved ambient air quality can be measured
as the value that individuals place on the opportunity to consume cleaner
air. The conceptually correct valuation of this opportunity requires the
identification of individuals' willingness to pay for cleaner air (or to be
compensated for deterioration of air quality).** Where possible,
willingness to pay is the measure of benefits that is adopted in this
analysis. Estimates of society's willingness to pay for cleaner air do not
exist for all benefit categories, however. Alternative measures are used
for those benefit categories included in the analyses for which estimates
of willingness to pay are unavailable.
* Descriptions of several potentially adverse impacts associated with the
alternative PM NAAQS are described in a separate report by another
contractor (1).
** The appropriateness of willingness to pay versus willingness to be com-
pensated depends on the property right endowments of receptors. Because
of income constraint considerations, willingness to be compensated may
be greater than willingness to pay.
2-8
-------
Similarly, an appropriate measure of the cost of pollution emissions
control can be measured as the value that society places on those goods and
services not produced as a result of resources being diverted to the pro-
duction of improved air quality. Again, the conceptually correct valuation
of these costs requires the identification of society's willingness to pay
for these foregone consumption opportunities that would otherwise be
available.
As the preceding discussion suggests, only impacts that result from a
reallocation of economic resources are included in the benefit and cost
estimates. The implementation of a PH NAAQS, however, may be accompanied
by price adjustments and subsequent short-run income transfer effects. For
example, a sudden demand for emissions control devices resulting from a
more stringent air quality standard may initially result in higher prices
for these devices. Increased income in the form of excess profits will
then accrue to the manufacturers of these devices; however, this favorable
effect occurs at the expense of those who must pay the higher prices. The
total effect is merely pecuniary.* These pecuniary effects are not and
should not be included in the benefit estimates that follow.
It is also possible that air quality standards may generate indirect
economic impacts. One example of an indirect effect is the transfer of
income to owners of recreational facilities as individuals whose health is
improved as a direct result of better air quality increase their use of the
recreational facilities. Indirect costs and benefits which represent real
(as opposed to pecuniary) effects should be included in benefit and cost
estimates. As a practical matter, however, it is often very difficult to
measure all indirect effects and to determine whether they are real or
pecuniary in nature. Consequently, the benefits and costs considered in
this analysis include only the direct effects of the alternative PM NAAQS.
The exclusion of possible indirect effects may mean that these estimated
* "Income transfers" refer to redistributions of income that do not result
directly from production of goods or services. These transfers are
referred to as "pecuniary".
2-9
-------
benefits and costs understate the total effects of the alternative air
quality standards.
Estimates of Benefits
The benefit-cost analyses described later in this section represent
relatively straightforward comparisons of the estimated benefits and costs
associated with alternative PM NAAQS. Because of this, the uncertainty
that surrounds both estimated benefits and costs may be obscurred. None-
theless, the validity of the benefit-cost analyses depends critically on
the accuracy of estimated benefits and costs. Any uncertainty embedded in
these estimates will carry over to the benefit-cost analyses.
The sources of uncertainty in the benefits estimates have already been
described in detail earlier in Section 1. These sources of uncertainty
fall into the following three general categories:
Scope and magnitude of air quality improvements.
Valuation of health and welfare improvements.
Evidence of health and welfare effects.
Coverage of effects/benefit categories.
Some of the degree of uncertainty can be estimated quantitatively. Ranges
of estimated benefits have already been reported in Section 1. Benefit-
cost analyses, however, are conducted only for the point estimates of
benefits. Nonetheless, the results of these analyses should be interpreted
in view of the uncertainty that does exist in the benefit estimates. Each
of the major sources of uncertainty listed above is reviewed briefly below
for the convenience of the reader.
Scope and Magnitude of Air Quality Improvements
Benefits are estimated based on projected improvements in air quality
associated with the alternative PM NAAQS. The projections of air quality
2-10
-------
improvements are also subject to uncertainty. This, in turn, creates an
additional source of uncertainty to the benefit estimates.
The uncertainty surrounding the estimates of air quality improvements
has already been described earlier in Section 1. The major sources of
uncertainty include the following:
Use of county-wide rollback to predict air quality changes
at the worst-case monitor.
Use of proportionality to predict air quality changes
elsewhere.
The failure to account for potential air quality improve-
ments resulting from emissions controls in adjacent
counties.
Approximations required in converting among different PM
measures (TSP, PM10, BS).
Use of 1977-78 as a base year for predicting air quality.
The implications of these uncertainties associated with estimates of air
quality improvements have already been discussed in detail in Section 1.
The results of plausibility checks on some of the assumptions embedded in
the air quality estimates have also been reported there.
Valuation of Health and Welfare Improvements
Two general categories of benefits attributable to improvements in
ambient air quality are included in the benefit-cost analyses.* These
benefit categories are:
* Studies based on property value models and wage models were also included
in the benefit analysis as plausibility checks. However, since it is not
possible to disentangle the health and soiling benefits from other
benefits estimated by these models, they are not included in the esti-
mates reported in this section in order to avoid possible double-
counting. The property value and wage models are described in Sections 5
and 6 of this report.
2-11
-------
Health effects
Soiling and material damage.
Estimates of benefits derived from these two categories were developed from
previously conducted studies which examine the health and welfare effects
of ambient particulate matter. These studies were selected after a
thorough review of the literature and a careful evaluation of their suita-
bility for use in estimating the benefits of the alternative PM NAAQS.
Health effects are further classified in terms of the risk of
mortality, acute morbidity, and chronic morbidity. Measures of increased
risk of mortality and morbidity as a result of physical effects caused by
deteriorations in air quality are developed. The value placed on reduc-
tions in mortality rates attributable to alternative PM NAAQS reflects
individuals' willingness to pay to reduce the risk of death and thus con-
stitutes a theoretically correct measure of benefits. Nonetheless, there
are large variations in the estimates of the value assigned to mortality
risk reductions.
Benefits associated with reduced morbidity are based on: reductions
in medical expenditures, the value of improved labor productivity due to
reduced work days lost, and the value placed on increased opportunities to
engage in non-labor activities.* These benefit estimates are likely to
underestimate willingness to pay for improved air quality, since they do
not include the value that individuals place on reductions in residual pain
and suffering.**
* In addition to lost time in the workplace, individuals may be forced to
curtail non-labor activities as a result of morbidity caused by pollu-
tion emissions. The lost opportunity to engage in these activities are
estimated at one-half of the wage-rate of affected individuals.
** Some medical expenditures are made to reduce pain and suffering induced
by pollution emissions. However, prior to receipt of medical care
services, and in some cases during and after provision of such services,
pain and suffering occur. The reduction in this "residual pain and
suffering" is not included in the benefit estimate.
2-12
-------
Benefits attributable to reduced soiling and materials damage in both
the household and industrial sectors rely primarily on economic models
which identify associations between air quality and goods exchanged in
private markets. By observing how consumption expenditure patterns of
typical households vary in response to different levels of pollution
emissions, it is possible to estimate the value these households place on
various improvements in ambient air quality. In the industrial sector,
benefits are estimated for two 2-digit SICs by employing econometric
techniques which relate soiling or contamination to production costs. In
general, these benefits reflect willingness to pay for improved air
quality.
Evidence of Health and Welfare Effects
Additional uncertainty is introduced to the benefit estimates because
of uncertain evidence of the magnitudes of health and welfare effects
associated with cleaner air. These sources of uncertainty include:
Conflicting evidence across studies.
Incomplete control for confounding factors.
Use of study results from one time and area to estimate
effects in another time and area.
No information on the size distribution of particles
present during the studies.
Alternative functional forms possible.
Sampling variation in coefficient estimates.
Degree of risk at lower-level PM concentrations.
The degree of uncertainty introduced by some of the above sources can be
estimated quantitatively. In these cases, the degree of uncertainty is
reflected in the ranges of the incremental benefit estimates reported in
Section 1.
2-13
-------
Coverage of Effects/Benefit Categories
The benefit estimates used in this analysis do not include a number of
areas where benefits may arise. Limitations in the definition of certain
data transformations, control strategy designs, and methodological issues
precluded the inclusion of potential benefits attributable to improved
visibility, possible climatic effects, and reductions in acidic deposi-
tions. In addition to these omitted benefit categories, potential benefits
accruing to some sectors of the economy are also excluded from estimates
used in the benefit-cost analysis. These omissions include potential
benefits due to reduced soiling and materials damages to commercial,
governmental, and institutional structures as well as some industrial
structures. Furthermore, the estimates that are included focus exclusively
on user value benefits and exclude potential non-user benefit categories
such as existence, bequest, and option values.* The omission of these
plausible benefit categories may cause the estimates used in this analysis
to understate the true level of benefits associated with a particular PM
NAAQS. **
of Costs
The costs of achieving and maintaining air quality levels associated
with the alternative PM NAAQS were estimated by another contractor (2). As
is the case with the benefit estimates, emissions control costs are not
estimated with certainty. Consequently, a proper interpretation of the
benefit-cost analyses requires an understanding of how emissions control
* See Sections 1 and 8 for further discussion of non-user benefit
categories.
** Benefits may also be understated even in those categories where benefits
are measured if economic decision-makers have made prior behavioral
adjustments in response to air quality changes. For example, observed
pollution-induced morbidity may be reduced if individuals curtail out-
door activities because of perceived health risks associated with poor
air quality. There is, however, a cost associated with this behavioral
adjustment namely, the cost of the opportunity to engage in the fore-
gone outdoor activities.
2-14
-------
costs are estimated so that uncertainties and potential biases introduced
by methodological compromises and data quality can be assessed.
Control Cost Estimation Procedure
A fully detailed description of the procedures employed to estimate
emissions control costs is beyond the scope of this report. The abbre-
viated description offered below is intended to serve as a basis for a
later discussion of uncertainty and potential biases surrounding the cost
estimates.
The approach employed to estimate emissions control costs can be
viewed as a six-step procedure as illustrated in Figure 2-1. First,
estimates of both emissions (ISP and PM10) and air quality levels nominally
representative of 1978 conditions were developed, by county areas (and
subcounty areas in selected cases). Estimates of TSP emissions were based
on data from EPA's National Emissions Data Systems (NEDS); PM10 emissions
were estimated from the TSP data by using a conversion process. TSP'air
quality data were obtained from EPA's Storage and Retrieval of Aerometric
Data System (SARDAD). Since no PH10 data were available, PM10 air quality
levels were derived from TSP data by using a 0.55 conversion factor.
In Step 2, a linear rollback approach was employed to project future
air quality levels. Future emissions rates were projected by accounting
for both the growth of new sources and the replacement of existing sources.
Future air quality levels were then projected based on the presumption that
all sources within an area contribute to air quality levels directly pro-
portional to their emissions rates. Based on these estimates of future air
quality levels, areas (counties or subcounty areas) projected to be in
nonattainment with the alternative PM NAAQS were identified.
An inventory of control options and associated emissions was then
developed for sources within areas projected to be in nonattainment with
the alternative air quality standards. Only a single control strategy
could be identified for many sources, however, and no control strategies
2-15
-------
STEP 1
Develop estimates of emissions and air
quality levels nominally representative
of 1978 conditions
STEP 2
STEP 3
Project future emissions growth to
predict areas in nonattainment
i
Develop inventory of control options for
emissions sources in nonattainment areas
STEP 4
Establish relationship between emissions
and air quality levels in nonattainment areas
STEP 5
Select control options for each area based on
"least-cost" approximation to estimate
emissions control costs (some areas in
residual nonattainment)
STEP 6
Estimate additional costs of achieving
full attainment
Figure 2-1. Basic Steps in Estimating Emissions Control Costs
2-16
-------
could be identified for some other sources. In addition, control options
were eliminated if the cost of emissions reductions was extraordinarily
high based on current experience or if they cost more but reduced emissions
less than another option. As a result of these combined factors, more than
SO percent of all sources were not considered as candidates for control in
some scenarios analyzed.
Relationships between source emissions rates and area air quality
levels were established next in Step 4. A modified linear rollback
technique was employed. That is, it was assumed that all sources within an
area contribute to air quality directly proportional to their emissions
rates and inversely proportional to their effective stack height.
The cost of emissions control was then estimated (Step 5) by selecting
control strategies that minimized costs (in terms of dollars per microgram
of air quality improvement) for each area in nonattainment. Additional
control options were applied incrementally until attainment with a given
air quality standard was achieved or available control options were
exhausted. This resulted in some areas being in "residual" nonattainment.
It should be noted that when the "least-cost" control strategies were
selected in Step 5, it was assumed that controls put in place at the
beginning of the implementation period were sufficient to reduce emissions
to required levels for the entire implementation period. This typically
results in greater than necessary controls of emissions during the early
phases of the implementation period since baseline ambient PM concentra-
tions typically increase over time because of growth in the number of
emissions sources.
In Step 6, the cost of eliminating residual nonattainment was
estimated based on the cost of achieving partial attainment (Step 5) and
the degree of residual nonattainment. In particular, the additional cost
was estimated by multiplying the national cost of partial attainment by the
national average ratio of the additional air quality improvement needed in
each residual nonattainment county to the air quality improvement already
2-17
-------
achieved in each such county. This additional cost was then added to the
cost of partial attainment in order to estimate the full cost of bringing
all areas into full attainment.
As was noted earlier, uncertainty and potential biases are introduced
to the control cost estimates from both methodological compromises and data
quality. In addition, there is a third issue the difference between
emissions control costs and economic costs that should be addressed in
assessing the appropriateness of the cost estimates for use in the benefit-
cost analyses. These issues are discussed briefly below. Where possible,
the directions of potential biases are examined.
Methodological Compromises
Because of the scope of the nationwide emissions control cost study,
and time phasing and resource considerations, it was necessary to make
several methodological compromises. Those compromises that introduce
uncertainty and potential biases to the cost estimates include:
Use of design monitors.
Use of the modified linear rollback procedure.
Consideration of a limited set of control options.
The use of a national average measure of residual non-
attainment in estimating the cost of eliminating residual
nonattainment.
Placing some areas under controls greater than necessary to
achieve attainment.
Each of these potential problems is discussed immediately below.
Design Monitors In order to simplify the analysis, the design value
was obtained as the reading of the monitor showing the highest PM
2-18
-------
concentration level in an area.* This may be a poor measure of actual air
quality on an area-vide basis since the highest monitor naturally reflects
PM concentration levels that are higher than at other points within the
area. Other things being the same, the use of the highest reading as the
design value will tend to cause an upward bias in the cost estimates since
the overall level of required control within an area will be overstated.
The magnitude of this bias, however, is not predictable in a straight-
forward fashion, given the use of the modified linear rollback procedure.
This problem is discussed immediately below.
Modified RollbackThe modified linear rollback assumption was
employed as an alternative to detailed dispersion modeling. As was
explained earlier, this procedure assumes that all sources in an area con-
tribute to the PM concentration levels recorded at the design monitor
directly proportional to their emissions rates and inversely proportional
to effective stack heights. Unfortunately, the direction of the bias
introduced to the control cost estimates because of this factor cannot be
predicted .a priori.
All emissions sources within an area may not contribute to concentra-
tion levels at the design monitor. Consequently, it may not be necessary
to control all sources; as a result, estimated control costs may be over-
stated. It may be the case, however, that sources proximate to the design
monitor with high control costs contribute heavily to the reading.
Estimated control costs could be understated if these sources were not
controlled (or not controlled enough) under the optimal control strategy.
It is also noted that the rollback procedure employed to project
future ambient air quality and the approach used to estimate emissions
control costs are inconsistent. In establishing the relationship between
* It should be noted that this is not necessarily indicative of the highest
concentration in an area. Limitations on the number of monitors and
their placement (e.g., highly elevated to avoid vandalism) can mean that
the design value is not indicative of the highest concentration. How-
ever, such values are presumed to be good indicators of higher concentra-
tions.
2-19
-------
emissions control strategies and improvements in air quality, it was
assumed that sources contribute to air quality inversely proportional to
their effective stack heights. Stack heights were not considered, however,
in projecting future air quality levels from the base year. The direction
of the bias to estimated control costs because of this inconsistency is not
known.
Li«ited Set of Control Options Two or three emissions control
options could be identified for some sources (e.g., boilers). For many
emissions sources, however, only one or no emissions control options could
be identified. If, indeed, other control options less costly than those
selected by the "least-cost" algorithm actually exist, estimated control
costs will be overstated.
Residual Nonattainaent After all "cost-effective" control strate-
gies were adopted, some counties were still found to be in residual non-
attainment with the alternative PH NAAQS. In these cases, the additional
cost of achieving full attainment was estimated based on the cost of
achieving partial attainment and the degree of residual nonattainment. The
estimate was developed by multiplying the national cost of achieving
partial attainment by the national average ratio of the additional air
quality improvements needed in each residual nonattainment county to the
air quality improvement already achieved in each such county. This addi-
tional cost was then added to the cost of partial attainment in order to
estimate the full national cost of bringing all areas into full attainment.
Since this additional cost is a national estimate, no regional breakdowns
of control costs were available for the type B scenarios (complete attain-
ment scenarios).
An alternative procedure for estimating the additional cost would have
been to assume that the additional air quality improvement needed in each
county would have a cost per microgram equal to the average or marginal
2-20
-------
cost* per microgram for the air quality improvement already achieved in
that county. This approach would have provided a more accurate estimate
and allowed a regional breakdown of costs. It was not undertaken due to
time constraints. The relationship between the two approaches will depend
on whether counties with high partial attainment costs are in about the
same degree of residual nonattainment as counties with low partial attain-
ment costs. If high cost counties generally experience more residual
nonattainment, then the procedure actually used may lead to a lower
estimate compared to using average cost. If high cost counties generally
experience less residual nonattainment, then the procedure used may lead to
an estimate above what would have been predicted by using average cost.
The estimates of the additional cost of achieving full attainment are thus
highly uncertain.
Control Strategies Below the Standard Emissions were projected to
1995 and the resulting air quality level was estimated accordingly.**
Then, controls necessary to achieve required air quality levels for the
entire implementation period (to 1995) were put in place in the initial
implementation year (1987 or 1989). In many cases, this resulted in more
control than necessary to achieve air. quality levels sufficient to meet the
alternative standards. This approach is likely to produce an upward bias
in the cost estimates used in the benefit-cost analyses. In fact, the
source of the bias is twofold: 1) capital equipment might be installed
earlier than necessary (and depreciated more quickly than need be)
resulting in an increase in the present value of estimated costs; 2) some
unnecessary operating and maintenance costs might be incurred in those
* Strictly speaking, marginal cost is the conceptually correct measure to
use in estimating the cost of eliminating residual nonattainment. Since
marginal cost usually exceeds average cost, use of the latter measure
would have biased estimated costs downward. It is noted, however, that
use of average cost might be justified based on reasoned judgment:
first, one reason that cost-effective control options were exhausted is
that many sources were not considered as candidates for control; second,
it is possible that new technologies may produce less costly control
options in the future.
** See Section 9 for a more detailed discussion.
2-21
-------
years during which the controls resulted in air quality levels lower than
those required by the alternative PM NAAQS.
Data Availability and Quality
A number of problems involving both the availability and quality of
data surfaced during attempts to estimate the emissions control costs
associated with alternative PM NAAQS. These data problems include the
following:
The NEDS emissions data base may not reflect actual 1978
conditions.
The point source data employed is both incomplete and
occasionally inaccurate.
Secondarily formed aerosols are not accounted for in the
emissions control analysis.
Air quality data is incomplete.
Each of these data problems is discussed briefly below.
The 1978 NEDS Data Base May Not Represent Actual Conditions The
NEDS 1978 data base is not completely updated annually. Because of this,
the data may reflect both emissions rates and control strategies that were
in place in some earlier year. This particular data problem might produce
either a downward or an upward bias in the cost estimates used in the
benefit-cost analyses. Estimated control costs could be biased downward if
the least-cost control algorithm selected low-cost sources for control that
actually had substantial controls already in place instead of higher-cost
sources which should have been controlled. Alternatively, control costs
could be overestimated if the 1978 NEDS data base overstates emissions
rates and the overall degree of control necessary to satisfy the require-
ments of the alternative PM NAAQS.*
* This latter problem would also cause estimated benefits to be overstated
since necessary air quality improvements would likewise be overstated.
2-22
-------
Incomplete and Inaccurate Point Source Data Of a total of 5.72
million tons per year (TPY) of TSP emissions included in the analysis for
purposes of projecting future air quality, only 3.55 million TPY could be
accounted for in the point source emissions data base. This discrepancy
was caused by both incomplete and/or inaccurate point source data. It is
important to note that this particular data problem is likely to cause
pollution abatement costs to be overestimated, and the magnitude of the
bias is potentially large. This overestimate of emissions control costs is
caused by the fact that roughly 40 percent of all source candidates for
emissions control were omitted from the feasible solution to the least-cost
algorithm. If any of these omitted point sources exhibit marginal control
costs less than those actually selected for control, estimated control
costs are biased upward.
SecondarilyFor »ed Aerosols A-re Not AccoTiTited for in the A^al
The emissions control cost analysis does not account for secondarily-formed
aerosols which can account for a large percent of total ambient PM10 in
some areas. The effect of not including these secondarily-formed emissions
is to overstate the degree of control possible for ambient PM10 through
traditional strategies, thus causing an underestimation of pollution
control costs.
Air Quality Data As was noted previously, PM10 air quality data are
not currently available. The PM10 air quality data used in the control
cost analysis was obtained by applying a 0.55 conversion factor to TSP air
quality measures in the base year. This conversion factor was estimated
through an analysis of data available from EPA's inhalable particulate
(PH15) network. It should be stressed, however, that the conversion factor
is likely to vary across geographic areas. Although the direction of
potential bias due to the use of the conversion factor cannot be predicted,
its use does introduce uncertainty to the cost estimates.
In addition, valid air quality data for many counties were unavail-
able. If valid design values were unavailable for counties designated as
being in attainment or unclassif iable, they were eliminated from the cost
2-23
-------
analysis. If counties were classified as being in nonattainment, default
design values were set at one |i>g/m above the standard violated. In other
cases, valid air quality data for one of the averaging periods were
unavailable and default design values were estimated through regression
analysis which related one averaging period to another. Again, these data
deficiencies introduced additional uncertainty to the cost estimates.
In some areas, valid air quality data were unavailable for 1978, but
observations were available for earlier years (e.g., 1975-1977). The
earlier observations were substituted as 1978 values in these cases. It is
noted that this approach potentially causes a mismatch between emissions
sources and air quality. The direction of potential bias that this factor
introduces to the cost estimates, however, cannot be predicted ji priori.
Emissions Control Costs and Economic Costs
The estimated emissions control costs should be viewed as approxima-
tions of the conceptually correct measure of costs. Recall that the
conceptually correct measure .of costs is reflected by society's willingness
to pay for the goods and services not produced as a result of economic
resources being diverted to the production of improved air quality. The.
estimated pollution control costs, however, are based on engineering
estimates of the costs of equipment and labor required to control emissions
to permissible levels.
These estimates will correspond exactly to the conceptually correct
measure of costs only if two conditions are satisfied. First, affected
polluters must be able to pass forward all emissions control costs to
consumers through price mark-ups without reducing the market quantity
demanded of goods or services they sell. Otherwise, the estimated direct
pollution control costs will overstate the true economic costs associated
with the alternative PM NAAQS. Second, the prices of pollution control
resources (e.g., pollution control equipment and labor) used to estimate
costs must correspond to the prices that would prevail if these factors
were sold in perfectly competitive markets. For example, if excess profits
2-24
-------
are earned on pollution control equipment, the estimated emissions control
costs will overstate the conceptually correct measure of costs.
The estimated pollution control costs reported in this section do not
account for possible long-run adjustments to the PM NAAQS. Based on the
findings of the Economic Impact Analysis (3). for example, it is antici-
pated that some plants will close under the burden of additional emissions
control costs. Moreover, it is expected that some of the remaining plants
will expand production to compensate at least partially for output lost due
to the closures. The failure to account for these adjustments is likely to
produce a net upward bias in the cost estimates.*
In addition, the cost estimates reported here do not include possible
transition costs that some industries may experience as adjustments to the
alternative PM NAAQS occur. For example, some temporary unemployment may
occur as some affected plants reduce production or close. This effect,
however, is expected to be mitigated at least partially by increased
employment in other plants. The results of the Economic Impact Analysis,
however, suggest that the net impact of these transition costs are not
likely to affect significantly the cost estimates used here in the benefit-
cost analyses.
Administrative, Monitoring, and Enforcement Costs
The administrative, monitoring, and enforcement (AME) activities that
accompany an air quality standard are not fully reflected in the emissions
control cost estimates provided by the cost contractor. These AME costs
Estimated costs should be reduced because of the closures since these
plants will not implement pollution control strategies. At the same
time, costs for the remaining plants should be adjusted upward as they
expand production. The net effect, however, is likely to reduce costs,
since the additional cost of expanding production for the remaining
plants presumably will be less than the extra costs that would be
incurred by plants which close. Otherwise, these latter plants would
remain competitive instead of closing down operations. See Reference (1)
for further discussion of this and other related issues.
2-25
-------
may be borne both by directly affected firms and institutions and different
levels of government. It is the latter AME costs that are not reflected in
the emissions control cost estimates.* The proper measure of these AME
costs should reflect the incremental costs of implementing and maintaining
a given alternative PM NAAQS.
Estimates of the long-run annualized AME costs for state and local
agencies, as well as for EPA Regional offices, indicate that the government
sector resource requirements should not exceed $50 million annually (1981
dollars).** Relative to the magnitudes reported for the benefits and costs
of control, these additional expected AME expenditures are small. As a
consequence, these costs are not likely to affect appreciably the outcome
of the benefit-cost analyses. Nevertheless, if a complete measure of costs
is to be developed, an adjustment to the cost of control estimates is
required. Based on the annualized AME costs reported above and on an
examination of historical data, an adjustment factor of 4 percent of the
discounted present value of engineering and operation costs appears
reasonable. Thus, final estimates of control costs used in the benefit-
cost analyses are increased by 4 percent over the discounted present values
of the estimated costs obtained from the cost contractor.
Control Cost Truncation.
The costs provided by the cost contractor are based on a 15-year
engineering life for the emissions control equipment beginning at the 1987
(TSP) or 1989 (PM10) implementation dates. In this section, these costs
are adjusted to reflect a shorter time horizon (i.e., to 1995) by aggre-
gating the annualized stream of costs from the implementation date only
through 1995 and then applying the appropriate discounting formula. The
* Some institutions affected by the alternative PM NAAQS are government
agencies (e.g., utilities owned by local governments and government-
maintained roads). Those AME costs that will be incurred internally by
these directly affected government institutions have already been
included in the cost estimates provided by the cost contractor.
** See Reference (3).
2-26
-------
benefit-cost analysis is truncated at 1995 because of the uncertainty in
projections of emissions levels beyond this date.
Costs Associated With Alternative PM NAAQS
Estimated costs for each alternative PM NAAQS for both the A and B
scenarios are summarized in Table 2-1.* As is the case with the benefit
estimates, these costs are "incremental" in the sense that they are calcu-
lated relative to the baseline controls. Although direct pollution control
Table 2-1
SUMMARY OF INCREMENTAL COST ESTIMATES
Standard
PM10 70 AAM/250 24-hr. , 7 yr.
PM10 55 AAM 7 yr.
PM10 55 AAM/250 24-hr. , 7 yr.
PM10 55 AAM/200 24-hr., 7 yr.
PM10 55 AAM/150 24-hr. , 7 yr.
PM10 48 AAM/183 24-hr., 7 yr.
TSP 75 AGM/260 24-hr. , 7 yr.
TSP 150 24-hr., 7 yr.
TSP 75 AGM/260 24-hr., 9 yr.
TSP 150 24-hr., 9 yr.
Point Estimate of Incremental Costs*
A Scenario
0.50
0.90
0.93
0.95
1.26
1.32
1.08
1.78
1.61
2.63
B Scenario
0.95
2.52
2.57
2.54
3.53
3.36
2.61
5.96
4.02
8.95
* Expressed as the 1982 discounted present value in billions of 1980
dollars. Figures include estimated AME costs.
Source: Reference (2).
* Recall that some areas remain in residual nonattainment under the A
scenario while all areas are in full attainment with the alternative
standards under the B scenario.
2-27
-------
costs do exhibit some sensitivity to various assumptions and conditions
which underlie their derivation, no attempt is made to develop a range of
incremental cost estimates which reflect this uncertainty.*
It is noted that the incremental costs associated with the most
stringent of the PM10 standards are greater than those associated with the
TSP 75 AGM/260 standard, even though the latter is listed as being more
stringent. Two explanations of this phenomenon are:
The counties projected to be in nonattainment with the PM10
standards sometimes differ from those expected to violate
the TSP standard.
PM10 and TSP control options differ because of different
emissions sources and different particle sizes.
The figures displayed in Table 2-1 represent estimates of the
incremental or additional costs of achieving and maintaining the air
quality levels associated with the alternative PH NAAQS. These estimates
do not include the costs of achieving and maintaining baseline air quality
levels. Estimates of baseline control costs for the 1987-1995
implementation period range from $31.4 billion to $40.4 billion annually,
expressed as the present discounted value of 1980 dollars in 1982.
Comparable cost estimates over the 1989-1995 implementation period range
from $21.8 billion to $28.0 billion.** The total costs associated with the
alternative PM NAAQS are estimated as the sum of appropriate estimated
baseline costs and the estimated incremental costs reported in Table 2.1.+
* The results of several sensitivity analyses are documented in Reference
(2).
** See Reference (4).
+ Sufficient tests for the feasibility have been briefly described earlier
in this section. These tests can be replicated by subtracting the
estimated total costs associated with each alternative standard from the
corresponding estimated incremental benefits reported in Tables 1-11
through 1-14 in Section 1. The midpoints of the ranges were used as
estimates of baseline control costs (i.e., $35.9 billion and $24.9
billion, respectively, for the 7- and 9-year implementation periods).
2-28
-------
Consistency of Benefits ««» Costs Estimates
Although the incremental benefit and cost estimates used in this
benefit-cost analysis are estimated using different methodologies, they are
consistent and comparable in the following respects:
Each is based on the same air quality information.
Each reflects the same time horizons.*
Each uses similar assumptions to project industry and popu-
lation growth from the implementation date through 1995.
Each uses a discount rate of 10 percent to obtain a present
discounted value and an estimate of annualized benefits and
costs.
Each are valued as 1980 dollars in 1982.
The air quality data used in the benefit analysis were generated as
part of the procedure employed to estimate the costs associated with each
standard. Thus, estimates of both benefits and costs are based on
identical assumptions regarding air quality levels.
Both benefits and costs accrue over time. When appropriate, similar
growth factors are projected to both benefits and costs in order to ensure
consistency between the two time streams.** For convenience, they are
expressed in terms of their present discounted values in 1980 dollars in
1982.
For each alternative PM standard under consideration, benefits and
costs are measured from the implementation date through 1995. It was
* As noted earlier, an adjustment factor is applied to the incremental
cost estimates in order to make them consistent with the timehorizon
used in the benefit analysis.
** Both analyses used projections of future industry and population growth.
Although the projections were obtained from different sources, they were
very similar in magnitude. For a comparative analysis, see Reference
(5).
2-29 -
-------
judged impractical to extend the benefit-cost analysis past this date,
given the modeling uncertainties which arise when growth factors are
incorporated into the analyses. Fortunately, the process of discounting
causes both benefits and costs which may accrue in the distant future to be
small relative to the present. Thus, the truncation of benefit and cost
estimates in 1995 may not greatly affect the outcome of the benefit-cost
analyses.
A constant real discount rate of 10 percent is used to discount the
time streams of benefits and costs back to 1982. This real discount rate
is prescribed by the Office of Management and Budget. Much debate
surrounds the proper choice of the discount rate, and there is no clear
concensus on this issue.* It is likely, however, that a real discount rate
of 10 percent is too high.
It is also important to recognize, however, that benefits and costs
compared in the benefit-cost analyses are inconsistent in some respects.
These inconsistencies include the following:
The costs of controlling emissions in nonattainment areas
are likely to improve air quality in adjacent areas; how-
ever, these air quality improvements are not estimated, and
thus the resulting benefits are omitted from the analysis.
Because of the limited scope of the benefit analysis, all
possible direct benefits (e.g., visibility benefits,
reduced soiling, and materials damages to the commercial
sector and some manufacturing industries) associated with
emissions controls have not been estimated, while, in
principle, all potentially significant direct costs
associated with the alternative standards are included in
the estimates.
Distinctions between particle sizes are not made in the
benefit analysis. The emissions control options selected
in the cost analysis, however, do depend on particle size
differences.
* For a discussion of this issue, see R. C. Lind et al. (6).
2-30
-------
The first two of the inconsistencies listed above tend to cause an over-
statement of the incremental costs relative to the incremental benefits
associated with the alternative PM NAAQS. The implications of the third
inconsistency, however, are less straightforward and deserve additional
comment.
Host of the studies used to estimate health effects are based on TSP
measures of air quality. As a result, the benefit analysis cannot distin-
guish between effects caused by different particle sizes. The benefits
associated with the alternative PH10 standards are estimated based on the
estimated TSP concentration levels associated with PH10 controls. Evidence
cited in the EPA/OAQPS Staff Paper, however, suggests that the PM10
fraction of TSP is primarily responsible for adverse health effects.
It is expected that PM10 controls will reduce both TSP and PM10 con-
centration levels. If PM10 controls reduce the PM10 fraction of TSP, the
estimated benefits associated with the alternative PM10 standards might be
understated relative to those associated with the TSP alternatives. An
analysis was conducted of the predicted air quality associated with several
standards. This analysis suggests that the average fraction of PM10 in TSP
will remain virtually unchanged after controls are implemented, both for
PM10 and TSP control strategies.* This result, however, should be inter-
preted in view of the limitations of the methods and data employed to
select emissions control options and to project future air quality levels.
Later in this section, comparisons between the alternative PM10 and
TSP standards are offered. These comparisons are based on approximate
differences in the stringency of standards (in terms of TSP) and not
differences in particle sizes. It is also noted that the estimated incre-
mental net benefits associated with PM10 standards might be understated
relative to those for the TSP standards, if PM10 control options are more
effective than TSP control options in reducing the PM10 fraction of TSP.
* See Section 1 for a more detailed discussion.
2-31
-------
LIMITATIONS AND QUALIFICATIONS OF BENEFIT-COST ANALYSIS
The preceding discussion in this section provides the necessary
background for an interpretation of the benefit-cost analyses that follow.
Before the results of these analyses are reported, however, it is appro-
priate to review some of the limitations of this approach as a means of
evaluating the relative desirability of the alternative PM NAAQS.
Many of the shortcomings of benefit-cost analysis have already been
described in the preceding text. A more complete list is as follows:
The distributive or equity impacts of the alternative PM
NAAQS are not evaluated within the framework of the
benefit-cost analyses.
The benefit-cost analyses rely on the validity and scope of
estimates of benefits and costs.
The benefit-cost analyses are limited in scope to an
evaluation of a limited set of ambient air quality
standards.
The feasibility of the alternative PM NAAQS has not been
determined.
Each of these points is discussed briefly below.
It has already been noted that the air quality standards will have
different impacts on members of society; those who enjoy the benefits of
the standards will not always be the same as those who bear the costs. The
benefit-cost analyses conducted in this study do not evaluate these
distributive or equity impacts. These impacts are described through a
separate analysis. However, no judgment on the relative desirability of
these distributive impacts is offered.
The results of the benefit-cost analyses depend, of course, on the
validity and scope of ,the estimates of both benefits and costs associated
with the alternative ambient air quality standards. The wide range of
variability in the estimated benefits associated with each air quality
2-32
-------
standard have already been described. Similarly, all costs associated with
the alternative PM NAAQS cannot be estimated with certainty. In some
cases, the definitions of both benefits and costs employed in the estimates
do not correspond exactly to the conceptually correct definitions.
Finally, the benefit and cost estimates are limited in scope in that they
do not inclnde all possible benefits and costs (e.g., indirect costs) that
may result from the alternative PH NAAQS.
The benefit-cost analyses described in this section are limited in
scope in that they consider only a limited selection of possible PM NAAQS.
As a result, the benefit-cost analysis will identify, with uncertainty,
only the most efficient PM NAAQS among those considered. The identifica-
tion of the most efficient possible standard requires an evalnation of all
possible PM NAAQS.
Even if a standard is efficient, it might not be associated with an
economically desirable allocation of resources if total costs exceed total
benefits. The feasibility of the alternative PM NAAQS has not been compre-
hensively evaluated because estimates of the benefits associated with
baseline controls are unavailable. Sufficient tests for feasibility, which
assume that no benefits are associated with baseline emissions controls,
yield generally inconclusive results.
The results of the benefit-cost analyses that follows should be inter-
preted in light of these limitations. Specifically, these analyses provide
a qualified assessment of the cost-effectiveness and relative economic
efficiency of the alternative PM NAAQS considered.
BENEFIT-COST ANALYSIS OF ALTERNATIVE PM NAAQS
The results of the benefit-cost analyses of the alternative PM NAAQS
are reported below in this subsection. The cost-effectiveness and relative
efficiency of each standard is determined through an analysis of the incre-
mental benefits and incremental costs associated with the alternative PM
2-33
-------
NAAQS. Finally, the distributive or equity impacts of these air quality
standards are described.
Benefit-Cost Analysis: Incremental Net Benefits
The results of the analyses of the cost-effectiveness and relative
economic efficiency of the alternative PM NAAQS are described below. Both
analyses are conducted under each of the six benefit aggregation proce-
dures. The six aggregation procedures and the studies included in them are
described in Section 1. Briefly, these aggregation procedures are:
Procedure A. Only quantitative studies approved by CASAC
(Mazumdar et al. and Ferris et al.) are used to estimate
benefits under this aggregation procedure. Estimates of
nonhealth benefits are not included. Double counting of
benefits is unlikely.
Procedure B. This aggregation procedure includes estimates
of benefits due to reduced acute morbidity (Samet et al.)
in addition to benefit estimates included in Procedure A.
Consequently, estimates of nonhealth benefits are not
included and double counting of benefits is unlikely.
Procedure C. The Samet et al. study is replaced by the
Ostro study for use in estimating benefits attributable to
reduced acute morbidity. This aggregation procedure also
includes benefits estimated by the Mathtech study of house-
hold soiling effects. Coverage of all benefit categories
is still incomplete. Double counting of benefits is
unlikely.
Procedure D. Studies by Lave and Seskin and by Lipfert are
used to estimate benefits associated with reduced mortality
instead of the Hazumdar et al. study. Otherwise, this pro-
cedure is identical to Procedure C.
Procedure E. The Ferris et al. study is replaced by the
Crocker et al. study for use in estimating benefits in the
chronic morbidity category. Expanded coverage is given to
household soiling and materials damage effects. A Mathtech
study of soiling and materials damage in parts of the
manufacturing sector is also included. Otherwise, this
procedure is the same as Procedure D.
Procedure F. This aggregation procedure includes the
Mazumdar et al. study in addition to the studies by Lave
and Seskin and by Lipfert for use in estimating benefits in
2-34
-------
the mortality category.' Both the Ostro and the Samet e_t
al. studies are used in the acute morbidity category and
the Crocker et al. study is used in the chronic morbidity
category. Studies by Watson and Jaksch, and Cummings et
al. are used for household soiling and damage effects. The
Mathtech study is used to estimate benefits due to reduced
soiling and materials damage for part of the manufacturing
sector. Some double counting of benefits among the subset
for which estimates are provided is likely under this
aggregation procedure.
Results are also reported for analyses conducted with and without the
imposition of the Staff Paper* lower bound and for full (B Scenario) and
partial (A Scenario) attainment states.
Comparisons of the alternative PH10 and TSP standards are provided
below. Recall that these comparisons are based on TSP stringency and not
particle sizes. Because of issues related to particle size differences, a
second evaluation of only alternative PM10 standards is also provided.
Cost Effectiveness: Dominant and Inferior Options
It is convenient to identify inferior air quality standards prior to
conducting the efficiency tests. Recall that an inferior standard requires
higher costs to achieve the same or smaller benefits than a dominant
alternative. Once inferior alternatives are identified, only those
standards that are cost effective need be evaluated in terms of efficiency.
The inferior options are identified in Tables 2-2 through 2-5. Of the 240
(i.e., 10 z 6 x 2 x 2) combinations of standards, aggregation procedures,
attainment states, and Staff Paper lower-bound conditions, 118 are
inferior.
Incremental Net Benefits
The results of the incremental net benefit analysis are reported in
Tables 2-2 through 2-5 and interpreted below. Any standard that yields
* The Staff Paper lower bound refers to the TSP annual arithmetic mean
(AAM) 110 |ig/m constraint applied to health studies.
2-35
-------
Table 2-2
INCREMENTAL NET BENEFITS FOR ALTERNATIVE PM10 AND TSP STANDARDS*
(B Scenarios)
A 1 * 4- -5 C* 4- A J
AI terna t ive ocanciara
PM10 70 AAM/250 24-hr. 7 yr.
PM10 55 AAM 7 yr.
PM10 55 AAM/250 24-hr. 7 yr.
PM10 55 AAM/200 24-hr. 7 yr.
PM10 55 AAM/150 24-hr. 7 yr.
PM10 48 AAM/183 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 7 yr.
TSP -7150 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 9 yr.
TSP -/150 24-hr. 9 yr.
A
0.29
-0.71
-0.75**
-0.69**
-1.5**
-1.3**
-0.48
-3.5**
-1.1
-5.6
)
B
1.6
1.6
1.6**
1.7
1.4**
1.8**
2.6
0.45**
3.2
-0.09
Iggregatioi
C
12
20
20**
20
24**
25**
27
30**
37
41
i Procedure
D
23
40
41**
42
49**
53**
54
65**
77
91
E
38
72
72**
74
87**
95**
98
120**
140
160
F
51
95
96**
98
120**
130**
130
160**
180
220
OJ
* 1982 discounted present values in billions of 1980 dollars at a 10 percent discount rate. Time
horizons for 7- and 9-year standards are, respectively. 1989-1995 and 1987-1995. Comparisons
between PM10 and TSP standards are in terms of TSP stringency, not particle sizes.
** These options are inferior. Other options provide the same or larger incremental benefits at
lower incremental cost.
-------
Table 2-3
INCREMENTAL NET BENEFITS FOR ALTERNATIVE PM10 AND TSP STANDARDS*
(A Scenarios)
A 1 * 4-« C * A A
PM10 70 AAM/250 24-hr. 7 yr.
PM10 55 AAM 7 yr.
PM10 55 AAM/250 24-hr. 7 yr.
PM10 55 AAM/200 24-hr. 7 yr.
PM10 55 AAM/150 24-hr. 7 yr.
PM10 48 AAM/183 24-hr. 7 yr.
TSP 75 AGM/260 24~hr. 7 yr.
TSP -/150 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 9 yr.
TSP -/150 24-hr. 9 yr.
A
0.38
0.20
0.18**
0.18**
-0.03**
-0.08**
0.16
-0.38**
0.11
-0.69
B
1.2
1.4
1.4**
1.5
1.5**
1.4**
1.7
1.6**
2.2
2.1
Aggregatioi
C
7.7
11
11**
11
13**
13**
13
16**
18
22
i Procedure
D
15
21
21**
22
25**
25**
26
32**
37
45
E
23
34
34**
35
41**
42**
43
54**
60
76
F
32
46
46**
48
57**
57**
59
74**
83
100
OJ
-J
* 1982 discounted present values in billions of 1980 dollars at a 10 percent discount rate. Time
horizons for 7~ and 9-year standards are, respectively, 1989-1995 and 1987-1995. Comparisons
between PM10 and TSP standards are in terms of TSP stringency, not particle sizes.
** These options are inferior. Other options provide the same or larger incremental benefits at
lower incremental cost.
-------
Table 2-4
INCREMENTAL NET BENEFITS FOR ALTERNATIVE PM10 AND TSP STANDARDS WITH LOWER BOUND APPLIED*
(B Scenarios)
ro
i
00
A 1 4- * * O A- A J
Ai terna t ive otanuara
PM10 70 AAM/250 24-hr. 7 yr.
PM10 55 AAM 7 yr.
PM10 55 AAM/250 24-hr. 7 yr.
PM10 55 AAM/200 24-hr. 7 yr.
PM10 55 AAM/150 24-hr. 7 yr.
PM10 48 AAM/ 183 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 7 yr.
TSP -/150 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 9 yr.
TSP -/150 24-hr. 9 yr.
-
A
-0.51
-2.1
-2.1**
-2.1**
-3.1**
-2.9**
-2.2**
-5.5**
-3.5
-8.4
B
-0.33
-1.9
-1.9**
-1.9**
-2.9**
-2.7**
-2.0**
-5.3**
-3.2
-8.1**
Aggregation
C
1.7
0.8
0.76**
0.83
0.12**
0.39**
1.2
-1.7**
0.97
-3.2
Procedure
D
13
16
16**
16**
15**
16**
16**
14**
22
18
E
28
38
38**
38**
39**
40**
41
41**
56
55
F
39
58
59**
59
65**
68**
69
78**
95
110
* 1982 discounted present values in billions of 1980 dollars at a 10 percent discount rate. Time
horizons for 7- and 9-year standards are. respectively, 1989-1995 and 1987-1995. The TSP AAM
lower bound of 110 jig/m is applied to all health studies. Comparisons between PM10 and TSP
standards are in terms of TSP stringency, not particle sizes.
** These options are inferior. Other options provide the same or larger incremental benefits at
lower incremental cost.
-------
Table 2-5
INCREMENTAL NET BENEFITS FOR ALTERNATIVE PM10 AND TSP STANDARDS WITH LOWER BOUND APPLIED*
(A Scenarios)
A 1 4- 4-* O 4- A J
AI ic run iivo atano&ra
PM10 70 AAM/250 24-hr. 7 yr.
PM10 55 AAM 7 yr.
PM10 55 AAM/250 24-hr. 7 yr.
PM10 55 AAM/200 24-hr. 7 yr.
PM10 55 AAM/150 24-hr. 7 yr.
PM10 48 AAM/183 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 7 yr.
TSP -/150 24-hr. 7 yr.
TSP 75 AGM/260 24-hr. 9 yr.
TSP -/150 24-hr. 9 yr.
A
-0.16
-0.56**
-0.59**
-0.61**
-0.92**
-0.98**
-0.74**
-1.4**
-1.2
-2.2**
1
B
-0.03
-0.43**
-0.46**
-0.48**
-0.79**
-0.85**
-0.61**
-1.3**
-1.0
-2.0** '
Aggregation
C
1.3
1.1
1.0**
1.1**
0.87**
0.80**
1.1
0.60**
1.3
0.60
Procedure
D
8.8
9.7
9.6**
9.6**
9.5**
9.4**
9.6**
9.3**
13
13
E
17
20
20**
20
20
20**
20**
22**
28
30
F
24
30
30**
31**
33**
34**
34
39**
47
55
(NJ
I
* 1982 discounted present values in billions of 1980 dollars at a 10 percent discount rate. Time
horizons for 7- and 9-year standards are, respectively, 1989-1995 and 1987-1995. The TSP AAM
lower bound of 110 ug/m is applied to all health studies. Comparisons between PM10 and TSP
standards are in terms of TSP stringency, not particle sizes.
** These options are inferior. Other options provide the same or larger incremental benefits at
lower incremental cost.
-------
positive incremental net benefits will produce an improvement in the
efficiency of resource allocation relative to baseline controls.
Alternatively, a standard associated with negative incremental net benefits
will produce an inefficient allocation of resources relative to baseline
controls. The standard generating the largest incremental net benefits is
the most efficient of the alternative standards evaluated.
The figures presented in Tables 2-2 and 2-3 represent estimates of
incremental net benefits corresponding to the full and partial attainment
scenarios (scenarios B and A), respectively, when no lower bounds are
applied to the health studies.* Comparable estimates when the Staff Paper
lower bounds are applied are reported in Tables 2-4 and 2-5 for each of the
two attainment scenarios.** The results reported in these latter two
tables should be interpreted in view of the comments offered in Sections 1
and 10 regarding the application of the lower-bound concentration levels to
health studies. Similarly, estimates of incremental net benefits for the B
scenarios should be interpreted in view of the uncertainty in the estimated
costs of achieving full attainment.
The results of the incremental net benefit analysis shed light on
three important issues regarding the alternative PM NAAQS. These issues
are:
* The figures reported in Table 2-2 are computed as the difference between
the incremental benefits reported earlier in Table 1-11 and the incre-
mental costs for the B scenario reported in Table 2-1. Comparable
figures in Table 2-3 are computed as the difference between estimated
incremental benefits in Table 1-12 and the incremental costs for the A
scenario in Table 2-1. The estimated incremental net benefits reported
in Tables 2-2 and 2-3 have been rounded to two significant digits.
** Net incremental benefits reported in Table 2-4 are computed as the
difference between the incremental benefits reported in Table 1-13 and
the B scenario costs reported in Table 2-1. The figures in Table 2-5
are computed in the same fashion, except that incremental benefits
reported in Table 1-14 and A scenario costs reported in Table 2-1 are
used. The estimated incremental net benefits reported in Tables 2-4 and
2-5 have been rounded to two significant digits.
2-40
-------
The relative economic efficiency of the alternative
standards.
The economic efficiency of complete versus partial attain-
ment.
The economic efficiency of the implementation periods for
the TSP options.
Some conclusions regarding each of these issues are provided immediately
below. These findings, however, should be interpreted in view of the
previously mentioned caveats regarding benefit-cost analyses.
Economic Efficiency of Alternative Standards
The estimated incremental net benefits associated with the alternative
PM NAAQS are sensitive to the aggregation procedures employed, the attain-
ment status considered, and the application of lower bounds to the health
studies. A general understanding of how these three conditions affect
estimated incremental net. benefits is helpful in evaluating the economic
efficiency of specific alternative standards.
Other things being the same, the estimated incremental benefits
associated with a given standard increase as more comprehensive aggregation
procedures are applied. The estimated costs associated with a given
standard, however, remain constant across aggregation procedures. Conse-
quently, the estimated incremental net benefits associated with all
standards increase as more comprehensive aggregation procedures are
employed. It is also noteworthy that, ceteris paribus. the estimated
incremental net benefits associated with more stringent standards typically
increase relative to those associated with less stringent standards as more
comprehensive aggregation procedures are applied.
The estimated incremental net benefits under the full attainment
scenario (B) are usually higher than those estimated under the partial
attainment scenario (A), other things being the same. This general obser-
vation, however, consistently does not hold for aggregation procedure A.
2-41
-------
This occurs because the functional forms used in the studies included in
aggregation procedure A predict diminishing marginal benefits to incre-
mental air quality improvements over certain concentration ranges.
The application of the Staff Paper lower bound TSP annual arithmetic
mean (AAM) concentration to health studies has the general effect of
reducing the estimated incremental benefits associated with all standards,
other things being the same. Since estimated costs do not depend on the
lower bound, estimated net incremental benefits naturally decrease when the
lower bound is applied. Moreover, for a given aggregation procedure and
attainment status, the application of the lower bound has the typical
effect of increasing the estimated economic efficiency of less stringent
standards relative to more stringent standards.
Given the general effects of aggregation procedures, attainment
status, and the application of the lower bound on estimated incremental net
benefits, the evaluations of the relative economic efficiency of specific
standards are more straightforward. Specifically, three alternative
standards the two TSP 9-year standards and the PM10 70 AAM/250 24-hour
standard stand out as being most efficient when the standards are
compared based on TSP stringency. Typically, the TSP standards are
preferred (based on the comparison of TSP stringency) when more comprehen-
sive aggregation procedures are applied, while the PM10 70 AAM/250 24-hour
standard is generally preferred when less comprehensive aggregation proce-
dures are employed, the lower bound is applied, and only partial attainment
is achieved.
The domain of economically preferred air quality standards is dis-
played in Table 2-6. The most efficient standard (i.e., the standard
associated with the largest positive estimated incremental net benefits) is
indicated for each of the several conditions under which incremental net
benefits are estimated.
As Table 2-6 indicates, none of the alternative PM NAAQS considered is
efficient when incremental net benefits are estimated with the TSP AAM
2-42
-------
Table 2-6
DOMAIN OF ECONOMICALLY PREFERRED STANDARDS*
Conditions
WITH LOWER BOUND:
Partial Attainment
(Scenario A)
WITO LOWER BODND:
Complete Attainment
(Scenario B)
WITHOUT LOWER BOUND:
Partial Attainment
(Scenario A)
WITHOUT LOWER BOUND:
Complete Attainment
(Scenario B)
Aggregation Procedure
A
No Standard
Efficient
No Standard
Efficient
PM10 70/250
PM10 70/250
B
No Standard
Efficient
No Standard
Efficient
TSP 75/260
TSP 75/260
C
PM10 70/250
or TSP 75/260
PM10 70/250
TSP 150
TSP 150
D
TSP 75/260
or TSP 150
TSP 75/260
TSP 150
TSP 150
E
TSP 150
TSP 75/260
TSP 150
TSP 150
F
TSP 150
TSP 150
TSP 150
TSP 150
* All preferred TSP standards identified are 9-year standards with the 1987-1995 time horizon. Condi-
tions "with lower bound" refer to cases in which the TSP AAM lower bound of 110 ug/m is applied to all
health studies. Comparisons between PM10 and TSP standards are in terms of TSP stringency, not
particle sizes.
-------
lower bound applied and under aggregation procedures A and B, regardless of
attainment status. In these cases, estimated incremental net benefits are
negative for all standards evaluated. A literal interpretation of the
estimates indicates that all considered standards are less efficient than
baseline controls.
When estimated incremental net benefits are computed under the same
conditions, except without applying the lower bound, the PM10 70/250
standard is the most efficient standard under aggregation procedure A,
while the TSP 75/260 9-year standard is most efficient in terms of TSP
stringency under aggregation procedure B. Both conclusions hold regardless
of attainment status.
Under aggregation procedure C, the PM10 70/250 standard is the most
efficient of all considered standards for the complete attainment
scenario, and is tied with the 9-year TSP 75/260 standard under partial
attainment if benefits are computed while applying the lowet bound to all
health studies. The TSP 150 9-year standard, however, is the economically
preferred option under aggregation procedure C when benefits are estimated
without the lower bound.
The two TSP standards are preferred under aggregation procedures D
through F, regardless of attainment status and the application of the lower
bound. The TSP 150 standard is the most efficient of standards considered,
except in procedure D when incremental net benefits are estimated with the
lower bound applied, and in procedure E with the lower bound applied and
under the complete attainment scenario. The TSP 75/260 standard is
preferred or equally efficient in these exceptions.
All of the comparisons between alternative PM10 and TSP standards
described above are based on TSP stringency and not particle sizes. The
estimated incremental net benefits associated with the alternative PM10
standards might be understated relative to those associated with the TSP
standards if PM10 control options are more effective than TSP control
options in reducing the PM10 fraction of TSP. If only the alternative PM10
2-44
-------
standards are evaluated for efficiency, the PM10 48 AAM/183 24-hour
standard usually replaces the two TSP standards as being economically
preferred.* The following cases are exceptions:
Either the PM10 55 AAM/200 or the PM10 55 AAM/150 standard
is preferred when incremental net benefits are computed
under partial attainment (A scenario) without applying the
Staff Paper lower bound for aggregation procedure B.
The PH10 55 AAM standard is preferred when incremental net
benefits are computed under partial attainment and
with the application of the lower bound for aggregation
procedure D.
There are also some cases in which other less stringent PM10 standards have
the same estimated incremental net benefits as the PM10 48 AAM/183
standard. The benefit-cost analysis cannot generally make distinctions
between the economic efficiency of alternative standards with the same
estimated incremental net benefits.
It is noteworthy that the intermediate (in terms of stringency) air
quality standards are generally not preferred, based on the analysis of
incremental net benefits. The least stringent standard, the PM10 70
AAM/250 24-hour standard, is usually preferred when relatively restrictive
assumptions are imposed on the estimated benefits. However, the most
stringent standards are usually most efficient when less restrictive
assumptions are imposed on estimated benefits. These two results typically
hold, regardless of whether all alternatives or only PH10 alternatives are
considered.
Then the standards which yield the highest net incremental benefit are
the least or most restrictive considered, it is desirable to evaluate a
wider range of alternative standards. Specifically, in cases where PM10
70/250 is preferred, more lenient alternatives should be examined to
* Tables 2-2 through 2-5 indicate that the PM10 48 AAM/183 standard is
inferior. The results of the cost-effectiveness analyses reported in
these tables, however, include the alternative TSP standards.
2-45
-------
determine if incremental net benefits peak at 70/250, or whether they would
be larger with a less stringent standard. In cases where TSP 150 (or the
PM10 48 AAM/183) standard is preferred, more stringent alternative
standards should be examined to see if incremental net benefits peak, or if
some more stringent standard is economically preferred.
The identification of the most efficient option is far more sensitive
to the application of aggregation procedures than to either the attainment
status or the application of the Staff Paper lower bound. Under aggrega-
tion procedure A (the least comprehensive aggregation procedure), the PM10
70 AAM/250 24-hour standard is always preferred among the standards
evaluated. In contrast, the two 9-year TSP standards are always most
efficient when alternative PM10 and TSP standards are compared in terms of
TSP stringency under aggregation procedures 0 through F.
As noted earlier in this section, applied benefit-cost analysis
typically can provide only a qualified economic assessment of regulatory
alternatives. The previously noted caveats regarding benefit, cost, and
air quality estimates as inputs to the analysis should be considered in
using the estimated incremental net benefits to assess the relative
efficiency of the alternative PM NAAQS. First, in order to determine if a
given standard is efficient (i.e.. relative to the baseline scenario), the
positive difference between estimated incremental benefits and costs should
be statistically significant. Second, in order to identify the most
efficient standard (among those evaluated), the difference between
estimated incremental net benefits for the preferred and next best standard
should also be statistically significant. However, because some of the
uncertainty in the estimated incremental benefits and costs cannot be
measured, no statistical tests have been conducted. It is noted that
differences in estimated incremental net benefits become greater, both in
absolute and relative terms, as more comprehensive aggregation schemes are
employed. These differences, however, do not imply the presence or the
absence of statistical significance.
2-46
-------
Economic Efficiency: Attainment Status and TSP Implementation Periods
Based on the results of the benefit-cost analysis, complete attainment
is generally preferred to partial attainment because, other things being
equal, estimated incremental net benefits are larger under the B scenario.
Partial attainment is preferred for all efficient standards evaluated,
however, if aggregation procedure A is adopted. These conclusions depend
on the accuracy of the estimates of the differential cost associated with
complete attainment. As noted previously, these cost estimates are highly
uncertain.
The results of the analyses also suggest that the 9-year implementa-
tion period for the TSP standards are associated with higher incremental
net benefits than the 7-year implementation period. Although there are
some conditions under which the estimated incremental net benefits are
larger for the 7-year standards than the 9-year standards, the issue of the
implementation period is moot in these cases because the TSP standards are
not efficient relative to the PM10 standards. That is, in those cases in
which the TSP standards are ranked as being more efficient than the PM10
standards, the 9-year implementation period is always more efficient than
the shorter period.
These conclusions imply the ability to measure statistically signifi-
cant differences in benefits and costs. As was the case for the analysis
of the relative efficiency of specific standards, differences in estimated
incremental benefits and costs become greater as more comprehensive aggre-
gation procedures are applied. However, the earlier caveat applies here as
well: relatively large differences do not imply statistically significant
differences; smaller differences do not necessarily imply the absence of
statistical significance.
Distribution of Incremental Net Benefits
As was discussed earlier, the benefit-cost analyses are employed to
assess the economic effects of a given PM NAAQS on resource allocation
2-47
-------
within the'economy. For purposes of this analysis, it is assumed that
those who benefit from a given PM NAAQS can potentially compensate those
who incur the costs of this PM NAAQS, and still retain some welfare gain.
However, full compensation is not actually paid, so that some individuals
will be worse off. In order to evaluate the full effect of a particular
standard on economic welfare, judgments must be made regarding the desira-
bility of the distribution of benefits and costs associated with the imple-
mentation of a given PM NAAQS. These judgments are not made within the
context of the benefit-cost analyses; rather, they are described below. In
particular, the following distributional issues are considered:
Net benefits across regions.
The sectoral distribution of impacts i.e., the distribu-
tions across and within industries.
Regional Distribution
It is not likely that net benefits associated with any given PM NAAQS
are distributed equally over the entire United States. Thus, data on the
distribution of benefits and costs can provide additional information
regarding the impacts of the alternative PM NAAQS.
The method employed to estimate the incremental control costs did not
permit a consistent estimation of costs by region for the complete attain-
ment scenario (i.e., the B scenario). However, incremental control costs
by region are available when residual nonattainment is permitted (i.e., for
the A scenario). Estimates of the incremental net benefits associated with
this scenario are calculated for the PM10 70 AAM/250 24-hour standard
across each of several regions for each of the six aggregation procedures.
The regional distribution for this standard is presented in Table 2-7,
where the incremental net benefits are reported by EPA administrative
region. The comparable distribution when the TSP AAM lower bound of 110
q
(ig/m is applied to all health studies is reported in Table 2-8.
2-48
-------
t\)
I
Table 2-7
INCREMENTAL NET BENEFITS BY REGION FOR THE PM10 (70, 250) SCENARIO A STANDARD*
EPA Region
I New England
II New York-New Jersey
III Middle Atlantic
IV South Atlantic
V East North Central
VI South Central
VII Midwest
VIII Mountain
IX South Pacific
X North Pacific
U.S.
Aggregation Procedure
A
0.0
0.0
0.07
0.0
-0.05
0.01
0.01
0.0
0.33
0.01
0.38
B
0.0
0.0
0.13
0.06
0.24
0.11
0.03
0.05
0.59
0.03
1.2
C
0.0
0.02
0.60
0.52
2.5
0.83
0.17
0.39
2.5
0.21
7.7
D
0.0
0.04
1.0
1.0
5.7
1.8
0.35
0.74
4.2
0.38
15
E
0.0
0.06
1.8
1.7
8.6
2.7
0.55
1.3
6.3
0.58
23
F
0.0
0.08
2.4
2.2
11
3.7
0.74
1.8
8.9
0.81
32
* 1982 discounted present values in billions of 1980 dollars at a 7-year time horizon (1989-95) and a 10
percent discount rate. Regional incremental net benefits may not sum to U.S. totals due to independent
rounding.
Sources: Regional incremental benefits, Section 11, Table 11-22; regional incremental costs, Reference
(5). Costs have been adjusted to reflect the 9-year implementation horizon, and AME costs at 4
percent have been added.
-------
Table 2-8
INCREMENTAL NET BENEFITS BY REGION FOR THE PM10 (70, 250) SCENARIO A STANDARD WITH LOWER BOUND APPLIED*
EPA Region
I New England
II New York-New Jersey
III Middle Atlantic
IV South Atlantic
V East North Central
VI South Central
VII Midwest
VIII Mountain
IX South Pacific
X North Pacific
U.S.
Aggregation Procedure
A
0.0
0.0
-0.01
-0.05
-0.24
-0.05
-0.01
-0.02
0.22
-0.01
-0.16
B
0.0
0.0
-0.01
-0.05
-0.24
-0.05
-0.01
0.0
0.33
-0.01
-0.03
C
0.0
0.0
0.03
0.0
-0.05
0.0
0.0
0.13
1.2
0.02
1.3
D
0.0
0.0
0.53
0.51
3.3
0.95
0.18
0.35
3.0
0.16
8.9
E
0.0
0.01
1.3
1.1
6.1
1.8
0.37
0.72
5.0
0.30
17
F
0.0
0.03
1.8
1.6
8.4
2.6
0.53
1.2
7.4
0.48
24
* 1982 discounted present values in billions of 1980 dollars at a 7-year time horizon (
percent discount rate. The TSP AAM lower bound of 110 ug/m3 is applied to all
(1989-95) and a 10
percent discount rate. The TSP AAM lower bound of 110 ug/m3 is applied to all health studies.
Regional incremental net benefits may not sum to U.S. totals due to independent rounding.
Sources: Regional incremental benefits. Section 11, Table 11-62; regional incremental costs. Reference
(5). Costs have been adjusted to reflect the 9-year implementation horizon, and AME costs at 4
percent have been added.
-------
The regional distribution of estimated incremental net benefits does
vary somewhat across aggregation procedures, but generally, the largest
shares of incremental net benefits are realized by East North Central and
South Pacific regions. For example, the estimates reported in Table 2-7
indicate that about 65 percent of the incremental net benefits for the PM10
standard accrue to these two regions when aggregation procedure E is
applied. The figure is also about 65 percent under procedure E. when the
lower bound is applied as in Table 2-8. As noted above, this result
assumes that not all counties are in complete attainment with the standard.
Consequently, this distribution cannot be used to make any inferences
regarding the distribution of net benefits when the condition of full
attainment is applied.
The behavior of the regional distribution of incremental net benefits
associated with more stringent air quality standards may also be of
interest. Regional distributions of estimated incremental net benefits for
the TSP 75 AGM/260 9-year standard are reported in Tables 2-9 and 2-10,
respectively, when incremental net benefits are computed without and'with
the TSP AAM lower-bound. The results indicate a slightly more even
regional distribution of incremental net benefits compared to the distribu-
tion associated with the more lenient PM10 standard. For example, the East
North Central and South Pacific regions receive only about 52 percent or 63
percent of total estimated incremental net benefits under aggregation
procedure E, depending on whether incremental net benefits are computed
without or with the Staff Paper lower-bound concentration level.
Sectoral Distribution
Incremental net benefits associated with alternative PH NAAQS are also
distributed over various economic sectors. For example, benefits and/or
costs may accrue in the household, manufacturing, commercial, institu-
tional, and government sectors. In this analysis, benefit estimates are
limited to the household sector and a small subset of the manufacturing
sector, while estimates of control costs include costs borne by private
industrial sources and government-owned utilities and roads. Thus, a
2-51
-------
I
(Jl
Table 2-9
INCREMENTAL NET BENEFITS BY REGION FOR THE TSP (75, 260) 9-YEAR SCENARIO A STANDARD*
EPA Region
I New England
II New York-New Jersey
III Middle Atlantic
IV South Atlantic
V East North Central
VI South Central
VII Midwest
VIII Mountain
IX South Pacific
X North Pacific
U.S.
Aggregation Procedure
A
0.0
-0.01
-0.03
-0.09
-0.24
0.02
-0.01
-0.02
0.48
0.01
0.11
B
0.05
0.03
0.20
0.10
0.44
0.22
0.10
0.09
0.92
0.09
2.2
C
0.41
0.37
1.9
1.6
5,6
1.6
0.92
0.92
4.2
0.75
18
D
0.80
0.85
3.7
3.1
12
3.5
1.8
1.7
7.2
1.5
37
E
1.4
1.6
6.6
5.4
20
5.6
3.0
2.9
11
2.5
60
F
2.0
2.1
9.1
7.4
27
7.7
4.2
4.1
16
3.4
83
* 1982 discounted present values in billions of 1980 dollars at a 9-year time horizon (1989-95) and a 10
percent discount rate. Regional incremental net benefits may not sum to U.S. totals due to independent
rounding.
Sources: Regional incremental benefits. Section 11, Table 11-38; regional incremental costs. Reference
(5). Costs have been adjusted to reflect the 9-year implementation horizon, and AME costs at 4
percent have been added.
-------
Ul
Table 2-10
INCREMENTAL NET BENEFITS BY REGION FOR THE TSP (75, 260) SCENARIO A STANDARD WITH LOWER BOUND APPLIED*
EPA Region
I New England
II New York-New Jersey
III Middle Atlantic
IV South Atlantic
V East North Central
VI South Central
VII Midwest
VIII Mountain
IX South Pacific
X North Pacific
U.S.
Aggregation Procedure
A
-0.02
-0.03
-0.23
-0.21
-0.65
-0.07
-0.08
-0.08
0.24
-0.04
-1.2
B
-0.02
-0.03
-0.23
-0.20
-0.65
-0.07
-0.08
'-0.06
0.39
-0.04
-1.0
C
0.0
0.0
-0.11
-0.11
-0.24
0.03
-0.03
0.11
1.6
0.03
1.3
D
0.04
0.13
0.91
0.72
4.8
1.4
0.27
0.45
4.3
0.27
13
E
0.22
0.49
2.7
1.9
9.9
2.9
0.79
1.1
7.6
0.64
28
F
0.68
0.95
4.7
3.7
16
4.6
1.8
2.1
12
1.4
47
* 1982 discounted present values in billions of 1980 dollars at a 9-year time horizon (1989-95) and a 10
percent discount rate. The TSP AAM lower bound of 110 (ig/m is applied to all health studies.
Regional incremental net benefits may not sum to U.S. totals due to independent rounding.
Sources: Regional incremental benefits, Section 11, Table 11-78; regional incremental costs, Reference
(5). Costs have been adjusted to reflect the 9-year implementation horizon, and AME costs at 4
percent have been added.
-------
complete analysis of the distribution of net incremental benefits by sector
is not possible. However, the results of the Economic Impact Analysis
(EIA) do provide some information on the distribution of estimated control
costs by economic sector (1).
Estimated direct emissions control costs are summarized by 2- and 4-
digit SIC codes in Table 2-11 for the 9-year TSP 75/260 standard under the
partial attainment scenario (estimated costs by industry are unavailable
for the complete attainment scenario). For most industries, the estimated
control costs associated with less stringent standards are lower than those
associated with the 9-year TSP 75/260 standard under the partial attainment
scenario. The estimated control costs for the complete attainment
scenario, however, are highly uncertain and much larger than those
associated with partial attainment. Consequently, the adverse impacts on
some industries could be more severe under complete attainment.
A high percentage of emissions control costs will be shared by the
manufacturing (SIC 20-39) and utility (SIC 49) sectors. The manufacturing
sector, for example, accounts for about 50 percent of all capital expendi-
tures under the TSP 75/260 standard. The utility sector will incur about
44 percent of estimated capital costs under this standard. Electric utili-
ties (SIC 4911) are expected to bear about 40 percent of total capital
costs.
The EIA uses the estimates of control costs to assess the adverse
impacts of the most restrictive standard (i.e., the current TSP combined
primary and secondary standards) on production, prices, plant closures, and
employment. The analysis includes 16 industries judged to be most
seriously affected by the current TSP standards. Each industry is analyzed
separately. Consequently, the effects in factor and output markets between
and among these industries are ignored. In addition, the EIA does not
consider possible positive economic impacts such as employment gains in the
pollution control equipment industry or gains due to innovation-inducing
effects of the regulations. However, the EIA does provide information
which is useful in describing the distribution of several economic effects
2-54
-------
Table 2-11
CAPITAL AND AFTER-TAX ANNUALIZED COSTS (ATAC) OF CONTROL,
BY INDUSTRY, FOR THE TSP (75, 260) 9-YEAR STANDARD*
(Scenario A)
SIC**
02
07
08
10
14
1422
1429
1442
16
20
2041
21
22
24
25
26
27
28
29
2951
30
32
3241
3274
3281
3295
33
3312
3321
3331
3332
3334
Capital
Cost
1.91
0.400
0.008
10.7
61.2
20.2
5.2
5.68
103
45.6
2.11
5.51
1.67
0.816
67.0
2.30
71.4
73.2
36.6
0.764
186
74.5
21.2
4.95
51.4
623
444
67.2
2.96
0.340
ATAC
0.193
0.042
0.001
1.37
9.92
3.66
0.870
1.40
0.125
11.1
4.64
0.220
0.632
0.424
0.091
7.87
0.250
8.11
11.9
7.27
0.090
23.6
8.73
2.58
0.562
6.87
72.4
53.4
6.77
0.623
0.067
0.166
SIC**
34
35
36
37
38
39
40
42
44
49
4911
4961
50
51
52
53
65
80
82
91
92
93
94
95
96
97
99
Total
Capital
Cost
27.2
8.65
8.16
39.4
6.84
6.29
1.41
6.40
0.214
1,130
1,020
65.5
1.06
42.7
0.414
0.230
1.14
11.2
10.4
2.65
0.727
1.43
1.52
0.391
14.2
2.64
45.5
2,580
ATAC
3.19
0.999
0.888
4.20
0.755
0.787
1.28
0.892
0.022
120
108
7.60
0.313
5.11
0.071
0.024
0.149
1.31
1.10
0.325
0.080
0.150
0.166
0.046
1.62
0.301
8.86
301
* Millions of 1980 dollars. Individual entries may not sum to totals due
to independent rounding.
** Sectors by SIC are: Agriculture (01-09); Mining (10-14); Construction
(15-19); Manufacturing (20-39); Service Utilities (40-49); Wholesale
Goods (50-59); Service Industries (80-89); Government Operations (90-
99).
Source: Reference (1).
2-55
-------
attributable to the current TSP standards. These effects are summarized
below.
For each industry, the type and magnitude of economic effect depends,
to a large extent, on the ability of firms within that industry to pass
control costs through to other economic agents. In some instances, these
agents may be firms in other industries. However, it is also possible that
agents in the household, commercial, institutional, and governmental
sectors may bear the costs that are passed through. On the basis of the
El A, it appears that most affected firms will absorb at least some of these
costs. Hence, the effect of the current TSP standards on prices or output
in the product markets of these firms will be mitigated. However, for some
of the controlled firms, the cost increase they must absorb may force them
to consider terminating operations.* While there will be some employee
displacement effects associated with these closures, the EIA concludes that
no net impact on industry employment or output is anticipated.
No significant foreign trade effects are anticipated for any of the 16
industries. The impact of current TSP standards on investment, producti-
vity, and innovation within any of these industries is likely to be adverse
but small. Finally, it should be noted that the economic impacts of less
stringent PH NAAQS generally will be less severe, but those associated with
full attainment could be much more severe.
CONCLUSIONS AND QUALIFICATIONS
The benefit-cost analyses reported earlier in this section were con-
ducted to assess the cost-effectiveness and economic efficiency of the
alternative standards. The economic efficiency of attainment states, and
* Emissions control costs may force some establishments to close in the
following industries: cement; crushed and broken stone; construction
sand and gravel; paving mixtures and blocks; cut stone and stone
products; crushed minerals and earths; grain and flour milling; and iron
foundry. The number of expected closures in each of these industries,
however, is a relatively small percent of total industry establishments.
2-56
-------
the implementation periods for the TSP standards, were also evaluated. In
addition, the distributional impacts associated with the alternative
standards were described. The conclusions offered from the interpretation
of these analyses, however, are subject to qualifications.
Conclusions
The major conclusions that emerge from the benefit-cost analyses are
described briefly below for each of the several aspects of the alternative
PM NAAQS that were evaluated.
Cost Effectiveness
The analysis of cost effectiveness indicates that 118 of the 240
combinations are inferior (i.e., cost more to achieve the same or smaller
benefits than other options).
Economic Efficiency of Alternative Standards
The major conclusions regarding economic efficiency of alternative
standards are as follows:
The ranking of standards in terms of economic efficiency is
especially sensitive to the different benefit aggregation
procedures employed.
Given the use of only CASAC-approved quantitative studies
that are amenable to benefit analysis and the imposition of
the Staff paper lower bound, none of the alternative
standards is efficient. Given a literal interpretation of
this result, an alternative less restrictive than the least
restrictive considered alternative (PM10 70 AAM/250 24-hour
expected value) is warranted in terms of economic
efficiency.
Given the use of the same CASAC-approved studies, but
without imposing the lower-bound concentration level, the
PM10 70 AAM/250 24-hour expected value alternative is pre-
ferred in terms of economic efficiency.
2-57
-------
Given the adoption of more comprehensive benefit
aggregation procedures, the two 9-year TSP alternatives of
75 AGM/260 24-hour (second high) and 150 24-hour (second
high) are preferred when all alternatives are compared in
terms of TSP stringency.
When only PM10 standards are evaluated for efficiency, the
PM10 48 AAM/183 24-hour standard is usually preferred.
Economic Efficiency of Partial Versus Complete Attainment
Two major conclusions result from the analysis of the economic
efficiency of attainment status. These conclusions are:
Given the use of only CASAC-approved quantitative studies
which are amenable to the benefit analysis, partial attain-
ment is preferred to full attainment.
Given the use of more comprehensive benefit aggregation
procedures, complete attainment is generally preferred to
partial attainment.
However, these conclusions depend importantly on the method and data used
to estimate the cost differential associated with complete attainment. As
noted previously, these cost estimates are subject to considerable
uncertainty.
Economic Efficiency of TSP Implementation Periods
Alternative TSP standards were evaluated for both 7-year (1989-1995)
and 9-year (1987-1995) implementation periods. The 9-year TSP implementa-
tion period is preferred in all cases in which either of the two TSP
standards were ranked as most efficient of the alternative standards
evaluated.
2-58
-------
Distributional Impacts
The distributional impacts of the alternative PM NAAQS are not
evaluated by the benefit-cost analyses. These distributional impacts were
described. Briefly summarizing, the following conclusions are offered:
Based on the findings of the economic impact analysis, no
major adverse impacts associated with the alternative
standards are expected. This conclusion, however, is based
on an analysis of estimated control costs under partial
attainment. Estimates of costs under full attainment are
much larger and highly uncertain. As a result, adverse
impacts could be more severe under complete attainment.
Under the less restrictive standards, much of the estimated
incremental net benefits will be realized in the East North
Central and South Pacific EPA administrative regions.
Under the more stringent standards, these two regions
receive a smaller but still considerable share of the
estimated incremental net benefits. The regional distribu-
tion of incremental net benefits is based on the partial
attainment scenario. The results of this analysis cannot
be used to project regional distributions under complete
attainment.
Qualifications
The general limitations of applied benefit-cost analysis and specific
qualifications to this analysis have already been described in this
section. The conclusions summarized immediately above should be assessed
in full view of these limitations and qualifications. Some of the specific
qualifications to this analysis are summarized briefly below.
The validity of the benefit-cost analyses depends on the estimates of
benefits, costs, and air quality. Sources of potential biases and
uncertainty in each of these estimates have already been described. It is
noted that these biases and uncertainties do bear directly on the conclu-
sions summarized above. These conclusions depend on the ability to detect
meaningful differences between benefits and costs, and differences in net
benefits across the alternative standards. As was stated previously, tests
for statistically significant differences have not been conducted because
2-59
-------
not all of the uncertainty associated with the estimated benefits and costs
is known.
In most cases, the alternative PM NAAQS ranked as most efficient are
either the least or most stringent of the standards evaluated. Because of
this, additional information on efficiency could have been obtained if a
wider range of standards had been evaluated. However, the analysis was
limited to options that were no less restrictive than the middle of the EPA
Staff Paper range (i.e., PM 70 AAM/250 24-hour expected value) and no more
restrictive than the current TSP secondary standard (150 24-hour second
high).
REFERENCES
1. Viola, J., R. Silverman, S. Schechter, J. Wagner, E. Hurley, J.
HcGovern and G. Scarborough. Technical Appendices, Economic Impact
Analyses of Alternative National Ambient Air Quality Standards for
Particulate Matter. Final Report to the U.S. Environmental Protection
Agency by Energy and Environmental Analysis, Inc., Arlington, VA,
December 1982.
2. Smith, A. E. and K. L. Brubaker. Costs and Air Quality Impacts of
Alternative National Ambient Air Quality Standards for Particulate
Matter. Technical Support Document. Final Report to the U.S.
Environmental Protection Agency by Argonne National Laboratory,
Argonne, IL, January 1983.
3. Vatavuk, W. Development of AME Estimates. U.S. Environmental
Protection Agency Memorandum dated April 13, 1981.
4. Walton, T. Baseline Costs. U.S. Environmental Protection Agency,
Memorandum dated August 2, 1982.
5. Laarman, J. Comparison of Mathtech with BEA Data on SIC Growth Rates.
Memorandum to T. Walton of U.S. Environmental Protection Agency dated
February 17, 1982.
6. Lind, R. C., K. J. Arrow, G. R. Corey, P. Dasgupta, A. K. Sen. T.
Stauffer, J. E. Stiglitz, J. A. Stockfisch, and R. Wilson.
Discounting for Time and Risk in Energy Policy. Baltimore: Johns
Hopkins University Press, 1982.
2-60
-------
|