United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-86-006
June 1986
yERA Regulatory Impact
Analysis for New
Source
Performance
Standards:
Industrial-
Commercial-
Institutional Steam
Generating Units of
Greater than 100
Million Btu/hr
Heat Input
-------
EPA-450/3-86-006
Regulatory Impact Analysis for New
Source Performance Standards:
Industrial-Commercial-lnstitutional Steam
Generating Units of Greater than 100
Million Btu/hr Heat Input
U. S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park. North Carolina 27711
June 1986
-------
V
This report has been reviewed by the Emission Standards and Engineering Division of the Office of Air Quality Planning
and Standards, EPA, and approved for publication. Mention of trade names or commercial products is not intended to
:onstitute endorsement or recommendation of use. Copies of the report are available through the Library Services Office
'MD-35), U.S. Environmental Protection Agency, Research Triangle Park, N.C. 27711, or from National Technical
Information Services, 5285 Port Royal Road, Springfield, Virginia 22161.
ii
-------
Contents
Page
ROLE OF THIS DOCUMENT vi
EXECUTIVE SUMMARY viii
I. Statement of Needs and Consequences 1-1
II. Alternatives Examined II-l
A. Introduction 11-1
B. No Further Regulation 11-1
C. Other Regulatory Approaches 11-2
D. Market Oriented Approaches I1-2
E. Regulatory Alternatives Within the 11-5
Scope of Present Legislation
III. Air Quality Data III-l
A. Introduction III-l
B. SO4 Air Quality Assessment III-l
C. SO4 Air Quality Results III-4
D. PM Air Quality Assessment III-6
E. SO2 Air Quality Assessment 111-6
IV. Engineering Costs IV-1
A. Introduction IV-1
B. Analytical Framework IV-1
V. Economic Impact Analysis V-l
A. Introduction V-l
B. Major Steam Users V-l
C. Selected Industries V-2
D. Conclusions V-3
-------
i i
Page
VI. Benefit Analyis VI-1
A. Introduction VI-1
B. Methodology V1-2
C. Study Selection, Application, Qualifications, VI-8
and Plausibility Checks
D. Benefit Analysis Estimates VI-32
VII. Benefit-Cost Analysis VII-1
-------
111
Tab! es
Number Page
1 National Cost Impacts of Regulatory xiv
Alternatives
2 Industrial Boiler NSPS Potential Benefit xvi
Categories
3 Annualized Benefits xviii
4 Average Benefits Per Ton xix
5 Incremental Benefits Per Ton xx
6 Net Benefit Analysis xx
7 Average Control Cost Per Ton and Benefit xxi
Per Ton
8 Incremental Control Cost Per Ton and Benefit xxii
Per Ton
I-l ASTRAP Matrix 111-3
1-2 1995 Base Case Sulfate Emissions 111-5
1-3 1995 Sulfate Emissions, LSF Alternative 111-7
1-4 1995 Sulfate Emission, Percent Reduction II1-8
Alternative
i
1-5 ASTRAP 1995 Estimated Visibility 111-9
1-6 Emissions Data Bases and Growth 111-10
V-l National Cost Impacts of Regulatory IV-4
A1ternatives
V-2 Average Control Cost Per Ton IV-5
IV-3 Incremental Control Cost Per Ton IV-5
VI-1 Industrial Boiler NSPS Potential VI-5
Benefit Categories
VI-2 Demand Equations for Which SO2 Is V1-26,27
A Significant Explantory Variable
VI-3 Annualized Benefits VI-33
VI-4 Average Benefits Per Ton, SO4 VI-34
-------
IV
Page
VI-5
Average Benefits Per Ton, PH
V1-35
VI-6
Average Benefits Per Ton, SO2
V1-36
VI-7
Average Benefits Per Ton, Total
VI-37
VI-8
Incremental Benefits Per Ton, Total
VI-38
VII-1
Net Benefit Analysis
VI1-2
VI1-2
Average Control Cost Per Ton and Benefit
Per Ton
VI1-3
VI1-3
Incremental Control Cost Per Ton and Benefit
Per Ton
VI1-3
-------
J
V
Figures
Number
VI-1 Description of Air Quality Scenario
VI-2 Basic Steps in Estimating Benefits
for An Individual Study
-------
vi
ROLE OF THIS DOCUMENT
This regulatory impact analysis (RIA) has been prepared to satisfy
Executive Order 12291 (1981) which requires preparation of an RIA for every
"major rule." The proposed regulation is a "major rule" because it could
result in industry-wide annualized costs, in the fifth year after the
standards go into effect, of more than $100 million dollars.
In compliance with the Executive Order, this study evaluates the pro-
jected consequences of alternative approaches to regulating new, modified,
and reconstructed industrial steam generating units. It considers benefits,
costs and economic impacts. The benefit-cost analysis portion of the RIA
has not played any role in developing the proposed standards.
There are several simplifying assumptions which inherently limit the
precision of this RIA. First, whereas the air quality modeling procedures
used are an improvement over that used in other similar studies, they
remain subject to simplifying assumptions regarding growth rates and stack
characteristics. Second, on the benefit analysis side, there are many
benefit categories not covered in this analysis and in the case of the
SO4 and PM benefits, the coverage is limited to the 31 eastern state
region. There are other assumptions used in this study which are detailed
in the appropriate sections of this RIA. While these assumptions are
necessary and reasonable, they limit the precision of the analysis.
-------
J
vi i
REFERENCES
Executive Order 12291 (1981). Federal Register, Vol. 48, No. 33. In
Executive Order 12291, the use of the words "major rule" refers to the
adverse economic consequences of a regulation.
-------
vi i i
EXECUTIVE SUMMARY
I. OVERVIEW
A. Background
New source performance standards have been proposed limiting emissions
of sulfur dioxide and particulate matter from new, modified or reconstructed
industrial-commercial-institutional steam generating units capable of
combusting more than 29 MW (100 million Btu/hour) heat input. Executive
Order 12291 states that a RIA must be performed on any regulation having
an annual effect on the economy of $100 million or more. This RIA was
prepared for the purpose of fulfilling this requirement.
It should be noted at the outset that in conducting this analysis a
number of analytical problems had to be overcome. Some of the analytical
assumptions made may affect the final results. Such assumptions are
noted in this Executive Summary and discussed in more detail in the text.
B. Summary of Findings
This analysis estimates the benefits and costs of two regulatory
alternatives:
1) standards based on the use of low sulfur fuel to reduce SO2
emissions
2) standards requiring a percent reduction in SOg emissions
The cost analysis estimated all known private SOg control costs that
would be incurred as a result of each of these regulatory alternatives.
The benefit analysis estimated the potential SO4 visibility improvement-
benefits, SO2 related agriculture benefits, residential materials damage
and morbidity benefits, and PM reduced morbidity and soiling benefits of
both regulatory alternatives. Potentially large benefit categories such
-------
ix
as commercial and industrial materials damage and acid deposition, were not
estimated. Other, most likely smaller, benefit categories such as climate
effects and improved transportation safety, have not been included either.
The net benefit analysis (benefits minus costs) yielded ambiguous
results. The estimated range of net benefits for the low sulfur fuel
alternative yielded positive results. The range of net benefits for the
percent reduction alternative spanned both positive and negative net
benefits. Annualized net benefits in the year 1990 for the low sulfur
fuel alternative, ranged from four million dollars to sixty-two million
dollars (1983 dollars). In 1990, the percent reduction alternative
yielded net benefits ranging from a negative forty-two million dollars to a
positive forty-three million dollars (1983 dollars).
II. STATEMENT OF NEED AND CONSEQUENCES
On August 21, 1979, a priority list for development of additional
new source performance standards was published in accordance with Sections
111(b)(1)(A) and 111(f)(1) of the Clean Air Act Amendments of 1977. This
list identified 59 major stationary source categories that were judged to
contribute significantly to air pollution. This list did not include
utility boilers since new source performance standards had previously
been developed for this source category. Fossil fuel-fired industrial
steam generating units ranked eleventh on the priority list of sources
for which new source performance standards would be established. Fossil
fuel-fired steam generating units were the highest ranked source of
particulate matter and sulfur dioxide emissions, and the second highest
ranked source of nitrogen oxide emissions when the priority list of
source categories was published.
The proposed standard could reduce emissions of sulfur dioxide from
a typical industrial steam generating unit by between 465 Mg/Year
-------
X
(510 tons/year) and 2,420 Mg/Year (2,670 tons/year), depending on unit
size and type of fuel fired.
The need for regulatory action arises from the failure of the market
system to deal effectively with air pollution as a negative externality.
Industrial boilers dispose of unwanted by-products by emitting them to
the atmosphere. The societal costs associated with those emissions, in
terms of health and welfare, are not fully paid by the emitters. The
market place does not function effectively or efficiently when either
total (private and social) costs and/or benefits are improperly accounted for.
III. ALTERNATIVES EXAMINED
Executive Order 12291 requires that the following alternatives be examined:
a) no further regulation,
b) other regulatory approaches,
c) market oriented approached
d) alternative approaches within the scope of present legislation.
Alternative a) -- was not found to be reasonable given the Administrator's
determination that the source category contributes significantly to air pollution
which could reasonably be expected to endanger public health and welfare.
Alternative b) -- Reliance on BACT and LAER determinations, in lieu of
NSPS for control of steam generating emissions, was found undesirable.
The primary reason for this is the determination that NSPS, in practice,
sets the "floor" for BACT and LAER determinations. In the absence of
NSPS, the BACT and LAER control levels would likely be less stringent.
Alternative c) -- Several market oriented approaches were considered
including pollution taxes, charges, and permits. The Clean Air Act
explicitly requires (Section III) that new, modified, and reconstructed
steam generating units control emissions to the level achievable by the
-------
xi
best demonstrated technological system of continuous emission reduction,
considering costs, nonair quality health, and environmental and energy
impacts. This effectively precludes the use of pollution taxes and
subsidies as alternatives to NSPS unless the exact level of taxes and
subsidies could be established that would achieve pollution abatement
equivalent to the level achievable by best demonstrated technology. In
the case of permits, the number of permits and emission allowances would
have to be correctly fixed in advance.
Alternative d) -- A number of regulatory alternatives based on the
use of low sulfur fuels to reduce SO2 emissions or the use of flue gas
desulfurization (F6D) systems to achieve a percent reduction in SO2
emissions were examined in developing the proposed NSPS. For purposes of
this RIA, two alternative control levels were examined and compared to
the regulatory baseline.
1) Regulatory Baseline
Defined by existing State Implementation Plans and the
existing NSPS for large steam generating units of more than 73
MW (250 million Btu/hour) heat input.
2) Low Sulfur Fuel
Based on the use of low sulfur fuel to reduce emissions to
344 ng SO2/J (0.8 lb S02/million Btu) heat input from oil
combustion and to 516 ng SO2/J (1.2 lb S02/million Btu) heat
input from coal combustion.
3) Percent Reduction
Based on the use of FGD systems to achieve a 90 percent
reduction in SO2 emissions from both oil and coal combustion.
This alternative would reduce emissions from coal combustion to
-------
xii
less than 258 ng SO2/J (0.6 lb S02/million Btu) heat input and
from oil combustion to less than 129 ng SO2/J (0.3 lb S02/mi 11 ion
Btu) heat input.
IV. AIR QUALITY DATA
Regional air quality analysis was conducted using transfer matrices
derived from the Advanced Statistical Trajectory Regional Air Pollution
(ASTRAP) model. ASTRAP is a regional scale SO2 model that considers
transport, dispersion, chemical transformation and removal of sulfur
oxides. As used in this analysis, ASTRAP models SO2 source-receptor
relationships in the 31 eastern state region. Sulfate reductions were
estimated for the 31 eastern states, for both regulatory alternatives
considered, and compared to the 1995 base case. The 1995 base case was
used, rather than the more appropriate 1990 base case, because it was
readily available in computer form and there is a less than 2% difference
in total emissions between the two cases. The percent reduction
alternative yielded the largest sulfate reductions. These estimated
sulfate levels were used to calculate visual range improvements due to
each regulatory alternative.
The SO4 changes for the 31 eastern states were translated into TSP
changes and PM benefits were calculated. No estimates of PM ambient air
reductions due to reduction of PM emissions were included in this analysis.
The SO2 air quality analysis was conducted primarily using the inventory
of new industrial boilers provided by the Industrial Fuel Choice Analysis
Model (IFCAM). SO2 reductions in 1990 were 170,000 tons for the low
sulfur fuel alternative and 240,000 for the percent reduction alternative.
In the 31 eastern states, for the two regulatory alternatives, SO2 reductions
-------
xi i i
were 120,000 tons and 175,000 tons respectively. The tons reduced in the
31-state region are relevant for both the SO4 and PM benefit estimates.
V. ENGINEERING COSTS
National cost impacts due to the regulatory alternatives for new
industrial fossil fuel-fired steam generating units were analyzed through
the use of the Industrial Fuel Choice Analysis Model (IFCAM). IFCAM is
an energy demand model developed to evaluate fuel choice decisions in the
industrial sector. IFCAM is a highly disaggregated process engineering
model designed to analyze factors affecting fuel choice decisions. Total
annualized costs and control costs are presented in Table 1.
VI. ECONOMIC IMPACT ANALYSIS
An economic analysis was performed to determine the potential industry-
specific economic impacts associated with the percent reduction regulatory
alternative. Since this control level was the most stringent analyzed, the
impacts from the low sulfur fuel alternative would necessarily be less severe
assuming the distribution of costs across affected industries are the same.
Seven industries most likely to experience severe impacts were
analyzed. Product prices were projected to increase from less than 0.01
to 0.5 percent in 1990 for six of the seven industries, assuming full cost
pass-through of increased steam costs. Cost as a percent of value added
was projected to increase by about 0.01 to 0.9 percent in 1990 for these
six industries, assuming full cost pass-through of increased steam costs.
The highest increases are projected for the beet sugar refining industry
due to the fact that the product price is low and steam costs represent an
unusually high proportion of manufacturing costs.
-------
xi v
Table 1
NATIONAL COST IMPACTS OF REGULATORY ALTERNATIVES1.2
(Millions 1983 $)
Regulatory Annualized Annualized
A1ternative Costs ($/yr) Control Costs ($/yr)5
Baseline 3,396 —
Low Sulfur Fuel-3 3,454 58
Percent Reduction^ 3,531 135
1 These annualized costs are in January 1983 dollars and are therefore s1ightl>
higher than the June 1982 costs reflected in the Preamble.
2 Assumed high oil penetration scenario - i.e., assumed lower oil prices than
in the high coal penetration scenario.
3 0.8 lb SOo/Million Btu - Oil
1.2 1b S02/Mi11i on Btu - Coal
4 90 Percent Reduction
^ These are the control costs associated with the steam generating unit NSPS.
They are measured as additions to the baseline costs.
-------
XV
In addition, it was concluded that a 90 percent reduction requirement
would not impair the ability of firms to raise sufficient capital to
construct fossil fuel-fired steam generating units.
VII. BENEFIT ANALYSIS
The benefit analysis quantifies some of the economic benefits attribut-
able to health and welfare improvements of SO2 reductions. Two regulatory
alternatives were analyzed — low sulfur fuel and percent reduction --
and compared to the regualtory baseline. Benefit coverage includes
visibility improvement due to sulfate (SO4) reductions, morbidity reductions
and household soiling reductions due to reduced levels of PM, and residential
materials damage, morbidity, and agricultural yield reductions due to SO2
reductions. Geographic coverage for SO4 and PM is the 31 eastern state
region. Geographic coverage for SO2 includes 113 areas where rrew industrial
boilers are projected to be built in 1990. All benefits are estimated
for the year 1990 -- that is, they represent the benefits that would
accrue from all boilers subject to the NSPS, that were on line in 1990.
Just as important as the benefit categories that are covered in part or in
total are those categories not covered at all. Table 2 lists all potential
benefit categories for the proposed standards and indicates which ones
are analyzed in this RIA. As can be seen, there are many categories not
included in this analysis. Although the inclusion of these benefit
categories would undoubtedly increase the benefit numbers, the magnitude
«f this increase is not known. Potentially significant benefits are
omitted from the nonhuman biological effects category (e.g., acid deposition)
and from the SO4 health effects category. In addition, for the benefit
categories that are included, there remains incomplete coverage. For
example, the SO2 benefit estimates include calculations for reductions
-------
XVI
Tabl e 2
Industrial Boiler NSPS
Potential Benefit Categories
SO?
th Effects
Mortality
Due
to
Chronic Exposure
X
Mortality
Due
to
Acute Exposure
X
Morbidity
Due
to
Chronic Exposure
~
Morbidity
Due
to
Acute Exposure
~
ing and Materials
Damage
- Residential Facilities *
- Commercial, Institutional X
and Industrial Facilities
Climate and Visibility Effects
- Local Visibility NA
- Non-Local Visibility NA
- Climate X
- Visibility at Parks NA
- Transportation Safety NA
Non-Human Biological Effects
- Agriculture *
- Forestry X
- Fishing X
- Ecosystem X
X are categories not analyzed in this study.
* are categories analyzed in this study.
NA are categories not applicable to the pollutant.
-------
xvi i
in residential materials damage, but no estimates are included for commercial,
institutional, or industrial facilities materials damage.
The benefit estimates are exhibited in Table 3. The benefits associated
with the percent reduction alternative exceed the benefits associated with
the low sulfur fuel alternative. This reflects the greater improvement in
air quality that can be achieved when percent reduction is required. Average
and incremental benefits per ton are displayed in Tables 4 and 5, respectively.
The average benefits per ton are extremely similar for the regulatory alter-
natives. Whereas the annualized benefits for the percent reduction alternative
are approximately 33% greater than benefits for the low sulfur fuel alternative,
the tons of SO2 reduced under the percent reduction alternative are also
approximately 30% greater than those under low sulfur fuel. Consequently,
the benefit per ton estimates are almost identical.
VIII. BENEFIT-COST ANALYSIS
A net benefit analysis (benefits minus costs) was conducted for the
two regulatory alternatives examined. As can be seen in Table 6, there
are cases where the annualized net benefits are negative. Table 7 presents
the average costs and benefits per ton and Table 8 presents the incremental
costs and benefits per ton for the percent reduction alternative over the
low sulfur fuel alternative.
The results of the benefit-cost analysis depend, of course, on the
validity and scope of the estimates of both benefits and costs associated
with the regulatory alternatives. The wide range of variability in the
estimated benefits is discussed in Chapter VI. Limitations also exist on
the cost side. While all known private control costs associated with the
regulatory alternatives have been estimated, complete coverage of all private
and social costs is uncertain. The results of the benefit-cost analysis are
therefore limited in scope.
-------
Table 3
Annualized Benefits*
(Thousands of 1983 dollars)2
Regulatory Alternatives
Industrial Boiler SO?
Benefit Category
Low Sulfur Fuel3
1 2 3
Percent Reduction3
1 2 3
SO4 Visibility
Local
8,000
21,000
41,000
12,000
31,000
61,000
Non-Local
3,000
11,000
25,000
4,000
15,000
35,000
PM Morbidity & Soiling
23,000
25 ,000
26,000
35,000
38,000
40,000
S02 Health
2,000
2,000
2,000
2,000
2,000
2,000
SO2 Welfare
Materials Damage
26,000
26,000
26,000
40,000
40,000
40,000
Agriculture
200
200
200
200
200
200
Total
62,000
85,000
120,000
93,000
126 ,000
178,000
1 The benefit assessment only includes a subset of potential benefits.
2 Benefits are for the year 1990, in 1983 dollars.
3 A range of benefits (1,2,3) is presented for S04 and PM to reflect sensitivity analysis on the air quality
modeling and economic valuation inputs.
-------
xi x
Tabl e 4
Average Benefits per Ton*
Totals (SOa, PM, and SO?)
Regulatory B/T
Alternative^ (1983 $)
Low Sulfur Fuel/1 450
Low Sul fur Fuel/2 640
Low Sulfur Fuel/3 930
Percent Reduction/1 470
Percent Reduction/2 660
Percent Reduction/3 950
1 The "tons of S0£ reduced" used to calculate B/T differ between SO* and
PM and SO2. This reflects differences in geographic coverage. SO4
and PM only cover 31 eastern state region.
2 A range of benefits (1,2,3) is presented for SO4 and PM to reflect
sensitivity analyses on the air quality modeling and economic valuation
inputs.
-------
XX
Table 5
Incremental Benefits per Ton*»2
Totals (S0A, PM, and SO^
B/T
(1983 $)
Percent Reduction/1 510
Percent Reduction/2 690
Percent Reduction/3 1000
1 Percent reduction over low sulfur fuel.
^ The "tons of SO2 reduced" used to calculate B/T differ between SO4
and PM and SO2. This reflects differences in geographic coverage.
SO4 and PM only cover the"31 eastern state region.
A range of benefits (1,2,3) is presented for SO4 and PM to reflect
sensitivity analysis on the air quality modeling and economic valuation
inputs.-
Table 6
Net Benefit Analysis
(thousands of 1983 dollars)
Regul atory
A1ternati ve*
Total
Annuali zed
Benefits
Total
Annualized
Control Costs
Annualized
Net Benefits
Low Sul fur Fuel /I
62,000
58,000
4,000
Low Sul fur Fuel /2
85,000
58,000
27 ,000
Low Sul fur Fuel /3
120,000
58,000
62,000
Percent Reduction/1
93,000
135,000
-42,000
Percent Reduction/2
126,000
135,000
-9,000
Percent Reduction/3
178,000
135,000
43,000
* A range of benefits (1,2,3) is presented for SO4 and PM to reflect sensitivity
analysis on the air quality modeling and economic valuation inputs.
-------
xxi
Table 7
Average Control Cost Per Ton and Benefit Per Ton!
(1983 $)
Average Average
Regulatory Alternative^ C/T B/T
Low Sulfur Fuel/1 340 450
Low Sulfur Fuel/2 340 640
Low Sulfur Fuel/3 340 930
Percent Reduction/1 560 470
Percent Reduction/2 560 660
Percent Reduction/3 560 950
* The "tons of SO2 reduced" used to calculate B/T differ between SO* and
PM and SO2. This reflects differences in geographic coverage. SO4 and
PM only cover 31 eastern state region.
^ A range of benefits (1,2,3) is presented for SO^ and PM to reflect
sensitivity analysis on the air quality modeling and economic valuation
inputs.
-------
xxi i
Tabl e 8
Incremental Control Cost Per Ton and Benefit Per Ton
(1983 $)
Incremental Incremental
C/T3 B/T3
I
Percent Reduction/1 1100 510
Percent Reduction/2 1100 690
Percent Reduction/3 1100 1000
• The "tons of SO2 reduced" used to calculate B/T differ between SO* and
PM and SOg. This reflects differencies in geographic coverage. SO4 and
PM only cover 31 eastern state region.
A range of benefits (1,2,3) is presented for SO^ and PM to reflect sensitivity
analysis on the air quality modeling and economic valuation inputs.
3 Percent reduction over low sulfur fuel.
-------
xxiii
IX. CAVEATS AND QUALIFICATIONS
Many caveats and qualifications with regard to the air quality
modeling and benefit analysis are presented in the body of the RIA. Some
of the more important qualifications are mentioned below along with the
potential impacts on the results of the analysis.
The most important qualification concerns the benefit coverage. As
can be seen in Table 2 there are many, potentially significant categories
omitted or only partially covered.' Thus, total benefits will actually be
greater than those reported, and net benefits will be larger. It is
impossible to know the magnitude of this effect with any precision.
Geographic coverage is also limited. Since SO4 and PM calculations
only include the 31 eastern states, potentially significant benefits from
the western states and Canada are omitted. The inclusion of these areas
in the benefit calculations could only increase the annualized benefits
and net benefits.
-------
1-1
I. STATEMENT OF NEEDS AND CONSEQUENCES
The proposed standards implement Section 111 of the Clean Air Act
and are based on the Administrator's determination that emissions from
industrial-commercial-institutional steam generating units cause, or
contribute significantly to, air pollution which may reasonably be expected
to endanger public health or welfare. The intent is to require new,
modified, and reconstructed industrial-commercial-institutional steam
generating units to control emissions to the level achievable by the best
demonstrated technological system of continuous emission reduction,
considering costs, nonair quality health, and environmental and energy
impacts. The proposed standards apply to steam generating units capable
of combusting more than 29 MW (100 million Btu/hour) heat input.
On August 21, 1979, a priority list for development of additional
new source performance standards (NSPS) was published in accordance with
Sections 111(b)(1)(A) and 111(f)(1) of the Clean Air Act Amendments of
1977. This list identified 59 major stationary source categories that
were judged to contribute significantly to air pollution that could
reasonably be expected to endanger public health or welfare. New source
performance standards had previously been developed for utility boilers
(see EPA 450/3-79-021). Fossil fuel-fired industrial steam generating
units ranked eleventh on this priority list of sources for which new
source performance standards would be established in the future. Fossil
fuel-fired industrial steam generating units were the highest ranked
source of particulate matter and sulfur dioxide emissions, and the second
highest ranked source of nitrogen oxide emissions when the priority list
of source categories not regulated by NSPS was published.
-------
1-2
Fossil and nonfossil fuel-fired steam generating units are significant
sources of emissions of three major pollutants: particulate matter (PM),
sulfur dioxide (SO2), and nitrogen oxides (NOx). As discussed in
"Priorities for New Source Performance Standards Under the Clean Air Act
Amendments of 1977" (EPA-450/3-78-019), in the absence of the proposed
standards, emissions from industrial-commercial-institutional steam
generating units with a heat input capacity of 3 through 73 MW (10 through
250 million Btu/hour) would contribute approximately 9 percent of the
total national SO2 emissions from major sources in 1990.
The proposed regulation would require all coal-fired industrial-
commercial-institutional steam generating units equipped with conventional
SO2 emission reduction technologies to achieve a 90 percent reduction in
SO2 emission and meet an emission limit of 516 ng SO2/J (1.2 lb S02/million
Btu) heat input. Coal-fired steam generating units using an emerging
technology for SO2 control would be required to achieve a 50 percent
reduction in SO2 emissions and meet an emission limit of 258 ng SO2/J
(0.6 lb S02/mi11ion Btu) heat input. Steam generating units firing
mixtures of coal with any other fuel (except oil) and having an annual
capacity utilization factor for coal of less than 30 percent (0.30) would
be required to meet an SO2 emission limit of 516 ng SO2/J (1.2 lb
S02/million Btu) heat input.
The proposed standards would require all oil-fired industrial-
commercial-institutional steam generating units equipped with conventional
SO2 emissions and meet an emission limit of 344 ng S02/million Btu) heat
input. Oil-fired steam generating units using an emerging technology to
control SO2 emissions would be required to achieve a 50 percent reduction
-------
1-3
in SO2 emissions and meet an emission limit of 172 ng SO2/J (0.4 lb
SOg/mi11ion Btu) heat input. Steam generating units firing oil and
meeting an SO2 emission limit of 86 ng SO^/J (0.2 lb S02/million Btu) or
less would not be required to achieve a further percent reduction in SO2
emissions. The emission limit for steam generating units firing a mixture
of coal and oil would be determined by proration.
The proposed standards could reduce emissions of particulate matter
from a typical oil-fired industrial-commercial-institutional steam
generating unit by about 49 Mg/year (54 tons/year).
In the fifth year after this NSPS becomes applicable, nationwide
emissions of SO2 could be decreased by between 240,000 and 281,000 Mg/year
(263,000 and 310,000 tons/year) compared with projected emission levels
under the regulatory baseline.
National ambient air quality standards have been established for SO2
because of its known adverse effects on public health and welfare. Im-
pacts of this pollutant have been documented in the SO2 Criteria Document
prepared under Section 108 of the Clean Air Act. These effects are the
primary basis for the determination that emissions from industrial-commercial-
institutional steam generating units will continue to be located in urban
areas where a large population will be exposed to the emissions. Therefore,
the present concentration of steam generating units in industrialized
urban areas will continue to contribute to local and regional air pollution.
For these reasons the source category of industrial-commercial-institutional
steam generating units was selected for development of standards of ,
performance.
-------
1-4
REFERENCES
U.S. EPA (1985). Standards of Performance for New Stationary Sources,
Industrial-Commercial-Institutional Steam Generating Units, Draft
Preamble.
-------
II-l
II. ALTERNATIVES EXAMINED
A. Introduction
This section briefly presents potential alternatives to the proposed
regulation. The E.O. requires that EPA examine regulatory alternatives
that would mitigate or eliminate the environmental problem as described
in the Statement of Needs and Consequences. The format of this section
follows the requirements of E.O. 12291 that the following alternatives be
examined:
a) no further regulation,
b) other regulatory approaches,
c) market oriented approaches,
d) alternative approaches within the scope of present legislation.
Although E.O. 12291 requires that all alternatives be examined, only the
most promising ones need be analyzed in detail.
B. No Further Regulation
A decision not to further regulate steam generating units would
result in a reliance on current control levels. The regulatory baseline
represents the current level of control required by existing State
implementation plans and the existing NSPS (40 CFR Part 60, Subpart D)
applicable to steam generating units of more than 73 MW (250 million
Btu/hour) heat input. If only current control levels were maintained,
SO2 emissions would be 240,000 to 281,000 Mg/year (263,000 and 310,000
tons/year) greater than they would be under the proposed regulation, in
the fifth year after proposal. The Administrator has determined that
this level of air pollution could reasonably be expected to endanger
public health and welfare. For this reason this source category was
-------
11-2
selected for development of standards of performance. This option is
discussed in later sections of the RIA as the baseline for the evaluation
of other regulatory alternatives.
C. Other Regulatory Approaches
The regulatory alternative to NSPS for controlling emissions from steam
generating units is reliance on Best Available Control Technology (BACT) and
Lowest Achievable Emission Requirement (LAER) in attainment and nonattainment
areas respectively. BACT and LAER determinations are made on a case by
case basis. Through review and analysis of current BACT and LAER levels
it has been determined that the applicable NSPS sets a "floor," or minimum,
on the level of control that is required. Therefore, the existence and
stringency of the applicable NSPS has a significant influence on the
eventual BACT or LAER determination. The importance of NSPSs to the BACT
and LAER process takes on even greater significance in practice, given
that the majority of BACT and LAER determinations appear to be set "by
default" equivalent to the applicable NSPS.
D. Market Oriented Alternatives
There are several market oriented approaches that can be considered
alternatives to regulating steam generating units under NSPS. These
approaches include pollution taxes, marketable permits, and subsidies.
1) Taxes
This policy would require a tax on each unit of pollution emitted.
Firms would then choose the amotmt of pollution abatement that minimized
their total cost, including the pollution tax. Pollution is abated until
the marginal cost of abatement is equal to the pollution tax. The regula-
tory agency would have to set the level of the tax in a manner that would
-------
n-3
result in the desired level of air quality. While conceptually possible,
and in terms of economic efficiency appealing, the pollution tax that
ensures the acceptable level of air quality would have to be done on a
case-by-case basis.
2) Permits
A permit system would allow a pollution source to purchase a
permit in order to emit a specific amount of a pollutant over a specified
period of time at a specified location. A fixed number of permits would
be issued and auctioned off to the highest bidders. Alternatively, the
permits could be distributed among the sources, who could then trade
these permits as they see fit. If the number of permits and emission
allowances were correctly fixed in advance, then the desired air quality
would be achieved. Again, this may be difficult in areas with numerous
and diverse sources.
3) Subsidies
A subsidy system pays sources for each unit of pollution that
they do not emit. This can take the form of direct payments or tax
credits. Subsidies and taxes are similar in that both increase the
opportunity cost of polluting. Each unit of pollution emitted has attached
to it the "cost" of the foregone subsidy. Thus, the subsidy is similar
to a tax, except in two respects:
(a) Administratively, there is the problem of determining the actual
abatement at each source. There must be a determination of what
pollution levels would have been in the absence of the subsidy,
and this determination must be adjusted as conditions change.
These determinations are difficult and may depend upon information
-------
11-4
from the polluters who obviously have the incentive to understate
their intended emissions.
(b) The long-run effects of subsidies may be quite different than
for permits or taxes since the former increase profit levels or
incomes for polluting industries and the latter decrease them.
Thus, in the long-run, subsidies would result in firms entering
the polluting industry, but permits or taxes would cause firms
to leave. Consequently, a fixed subsidy rate would tend to
create pollution levels beyond those that are economically
efficient. Also, subsidies in the form of investment tax
credits, would likely increase dependence on capital intensive
solutions.
4) Market Oriented Approaches and Clean Air Act Requirements. The
Clean Air Act (Section III) explicitly requires that new, modified, and
reconstructed stationary sources such as industrial-commercial-institutional
steam generating units control emissions to the level achievable by the
best demonstrated technological system of continuous emission reduction,
considering costs, nonair quality health, and environmental and energy
impacts. This precludes the use of pollution taxes as alternatives to
NSPS unless the exact level of taxation could be established that would
achieve pollution abatement equivalent to the level achievable by best
demonstrated technology. The use of subsidies and permits are also
rejected because of the Act requirements. In the case of permits, the
number of permits and emission allowances would have to be correctly
fixed in advance, that would achieve the same level of emission control
as would otherwise be achieved by the best demonstrated technology.
-------
11-5
E. Regulatory Alternatives Within the Scope of Present Legislation
A range of regulatory alternatives has been carefully evaluated
based on the use of low sulfur fuels to reduce SO2 emissions or the use
of flue gas desulfurization (FGD) systems to achieve a percent reduction
in SO2 emissions. A total of six alternative control levels were examined
and compared to the regulatory baseline. The six regulatory alternatives
were each analyzed under two different energy scenarios. This energy
sensitivity analysis was performed because the magnitude of the economic,
environmental, and energy impacts associated with various regulatory
alternatives for new steam generating units is primarily a function of
the type of fuel selected for each of the units. The type of fossil fuel
selected for each new steam generating unit is a function of the projected
prices for coal, oil, and natural gas, as well as the costs associated with
environmental regulations. For more information on the energy prices
used in this analysis see, "Regional Forecasts of Industrial Residual
Fuel Oil and Natural Gas Prices," July 1984. The high oil penetration
scenario reflects the Agency's best projection of future coal, oil, and
natural gas prices. Oil prices are relatively low and natural gas prices
are generally at or above the price of low sulfur residual oil. Under
this energy price scenario residual oil and natural gas compete for the
industrial steam generating unit energy market, with residual oil achieving
a slightly larger share. Coal does not effectively compete in this
market.
In response to concerns that this energy pricing scenario might
underestimate coal penetration in the new industrial steam generating
-------
11-6
unit energy market, an alternative energy scenario was developed to yield
higher coal penetration. In this energy scenario, coal prices were
assumed to remain the same as those discussed for the high oil penetration
energy scenario. The oil prices used were higher than those used in the
high oil penetration scenario and were also used to forecast natural gas
prices. In this high coal penetration scenario coal and natural gas
compete for the steam generating unit energy market with coal capturing a
slightly larger share than natural gas. Oil does not effectively compete
in this market due to the relatively low coal and natural gas prices.
For a more detailed discussion of the regulatory alternatives and the
energy scenarios see the Preamble (EPA-450/3-82- ).
Due to time and resource constraints, for purposes of this RIA two
regulatory alternatives were examined. These regulatory alternatives
were the two most viable candidates to serve as the basis for the proposed
NSPS. Each alternative is analytically compared to the current
regulatory baseline. Both regulatory alternatives were analyzed under the
high oil penetration scenario. This energy scenario was chosen because
it most accurately reflects current projections of future energy prices.
1. Regulatory Baseline
The regulatory baseline is defined by existing State Implemen-
tation Plans and the existing NSPS (40 CFR, Part 60, Subpart D)
for large steam generating units of more than 73 MW (250 million
Btu/hour) heat input.
2. Low Sulfur Fuel
This alternative is based on the use of low sulfur fuel to
reduce emissions to 344 ng SO2/J (0.8 lb S02/million Btu) heat
-------
n-7
input from oil combustion and to 516 ng SO2/J (1.2 lb S02/million
Btu) heat input from coal combustion.
Percent Reduction
This alternative is based on the use of FGD systems to
achieve a 90 percent reduction in SO2 emissions from both oil and
coal combustion. This alternative would generally reduce emissions
from coal combustion to less than 258 ng/J (0.6 lb S02/mi11ion
Btu) heat input and from oil combustion to less than 129 ng/J
(0.3 lb S02/million Btu) heat input.
-------
11-8
REFERENCES
Regional Forecasts of Industrial Residual Fuel Oil and Natural Gas Prices
July 1984 Energy and Environmental Analysis, Inc. for the U.S. Environ-
mental Protection Agency.
-------
Ill—1
III. AIR QUALITY DATA
A. Introduction
The estimation of economic benefits for the two regulatory
alternatives requires the calculation of changes in population exposure
from the baseline to each alternative. This, in term, requires knowledge
of projected air quality changes.
This section provides a description of the three types of air quality
data used for the benefit analysis. The first is SO4. data used for visual
range improvement benefits. The geographic coverage is 31 eastern states.
The air quality changes result from industrial boiler compliance with the two
regulatory alternatives. PM air quality data is used to estimate PM
related morbidity risk reduction and residential soiling reduction benefits.
The geographic coverage'for PM benefits is also the 31 eastern state
region. This second type of air quality data uses the first type of air
quality data as an input but transforms the changes in SO4. to changes in
TSP. The third type of air quality data is SO2 information used to
estimate materials damage reductions, agricultural yield reductions, and
morbidity reduction benefits. The geographic area for the SO2 ambient
air quality data is 113 areas, spread across the country, where new
industrial boilers are projected to be located in 1990.
B. SO4. Air Quality Assessment
The standards analyzed in this RIA all relate to the local scale
impacts of sulfur oxide sources. However, it is generally recognized
that SO2 emissions also have regional scale effects. In the atmosphere
SO2 can be transformed into sulfates (SO4) and both can be transported
for long distances (>300 km). As a part of the overall RIA the effect of
SO2 emissions reductions on sulfates and visual range were modeled on a
-------
111-2
regional scale. The modeling of source-receptor relationships on any
geographic scale and for any pollutant is difficult and involves some
uncertainty. Modeling of regional scale transport, dispersion, chemical
transformation and removal of sulfur oxides is quite difficult and involves
a number of uncertainties. Nevertheless over the past several years a
number of models have been developed. Although no models have been
tested using an adequate array of air quality meteorology data, comparisons
of model results with long-term data suggest that these models are now at
a stage of development that they can be used to provide some insight into
the magnitude and nature of air quality changes that might result from
emission reduction scenarios. For purposes of this RIA, transfer matrices
were derived from the Advanced Statistical Trajectory Regional Air Pollution
(ASTRAP) model and visual range was estimated outside the model/matrix
using the matrix generated SO4. concentrations. ASTRAP is a regional
scale SO2 model that considers transport, dispersion, chemical transfor-
mation and removal of sulfur oxides. As used in this analysis, ASTRAP
models SO2 source -receptor relationships in the 31-eastern state
region. A description of the ASTRAP matrix follows in Table 111-1.
Transfer matrices assume that concentrations at a given receptor
equal the sum of all partial contributions. Thus, the concentration C at
receptor A could be represented as:
CA = TA1 * E1 + TA2 * E2 + ••• + Taj * Ej
where Ej is the emission rate from the source region and T/y is a propor-
tionality coefficient relating region T emissions to region A concentrations.
An array of these T^j coefficients can be generated from running the full model
and constitutes a transfer matrix. Transfer matrices are extremely inexpensive
and easy to apply compared to full regional scale model runs. However, they
-------
Ill
TABLE
ASTRAP
-3
III-l
Matrix1
CRITERIA
ASTRAP
1)
Model Domain
48 States, plus Canada
2)
T ransformation
Rates2
Winter 0.5^
Spring 1.0
Summer 2.0
Fall 0.7
3)
Return F1 ows
Addresses return flows from Atlantic
4)
Treatment of
Emi ssion
Releases
Considers low level releases as well as
elevated releases'
5)
Background Factors
None
6)
Meteorology
1976 - 1981
7)
Western State
Emi ssions
(Outside 31-
State Region)
1995 projected; unchanged by alternative
8)
Canadian Emissions
1995 projected; unchanged by alternative
1 Advanced Statistical Trajectory Regional Air Pollution.
2 Expressed as % of SO2 converted to sulfate per hour.
^ Approximate daily averages. Actual rates exhibit strong diurnal variation.
-------
111-4
also carry with them a number of limitations in addition to those inherent to
the models themselves:
o The matrix used employs state level source and receptor regions.
Therefore, there is some "averaging out" of differences which might
be apparent with the finer grids used in the models themselves.
This may be a significant limitation in this analysis because the
emission changes being modeled vary from source to source,
o The models/matrices used are linear. There is substantial dis-
agreement as to whether the processes being modeled are in fact
linear. If non-linear processes were found then the estimates
used here would not be accurate,
o In modeling SO4 levels resulting from the two regulatory
alternatives only emission changes in the eastern 31 state region
were modeled. Western U.S. and Canadian emissions were modeled
at their 1995 base level. This was done because the data was
readily available for the 31 state region.
In considering the benefits calculations and results, the reader
should keep these limitations in mind.
A more complete discussion of regional modeling and transfer matrices
can be found in the Memorandum of Intent on Transboundary Air Pollution
(1982).
C. SO4 Air Quality Results
The 1995 base case sulfate emissions, estimated from the ASTRAP
matrix, for the 31 eastern states, are displayed in Table 111-2. The base
case sulfate estimates represent the estimated sulfate levels associated
with normal growth and retirement of sources but with no change in the
-------
111-5
TABLE 111-2
1995 Base Case Sulfate Estimates
(pg/m3 SO4)
STATE ASTRAP
ug/m3
A1 abama 5.2
Arkansas 3.1
Connecticut 8.2
Delaware 11.5
Florida 4.0
Georgia 5.7
111 inois 5.5
Indiana 8.1
Iowa 2.7
Kentucky 8.5
Louisiana 2.6
Maine 3.6
Mary! and 10.9
Massachusetts 6.5
Michigan 4.5
Minnesota 1.4
Mississippi 3.7
Missouri 4.0
New Hampshire 5.6
New Jersey 10.5
New York 7.5
North Carol ina 7.8
Ohio 9.6
Pennsylvania 11.3
Rhode Island 7.1
South Carolina 6.8
Tennessee 6.9
Vermont 5.1
Virginia 9.3
West Vi rgi ni a 11.5
Wisconsin 3.1
-------
111-6
regulatory scenario. Table 111-3 represents the reduction in sulfate
estimates due to the low sulfur fuel regulatory alternative, along
with the percent change from the 1995 baseline. Table 111-4 represents
the same information for the percent reduction alternative. As expected,
both alternative control levels result in SO4 reductions, with the percent
reduction alternative resulting in the largest reduction from the 1995 basecase.
The estimated sulfate levels were in turn used to calculate visual
range. This calculation uses the results from the ASTRAP model (Tables
111-3 and 111-4) with an average nonsulfate/sulfate light extinction ratio
of 1.25. The estimated 1995 baseline visibility is presented in Table
111-5 along with the visibility improvements due to the low sulfur fuel
and percent reduction regulatory alternatives.
0. PM Air Quality Assessment
The air quality information developed for the SO4 benefit assessment
is also used for the PM benefit assessment. Specifically, the SO4 changes
for each state described in the last section are used. The SO4 changes
are translated into TSP changes by multiplying by 1.4 for the low estimate,
1.5 for the middle estimate and 1.6 for the high estimate. These factors
(Bachmann, 1985) are used to account for the weight of ammonia and water
change in TSP weight associated with the change in SO4. No estimates of
direct PM emission reductions for alternative standards due to SO2 controls
such as scrubbing were calculated. Therefore no estimates of PM ambient
air reductions due to reduction of PM emissions are included in the analysis.
E. SO2 Air Quality Assessment
This section describes the principal procedures and assumptions used to
calculate the SO2 air quality impacts of the two regulatory alternatives.
The first subsection describes the development of the emissions inventory,
-------
111-7
TABLE 111-3
1995 Sulfate Emissions
Low Sulfur Fuel Alternative
(pg/m3 and % change from 1995 Base Case)
STATE
ASTRAP
gg/m3
% Change
A1 abama
5.2
- .8
Arkansas
3.1
-1.3
Connecticut
8.2
- .2
Del aware
11.5
- .2
F1 orida
3.9
- .6
Georgia
5.7
- .5
111inois
5.5
- .3
Indiana
8.1
- .2
Iowa
2.6
- .8
Kentucky
8.5
- .3
Louisiana
2.5
-3.9
Maine
3.6
- .1
Maryland
10.9
- .2
Massachusetts
6.5
- .2
Michigan
4.5
- .3
Minnesota
1.4
- .3
Mississippi
3.6
-1.6
Missouri
4.0
- .9
New Hampshire
5.6
- .2
New Jersey
10.5
- .2
New York
7.5
- .2
North Carolina
7.8
- .4
Ohio
9.6
- .3
Pennsylvania
11.3
- .2
Rhode Island
7.1
- .2
South Carolina
6.8
- .5
Tennessee
6.9
- .5
Vermont
5.1
- .2
Vi rginia
9.3
- .3
West Vi rginia
11.5
- .3
Wi sconsin
3.1
- .3
-------
111-8
TABLE 111-4
1995 Sulfate Emissions
Percent Reduction Alternative
(yg/m3 and % change from 1995 Base Case)
STATE
pg/m3
ASTRAP
% Change
A1 abama
5.1
-1.0
Arkansas
3.1
-1.4
Connecticut
8.2
- .4
Del aware
11.5
- .4
F1 orida
3.9
- .8
Georgia
5.7
- .6
111inois
5.5
- .5
Indiana
8.1
- .4
Iowa
2.6
-1.0
Kentucky
8.5
- .5
Louisiana
2.5
-4.0
Maine
3.6
- .3
Mary! and
10.9
- .4
Massachusetts
6.5
- .4
Michigan
4.4
- .6
Minnesota
1.4
- .7
Mississippi
3.6
-1.7
Missouri
4.0
-1.0
New Hampshire
5.6
- .3
New Jersey
10.5
- .4
New York
7.5
- .4
North Carolina
7.8
- .7
Ohio
9.6
- .4
Pennsylvania
11.3
- .4
Rhode Island
7.0
- .4
South Carolina
6.7
- .7
Tennessee
6.9
- .6
Vermont
5.1
- .3
Vi rginia
9.2
- .5
West Virginia
11.5
- .4
Wisconsin
3.1
- .7
-------
111-9
TABLE 111-5
ASTRAP 1995 Estimated Visibility*
(medium annual visual range in km)
State
Baseline
LSF
Percent Reduction
A1abama
13.4
13.5
13.5
Arkansas
12.6
12.7
12.7
Connecticut
27.5
27.5
27.5
Del aware
24.9
24.9
24.9
F1 orida
17.3
17.3
17.3
Georgia
16.9
17.0
17.0
111inois
17.7
17.7
17.7
Indiana
;16.8
16.8
16.8
Iowa
19.2
19.3
19.3
Kentucky
18.7
18.7
18.7
Louisiana
12.2
12.4
12.4
Maine
35.1
35.1
35.1
Maryl and
24.7
24.7
24.7
Massachusetts
29.3
29.3
29.3
Michigan
18.4
18.4
18.5
Mi nnesota
21.9
21.9
21.9
Mississippi
13.0
13.0
13.0
Missouri
16.7
16.8
16.8
New Hampshi re
31.3
31.4
31.4
New Jersey
25.3
25.3
25.3
New York
26.8
26.8
26.8
North Carol ina
16.8
16.8
16.9
Ohio
16.7
16.7
16.8
Pennsylvani a
19.1
19.1
19.1
Rhode Island
28.1
28.1
28.1
South Carolina
16.8
16.8
16.8
Tennessee
15.4
15.4
15.4
Vermont
31.9
31.9
31.9
Vi rginia
24.3
24.3
24.3
West Vi rginia
19.6
19.6
19.6
Wisconsin
18.4
18.4
18.5
1 Assumes an average nonsul fate/sul fate light extinction ratio of 1.25.
-------
111-10
the second the determination of the study areas, and the third the modeling
approach.
1. Emissions Inventory
a) Data Bases
Four data bases were used to develop the emissions inventory:
o A point source inventory based on MAP3S (referred to
as the EES inventory),
o The unit inventory of utilities from the Advanced Utility Simulation
Model (AUSM).
o The inventory of new industrial boilers provided by
Energy and Environmental Analysis (EEA) based on the
Industrial Fuel Choice Analysis Model (IFCAM),
o The current county-level area source inventory from
the National Emissions Data System (NEDS).
Table 111-7 shows which data base was used as the basis for the inventory for
different sources.
TABLE 111-6
Emissions Data Bases and Growth
Source Data Base
Type Category Age EES EEA Growth
(MAP3S) AUSM (IFCAM) NEDS Applied?
Point Utility Existing x x No
New x No
Industrial Existing x No
boilers
New x No
Other - x Yes
Area All x Yes
-------
III-ll
b) Inventory Development
The development of the inventory can be considered as having taken
place in a series of steps.
Utility data was taken from of the AUSM2 unit inventory. AUSM
contains data on individual utility boilers. Among other data, AUSM
provides:
o Location by state and county,
o Plant name,
o Startup year if unit is a planned new unit,
o Retirement year, and
o Fuel and fuel consumption data.
However, AUSM did not include the stack parameters required for modeling nor
a complete set of emissions.
1. Stack parameters (height, temperature, diameter, volume flow) and
industrial steam generating unit emissions were added to the AUSM utility
data. The values used were based on information in the Utility Boiler
Background Information Document (EPA 450/3-79-021) and EPA engineering
judgement. Emissions were based on a model 500 MMBtu/hour boiler and were
made proportional to capacity. It was assumed that Subpart D applied to
utility boilers with startup dates between 1977 and 1986, inclusive, and
that Subpart Da applied to utility boilers with startup dates of 1987 and
beyond.
The MAP3S inventory was developed by Brookhaven National Laboratory
under the MAP3S/RAINE program. The point source data in MAP3S are based
primarily upon data taken from EPA's NEDS. Some of the NEDS data were
revised or supplemented during the development of MAP3S (Benkovitz, 1980).
In developing the EES inventory, missing data were defaulted and inconsistent
-------
111-12
data were changed as described in Brubaker and Smith, 1984. The nominal
year of this inventory is 1978.
2. The EES point inventory was used as the basic data set for
sources other than utilities and industrial boilers. Where possible,
utility boilers in the EES inventory were replaced by utility boilers
from AUSM. These matches were made based on state, county, plant name,
and boiler size. Where matches could not be made, the EES data was
retained. The retirement or startup year were retained with the utility
boiler data based on AUSM.
3. The EEA industrial boiler data was processed to add stack
parameters (height, temperature, volume flow) based on information in the
Summary of Regulatory Analysis (EPA 450/3-86-) and EPA engineering judgement.
Costs and emissions were based on data in the EEA/IFCAM boiler file.
Boilers burning natural gas in all three cases (base case, low sulfur
fuel, and percent reduction were deleted because their SO2 emissions are
negl igibl e.
4. The EEA files locate boilers only by federal region. They also
associate each boiler with a 2-digit SIC category. Within a particular
SIC category and region, the EEA boilers were sited at existing plants.
This siting was accomplished in stages. First, the total coal and oil
consumption (Btu's/year) in industrial boilers was determined for each
plant in the region/SIC. Then each EEA boiler was randomly assigned to a
specific plant with the assignment weighted by the consumption of that
boiler's base-case fuel type. Thus, if plant l's industrial boilers
fired twice as many Btu's of coal as the boilers in plant 2, plant 1
would be twice as likely as plant 2 to be chosen as the site of a particular
EEA boiler fueled by coal in the base case.
-------
111 -13
5. The sources in the EEA/MAP3S were screened to eliminate those
with a maximum 1-hour SO2 concentration exceeding 4500 gg/m. Such large
maximum concentrations are believed to be due to problems in the data
base.
6. The county-level area source emissions from NEDS were used
without change.
At this point, the inventory can be considered as consisting of the
following:
0 New industrial boilers from EEA/IFCAM with stack parameters added
and sited randomly by region/SIC,
0 Utility boilers from AUSM2 with stack parameters and industrial
steam generating unit emissions added and including retirement
and startup years,
0 Some additional utility boilers that could not be matched to AUSM
from the EES inventory,
0 Other point sources from the EES inventory based on MAP3S with a
1-hour SO2 impact <4500 gg/nv*, and
0 County-level area source emissions from NEDS.
These emissions were projected to 1990 and the 113 study areas were
identified prior to modeling.
C. Projected Emissions
In projecting emissions to 1990, no growth rate was applied to
the following source categories (See Table II1-7):
0 Utility boilers because growth was assumed to be represented
adequately by the AUSM units scheduled to start-up before 1990,
0 Industrial boilers because growth was assumed to be represented
adequately by the boilers in the EEA/IFCAM files, and
-------
II1-14
o Primary copper, lead, and zinc smelters for which no new sources
are anticipated.
All other point source growth was assumed to occur in place. Growth
and retirement were projected on a source-by-source basis using two-digit
SIC state-specific growth rates and two-digit SIC national-average retirement
values (Smith and Brubaker, 1983). The projected growth and retirement
were required to meet applicable NSPSs. If a source's base-year emissions
exceeded 100 T/yr, the projected growth and retirement were assumed to
have to meet the BACT. NSPS's were taken from EPA's Cost-Effectiveness
File (CEF) and BACT was assumed to be the average control of the SO2
NSPS's in the CEF (97.5% removal). Area sources were grown at state-specific
population growth rates.
2. Determining Study Areas
The 113 study areas were chosen to include all the plants at which
EEA/IFCAM boilers had been sited. The study areas were determined
manually. An initial attempt was made to locate each EEA/IFCAM at the
center of a 60 km by 60 km square receptor grid with 6 receptors along
each edge giving a total of 36 receptors and a grid spacing of 12 km,
oriented with two edges running north-south. In many cases, however,
this initial location was modified to reduce the number of receptors
over water and to avoid overlapping grids. The receptor grids contained
from 24-48 receptors, had receptor spacings of 10-20 km, and had dimensions
of from 48 to 150 km.
The inventory associated with each grid for modeling purposes included
the point and area sources in all the counties within about 40-60 km of
the outer edges of the receptor grid and thus included most of the sources
which would have an appreciable impact within a particular receptor grid.
-------
111-15
It should be noted that a particular source, including new EEA/IFCAM
boilers, could be included in the inventories of 3 or perhaps more study
areas. However, no double counting occurred because there was no overlap
of the receptor grids.
3. Air Quality Modeling
a) Meteorological Data
Meteorological data specific to each of the 113 areas in the study
was not readily available. In lieu of site-specific data, eight locations
in the continental United States were selected so as to incorporate all
major geographic variables that might have a large, systematic effect on
local climate. Hourly meteorological data were obtained from the National
Weather Service for these locations, and were processed using a modified
version of the U.S. EPA meteorological data processing program (CRSTER)
to produce a dataset suitable for use in air quality modeling. In order
to run the air quality model for some given analysis area, the most
suitable of these eight meteorological data sets was selected, based upon
geographic similarity and proximity, and used without further modification.
The eight locations used in this study were San Francisco, CA; Albuquerque,
NM; Ellsworth AFB, SD; Columbus, OH; Shreveport, LA; Philadelphia, PA;
Washington, DC; and Jacksonville, FL.
An entire year of hourly meteorological data was available for each
model run. In order to reduce the computing time, so that all areas
could be handled within the scope of the program, only every fifth day
was actually modeled. Thus, annual or seasonal extreme values may have
been underestimated and mean values are subject to uncertainties arising
from this sampling procedure.
-------
111-16
b) Air Quality Model
In order to make reasonable estimates of the required air quality
variables for each of the 113 areas without the expenditure of unreasonable
amounts of computer time, an approximate procedure was used for the
estimating of the contributions from the minor background point sources.
Briefly, all "major" point sources, defined for this work as those
with emissions exceeding 300 T/year, were modeled as point sources. The
remaining minor point sources were treated as clusters, identified based
on physical proximity, and replaced by an el 1iptical-shaped "area"
source.
The calculation of the 1-hour SO2 concentration at a receptor from a
point source was done using a recently-developed expression for the mean
concentration, given that the wind direction is only known to within 10
degrees. The formula is based on the standard Gaussian model. The
dispersion coefficients and plume rise values used were computed using
USEPA-approved algorithms.
The emission inventory actually used in a model run consisted, then,
of 1) new and old utilities as points, 2) new EEA/IFCAM boilers as points,
3) old major point sources (those with annual emissions in excess of 300
T/year), 4) elliptical area sources (simulating the effects of minor point
sources, i.e., those with annual emissions less than 300 T/year) and
5) true area sources, whose emissions were arbitrarily allocated to those
of the elliptical areas in proportion to their emission rates.
-------
111-17
REFERENCES
Memorandum of Intent on Transboundar7 Air Pollution (1982). Atmospheric
Sciences and Analysis Work Group 2 Final Report, November 1982.
Bachmann, John (1985). Memorandum to Tom Walton. Subject: Calculation
of Visual Range Changes from Predicted Sulfate Changes. June 10, 1985.
Benkovitz, C. M. (1980). Emission Inventory Progress Report, Brookhaven
National Laboratory report 8NL-51378 UC-11, New York.
Brubaker, K. L., and A. E. Smith (1984). Air Quality Analyses in Support
of Regulatory Impact Analysis for S0?~NAAQS - Technical Notes,
unnumbered, Argonne National Laboratory report, Argonne, Illinois.
Smith, A. E., and K. C. Brubaker (1983). Costs and Air Quality Impacts of
Alternative National Ambient Air Quality Standards for Particulate
Matter - Technical Support Document, Argonne National Laboraotyr Report
ANL/EES-TM-229.
-------
IV. ENGINEERING COSTS
A. Introduction
National cost impacts on new industrial fossil fuel-fired steam
generating units were analyzed through the use of a computer model called
the Industrial Fuel Choice Analysis Model (IFCAM). This cost analysis is
described in detail in "Projected Impacts of Alternative Sulfur Dioxide
New Source Performance Standards for Industrial Fossil Fuel-Fired Boilers,"
March 1985. IFCAM is an energy demand model developed to evaluate fuel
choice decisions in the industrial sector over a five to 15-year forecast
horizon. IFCAM is a highly disaggregated process engineering model
designed to analyze factors affecting fuel choice decisions in industrial
boiler energy applications.
IFCAM assesses the impacts of four sets of factors affecting industrial
fuel choice: fuel prices, State and Federal energy and environmental
policy proposals, the costs associated with firing alternate fuels, and
other key model parameters such as the expected size distribution of new
industrial boilers. The model is capable of estimating the energy,
environmental, and cost impacts of alternative air emission regulations
for new industrial boilers.
B. Analytical Framework
IFCAM focuses on boiler fuel choice decisions between coal, oil, and
natural gas. The model determines fuel choices for boilers in each of
the 10 Federal regions using projected industrial fuel prices and total
fossil fuel demand that primarily have been generated by the U.S. Department
of Energy (DOE). For purposes of this analysis, the model must distinguish
between new industrial boilers, other process heaters, and existing
-------
IV-2
industrial boilers. Since this is a NSPS, the only category relevant to
this analysis is the new industrial boilers.
A very disaggregate representation of industrial energy uses is
incorporated in IFCAM. Energy use is disaggregated into nine industrial
subsectors having different growth rates and regional dispersions. Each
of these industrial subsectors is further divided into boiler and process
heat applications.
Industrial boiler energy use is disaggregated by new and existing
boilers of different sizes and capacity utilizations rates. New and
existing boilers are classified on the basis of the type of fuel they
were designed to fire and whether they can be retrofitted to fire
alternative fuels. Eight size classes are delineated for boilers. There
are five capacity utilization rate categories for boilers.
The final level of disaggregation incorporates regional detail.
Industrial boiler fuel use is divided among Air Quality Control Regions
(AQCR's). Since a large number of regions are considered by IFCAM, the
model can portray the variability in the environmental control costs of
using the various fuels to fire boilers located throughout the country.
Based on the characteristics of each combustor (size, operating
rate, pollution control requirements, etc.), capital and nonfuel operating
costs are generated for the options of choosing oil, gas, or coal. The
fuel type associated with the lowest after-tax present value (including
expected fuel expenses) of the total cost of generating energy over the
investment period is selected. As a final output, the model generates
total costs (capital, annual operating, maintenance and fuel expenses) of
using the selected fossil fuel to generate steam.
-------
IV-3
Table IV-1 presents the total annualized costs and control costs
for the baseline and two regulatory alternatives examined in this RIA.
Each alternative is analyzed using the high oil penetration energy scenario.
The incremental control costs over baseline for the low sulfur fuel
alternative are $58 million. For percent reduction, the incremental
control costs over baseline are $135 million. These costs will be used
in the following sections to analyze economic impacts and to calcualte
net benefits. Table IV-2 presents average cost per ton data, calculated
over the regulatory baseline and Table IV-3 presents incremental cost per
ton data for the percent reduction alternative calculated over the low
sulfur fuel alternative.
-------
IV-4
Table IV-1
NATIONAL COST IMPACTS OF REGULATORY ALTERNATIVES1'2
(Millions 1983 $)
Regul atory
A1 ternative
Annualized
Costs ($/yr)
Annualized
Control Costs ($/yr)5
Baseline
Low Sul fur Fuel3
Percent Reduction4
3,396
3,454
3,531
58
135
1 These annualized costs are in January 1983 dollars and are therefore slightly
higher than the June 1982 costs reflected in the Preamble.
2 Assumed high oil penetration scenario.
3 0.8 lb SOp/Million Btu - Oil
1.2 1 b S02/Mi11i on Btu - Coal
4 90 percent reduction.
5 These are the control costs associated with steam generating unit NSPS.
They are measured as additions to the baseline costs.
-------
IV-5
Table IV-2
Regul atory
Alternative
Average Control Cost Per Ton
Annualized
Costs (thousands
of 1983 $)
Tons of
SO2 Reduced
(thousands)
C/T
Low Sul fur
Percent Reduction
58,000
135,000
170
240
340
560
Table IV-3
1
Incremental Control Cost per Ton
Incremental Incremental
Costs (thousands Tons SO2
of 1983 $) Reduced (thousands) C/T
Percent Reduction 77,000 70 . 1100
1 Percent reduction over low sulfur fuel.
-------
IV-6
REFERENCES
Projected Impacts of Alternative Sulfur Oioxide New Source Performance
Standards for Industrial Fossil Fuel-Fired Boilers (March 1985).
Energy and Environmental Analysis, Inc. for the U.S. Environmental
Protection Agency.
-------
V-l
V. ECONOMIC IMPACT ANALYSIS
A. Introduction
A detailed analysis was undertaken to assess the potential industry-
specific economic impacts associated with the most stringent regulatory
alternative—that requiring a percent reduction in SO2 emissions. See
"Projected Impacts of Alternative Sulfur Dioxide New Source Performance
Standards for Industrial Fossil Fuel-Fired Boilers," March 1985. The
industry-specific impacts associated with the low sulfur fuel alternative,
would therefore be less than those discussed below.
The potential industry-specific economic impacts were analyzed in two
phases. The first phase focused on aggregate economic impacts for major steam-
using industries and the second phase focused on selected industries that were
considered likely to experience the most severe economic impacts.
B. Major Steam Users
Potential impacts on steam costs and product prices were estimated based
on industry-wide averages for eight large industry groups. The groups selected
for analysis account for approximately 70 percent of domestic industrial steam
consumption. These eight industry groups were: food; textiles; paper; chemicals;
petroleum refining; stone, clay and glass; iron and steel; and aluminum.
Assuming full cost pass-through of increased steam costs, product
prices in the major industry groups would increase by less than 0.03
percent. This potential impact represents a maximum product price increase
-because of the full cost pass-through assumption. In some instances,
increased steam costs would not be completely passed through to product
prices, and therefore, the impact on product prices would be less.
-------
V-2
C. Selected Industries
Seven industries, most likely to experience severe impacts, were
selected due to the steam-intensive nature of their operation or the low .
utilization of their steam generating unit capacity. These industries
were: beet sugar refining; fruit and vegetable canning; rubber reclaiming;
automobile manufacturing; petroleum refining; iron and steel manufacturing;
and liquor distilling.
The economic impact analysis examined potential impacts on prices,
value added, profitability, and capital availability. This analysis was
based on "model" plants and "model" firms representative of each industry.
Product prices were projected to increase from less than 0.01 to 0.5
percent in 1990 for all except the beet sugar refining industry, assuming
full cost pass-through of increased steam costs. For these same seven
industries, value added was projected to increase by about 0.01 to 0.9
percent in 1990 for all except the beet sugar refining industry, assuming
full cost pass-through of increased steam costs.
For both product price and value added impacts, the highest increases
are projected for the beet sugar refining industry. In the case of
product prices, this is due to the fact that the product price is low and
steam costs represent an unusually high proportion of manufacturing costs
in the beet sugar refining industry, compared to the other industries
examined. Similarly, value added impacts are higher since steam costs
represent an unusually high proportion of the non-raw material costs of
manufacturing the product in the beet sugar refining industry, compared
to the other industries examined.
-------
V-3
Based on the same scenarios outlined above, but assuming "full cost
absorption" of increased steam costs, return on assets was projected to
decrease by 0.03 to 2.8 percentage points. Again, these potential impacts
represent "worst case" projections of the assumption of full cost absorption
of the increased steam costs.
Cash flow coverage ratios and book debt/equity ratios showed
essentially no change for any of the model firms under two different
debt/equity financing strategies. Consequently, a percent reduction
requirement would not impair the ability of firms to raise sufficient
capital to construct fossil fuel-fired steam generating units.
D. Conclusions
The industry-specific economic impacts analysis indicates that a
percent reduction requirement would generally increase product prices and
value added by less than 1 percent if all steam cost increases were passed
through to product prices. In addition, assuming absorption of all steam
cost increases, return on assets would generally decrease by about 3
percentage points or less. Cash flow coverage and book debt/equity ratios
showed essentially no change. Therefore, a percent reduction requirement
would not impose any capital availability constraints on firms.
As mentioned earlier, a percent reduction requirement is the most
stringent regulatory alternative considered. Consequently, the industry-
specific economic impacts associated with the low sulfur fuel regulatory
alternative would be less severe than those discussed above.
-------
REFERENCES
Projected Impacts of Alternative Sulfur Dioxide New Source Performance
Standards for Industrial Fossil Fuel-Fired Boilers (March 1985).
Energy and Environmental Analysis, Inc. for the U.S. Environmental
Protection Agency.
-------
VI-1
VI. BENEFIT ANALYSIS
A. Introduction
The economic benefits associated with the Industrial Boiler SO2 NSPS
reflect the willingness to pay by individuals to be in a preferred state
(i.e., cleaner air) as opposed to a less preferred state (i.e., dirtier air)
of the world. In the present analysis, the less preferred state is associated
with pre-NSPS conditions. In particular, it is assumed that industrial
boilers for which the NSPS are applicable are subject to controls currently
required by representative State Implementation Plans (SIPs). Conversely,
the post NSPS scenarios reflect £he two regulatory alternatives discussed
previously-low sulfur fuel and percent reduction.
Although the NSPS are, by design, technology oriented, it is the case
that emissions reductions achieved translate into improvements in air
quality. It is the reduction in concentrations associated with the NSPS
that are the source of the economic benefits. Figure VI-1 portrays the
situation. At time to, ambient concentrations in a given area may be Cg.
In the absence of additional controls beyond those currently in place, net
growth of industrial boilers in the area will lead to higher total emis-
sions and consequently greater ambient concentrations. The point (tj, Ci)
is an example of this possibility. If additional control programs are
instituted, emissions will be reduced over some time path, and ambient air
quality will improve. This is shown by point (ti C2) in Figure VI-1.
The area which represents the improvement in air quality is the source of
the benefit estimates.
-------
VI-2
1
Conceatrttioat
C1
M
C2 "
(t0»CQ)
(tj t Cj)
-------
VI-3
studies which address some aspects of the health or welfare implications
of ambient SO2, SO4 and PM. This approach, which involves transformation
and extrapolation of existing research and studies, cannot be accomplished
in a thorough and comprehensive manner without first recognizing the many
technical problems associated with drawing inferences from studies not
necessarily designed for the purposes of this analysis. The problems
stem from a variety of sources, which include limited and sometimes
conflicting scientific information, paucity of data, and analytic techniques
which have not always been thoroughly tested. The technical problems
which exist result in uncertainty regarding the magnitude and precision
of the empirical economic benefit estimates. In order to deal explicitly
with this uncertainty, the extrapolation approach to benefit estimation
requi res:
0 Identification and use of the best data currently available;
0 Accomplishment of sensitivity analysis when alternative data
or assumptions exist;
0 Development of ranges of estimates to demonstrate the level of
uncertainty associated with different assumptions.
The above elements serve as the basis for the analytic strategy
that is used to develop estimates of the benefits for the steam generating
unit NSPS. The approach begins with a thorough literature search for
existing studies that could possibly be used in the extrapolation process.
The quantitative relationships which are contained in or derived from
the best of the available studies are then used to develop the benefit,
estimates presented in this analysis. A summary description of the
approach used is presented in subsequent paragraphs.
-------
VI-4
All categories of potential benefits that might result from
control strategies needed to meet the alternative steam gene-
rating unit SO? NSPS are identified. A review of the SO2 Criteria
Document and other reports provided a comprehensive listing of
possible adverse effects. A listing of effects categories is
presented in Table VI-1 below.
The existing research literature on the potential effects is
identified, classified and reviewed. All identified studies
are screened on the basis of several criteria, the most notable
of which are analytic quality and potential for extrapolation of
estimates for benefits analysis (e.g., requisite air quality
data available). As a result of this screening analysis, and
because of time and resource contraints, it is determined that
only some of the categories shown in Table VI-1 can be
estimated. Table VI-1 organizes the potential benefits by
pollutant and effects categories.
To comply with the steam generating unit NSPS, either
compliance fuel is burned or sulfur dioxide emission controls
are applied. These in turn reduce concentrations of SO2 directly
and sulfates and particulate matter indirectly (by reducing the
SO2 precursor). Furthermore, when scrubbers are used to reduce
SO2 emissions from oil-fired boilers, particulate matter emissions
may also be reduced below baseline levels. Consequently, compliance
with the Industrial Boiler NSPS may result in indirect and direct
reductions in PM concentrations. However, for purposes of this
study PM benefits are only estimated for the indirect reductions.
-------
VI-5
Table VI-1
Industrial Boiler NSPS
Potential Benefit Categories
SOz
Health Effects
- Mortality Due to Chronic Exposure X
- Mortality Due to Acute Exposure X
- Morbidity Due to Chronic Exposure *
- Morbidity Due to Acute Exposure *
Soiling and Materials Damage
- Residential Facilities *
- Commercial and Industrial Facilities X
Climate and Visibility Effects
- Local Visibility NA
- Non-Local Visibility NA
- Climate X
- Visibility at Parks NA
- Transportation Safety NA
Non-Human Biological Effects
- Agriculture *
- Forestry X
- Fishing X
- Ecosystem X
X are categories not analyzed in this study.
* are categories analyzed in this study.
NA are categories not applicable to the pollutant.
-------
VI-6
Direct PM benefits are attributed to the particulate matter
steam generating unit NSPS (see EPA-450/3-82-006a).
The coverage of benefits is also limited in other ways.
The benefits are often incomplete for those effects categories
where benefits are measured. For example, the SO2 materials
damage benefits only cover the residential sector. Commercial,
industrial, and governmental facilities are omitted. Finally,
as noted previously, there are other categories of effects
where no benefit estimates are made. See Table VI-1.
Benefits estimates are developed using the quantitative
relationships from each individual study and the air quality
improvements postulated for each alternative standard. These
estimates are accomplished according to a four-step
procedure as shown in Figure V1-2. The first step is to
identify the magnitude of the ambient air quality improvement
that is estimated to occur in each area and year. This is the
improvement achieved due to implementation of a particular
standard, relative to a baseline situation reflecting controls
already in place.
The second step involves estimating the health and welfare
improvements that might 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 SO2 and
SO4. Note that estimates are generally required for each area and
year in which there is an air quality improvement.
-------
VI-7
STEP
1
STEP
2
STEP
3
Estimate economic value of the health
or welfare improvement for i, t
Estimate health or welfare
improvement in area i at year t
Identify air quality improvement
in area i at year t
STEP
4
Aggregate over i to obtain
regional totals
Figure VI-2. Basic steps in estimating benefits for an individual study.
-------
VI-8
The third step is to impute an economic value to the estimated
changes in health and welfare status (Step 3a). For some classes
of benefits it is possible to estimate economic values directly
from the air quality improvement (Step 3b). The household sector
materials benefit estimate is an example of this approach. Studies
of this type allow direct estimates of the perceived economic value
of environmental improvements.
In Step 4, benefits for each area are summed to obtain totals
commensurate with the control strategy and air quality data.
4) Total incremental benefit estimates are developed by combining or
aggregating estimates from the appropriate effects categories.
Total incremental benefit estimates for each of the alternative
standards under consideration are required in order to complete a
benefit-cost analysis of the various policy alternatives. The
estimates are incremental in the sense they are derived from the
air quality change associated with going from present existing SIPs
to controls required for the alternative NSPS analyzed.
C. Study Selection, Application, Qualifications and Plausibility Checks
The procedures used to select studies as a basis for the benefit calculations
are identified in this section. The procedures used for applying the selected
studies to generate benefit estimates are also described. In addition, the
qualifications associated with each application and the plausibility checks for
resulting benefit estimates are discussed. The section is organized by the
pollutant and effect category for which benefits are estimated.
-------
VI-9
1. SO4 Related Benefits
a) Visual Range
1) Study Selection
Because visibility is not directly traded in existing markets,
benefit estimation methods require the establishment of a hypothetical
market or identification of complementarity between visual range and
existing markets. The former method manifests itself in contingent
ranking and contingent valuation surveys eliciting willingness-to-pay
estimates from survey respondents. The latter method generally infers a
relationship between visual range and residential property markets or
visual range and the cost of travel to alternative sites (e.g., national
parks).
Several visual range studies have been accomplished in the western
U.S. The contingent valuation and ranking studies generally focused on
user and existence values at national parks (Rowe et al. 1980, Schulze
1981, Rae 1982). The western visual range studies also included contingent
property value studies for two urban areas: the South Coastal Air Basin
(Brookshire, et al. 1979, 1980) and the San Francisco Bay Area
(Loehman, et al. 1981). However, visual range baselines are vastly
different between the eastern and western U.S., as are the vistas.
Because studies are available for the eastern U.S. and that is
the geographic focus of the benefit analysis, the western studies are
not used. To date, three eastern visibility studies have been conducted.
One study (Randall, et al. 1981) assessed the willingness-to-pay for
visual range improvement in the Chicago area. Another evaluated the
willingness-to-pay for visual range improvement in the Cincinnati area
(Rae, 1984). A third (Tolley et al. 1984) assessed the willingness-to-pay
-------
VI-10
for visual range improvement in each of six cities (i.e., Atlanta,
Boston, Cincinnati, Miami, Mobile, and Washington, D.C.) as well as the
rest of the east. All of the eastern visibility studies used contingent
valuation or contingent ranking methods.
The study by Tolley et al. is chosen for this analysis. It covers
more cities and has a larger sample size than the other studies. It
also enables an assessment of visual range changes "locally" as well as
other parts of the east (non-local changes).
2) Application
The visual range benefit assessment involves two sets of procedures.
The first involves using the visual range estimates for each of the
31 states (Chapter 3, Table 111-5) to develop estimates of "local" visual
range change and "non-local" visual range change. The estimated annual
visual range change for each state is used for the "local" change.
The area weighted average visual range change for the 31 eastern states
covered in the SO4 air quality analyses is the proxy for visual range
changes in other parts of the east or "non-local" visual range change.
The area weighted average visual range for the thirty-one state regi
31
is computed by the formula = £ VjAj where:
i=l A31E
V31E = 31 eastern state area weighted average visual range.
V-j = visual range for state i.
Ai = geographic area of state i.
A31E = total geographic area of the 31 states.
The Statistical Abstract provided population estimates for 1980 and 1990
as well as household estimates for 1980 for each of the 31 eastern
-------
VI-11
states. To calculate the predicted number of households for each of
the 31 eastern states for 1990 the following steps were taken. First
the 1980 ratio of households to population for each state was determined
using the formula Rs80 = Hqftn where:
ps80
Rs80 = ratio of households to population in state s in 1980
Hs80 = numbers of households in state s in 1980
Ps80 = population in state s in 1980
The number of households for 1990 and each state were then estimated
using the formula HS9q = Psgo Rs80 where:
Hst = household in statue s in year 1990
Pst = population in state s in year 1990
Rs80 = ratio of 1980 households to population in state s
The second set of procedures entails developing visual range improvement
valuation coefficients from the Tolley et al. study and applying them to
the predicted visual range changes. The Tolley et al. study calculates
the mean bids for the willingness-to-pay for a 10 mile (16.1 kilometer)
improvement in "local" visual range and changes in the other parts of
the east. The bids vary widely within and across the six study areas.
The mean bid for a 10 mile improvement in "local" visual range goes from
$3.54/year/household/kilometer in Cincinnati to $14.81/year/household/
kilometer in Washington, D.C. To reflect some of this variability, range
and midpoint benefit estimates are developed. The lower estimate applies
the Cincinnati mean bid to each of the 31 states (i.e., mean bid times
the air quality change in each state times the population in each state)
as a measure of the willingness-to-pay for "local" visibility changes.
Similarly, the Washington, D.C. mean bid is applied to each state to
-------
V1-12
generate the higher estimate. The mid-point estimate is derived by
applying valuation coefficients to each of the 31 states in accordance
with their nearness to one of the study cities. For example, the mean
bid for Washington, D.C. is apportioned to Virginia, Maryland, Delaware,
etc.
For the changes in visual range in other parts of the east a similar
procedure is used to reflect a range and midpoint estimate. However,
it was necessary to compare two types of bid information in Tolley
et al. (1984). In that study, the exact questions asked survey respondents
were what would you be willing-to-pay for a visual range improvement
locally and for a visual range improvement in the east. There is overlap in
J
these revealed preference bids because the "local" area is a part of the
east. To account for this overlap, the "local" bid is subtracted from
the eastern bid to yield the estimate of willingness-to-pay for visual
range improvements in other parts of the east. For example, in Cincinnati,
the mean bids derived from the Tolley et al. study are $3.54 for "local"
and $4.57 for eastern visual range improvements. The value used in this
analysis for visual range improvement in other parts of the east is
$1.03 (i.e., $4.57 - $3.54). The $1.03 value from Cincinnati is used to
generate the lower benefit estimate for visual range improvements in
other parts of the east while a value of $7.44 (i.e., $22.25 - $14.81)
is used to generate the higher estimate.
An example of the procedure to calculate the low estimate for local
benefits for the low sulfur fuel regulatory alternative is described below.
Given the following data for Alabama: 1980 population - 3,890,000; 1980
number of households - 1,342,000; 1990 projected population - 4,214,000,
-------
VI-13
we estimate the number of households in 1990 to be 1,454,000. The low
estimate for local benefits is the low predicted change in visual range
(.1 km) times 1,454,000 households times $3.54 per household per kilometer
change (1982 dollars) times the 1982 to 1983 Consumer Price Index inflator
(1.0311311) or $530,740. This number is the estimate of benefits in 1990
expressed in constant 1983 dollars for local visibility for Alabama.
3) Qualifications and Plausibility Checks
There are certain factors which suggest that the estimates of visual
range improvement benefits are understated. People in the west may be
willing-to-pay for visual range improvements in the east, especially at
the national parks. They have not been included in the benefit assessment.
Likewise, the control strategies implemented in the 31 eastern states
could result in visual range improvements in Canada and certain western
states. These too are omitted from this analysis. Finally, a common
result of the visual range studies conducted to date is that the
willingness-to-pay for an additional mile of visual range diminishes as
visual range increases. In other words people appear to be willing-to-pay
more for the first mile of visual range improvement than they do for the
second, or third. Average visual range in the eastern U.S. is about 10
miles (16 kilometers). The Tolley et. al. study posited the willingness-
to-pay for a visual range improvement of 10 miles (from a base of 10
miles to a visual range of 20 miles). Consequently, the willingness-
to-pay for that first mile of visual range improvement could be greater
than our estimate which assumed a constant bid averaged over the 10 mile,
change.
However, there are other factors which suggest the estimates of
visual range improvement benefits may be overstated in some respects.
-------
• VI-14
For example, the legitimacy of the benefit estimates is dependent in
part on the ability of people to perceive the projected visual range
improvements. The perception threshold for a change in visual range is
two to five percent for a single visibility event (EPA Report to Congress,
1979). Depending on the variance of visual range improvements through
the year, the perception factor could influence the benefit estimate.
For purposes of this analysis, no threshold was imposed on visibility
perception.
The estimates of willingness to pay for non-local changes are based
on an extrapolation of the difference between local and eastern bids per
mile change in visual range for a ten mile change and applied to an average
change in visual range for the thirty-one states. This extrapolation
presumes that the spatial distribution of visual range changes is unim-
portant and that the average change is what is valued in the bids. This
assumption is open to question and less confidence should be placed in
non-local benefit estimates than in the local benefit estimates.
Other qualifications to the visual range benefit estimates include
concern about extrapolating the contingent valuation results from six
cities to other areas of the east. This extrapolation includes urban to
rural extrapolation which may be questionable.
Attempts to assess the plausibility of the benefit estimates are
limited. A presurvey study in Cincinnati (Rae, 1984) resulted in mean
bid estimates similar to those developed by Tolley et. al. However, due
to possible information bias problems, the final survey by Rae in Cincinnati
yielded mean bid estimates which are much higher than the one used in
this study.
-------
V1-15
2. PM
The PM benefits developed in this analysis are derived from willing-
ness-to-pay estimates in the form of dol 1 ars/yg/m^ reduction in TSP. These
estimates in turn are derived from earlier Agency directed studies, including
the PM NAAQS RIA. In this section the study selection, application,
qualification and plausibility check procedures used in the PM NAAQS work
are briefly summarized. In addition the procedures, qualifications, and
plausibility of using those studies to develop estimates of PM benefits
associated with the steam generating unit NSPS are described,
a) Chronic Morbidity
1) Study Selection
The epidemiological studies amenable for benefit analysis in this
category include the Ferris et al. (1962, 1973, and 1976) and Crocker
et. al. (1979) research. The 1ongitutional studies by Ferris et al.
(1962, 1973, and 1976) in Berlin, New Hampshire were classified by
the criteria document and staff paper as useful for quantitative
purposes in PM standard setting. The cross-sectional study by Crocker
et al., although analytically appealing from a benefit analysis
perspective, was viewed as not quantitative for purposes of standard
setting during the development of the SO2 Criteria Document. The Crocker
study yields higher benefit estimates. However, the Ferris et al.
study is adopted for purposes of estimating PM benefits associated
with the SOg NSPS.
2) Application
In the PM NAAQS study, air quality concentration thresholds are
imposed to avoid going outside the TSP concentration range used by
-------
VI-16
Ferris et al. in estimating the concentration response relationship.
The health end point used by Ferris et al. is incidence of chronic
respiratory disease. After applying the Ferris et al. concentration-
response relationship, the Health Interview Survey is used to establish
a relationship between the change in chronic respiratory disease incidence
and work loss or restricted activity days. The geographic unit of
analysis is the county. The valuation of chronic respiratory disease
damages included an assessment of lost productivity (i.e., foregone
average daily wage), and increased medical care service charges (i.e.,
direct medical expenditures). Of course, the benefit of decreased PM
concentrations are evaluated as the reduction in the aforementioned
damages. Benefits are not calculated for reductions in pain and
suffering prior and subsequent to the receipt of medical care services.
The average per capita benefit per pg/m^ reduction in annual TSP for
the counties used in the PM NAAQS analysis ranged from $0.12 to $0.18
for direct medical expenditures, $0.15 to $0.21 for reduced work loss
days per employed worker, and $0.33 to $0.49 per capita reduction in
reduced activity days.
3) Qualifications and Plausibility Checks
Applying a study based on a small area in New Hampshire to estimate
benefits in 31 eastern states is not without uncertainty. However, the
fact that the study is judged to be quantitative for purposes of
standard setting does suggest some underlying credibility in the
benefit estimates. Even so, the benefit estimates for this PM category
are limited. For example, no valuation factor is applied to the
residual pain and suffering category.
-------
VI -17
b) Acute Morbidity
1) Study Selection
The acute morbidity studies amenable for benefit analysis include
the Samet et. al (1981) longitudinal epidemiological study in Steuben-
ville, Ohio and the cross sectional microepidemiological study by
Ostro (1983). The Samet et.al. study focused on the relationship of
emergency room visits to acute respiratory disease. The study was
classified by the PM staff paper as useful for determining concentra-
tion-response relationships in the standard setting. The Ostro study
focused on broader health end points which related work loss and
restricted activity days due to respiratory disease to TSP concentra-
tions. The Ostro study was not available in time for inclusion in
the PM Criteria Document or Staff Paper; but like some other studies
published later was regarded highly enough that the Administrator
asked for public comment (PM NAAQS Proposal Preamble) on using the
study in standard setting.
The Ostro study provides a more complete geographic and health
end point assessment than the Samet et. al. study and estimates work
loss and restricted activity days directly. Hence, it is selected
for use in estimating PM benefits associated with the steam generating
unit NSPS.
2) Application
Like the chronic morbidity benefits, the geographic unit of analysis
is the county. Similarly, the damage categories are lost productivity,
reduced non-work day activities, and increased medical care service charges.
Furthermore, benefits are not calculated for reduction in pain and suffering
prior and subsequent to the receipt of medical care services.
-------
VI-18
3) Qualifications
Like all epidemiological studies, there is the question of inferring
causality from the statistically significant relationship. However, unlike
many other epidemiological studies, Ostro has relatively better control
for potentially confounding factors although some such as other pollutants
are omitted. Ostro uses annual average TSP concentrations as the measure
for exposure for the two-week period for which acute illness is examined.
The estimated decrease in work loss days due to respiratory causes for a
10 yg/m3 decrease in TSP is 3.6%. The estimated decrease in restricted
activities days due to respiratory causes is about 6.7% for a 10 yg/nv*
decrease in TSP.
c) Household Soiling
1) Study Selection
The studies available for benefit assessment include the two Phila-
delphia studies by Watson-Jaksch and Cummings et al. (1981) and the 24
SMSA longitudinal study by Mathtech (1982). The first two studies are
developed from a soiling survey (Booze, Allen, Hamilton 1970) designed
to determine out-of-pocket cost for air pollution induced soiling.
Watson-Jaksch (1978) analyzed the survey results and found that although
cleaning expenditures did not change a welfare loss could also be incurred.
Cummings et al (1981) analyzed the survey and found that by imputing a
value for the time of the do-it-yourself cleaning activities, which varied
with pollution levels, welfare losses could also be incurred.
The Mathtech study (1982) is referenced later in the SO2 section on
materials damage. As will be indicated there, the approach incorporates a
physical damage function indirectly in the structure of an economic demand
-------
VI-19
and supply model. Within that framework, household behavior in terms of
soiling perception, cleaning activity, and expenditures to maintain a
given degree of cleanliness is estimated econometrically. Prices, environ-
mental variables, and sociodemographic characteristics are inputs to the
model.
The Mathtech model is chosen because it is based on more recent
data, has broader geographic coverage, and uses the theoretically correct
measure of benefits, willingness-to-pay.
2) Application
To apply the model in the PM NAAQS (Mathtech, 1983) analysis, values
of prices, environmental variables, and socio-demographic characteristics
for 491 counties are put into the estimated model at the county level.
The benefits are determined by changing the value of the environmental
variable (i.e., TSP) to reflect alternative SO2 NSPS or conditions
with and without TSP air quality improvements.
3) Qualification and Plausibility
The Mathtech model was peer reviewed in a public forum and a panel of
experts found the model and empirical methods to be sound (Mathtech, 1982).
The predicted level of benefits was judged to be reasonable as a percent
of household income and relative to previously developed estimates. The
model is also referenced in the Criteria Document. However, the framework
of the Mathtech household model does not capture soiling related benefits
such as the time cost incurred for the do-it-yourselfers or the residential
location decision. The Mathtech model is limited in other ways. For
example, the level of detail for the data inputs of the econometric esti-
mation could have been more refined in the absence of the disclosure
-------
VI-20
restrictions on the Consumer Expenditure Survey. Application of a model
based on 24 SMSAs to 491 counties is not without analytic uncertainty,
d) Extrapolation
1) Method
The studies selected for estimating PM benefits associated with re-
ductions in chronic morbidity, acute morbidity, and household soiling were
used in a related PM generic NSPS study (Mathtech 1984). Rather than
applying these same studies at a micro level, average benefits estimated
for this PM generic NSPS study are extrapolated to the PM changes associated
with the two regulatory alternatives. Of the counties covered in the re-
lated PM generic NSPS study 491 are in the 31 eastern states addressed in
this analysis. The baseline for the PM generic NSPS analysis is current
PM SIP levels. Projected air quality is with and without 16 currently
promulgated PM NSPSs. The chronic and acute morbidity and household
soiling benefits accruing in the year 1990, to the population projected to
reside in those 491 counties, are calculated for purposes of this analysis.
For each county, the estimated benefits are divided by population to yield
a benefit/person figure. This, in turn, is divided by the annual change
in air quality to yield a benefit/percent pg/m3 reduction TSP figure.
These figures are summed and divided by the number of counties (i.e. 491)
to give the S/person/gg/m3 reduction TSP estimate used in this analysis.
2) Qualifications and Plausibility Checks
To the extent the air quality scenario in the PM generic NSPS
analysis is different from that projected in the SO2 NSPS, analysis
* For example, controlling SO2 emissions results in lower fine particle (i.e.,
£2.5 um) levels, while NSPS controls for PM emissions result in lower
concentrations of other particle sizes in addition to fine.
-------
VI-21
biases of an unknown direction will result. Furthermore, population
differences between the 31 eastern states and the 491 counties forming
the basis of this analysis could also impart biases of an unknown
di rection.
3. SO2 Benefit Assessment
For purposes of this analysis, the scope of coverage of SO2 benefits
includes materials damage (household sector), acute and chronic morbidity,
and agriculture benefits (soybeans, wheat, and oats),
a) Materials Damage
1) Study Selection
There are two elements that are important in selecting models
that consider materials damage. First, one must evaluate the validity
of statistical relationships between measures of physical damage (e.g.,
corrosion rates) for a given material and plausible explanatory factors
such as SO2 concentrations and relative humidity. Second, one must
consider the soundness of the procedures used to impart economic value
to the physical damage that is caused.
The first topic requires a review and critique of the literature on
physical damage functions. These functions have been reviewed in the
Criteria Document and the general consensus is that fairly reliable damage
functions have been developed for ferrous metals and zinc. However, cur-
rently available damage functions for exterior paints, stone, masonary,
concrete, textile, leather, and paper are considered not to be as well
specified.
There are three general approaches that have been used to value
physical damage. Two of the approaches use the physical damage functions
directly. These approaches have been classified as the value of lost
-------
VI-22
material method and the cost of ameliorative/preventive action method.
Examples of the first approach include Salmon (1970) and SRI (1981).
Examples of the second valuation approach include Fink et al. (1971) and
TRC (1981). The third approach does not rely on physical damage functions
directly. Instead, economic demand and supply surveys are estimated as
factors (Mathtech, 1982). Note however, that the physical damage
functions can provide supporting information on the types of behavioral
responses implied by the economic analysis.
A prime concern in selecting benefit models for valuing material
damage is the data requirements. Both valuation methods that rely
explicitly on physical damage functions require information on materials
inventory, exposure and distribution. These data are most likely
available only with significant approximation and are most certainly
available with any degree of accuracy only for selected cities where
materials surveys have been conducted. It is principally for this
reason that the calculation of benefits from reduced materials damage
is limited to the demand/supply model as estimated by Mathtech.
Although the data used in the Mathtech analysis are limited in some
respects and interpretations of some of the implied behavioral
responses have been questioned, the theoretical and empirical methods
have been judged by several groups to be sound. The Mathtech model
is an economic demand model which describes relationships between SO2
and various goods and services that may be purchased by a consumer.
Thus, the concentration-response function relates SO2 directly to
economic parameters; there is no intermediate calculation of physical
damage. In addition to reduced data requirements, the Mathtech model
-------
V1-23
of materials damage in the household sector has other positive attributes
that are not shared by the other valuation approaches. The method of
benefit calculation is based on willingness to pay which is the theo-
retically correct procedure for measuring welfare change. Also, the
Mathtech model permits consideration of substitution possibilities.
2) Application
Demand equations are estimated for seven commodity groups: food,
shelter, household operations, home furnishings, clothing, transportation,
and personal care. Two additional plausible demand categories, property
expenditures and recreation are not included because of data problems.
For each of the seven commodity groups, from 2 to 5 goods are separately
identified and demand equations are also estimated for each of these goods.
For example, separate equations are estimated for household textiles,
furniture, appliances, and houseware in the home furnishings commodity
group. In all, 21 demand equations for goods are estimated.
Data for the original estimation of the household sector model were
obtained from several sources. The expenditure data for individual goods
and specific SMSA's are taken from the U.S. Bureau of Labor Statistics,
Consumer Expenditure Survey (1978). Retail price data for the same
SMSA's and goods are also from a Bureau of Labor Statistics publication
(1973). Air quality data are developed from data maintained by EPA on
the SAROAD data base. The data base that is constructed covers 24
SMSA's with budget and environmental data representative of conditions
in 1972 and 1973. It is assumed that the preferences revealed in the
1972-73 data can be applied to the time period and areas considered in
the case study analysis. That is, consumer preferences and relative
prices remain unchanged.
-------
V1-24
The procedures used to develop estimates of benefits are quite com-
plex, involving simultaneous evaluation of systems of demand equations.
The major steps of the calculation routine are highlighted below.
The first step involves aggregation of market data to form price and
quantity indices for the 21 goods for which separate demand equations are
to be estimated.
Second, the 21 goods are grouped into seven commodity groups. The
seven commodity groups are consistent with the major divisions defined by
the Bureau of Labor Statistics for household budget allocations.
As noted earlier, each .-of the seven commodity groups contains from 2
to 5 goods. The third step thus involves the simultaneous estimation of
demand equations for each good in a given commodity group. It is at this
stage that the SO2 variables are entered to determine whether they are
relevant explanatory variables. In addition, several socioeconomic vari-
ables are also considered to determine whether they help explain the varia-
tion in demand.
The econometric modeling of the goods demand equations indicated that
S02 is a significant explanatory variable in three of the equations: home
repair activities; textiles excluding clothing; and transportation ser-
vices. These equations are reproduced in Table VI-2. The measure of SO2
found to be most robust is the 24-hour second high for the monitor re-
cording the maximum 24-hour second high concentration among all receptors.*
* Additional work is underway to estimate the demand equations using annual
average values for the pollutant variables and to test specifically for
non linearities in the pollution effects on demand.
-------
VI-25
Table VI-2
DEMAND EQUATIONS FOE WHICH SOj IS A SIGNIFICANT EXFLANATOKI VARIABLE
1. Hoaa Repair
X. - 0.23288 + 0.767 • (-33.314 + 0.0328* • SO,) • Z.
- 0.23288 • (-133.27 ~ 0.11979 • TSP) • Zj
vhara Xj ¦ the eh ax a of hoae repair expeaditoraa relative to all
ezpeaditaxea fox aheltex.
Zj ¦ the ratio of the boat repair priea iadaz to aheltex
ezpeaditaxea.
Zj ¦ the ratio of the atility pxice iadaz to aheltex
ezpeaditaxea.
SOj ¦ tha 24 hoax sacoad high aeaaaxe of S0^ for the aoaitor
xecordiag tha aaziaaa valae of thia variable vithia tha
caaa atady axea.
TSP ¦ tha 24 hoax aaeoad high aaaaaxa of Total Saapaadad Par-
ticalatea for tha aoaitor raeordiag tha aaziaaa rila* of
thia variable vithia tha caaa atady araa. Thia valaa ia
dafaoltad to tha earraat pxiaary NAAQS.
2. Taztilaa azeladiag clothiag
X- - 0.104436 + 0.89334 • (3.23302 + 0.020286 • SO.) • Z,
- 0.106436 • ((-21.3796 - 37.8934 • FAMSZ + 33.373 *
REGION) • Z4 * (-30.0434 ~ 10.3393 • FAMSZ + 2.4711 •
REGION) • ZPS - 7.3 3 839 • Ztf)
vhexe Xj - tha ahara of taztila azpaaditaraa relative to all
azpaaditaxaa for hoaa faraiahiaga.
Zj ¦ tha ratio of the taztila price iadez to hoae faraiahiag
ezpaaditaxaa.
Zj ¦ tha ratio of tha faraitare price iadez to hoae
faraiahiag azpaaditaraa.
Zj m tha ratio of tha appliaace price iadaz to hoaa
faraiahiag azpaaditaraa.
-------
VI-26
Table VI-2 (continued)
DEMAND EQUATIONS FOR VHICH SQj IS A SIGNIFICANT EXFLANATORI VARIABLE
Z6
a
ths ratio of the hoassvares pries iadsz to boat
foraishiag szpoaditarsa.
so2
m
ths 24 hoar sscoad high asasare of SO2 for the aoaitor
recordiag the aaziaaa valae of this variable vithia the
cast atadj int.
FAMSZ
m
faaily six*.
REGION
daaay variable for locatioa of eass stady arta
(1-Northoaat or North Ceatral, (Mother* iae).
3. Traaaporatioa Services
X3
m
0.53446 + 0.46534 • (-675.609 + 37.1853 • SO,) • Z.
- 0.53466 • (-377.413 + 28.3191 • TEMP) • Z,
ihtrt Xj
m
ths ahars of gaaolias ezpeaditarea relative to all
szpoaditarsa for traaaportatioa.
h
m
ths ratio of ths gaaolias pries iadez to all
traaaportatioa szpoaditarsa.
Z8
m
ths ratio of ths pries iadsz for ths "other"
traaaportatioa dsaaad category to all traaaportatioa
szpoaditarsa.
m
ths 24 hoar sscoad high asaaars of SOj for the aoaitor
recordiag the aaziaaa vale* of this variable vithia the
ease stady area.
TEMP
-
-------
VI-27
It is important to note that these equations are estimated as part of
a system of demand equations, and by themselves do not lead directly to a
measure of benefits for reductions in SO2. In fact, a series of steps is
required before benefits can be ascertained.
Given the demand equations for goods, the next step involves using
these equations to define price and quantity indices for the commodity
category in which they have been placed. Note, in particular, that the
price index for a commodity will be sensitive to the level of SO2 that is
used to evaluate the goods demand equations. Since demand for only three
goods is found to be associated with SO2, only three of the commodity price
indices will be dependent on SO2. However, since the seven commodity
groups also form a demand system, the SO2 variable indirectly plays a role
in the overall budget allocation across commodities.
With aggregate price and quantity indices defined for the commodities
the next step involves the simultaneous estimation of the commodity demand
functions. With the parameters of the comnodity demand functions identi-
fied, it is now possible to calculate benefits. The goal is to use the
commodity demand functions to define an expenditure function that is a
function of SO2. This is accomplished in the following way.
First, the parameters of the system of commodity demand functions are
used to define a compensated commodity demand system. The compensated
curves result from a consumer optimization problem in which the consumer
minimizes expenditures subject to a constant level of utility. The expen-
diture function is then formed as the inner product of the commodity
compensated demand function and the associated price index for all seven
-------
V1-28
commodities. This procedure is performed with SO2 first at baseline
(pre-control) levels and then at the two post-control levels - low
sulfur fuel and percent reduction. The difference between the two
expenditure functions is a measure of the compensating variation, a
welfare measure that is consistent with the correct concept of a
benefits measure: willingness to pay.
3) Qualification and Plausibility Checks
Using modeled air quality receptors rather than monitored data as
was used in the original study may bias the benefit estimate upward.
Only residential material damage is estimated. Because of the pollution
index used in the underlying study (24-hour second high for an SMSA)
and the geographic unit of the underlying study (SMSAs), the impact of
applying these coefficients to a different area (a rectangular grid
around an industrial boiler) is unclear. For the 113 study areas analyzed
benefits calculated on a per-household basis range from 0 to $15.55
per year for both the low sulfur fuel and percent reduction regulatory
alternatives. This range reflects the population distribution among
113 study areas,
b) Morbidity
1) Study Selection
A large number of studies were reviewed to identify sources for
benefit functions relating SO2 to morbidity. Based on the type of
exposure, the studies can be divided into three categories:
Short-term (1-hour),
Daily (24-hour)
-------
V1-29
Short-term Exposure
For the short-term exposure category, clinical studies are
utilized. In the relevant range of SO2, the major effects occur in
asthmatics. Studies of this group examine effects on both lung function
changes are difficult to value. There are no estimates Of willingness
to pay for these changes or of the accompanying medical costs. Therefore,
the analysis considers only effects on symptoms.
The full range of clinical studies of effects of SOg on symptom
prevalence is utilized. The lower-bound estimate is based on the
pooled results of eight studies. The upper-bound estimate is based on
work of Kirkpatrick et al. (1982).
Daily Exposure
No microepidemiology studies were found to be directly usable for
estimating effects of daily exposure. Among the limitations of the
studies were failure to derive quantitative concentration-response
functions and lack of control of confounding factors.
Several macroepidemiology studies examine the relationship between
SOg and the number of emergency room visits in an area using time-
series regression analysis. The studies of Mazumdar and Sussman (1981)
and Graves et al. (1980) were selected to develop a range of benefit estimates.
2) Application
For the clinical studies a minimum estimate of zero is used. The
maximum estimate for each day's symptoms is $25 based on unpublished work
by Dickie and Gerking and Tolley and Babcock.
3) Qualifications and Plausibility Checks
To estimate the benefits using clinical studies, assumptions about
-------
V1-30
activity levels are derived form the original PM exposure analysis. Doing
this implicitly assumed the same activity patterns for asthmatics as the
rest of the population. Inhaled dose is estimated using an assumed
ventilation rate of 40 liters per minute and a peak concentration of twice
the hourly concentration. The symptom valuation is an arbitary escalation
of a study valuing similar symptoms for the general population rather than
for asthmatics,
c) Agriculture
1) Study Selection
The selection of crops for analysis is governed by the following
criteri a:
o Crop sensitivity to exposure to SO2;
0 Information availability for the formulation of a
dose-response function; and
0 Significant economic value of the crop.
Based on application of these criteria, soybeans, wheat, and oats are selected.
For calculation of benefits, the parameters which define the response
of a crop to SO2 should reflect its economic value. Yield can be given a value
in economic terms. While foliar and yield effects may occur together and may,
in some instances, be highly correlated, yield response cannot generally be
accurately extrapolated from observed foliar effects of variations in SO2.
Therefore, study selection is limited to studies of yield effects. The
additional criteria for study selection are: study focus is on ambient SO2
exposure and other ambient environmental factors; control for the impact of
other confounding environmental conditions on the concentration relationship;
and plausible and quantifiable results which can be used to develop
concentration response functions.
-------
VI-31
Application of these criteria yielded a very limited number of studies
for each crop. The major study used for soybeans is from Sprugel et al.
(1980). For oats and wheat, the principal source is a series of reports by
Guderian and Stratmann (1962 and 1968).
2) Application
Average air quality for each receptor area for the same part of the
growing season as the underlying study uses is applies. The estimate is
restricted to air quality changes within the underlying study range. Zero
is used as the minimum estimate. The maximum estimate applies the dose
response function of the study varietal to all of that crop. One-half the
maximum estimate is used for the middle estimate.
3) Qualifications and Plausibility Checks
The analysis values yield changes at recent state market prices. No
attempt is made to adjust for price changes that might result from nationwide
yield changes. No changes in cropping patterns are factored into the
analysis. No attempt is made to value contributions to yield for sulfur
deficient soils.
The yield functions only consider average sulfate concentrations for
the growing seasons. This ignores impacts of the time pattern of exposure
(e.g. size and timing of peaks), the impact of other pollutants, and
interactions with other determinants of yield such as temperature, light,
soil, and agricultural practice.
The choice of functional form for the yield function was necessarily
made with a very limited set of data points. The yield changes are sensitive
to the functional form.
-------
V1-32
0. Benefit Analysis Estimates
The benefit estimates are exhibited in Table VI-3. Ranges are presented
for SO4 and PM estimates to reflect the uncertainty concerning air quality
and economic valuation, as discussed in III-B and VI-C, respectively. The
benefit categories estimated include increased local and non-local visibility
due to SO4 reductions, reduced chronic and acute morbidity, and reduced
household (residential) soiling due to PM reductions, and decreased household
sector materials damage due to SOg reductions. The SO4 and PM benefit
estimates are for the 31 eastern state region, while the SOg benefit estimates
are nationwide. The benefits associated with the percent reduction regulatory
alternative exceed the benefits associated with the low sulfur fuel alternative.
This reflects the greater improvement in air quality that can be achieved
when percent reduction is required.
Tables VI-4 - VI-6 report the average benefits per ton of SOg reduced.
For SO4 and PM, only the tons reduced from the 31 eastern states are considered.
For SOg, tons reduced from all 113 areas where new industrial boilers were
placed, are calculated.
The benefits per ton are extremely similar for both low sulfur fuel
and percent reduction, as can be seen in Table V1-7. The annualized
benefits and the tons of SOg reduced for the percent reduction alternative
are approximately 30 - 33% greater than those for the low sulfur fuel
alternative. Consequently, the benefit per tons estimates are almost
identical. Table VI-8 presents incremental benefits per ton of percent
reduction over the low sulfur fuel alternative.
-------
Table VI-3
Annualized Benefits*
(Thousands of 1983 dollars)^
Regulatory Altematives^
Industrial Boiler SO2
Benefit Category
Low
1
Sulfur Fuel
2
3
Percent Reduction
1 2 3
SO4 Visibility
Local
8,000
21,000
41,000
12,000
31,000
61,000
Non-Local
3,000
11,000
25,000
4,000
15,000
35,000
PM Morbidity & Soiling
23,000
25,000
26,000
35,000
38,000
40,000
S02 Health
2,000
2,000
2,000
2,000
2,000
2,000
SO2 Welfare
Materials Damage
26,000
26,000
26,000
40,000
40,000
40,000
Agriculture
200
200
200
200
200
200
Total
62,000
85,000
120,000
93,000
126,000
178,000
* The benefit assessment only includes a subset of potential benefits.
2 Benefits are for the year 1990, in 1983 dollars.
A range of benefits (1,2,3) is presented for SO^ and PM to reflect sensitivity
analysis on the air quality modeling and economic valuation inputs.
-------
V1-34
Table VI-4
Average Benefits per Ton
SOg (Total Visibility)
Regul atoryl
Alternative
Low Sul fur Fuel /I
Low Sulfur Fuel/2
Low Sul fur Fuel /3
Annualized
Benefits
(thousands of 1983$)
11,000
32,000
66,000
Tons S02
Reduced
(thousands)
120
120
120
B/T
90
270
550
Percent Reduction/1
Percent Reduction/2
Percent Reduction/3
16,000
46,000
96,000
175
175
175
90
260
550
* A range of benefits (1,2,3) is presented for SO^ to reflect sensitivity
analysis on the air quality modeling and economic valuation inputs.
2 Tons reduced calculated only for 31 eastern state region.
-------
VI-35
Table VI-5
Average Benefits per Ton
Particulate Matter
Regul atory1
Alternative
Annualized
Benefits
(thousands of 1983$)
Tons SO2
Reduced
(thousands)
B/T
Low Sul fur Fuel /1
Low Sul fur Fuel /2
Low Sul fur Fuel /3
23,000
25,000
26,000
120
120
120
190
210
220
Percent Reduction/1
Percent Reduction/2
Percent Reduction/3
35,000
38,000
40,000
175
175
175
200
220
230
1 A range of benefits (1,2,3) is presented for PM to reflect sensitivity
analysis on the air quality modeling and economic valuation inputs.
2 Tons reduced calculated only for 31 eastern state region.
-------
VI-36
SO;
Table VI-6
Average Benefits per Ton
Annualized Tons SO2
Regulatory Benefits Reduced
Alternative (thousands of 1983$) (thousands) B/T
Low Sulfur Fuel 28,000 170 170
Percent Reduction 42,000 240 180
1 Tons reduced calculated for all 113 areas analyzed.
-------
V1-37
Table VI-7
Average Benefits per Ton
Totals (SOg, PM, and SO?)
Regulatory* B/T
Alternative (1983 $)
Low Sulfur Fuel/I 450
Low Sul fur Fuel/2 640
Low Sulfur Fuel/3 930
Percent Reduction/1 470
Percent Reduction/2 660
Percent Reduction/3 950
* A range of benefits (1,2,3) is presented for S04 and PM to reflect
sensitivity analysis on the air quality modeling and economic
valuation inputs.
Table VI-8
Incremental Benefits per Ton*
Totals (SO^, PM, and SOg)^
B/T
(1983 $)
Percent Reduction/1 510
Percent Reduction/2 690
Percent Reduction/3 1000
1 Percent reduction over low sulfur fuel.
^ A range of benefits (1,2,3) is presented for S04 and PM to reflect
sensitivity analysis on the air quality modeling and economic valuation
inputs.
-------
V1-38
Bib!iography
Bachmann, J. Memorandum to Tom Walton, "Estimate of Sulfate/Fine Mass
Relationship," 1985.
Brookshire, 0., d'Arge, R., Schulze, W., and Thayer, M., (1979). Experiences
in Valuing Non-Market Goods: Alternative Benefit Measures of Air
Pollution in the South Coast Air Basin of Southern California,
EPA-600/6-79-0016.
Brookshire, D., Thayer, M., Schulze, W., and d'Arge, R., (1980). "Valuing
Public Goods: A Comparison of Survey and Hedonic Approaches," Resource
and Environmental Economic Laboratory, University of Wyoming, later
published in American Economic Review, 72 (March).
Crocker, T.D., Schulze, W., Ben-David, S. and Kneese, A., (1979). "Methods
Development for Assessing Air Pollution Control Benefits," Vol. I,
Economics of Air Pollution Epidemiology, EPA-600/5-79-001a, U.S.
Environmental Protection Agency, Washington, D.C.
Cummings, R., Burness, H. and Norton, R. Methods Development for Environmental
Control Benefits Assessment, Volume V: Measuring Household Soiling
Damages from Suspended Air Particulates, A Methodological Inquiry. Draft
Report, January 1981.
Dickie, M. and Gerking, S. Written correspondence to Allen Basala, "Willingness
to Pay to Avoid Symptoms," 1980.
Ferris, B.G., Jr., Higgins, I., Higgins, M.W. and Peters, J.M., (1973). Chronic
Non-Specific Respiratory Disease in Berlin, New Hampshire, 1961-1967:
A Follow-up Study. American Review of Respiratory Disease, 107-110-122.
Ferris, G.B., Jr. and Anderson, D.O., (1962). Prevalence of Chronic Respiratory
Disease in a New Hampshire Town. American Review of Respiratory Disease,
86:165-177.
Fink, F.W., Buttner, F.H., and Boyd W.K., (1971). Technical-Economic
Evaluation of Air Pollution Corrosion Costs on Metals in the U.S.
APTD-0654. U.S. Environmental Protection Agency, Research Triangle
Park, N.C.
Graves, P.E., Krumrn R.J., and Violette, D.M., (1980). Estimating the Benefits
of Improved Air Quality. Report for Meeting of Benefit Methodology Panel,
National Commission on Air Quality, December.
Guderian, R. and Stratman, H., (1962). Field Experiments to Determine the
Effects of SO2 on Vegetation - Part I: Survey of Method and Evaluation
of Results. Research Reports by the State of North Rhine-Westphalia,
No. 1920. West German Press, Cologne and Opladen.
-------
V1-39
Hobart, J.M., R.L. Horst, Jr., K.M. Brennan, E.H. Manuel, Jr., et al.
(1984). Benefit Analysis of National Air Quality Standards for Sulfur
Dioxide. Report by Mathtech, Inc. for EPA, Office of Air Quality
Planning and Standards, Research Triangle Park, NC, December.
Horst, R.L., J.M. Hobart, E.H. Manuel, Jr., K.M. Brennan, M.C. Duff, and
G.D. Labovich (1985). Benefit-Cost Analysis of Selected New Source
Performance Standards for Particulate Matter. Report prepared by
Mathtech, Inc. for the Environmental Protection Agency under Contract
No. 68-02-3553, July.
Kirkpatrick, M.B., Sheppard, D., Nadel, J.A., and Boushey, H.A., (1982).
Effect of Oronasal Breathing Route on Sulfur Dioxide-Induced Bron-
choconstriction in Excercising Asthmatic Subjects. American Review
of Respiratory Diseases, 125:627-631.
Loehman, E., Boldt, D., and Chaikin, K., (1981). Measuring the Benefits of Air
Quality Improvements in the San Francisco Bay Area, SRI 8962.
Manuel, E.H., Jr., R.L. Horst, Jr., K.M. Brennan, et al. (1982). Benefits
Analysis of Alternative Secondary National Ambient Air Quality Stand-
ards for Sulfur Dioxides and Total Suspended Particulates, Volume II.
Final report by Mathtech, Inc. for EPA, Office of Air Quality Planning
and Standards, Research Triangle Park, North Carolina, May.
Mathtech, Inc., (1982). Benefits Analysis of Alternative Secondary National
Ambient Air Quality Standards for Sulfur Dioxide and Total Suspended
Particulates. Final report prepared for U.S. Environmental Protection
Agency under Contract No. 68-02-3392. Research Triangle Park, N.C.
Mathtech, Inc., (1983). Benefit and Net Benefit Analysis of Alternative
National Ambient Air Quality Standards for Particulate Matter. Prepared
for the U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards under Contract No. 68-02-3826.
Mathtech, Inc., (1984). Benefit Cost Analysis of Selected New Source
Performance Standards for Particulate Matter. Draft Final Report
Proposed for U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards under Contract No. 68-02-3553, Resarch Triangle
Park, N.C.
Mazumdar, S. and Sussman, N., (1981). Relationships of air pollution to
health: results from the Pittsburgh study. Proceedings of the 74th
Annual Meeting, Air Pollution Control Association, Philadelphia, Pa.
June 21-12. 1981.
Ostro, B. The Effects of Air Pollution on Work Loss and Morbidity, Journal
of Environmental Economics and Management 10:71-78. December 1983.
-------
VI-40
Rae, D., (1982). "The Value to Visitors of Improving Visibility at Mesa
Verde and Great Smokey National Parks," a paper presented at a National
Park Service Conference, Keystone Colorado, May, later published in
Robert Rowe and Lauraine Chestnut, eds., Managing Air Quality and Scenic
Resources at National Parks and Wilderness Areas, Westview Press, Boulder,
Colorado, 1983.
Rae, D., Charles River Associates, (1984). Benefits of Visual Air Quality In
Cincinnati Results of a Contingent Ranking Survey, draft report,
Electric Power Research Institute, Project #1742.
Randall, A., Hoehn, J., and Tolley, G. "The Structure of Contingent Markets:
Results of a Recent Experiment," a paper presented at the American
Economic Association Meetings, December 1981.
Rowe, R.D., d'Arge, R., and Brookshire, D., (1980). An experiment on the economic
value of visibility. Journal of environmental economics and management
7:1-19.
Salman, A. Systems Analysis of the Effect of Air Pollution on Materials,
Final Report for the National Air Pollution Control Association under
Contract No. CPA-22-69-1 13. January 1970.
Samet, J.M., Speizer, F.E., Bishop, Y., Spengler, J.D., and Ferris, Jr., B.G.,
(1981). The relationship between air pollution and emergency room
visits in an industrial community. J. Air Pollut. Control Assoc.
31:236-240.
Schulze, W.D., Brookshire, D.S., Walther, E., and Kelley, K., (1981). Methods
development for environmental control benefits assessment. Vol. X:
The benefits of preserving visibility in the national parklands of the
Southwest. Washington, D.C.: U.S. Environmental Protection Agency,
Office of Research and Development.
Sprugel, D.G., Miller, J.E., Muller, R.N., Smith, H.J., and Xerikos, P.B.,
(1980). Sulfur Dioxide Effects on Yield and Seed Quality in Field-
Grown Soybeans. Phytopathology, 70:1129-1133.
SRI, Inc. An Estimate of the Non-Health Benefits of Meeting the Secondary
National Ambient Air Quality Standards. Prepared for the National
Commission on Air Quality, January 1981.
Tolley, G., Randall, A., Bloomquist, G., Fabian, R., Fishelson, G., Frankel,
A., Hoehn, J., Krumm, R., Mensah, E., and Smith, T. Establishing
and Valuing the Effects of Improved Visibility in Eastern United States,
Office of Research and Development, Environmental Protection Agency,
March 1984.
Tolley, G. and Babcock, C., (1986). Draft - Valuation of Reductions in
Human Health Symptoms and Risks. USEPA Grant No. CR-811053-01-0.
-------
VI-41
TRC (1981). Benefit Model for Pollution Effects on Material. Final report
prepared for U.S. Environmental Protection Agency under Contract No.
68-02-3447. Research Triangle Park, N.C. 27711.
U.S. Bureau of Labor Statistics (1973). Average Retail Prices of Selected
Commodities & Services. U.S. Government Printing Office, Washington, OC.
U.S. Bureau of Labor Statistics (1978). Consumer Expenditure Survey:
Integrated Diary and Interview Survey Data, 1972-73. Bulletin 1992,
U.S. Government Printing Office, Washington, D.C.
U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards. Protecting Visibility, An EPA Report to Congress. EPA-
450/5-79-008. October 1979.
U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards. Review of the National Ambient Air Quality Standards for
Sulfur Oxides: Assessment of Scientific and Technical Information,
OAQPS Staff paper. November 1982.
Watson, W.D., Jr., and Jaksch, J.A., (1978). Household cleaning costs and
air pollution. Presented at the 71st Annual Meeting, Air Pollution
Control Association, Houston, Texas, June 25-30. Paper No. 78-52.3.
-------
VII-1
VII. BENEFIT-COST ANALYSIS
This chapter compares the cost of control for the alternative regulations
to the benefits estimated for SO4, PM, and SOg. Table VII-1 presents the re-
sults of the net benefit analysis (benefits minus costs). Table VI1-2 and
Table VI1-3 present a comparison of average and incremental benefit per ton
and cost per ton estimates, respectively.
As can be seen by examining Table VII-1, there are cases where the
annualized net benefits are negative. In drawing any conclusions from this
table, the reader should keep several factors in mind. Foremost, there are
many benefit categories not included in this analysis. Table VI-1, page
VI-5 lists all potential benefit categories for the proposed standards and
indicates which ones are analyzed in this RIA. Of the benefit categories
that are included, there remains incomplete coverage. For example, the SO2
benefit estimates only include calculations for reductions in residential
materials damage. No calculations are included for commercial or industrial
facilities. Potentially significant benefits are not included from the
non-human biological effects category (e.g., acid deposition) and from the
SO4 mortality category.
In addition, geographic coverage is limited. The SO4 and PM benefit
calculations only include the 31 eastern states. Potentially significant
benefits from the western states and from Canada are not included.
Another important consideration is that the RIA compares a range of
benefit estimates to a single cost estimate. This inherently implies a
greater confidence level in the cost estimate than may be justified. If
costs were also presented as a range of estimates, to reflect uncertainties
in the underlying assumptions, the comparison of benefits and costs would be
more valid.
-------
VI1-2
Table VII-1
Net Benefit Analysis
(thousands of 1983 dollars)
Regul atoryl
A1ternative
Total
Annualized
Benefits
Total
Annualized
Control Costs
Annualized
Net Benefits
Low Sulfur Fuel/I
62,000
58,000
4,000
Low Sul fur Fuel/2
85,000
58,000
27,000
Low Sul fur Fuel /3
120,000
58,000
62,000
Precent Reduction/1
93,000,
135,000
-42,000
Percent Reduction/2
126,000
135,000
- 9,000
Percent Reduction/3
178,000
135,000
43,000
1 A range of benefits (1,2,3) is presented for SO4 and PM to reflect
senstivity analysi on the air quality modeling and economic valuation
inputs.
-------
VII-3
Table VI1-2
Average Control Cost Per Ton and Benefit Per Ton
(1983 $)
Average Average
Regulatory Alternative* C/T B/T
Low Sulfur Fuel/1 340 450
Low Sulfur Fuel/2 340 640
Low Sulfur Fuel/3 340 930
Percent Reduction/1 560 470
Percent Reduction/2 560 660
Percent Reduction/3 560 950
1 A range of benefits (1,2,3) is presented for S04 and PM to reflect
sensitivity analysis on the air quality modeling and economic valuation
inputs.
Table VII-3
Incremental Control Cost Per Ton and Benefit Per Ton*
(1983 $)
Incremental ^ Incremental 2
C/T B/T
Percent Reduction/1 1100 510
Percent Reduction/2 1100 690
Percent Reduction/3 1100 1000
* A range of benefits (1,2,3) is presented for SO^ and PM to reflect
sensitivity analysis on the air quality modeling and economic valuation
inputs.
2 Percent reduction over low sulfur fuel.
-------
VI1-4 " ' '
Due to the limited benefit coverage and geographic/temporal scope
of this RIA, it is impossible to draw sound conclusions as to whether
the proposed standards will lead to a welfare improvement for society
relative to the baseline situation. This RIA has not been designed or
intended to portray a complete picture of all potential benefits. No
attempt has been made to quantitatively calculate all benefit categories.
Thus, total benefits will actually be greater than those reported, and net
benefits will be larger. While the direction of change in the benefit
calculation, due to the addition of new benefit categories is known, the
magnitude of the increase is unknown.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA 450/3-86-006
3. RECIPIENT'S ACCESSION NO.
4. title and subtitle Regulatory Impact Analysis for New
Source Performance Standards: Industrial-Commercial-
Institutional Steam Generating Units of Greater Than
100 Million Btu/hr Heat Input
5. REPORT DATE
June 1986
6. PERFORMING ORGANIZATION COOE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
DAA for Air Quality Planning and Standards
Office of Air and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Draft
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES
16. ABSTRACT
New source performance standards limiting emissions of sulfur dioxide (SO-) have been
proposed for new industrial-commercial-institutional steam generating units with heat
input capacities above 100 million Btu/hour. Executive Order 12291 states that a
regulatory impact analysis (RIA) must be performed on any regulation having an annual
effect on the economy of $100 million or more. This RIA was prepared to fulfill this
requirement.
The analysis estimates the benefits and costs of two regulatory alternatives: low
sulfur fuel-based standards and standards requiring a percent reduction in SO2
emissions. This document describes the methodology used in estimating the benefits
and presents the results of the analysis.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENOED TERMS
c. cosati licliI/Group
Air pollution
Standards of performance
Steam generating units
Industrial boilers
Major rule
Benefits analysis
Air pollution control
SO2 benefits
13B
18. DISTRIBUTION STATEMENT
Release unlimited
19. SECURITY CLASS (This Report;
Unclassified
21. NO. OF PAGES
108
20. SECURITY CLASS (This page I
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) previous f.oition is obsolete
------- |