Benefit Analyses
of Alternative
SAMI Strategies:
Selected Health
and Welfare
Methods and
Analysis Results
April 2002
Prepared for
Office of Air Quality Planning and
Standards
U.S. Environmental Protection
Agency
Research Triangle Park. NC"
Prepared by
Abt Associates Inc.
4800 Montgomery' Lane
Bethesda. MD 20814-5341
Work funded through
Contract No. 68-D-98-001
Work Assignment 4-74
Lisa Conner. Work Assignment Manager
Nancy Riley. Project Officer
-------
Disclaimer
This document was developed by Abt Associates Inc. under technical direction from U.S. EPA's
Office of Air Quality Planning and Standards. The analysis and conclusions presented in this report are
those of the authors and should not be interpreted as necessarily reflecting the official views or policies of
the U.S. EPA. The analysis is useful to derive estimates of air quality, costs, benefits, and/or economic
impacts. However, the analysis inputs and outputs associated with any emissions source, county, or local
area are subject to significant uncertainties and should not be used to predict attainment status, costs,
benefits, and/or economic impacts at this level of detail.
-------
Acknowledgments
The Work Assignment Manager, Lisa Conner, as well as Bryan Hubbell of the U.S.
Environmental Protection Agency, provided a variety of constructive suggestions, comments, and
technical direction at all stages of work on this report.
-------
Benefit Analyses of Alternative SAMI Strategies:
Selected Health and Welfare Methods and Results
-------
Table of Contents
Executive Summary ES-1
1. Introduction 1-1
2. Development of PM Air Quality Inputs For Use in the Benefits Analysis 2-1
3. General Issues in Estimating Health and Welfare Benefits 3-1
3.1 Population Projections 3-1
3.2 Change Over Time in Benefit Value in Real Dollars 3-3
3.3 Adjusting Benefit Estimates from 1990 Dollars to 2000 Dollars 3-4
3.4 Characterization of Uncertainty 3-4
3.4.1 Alternative Calculations 3-6
3.4.2 Sensitivity Analyses 3-8
3.4.3 Statistical Uncertainty Bounds 3-9
3.4.4 Unquantified Benefits 3-10
4. Health Benefits . 4-1
4.1 Premature Mortality 4-2
4.1.1 Short-Term Versus Long-Term Studies 4-3
4.1.2 Degree of Prematurity of Mortality 4-3
4.1.3 Estimating PM-Related Premature Mortality 4-3
4.1.4 Valuing Premature Mortality 4-7
4.2 Chronic Illness 4-12
4.2.1 Chronic Bronchitis 4-12
4.3 Acute Illnesses and Symptoms Not Requiring Hospitalization 4-16
4.3.1 Acute Bronchitis 4-16
5. Results 5-1
6. Unquantified Benefits From Other Pollutant Reductions 6-1
6.1 Other Pollutants (NO;, SO:, CO, and HAPs) 6-1
6.2 Extrapolation of Benefits from Other Studies 6-4
7. References 7-1
Appendix A: Results for Sensitivity Analyses A-l
Appendix B: Particulate Matter C-R Functions B-l
B . 1 Mortality B-l
B . 1.1 Mortality (Krewski et al., 2000) Based on ACS Cohort: Mean PM, 5 B-l
B.1.2 Mortality (Krewski et al., 2000), Based on Six-City Cohort: Mean PM; 5 .. B-2
B .1.3 Mortality (Dockery et al., 1993), Based on Six-City Cohort: Mean PM; 5 ... B-3
B .2 Chronic Morbidity B-4
B .2.1 Chronic Bronchitis (Abbey et al., 1995b, California) B-4
B .3 Acute Morbidity B-6
B.3.1 Acute Bronchitis C-R Function (Dockery et al., 1996) B-6
-------
List of Exhibits
Exhibit ES-l Annual Average PM2.5-Related Health Endpoints ES-2
Exhibit ES-2 Estimated Annual Average PM2.5-Related Health Effects Associated with Air Quality
Changes Resulting from the SAMI Control Scenarios ES-3
Exhibit ES-3 Estimated Annual Average PM2.5-Related Health Benefits Associated with Air Quality
Changes Resulting from the SAMI Control Scenarios ES-3
Exhibit 2-1 Changes in Annual Mean PM2.5 Due to SAMI Emission Control Scenarios 2-1
Exhibit 3-1 General Issues in Estimating Health and Welfare Benefits 3-1
Exhibit 3-2 Income Growth Adjustment Factors 3-3
Exhibit 3-3 Consumer Price Indexes Used to Adjust WTP-Based and Cost-of-Illness-Based Benefits
Estimates from 1990 Dollars to 2000 Dollars 3-4
Exhibit 3-4 Key Sources of Uncertainty in the Benefit Analysis 3-6
Exhibit 3-5 Alternative Benefits Calculations for the SAMI Benefit Analyses 3-8
Exhibit 3-6 Sensitivity Analyses for the SAMI Benefit Analyses 3-8
Exhibit 4-1 PM-Related Health Endpoints 4-1
Exhibit 4-2 Unit Values for Economic Valuation of Health Endpoints (2000 $) 4-2
Exhibit 4-3 Mortality Lag Structures Examined in Sensitivity Analyses 4-7
Exhibit 4-4 Summary of Mortality Valuation Estimates 4-9
Exhibit 4-5 Potential Sources of Bias in Estimates of Mean WTP to Reduce the Risk of PM Related
Mortality Based on Wage-Risk Studies 4-11
Exhibit 5-1 Unquantified Endpoints Associated with Pollution Reductions Associated with the SAMI
Emission Control Scenarios 5-3
Exhibit 5-2 Baseline Percentages 5-4'
Exhibit 5-3 Total Quantified Benefits of SAMI: B1 Scenario in 2010 5-4
Exhibit 5-4 Total Quantified Benefits of SAMI: B1 Scenario in 2040 5-5
Exhibit 5-5 Total Quantified Benefits of SAMI: B3 Scenario in 2010 5-6
Exhibit 5-6 Total Quantified Benefits of SAMI: B3 Scenario in 2040 5-7
Exhibit 5-7 Total Health Related Benefits in 2040 by State 5-8
Exhibit 5-8 Alternative Benefit Calculations for the 2010 SAMI "Bl" Scenario 5-9
Exhibit 5-9 Alternative Benefit Calculations for the 2040 SAMI "B1" Scenario 5-10
Exhibit 5-10 Alternative Benefit Calculations for the 2010 SAMI "B3" Scenario 5-11
Exhibit 5-11 Alternative Benefit Calculations for the 2040 SAMI "B3" Scenario 5-12
Exhibit 5-12 Alternative Mortality Calculations for the 2010 SAMI "Bl" Scenario 5-12
Exhibit 5-13 Alternative Mortality Calculations for the 2040 SAMI "Bl" Scenario 5-13
Exhibit 5-14 Alternative Mortality Calculations for the 2010 SAMI "B3" Scenario 5-13
Exhibit 5-15 Alternative Mortality Calculations for the 2040 SAMI "B3" Scenario 5-14
Exhibit 6-1 Comparison of Annual PM: ? and Other PM-related Benefits From Previous Analyses ... 6-5
Exhibit 6-2 Comparison of Annual PM, 5 and Ozone Benefits From Previous Analyses 6-6
Exhibit A-1 Sensitivity Analysis Results for the 2010 SAMI "Bl" Scenario A-l
Exhibit A-2 Sensitivity Analysis Results for the 2010 SAMI "B3" Scenario A-l
Exhibit A-3 Sensitivity Analysis Results for the 2040 SAMI "BI" Scenario A-2
Exhibit A-4 Sensitivity Analysis Results for the 2040 SAMI "B3" Scenario A-2
Exhibit A-5 Sensitivity Analysis: Effect of Thresholds on Estimated PM-Related Mortality Based on
Krewski et al. (2000) - Mean, All-Cause for the 2010 SAMI "Bl" Scenario A-3
Exhibit A-6 Sensitivity Analysis: Effect of Thresholds on Estimated PM-Related Mortality Based on
Krewski et al. (2000) - Mean, All-Cause for the 2010 SAMI "B3" Scenario A-4
Exhibit A-7 Sensitivity Analysis: Effect of Thresholds on Estimated PM-Related Mortality Based on
Krewski et al. (2000) - Mean, All-Cause for the 2040 SAMI "Bl" Scenario A-5
Exhibit A-8 Sensitivity Analysis: Effect of Thresholds on Estimated PM-Related Mortality Based on
Krewski et al. (2000) - Mean. All-Cause for the 2040 SAMI "B3" Scenario A-6
-------
Executive Summary
The Southern Appalachian Mountains Initiative (SAMI) commissioned an "integrated
assessment" to assess the environmental effects and selected socioeconomic costs and benefits of SAMI-
designed reduction strategies. Though four socioeconomic topics were considered in the assessment,1 the
topics were very narrow in scope in terms of the possible set of avoided health and welfare effects
achievable under each control strategy. In fact, human mortality was the only health effect considered in
the integrated assessment. This analysis is meant to provide an alternative evaluation of the same SAMI
control strategies through the estimation of health benefits associated with SAMI-related pollution
reductions.
The SAMI-designed emission reduction strategies proposed progressively more stringent
emission reduction controls in each of five major source categories (utility, industrial, highway vehicle,
non-road engines, and area sources) for 2010 and 2040. However, because human mortality was the only
health effect considered, and because the relationship between mortality and exposures to air pollution is
based on an annual average measure of PM2.5, SAMI did not estimate the lull impact of each control
strategy on the range of potential air pollutants. Instead, when converting these strategies into input data
for the socioeconomic analyses, the SAMI atmospheric modeling contractor only provided Abt
Associates with annual average PM2.5 profiles for each of three future control strategies. The first, and
least stringent, strategy is referred to as the "A2" scenario, which serves as the base level of future-year
air quality from which changes in air quality conditions are calculated. The remaining two scenarios are
named the "Bl" scenario and the "B3" scenario. These strategies serve as the control scenarios, where B1
is more stringent than A2, and B3 is more stringent than Bl.
We base our analysis on the assumptions and models that have been approved by the EPA
Science Advisory Board and are typically utilized by EPA to assess national regulatory programs.
Specifically, this analysis relies upon the methods used in the analysis of EPA's Heavy Duty
Engine/Diesel Fuel Rule and described in detail in the Heavy Duty Diesel Technical Support Document
(Abt Associates, 2000). We estimate not only human mortality, but other health effects associated with
exposures to annual measures of PM2.5. To account for unquantifiable benefits associated with the range
of potential SAMI-related air quality improvements, we consider, qualitatively, the benefits associated
with exposures to daily measures of PM2.5, PM10 and particulate matter between 2.5 and 10 microns
(coarse PM10), NOx, S02, ozone, and others.
Estimation of Health Effects
We estimated PM2.5-related health effects using the Criteria Air Pollutant Modeling System
(CAPMS). CAPMS is a population-based system for modeling exposures of populations to ambient
levels of criteria pollutants that we use to estimate health benefits. CAPMS divides the United States into
eight kilometer by eight kilometer grid cells and estimates the changes in incidence of adverse health
effects associated with given changes in air quality in each grid cell. Total incidence changes are the sum
of grid cell-specific changes.
The SAMI annual average PM2.5 data came to Abt Associates at the modeled grid cell level; a
nested grid structure comprised of an inner set of 12x12 km grid cells and an outer set of coarser 24x24
km grid cells. The data represented predicted annual average values at the center point of each grid cell.
The SAMI modeling domain covered a geographical range that extended beyond the eight state SAMI
1 The topics covered by the assessment include fishing, hiking/enjoying scenery, stewardship sense of place, and lifestyle
changes.
Abt Associates Inc. ES-I April 2002
-------
region while not entirely covering the extent of the eight SAMI states. Coverage of the 8 state SAMI
region was about 85%. This analysis limited the consideration of annual PM2.5 data to those grid cells
whose centers fell within the eight state region. We then assigned each 8x8 km CAPMS grid cell to the
nearest SAMI grid cell by calculating the shortest distance between the center of the CAPMS grid cell to
the center of a SAMI grid cell.
Using the suite of health effect studies considered in the Heavy Duty Diesel analysis as our guide,
we identified three annual average PM2.5-related endpoints for inclusion in the analysis; mortality
associated with long-term PM2.5 exposures, chronic bronchitis, and acute bronchitis. Exhibit ES-1
contains details about each health effect, the study upon which the concentration-response function is
based, and its associated valuation.
Exhibit ES-1 Annual Average PM2.5-Related Health Endpoints
Endpoint
Population
[ Study
Mean Estimate'
I Uncertainty Distribution"
Mortality
Associated with
long-term exposure
Ages 30+
Krewski et al. (2000).
reanalysis of Pope et al.
(1995) using the annua! mean
and all-cause mortality
S6.324 million
per statistical life
Weibull distribution, mean = $6,324
million; std. dev. = 4.27 million.
Chronic Illness
Chronic Bronchitis
>26
Abbey et al. (1995b)
S340.568 per case
A Monte Carlo-generated distribution,
based on three underlying distributions.
Respiratory Symptoms/Illnesses Not Requiring Hospitalization
Acute bronchitis
Ages 8-12
Dockery et al. (1989)
S59.29 per case
Continuous uniform distribution over
fS 17.13. S101.451.
"The derivation of each of Ibe estimates is discussed in the main text. All WTP-based dollar values were obtained by multiplying
rounded 1990 S values used in the $812 Prospective Analysis by 1.318 to adjust to 2000 S.
Health Effect Results: Incidence and Valuation
The total dollar benefit associated with a given endpoint depends on how much the endpoint will
change (e.g., how many premature deaths will be avoided) and how much each unit of change is worth
(e.g., how much a premature death avoided is worth). Exhibit ES-2 summarizes the mean changes in
incidence associated with each'SAMI control scenario in each future year. Exhibit ES-3 summarizes the
mean valuation (in 2000$) associated with the changes in incidence across all endpoints (mortality,
chronic bronchitis, and acute bronchitis) for each SAMI control scenario in each future year.
We note that the benefits presented in Exhibit ES-3 include an adjustment for the impact of
expected growth in real income on future year benefit estimates. The factors were calculated by EPA for
use in the Heavy Duty Standards RIA (U.S. EPA, 2000), and are discussed fully in the main text.
Abt Associates Inc.
April 2002
-------
Exhibit ES-2 Estimated Annual Average PM2.5-Related Health Effects Associated with Air Quality
Changes Resulting from the SAM1 Control Scenarios
Endpoint
Mean Avoided Incidence (cases/year)
2010 BJ
2040 B1
2010 B3
2040 B3
Mortality
1.662
4.273
6,155
8.007
Chronic Bronchitis
1,258
3,303
4,531
6,051
Acute Bronchitis
3,464
8,952
12,192
16,177
Exhibit ES-3 Estimated Annual Average PM2.5-Related Health Benefits Associated with Air
Quality Changes Resulting from the SAMI Control Scenarios
Endpoint
Mean Monetary Benefits (millions 2000$)
2010 B1
2040 B1
2010 B3
2040 B3
Mortality''
SI 1,114
$33,332
S41,163
S62.457
Chronic Bronchitis
S483
SI.508
SI.740
S2.763
Acute Bronchitis
SO.2
S0.6
SO.8
Sl.l
Total
.SI 1.597 + B"
S34.841 + B
S42.904 + B
S65.221 4 B
" Calculated using a 3% discount rate in the mortality lag adjustment. See main text for the discussion on mortality lags.
"B represents benefits associated with the SAMI control scenarios but not captured by health effects estimated using the
available annual average PM2.5 data. These benefits include avoided health effects associated with reductions in daily PM2.5.
PM10 and coarse PM10. ozone. N02, S02, and CO, hazardous air pollutants, and nitrogen deposition. B also represents benefits
associated with improvements in visibility and recreational fishing, calculated for the SAMI Integrated Assessment.
Uncertainty
As with any complex analysis such as this one, there are a wide variety of sources for uncertainty.
Some key sources of uncertainty in each stage of the benefits include:
• gaps in scientific data and inquiry;
• variability in estimated relationships, such as C-R functions introduced through differences in
study design and statistical modeling;
• errors in measurement and projections for variables such as population growth rates;
• errors due to misspecification of model structures, excluded variables, and simplification of
complex functions;
• biases due to omissions or other research limitations.
The above benefits are considered primary estimates for this analysis, based on the best available
scientific literature and methods. Where possible, we attempt to provide estimates of the effects of
uncertainty about key analytical assumptions. In the main text, we address uncertainty by presenting
alternative calculations, sensitivity analyses, and probabilistic assessments associated with the annual
average PM2.5-related health effects. They include:
Abt Associates Inc
ES-3
April 2002
-------
• Alternative Calculations - Estimates of mortality based on alternative studies; Valuation of
avoided premature mortality inpidence based on statistical life years; Age-based adjustments to
the value of a statistical life lost; Estimation and valuation of reversals in chronic bronchitis.
• Sensitivity Analyses - Calculation of the impact varying threshold assumptions have on the
estimation of mortality incidence; Calculation of the impact different lag structures have on the
estimation of benefits associated with avoided mortality incidence.
• Statistical Uncertainty Bounds - The total dollar benefit associated with a given endpoint
depends on how much the endpoint will change due to the assumptions in the control scenarios
(e.g., how many premature deaths will be avoided) and how much each unit of change is worth
(e.g., how much a premature death avoided is worth). Based on these distributions, we use Monte
Carlo methods to provide estimates of the 5,h and 95th percentile values of the distribution of
estimated health effect endpoint incidence and valuation.
Unqualified Benefits From Other Pollutant Reductions
One significant limitation of the SAMI health benefits analyses is the inability to quantify many
of the adverse effects associated with exposures to pollutants other than annual average PM2.5. Though
estimates of PM2.5-related mortality and chronic and acute bronchitis may have captured the bulk of the
economic benefits associated with reducing emissions in the SAMI region, we still miss a variety of
potential benefits because there are a limited number of epidemiological studies based on annual PM2.5.
Benefits missed in the SAMI analysis likely include:
• Other PM Effects (daily PM2.5, PM10 and coarse PM10) - In analyses conducted for the EPA,
benefit estimates related to hospital admissions, emergency room visits, lower and upper
respiratory symptoms, work loss days, MRADs, and recreational visibility improvements have
equaled between 3 to 5% of benefits related to annual average PM2.5 effects.
• Ozone Effects - Across the same EPA analyses, benefits of ozone related hospital admissions,
emergency room visits, MRADs, decreased worker productivity and agricultural crop losses have
equaled between 2 to 24% of benefits related to annual average PM2.5 effects.
• N02, S02, and CO Effects - These pollutants are generally related to a small subset of effects;
the most serious of which is perhaps hospitalization for heart-related problems. There have been
studies finding some evidence that N02 and CO are linked to mortality but it is difficult to
determine if these effects are in addition to effects associated with PM and ozone.
• Effects of Air Toxics - Air toxics encompass a broad range of harmful chemical compounds
that are either released directly into the air or formed in secondary reactions in the air, water, and
soil. Exposure to air toxics can result in cancer, noncancer health effects, and ecological damage.
The large number of air toxics, and the difficulties associated with estimating the impact of
changes in emissions of air toxics, make these effects extremely hard to quantify.
• Nitrogen Deposition Effects - Excess nutrient loads, especially that of nitrogen, are responsible
for a variety of adverse consequences to the health of estuarine and coastal waters. These effects
include: toxic and/or noxious algal blooms such as brown and red tides; low (hypoxic) or zero
(anoxic) concentrations of dissolved oxygen in bottom waters; the loss of submerged aquatic
vegetation due to the light-filtering effect of thick algal mats; and fundamental shifts in
phytoplankton community structure.
Abt Associates Inc.
ES-4
April 2002
-------
1. Introduction
In 1992, the Southern Appalachian Mountains Initiative (SAMI) was created to "identify and
recommend reasonable measures to remedy existing and prevent future adverse effects from human-
induced air pollution on the air quality related values of the Southern Appalachians, primarily those of
Class I parks and wilderness areas, weighing the environmental and socioeconomic implications of any
recommendations"(SAMI 2001). To do this, SAMI commissioned an "integrated assessment" to estimate
the environmental effects and selected socioeconomic costs and benefits of SAMI-designed emissions
reduction strategies. Four socioeconomic topics are covered by the integrated assessment, including: 1)
fishing; 2) hiking/enjoying scenery; 3) stewardship/sense of place; and 4) lifestyle changes.
The SAMI emission strategies were designed to propose progressively more stringent emissions
reduction controls in each of five major source categories (utility, industrial, highway vehicle, non-road
engines, and area sources) for 2010 and 2040. Converting these strategies into input data for the
socioeconomic analyses, the SAMI atmospheric modeling contractor generated air quality profiles for
each of three future control strategies. Each strategy represented a series of different assumptions,
including different applied control technologies, implementation of regulations and incentives, and
demand for goods and services. The first, and least stringent, strategy is referred to as the "A2" scenario,
or the reference strategy. This scenario served as the base level of future-year visual air quality from
which changes in air quality conditions are calculated.2 The remaining two scenarios are named the "Bl"
scenario and the "B3" scenario. These strategies served as the control scenarios, where Bl was more
stringent (i.e.. lower emissions) than A2, and B3 was more stringent than Bl. A complete description of
the assumptions present within each of these emission reduction strategies can be found in the 2001
SAMI Interim Report.
Though four socioeconomic topics were considered in the integrated assessment, the topics were
very narrow in scope in terms of the possible set of avoided health and welfare effects achievable under
each control strategy.3 In fact, human mortality was the only health effect considered until it was
removed from the SAMI integrated assessment. Before it was removed, however, the SAMI air modeling
contractor provided air quality inputs for the mortality analysis. Yet, because the relationship between
mortality and exposures to air pollution is based on an annual average measure of PM,,, the atmospheric
modelers did not estimate the full impact of each control strategy on the range of potential air pollutants.
Instead, only annual PM: < was provided, though improvements in ozone, sulfur dioxide (SO,), nitrogen
oxide (NOx), PMI0 and its associated coarse fraction (coarse PM,0) were likely to occur.
This analysis is meant to provide an alternative evaluation of the same SAMI control strategies
through the estimation of the health benefits associated with reductions in annual PM2 5. This includes not
only human mortality, but other health effects associated with exposures to annual measures of PM2 5.
We base our analysis on the assumptions and models that have been approved by the EPA Science
Advisory Board and are typically utilized by EPA to assess national regulatory programs. Specifically,
this analysis relies upon the methods used in the analysis of EPA's Heavy Duty Engine/Diesel Fuel Rule
: Benefits are calculated based on measured changes in air quality between a future-year base-case scenario and a future-
year control scenario.
' The SAMI Integrated Assessment estimated recreational visibility benefits related to visual air quality improvements
associated with the SAMI emission control scenarios. The Integrated Assessment also estimated recreational fishing benefits related
to water qualitv improvements associated with the SAMI emission control scenarios. Both of these benefit categories are added to
the health benefits we calculate in this analysis to capture the magnitude of total benefits associated with the SAMI emission control
scenarios. Section 5 presents total benefit results.
Abt Associates Inc.
1-1
April 2002
-------
and described in detail in the Heavy Duty Diesel Technical Support Document (Abt Associates, 2000).
To account for unquantifiable benefits associated with the potential reduction, we also present a
qualitative discussion of benefits associated with exposures to daily measures of PM2 5, PM,0 and
particulate matter between 2.5 and 10 microns (coarse PM,0), NOx, S02, ozone, and others.
Section 2 describes the method used to develop the PM air quality inputs for use in the benefits
analysis. Section 3 describes general issues arising in estimating and valuing changes in adverse health
effects associated with changes in PM. Section 4 describes in some detail the methods used for
estimating and valuing adverse health effects, while Section 5 presents the results of these analyses.
Section 6 presents a discussion of the unquantified benefits from other pollutant reductions that are likely
associated with the SAMI emission control scenarios.
This document also has two appendices. Appendix A presents the physical and monetary benefits
associated with sensitivity calculations for the SAMI emission control scenarios not considered in the
primary analysis. Appendix B presents the PM C-R functions used in this analysis.
Abt Associates Inc.
1-2
April 2002
-------
2. Development of PM Air Quality Inputs For Use in the Benefits Analysis
The following section summarizes how we use the SAMI annual average PM2 5 air quality model
results in conjunction with the Criteria Air Pollutant Modeling System (CAPMS) to estimate PM2 5
exposure.
CAPMS is a population-based system for modeling exposures of populations to ambient levels of
criteria air pollutants that we use to estimate health benefits.4 CAPMS divides the United States into eight
kilometer by eight kilometer grid cells, and estimates the changes in incidence of adverse health and
welfare effects associated with given changes in air quality in each grid cell. We then calculate the total
incidence change as the sum of grid-cell-specific changes.
Contractors for the SAMI integrated assessment forecasted annual average PM2, data associated
with each of the future-year (2010 and 2040) emission control scenarios; EPA provided this data, in
spreadsheet format, to Abt Associates. Data was provided at the modeled grid cell level; a nested grid
structure comprised of an inner set of 12x12 km grid cells and an outer set of coarser 24x24 km grid cells.
Location information (i.e., latitude and longitude) was also provided. The data represented predicted
annual average values at the center point of each grid cell.
The modeled air quality data predicted for the SAMI analysis, and used here, covers a
geographical range that extends beyond the eight state SAMI region while not entirely covering the extent
of the eight SAMI states. In fact, coverage of the 8 State SAMI region is about 85%. Some counties did
not appear in the database because they were outside the 12x12 or 24x24 grid cell areas. These excluded
areas include: southern Alabama and Georgia and western Tennessee and Kentucky. This analysis
limited the consideration of annual PM: ? data to those grid cells whose centers fell within the eight state
SAMI region. We then assigned each 8x8 km CAPMS grid cell to the nearest SAMI grid cell by
calculating the shortest distance between the center of the CAPMS grid cell to the center of a SAMI grid
cell.
Exhibit 2-1 presents the population-weighted average change in annual mean PM2.5 between
each of the SAMI emission control scenarios and their respective baselines. For the sake of comparison,
we have also included the same measure of PM2.5 from a recent analysis, the Heavy Duty Engine/Diesel
Fuel Rule Analysis.
Exhibit 2-1 Changes in Annual Mean PM2.5 Due to SAMI Emission Control Scenarios
Statistic
Annual Mean PM2.5 by Scenario (ug/m3)
2010 B1
2040 B1
2010 B3
2040 B3
2030 HDD Analysis
Population-Weighted Average
Change from Baseline
1.14
2.43
4.22
4.54
0.65
4 CAPMS does not model indiv idual exposures to these pollutants. For a complete description of CAPMS. please refer to
the Heavy Duty Diesel Technical Support Document (Abt Associates, 2000).
Abt Associates Inc.
2-1
April 2002
-------
3. General Issues in Estimating Health and Welfare Benefits
Changes in PM result in changes in a number of health effects, or ".endpoints," that society
values. The Heavy Duty Diesel Analysis Technical Support Document (HDD TSD) (Abt Associates,
2000) discussed key issues in the estimation of adverse health effects and in the valuation of health and
welfare benefits. For common issues between the Heavy Duty Diesel Analysis and the SAMI analysis,
we refer the reader to the HDD TSD. Exhibit 3-1 lists these common issues. For issues specific to the
estimation of benefits for the SAMI analysis, we include their discussion below.
Exhibit 3-1 General Issues in Estimating Health and Welfare Benefits
1. Estimating Adverse Health Effects
-The basic concentration-response model.
-Calculation of adverse health effects with CAPMS.
-Overlapping health effects.
-Baseline incidences.
-Thresholds.
-Application of a single C-R function everywhere.
-Estimating pollutant-specific benefits using single pollutant vs. multi-pollutant models.
2. Valuing Changes in Health and Welfare Effects
-Willingness-to-Pay estimation.
-Aeareeation of monetized benefits.
Because this analysis is limited to the evaluation of effects related to PM2.5, we have also
tailored the characterization of uncertainty to reflect uncertainties associated with PM2.5-related health
effects. We again refer the reader to the HDD TSD for the discussion of uncertainty characterization
associated with daily average PM2.5- and PMlO-related endpoints.
3.1 Population Projections
Benefits for the SAMI analysis are based on health effect incidence changes due to predicted air
quality improvements in the years 2010 and 2040. Integral to the estimation of such benefits is an
accurate estimate of future population projections.
The underlying data used to create county-level 2040 population projections is based on: (1) 1990
county-level population statistics for all U.S. counties collected by the U.S. Census (Wessex, 1994), and
(2) future-year state and metropolitan area population estimates provided by the Bureau of Economic
Analysis (1995). Growth factors are calculated using the BEA data and are applied to the 1990 county-
level populations.
A growth factor is calculated by taking the ratio of an estimated region's 2030 population divided
by the 1990 population for that same area. Population estimates for the years 1990-93, 2000, 2005, 2010,
2015, 2025 and 2045 were collected by the BEA. A 2040 population estimate was not provided. Instead,
2040 state and metropolitan area populations were interpolated linearly using estimates from the years
2025 and 2045.
Abt Associates Inc.
3-1
April 2002
-------
Growth factors are calculated for both urban areas and rural areas. An urban area is defined as a
county that falls within a metropolitan area. This includes metropolitan statistical areas (MSAs), primary
metropolitan statistical areas (PMSAs), consolidated metropolitan statistical areas (CMSAs), and New
England county metropolitan areas (NECMAs), as defined by U.S. Census Bureau.5 In this section,
however, all metropolitan areas are referred to as MAs. A rural area is defined as a county that falls
outside the defined metropolitan areas.
Urban areas grow according to the growth rate calculated for the particular metropolitan area
within which they are located. This adjustment is very straightforward, simply taking the ratio of future
year to base year metropolitan area population and multiplying that factor by the base year county
population. The equation is:
— FutureMAPop,
FutureCountyropi = 1990C ountyPopi —
1990 MAPop,
where:
FutureCountyPopj = projected 2010 or 2040 population in urban county i
1990CountyPop, = actual 1990 population for county i
FutureMAPop, = projected 2010 or 2040 population in metropolitan area for county i
1990MAPopi = actual 1990 population for metropolitan area for county i.
Rural areas grow according to the growth rate calculated for the particular state within which they
are located, adjusted to subtract out metropolitan area populations. Before the ratio of future year to base
year state population is calculated, the population attributed to all metropolitan areas located within that
state is subtracted from the future year and base year population totals. Once this metropolitan area
adjustment has been made, the rural growth factor is multiplied by the base-year population in all non-
MA counties to get future-year population projections.
To calculate the future year population, we use the following equation:
(FutureStatePopi -1 Future MAPop,)
FutureCountyPop, = \
-------
not have future year county population projections with which to apportion future year metropolitan area
populations. To remedy this, we apply the same percent of the population a given county contributes to a
metropolitan area in 1990 to 2010 and 2040 metropolitan areas when apportioning populations between
states.
The above procedure refers to population estimates at the county level. CAPMS, however,
apportions population estimates to the CAPMS grid cell level. To do this, CAPMS uses census-derived
1990 block group population estimates. Each block group has a centroid. For each centroid that is
located within a CAPMS grid cell, the grid cell is assigned that population. To inflate 1990 population
estimates to a future year estimation of population within a CAPMS grid cell, county level ratios,
calculated using the county level estimates described above, are applied to CAPMS grid cells that fall
within a particular county. There are a few inaccuracies with this procedure. CAPMS grid cells and
census block groups do not share similar borders. When a block group centroid is assigned to a CAPMS
grid cell, there may be some overlap with other grid cells. The total block group population, however, is
assigned only to the CAPMS grid cell in which it is located. A similar issue exists when assigning
county-level ratios to CAPMS grid cells. The county in which a grid cell is located is determined by the
grid cell center. However, the grid cell center may overlap with other counties. Both issues may lead to
the assignment of populations or adjustment factors to the wrong area. The overall magnitude of the
discrepancy, however, is slight because of the small area each of the block groups and grid cells represent.
3.2 Change Over Time in Benefit Value in Real Dollars
The value placed on benefits, or willingness to pay (WTP), for health-related environmental
improvements (in real dollars) could change between now and the years 2010 and 2040. If real income
increases between now and the year 2040, for example, it is reasonable to expect that WTP, in real
dollars, would also increase. Based on historical trends, the U.S. Bureau of Economic Analysis projects
that, for the United States as a whole as well as for regions and states within the U.S., mean per capita real
income will increase. For the U.S. as a whole, for example, mean per capita personal income is projected
to increase by about 16 percent from 1993 to 2005 (U.S. Bureau of Economic Analysis, 1995).
The monetary benefits presented in this Technical Support Document (TSD) have been adjusted
to account for changes over time in real income. A complete description of the theory behind the income
growth adjustment and its application to the benefits analysis can be found in Chapter 7 of the Heavy
Duty Diesel Regulatory Impact Analysis (EPA 2000b). Exhibit 3-2 displays the adjustment factors used
in this analysis.
Exhibit 3-2 Income Growth Adjustment Factors
Year
Endpoint
Minor Illness
Severe/Chronic Illness
Mortality
2010
1.0380
1.1274
1.1124
2040
1.0949
1.3407
1.2972
Abt Associates Inc.
3-3
April 2002
-------
3.3
Adjusting Benefit Estimates from 1990 Dollars to 2000 Dollars
This section describes the method used to convert benefits estimates into constant dollars. In past
RIA analyses performed for the EPA, cost and benefit estimates have been presented in constant 1990
dollars. Benefits estimates in this analysis, however, are presented in constant 2000 dollars. To adjust
benefits estimates from 1990 dollars to 2000 dollars, the method of adjustment depends on the basis of the
benefits estimates.6 For the SAMI analysis, all of the benefit estimates are based on direct estimates of
WTP.
Benefit estimates based directly on estimates of WTP have been adjusted using the CPI-U for "all
items." The CPI-Us, published by the U.S. Dept. of Labor, Bureau of Labor Statistics, can be found in
Council of Economic Advisers (2000, Table B-58). An overview of the adjustments from 1990 to 2000
dollars for WTP-based valuations is given in Exhibit 3-3.
Exhibit 3-3 Consumer Price Indexes Used to Adjust WTP-Based and Cost-of-Ulness-Based Benefits
Estimates from 1990 Dollars to 2000 Dollars
1990
(J)
2000
(2)
Adjustment
Factor"
(2)/(D
Relevant Endpoints
CPI-U for "All Items" k
130.7
172.2
1.3175
WTP-based valuation:
1. Statistical lives savedc
2. Chronic bronchitis
3. Morbidity endpoints using WTP J
* Benefits estimates in 1990 dollars are multiplied by the adjustment factor to derive benefits estimates in 2000 dollars.
"Source: Dept. of Labor. Bureau of Labor Statistics; reported in Council of Economic Advisers (2000. Table B-58)
' Adjustments to 1990 S were originally made by Industrial Economics Inc. using the CP1-L1 for "all items" (IEcl992).
J Adjustments of WTP-based benefits for morbidity endpoints to 1990 S were originally made by Industrial Economics Inc. (1993)
using the CPI-U for "all items."
3.4 Characterization of Uncertainty
In any complex analysis using estimated parameters and inputs from numerous different models,
there are likely to be many sources of uncertainty. This analysis is no exception. There are many inputs
that are used to derive the final estimate of benefits, including emission inventories, air quality models
(with their associated parameters and inputs), epidemiological estimates of C-R functions, estimates of
values, population estimates, income estimates, and estimates of the future state of the world, i.e.
regulations, technology, and human behavior. Each of these inputs may be uncertain, and depending on
their location in the benefits analysis, may have a disproportionately large impact on final estimates of
total benefits. For example, emissions estimates are used in the first stage of the analysis. As such, any
uncertainty in emissions estimates will be propagated through the entire analysis. When compounded
with uncertainty in later stages, small uncertainties in emissions can lead to much larger impacts on total
benefits.
h For example, benefit analyses in the past have included estimates based on direct estimates of WTP, cost of illness, and
earnings. For each, the adjustment from 1990 to 2000 dollars is different.
Abt Associates Inc.
3-4
April 2002
-------
Exhibit 3-4 summarizes the wide variety of sources for uncertainty in this analysis. Some key
sources of uncertainty in each stage of the benefits analysis are:
• gaps in scientific data and inquiry
• variability in estimated relationships, such as C-R functions, introduced through differences in
study design and statistical modeling
• errors in measurement and projection for variables such as population growth rates
• errors due to misspecification of model structures, excluded variables, and simplification of
complex functions
• biases due to omissions or other research limitations.
Our approach to characterizing model uncertainty in the estimate of total benefits is to present a
primary estimate, based on the best available scientific literature and methods, and to provide estimates of
the effects of uncertainty about key analytical assumptions. However, in some cases, it was not possible
to quantify uncertainty. For example, many benefits categories, while known to exist, do not have
enough information available to provide a quantified or monetized estimate. The uncertainty regarding
these endpoints is such that we could determine neither a primary estimate nor a plausible range of values.
Another source of uncertainty related to the SAMI analysis is the limited air quality data available
with which to conduct the benefit analysis. Only those health effects associated with annual mean PM25,
and for which a C-R function was available, were included in the analysis. However, the SAMI control
scenarios, "Bl" and "B3", project that there will also be reductions in PMI0, ozone, NOx and SO;. To the
extent that health effects are related to these other pollutant reductions, the current SAMI analysis will
underestimate the total benefits associated with the alternative SAMI pollution control scenarios. The
extent to which benefits are underestimated, however, is impossible to determine, however Section 6
discusses in detail the health effects associated with exposures to these additional pollutants and their
relative contribution to total benefits in other policy analyses.
This report also addresses uncertainty by presenting alternative calculations, sensitivity analyses,
and probabilistic assessments associated with the annual average PM; 5-related health effects. We discuss
each approach in turn.
Abt Associates Inc.
3-5
April 2002
-------
Exhibit 3-4 Key Sources of Uncertainty in the Benefit Analysis
1. Uncertainties Associated With CowennytjottResEonseJunctw^
-The value of the PM-coefficient in each C-R function.
-Application of a single C-R function to pollutant changes and populations in all locations.
-Similarity of future year C-R relationships to current C-R relationships.
-Correct functional form of each C-R relationship.
-Extrapolation of C-R relationships beyond the range of PM concentrations observed in the study.
-Application of C-R relationships only to those subpopulations matching the original study population.
2. Uncertainties Associated With PM Concentrations
-Responsiveness of the models to changes in precursor emissions resulting from the control policy.
-Projections of future levels of precursor emissions, especially ammonia and crustal materials.
-Model chemistry for the formation of ambient nitrate concentrations.
3. Uncertainties Associated with PM Mortality Risk
-No scientific literature supporting a direct biological mechanism for observed epidemiological evidence.
-Direct causal agents within the complex mixture of PM have not been identified.
-The extent to which adverse health effects are associated with low level exposures that occur many times in the year versus peak
exposures.
-Possible confounding in the epidemiological studies of PM; <. effects with other factors (e.g.. other air pollutants, weather,
indoor/outdoor air, etc.).
-The extent to which effects reported in the long-term exposure studies are associated with historically higher levels of PM rather
than the levels occurring during the period of study.
-Reliability of the limited ambient PM, < monitoring data in reflecting actual PM; < exposures.
4. Uncertainties Associated With Possible Lagged Effects
-The portion of the PM-related long-term exposure mortality effects associated with changes in annual PM levels would occur in a
single year is uncertain as well as the portion that might occur in subsequent years.
5. Uncertainties Associated With Baseline Incidence Rates
-Some baseline incidence rates are not location-specific (e.g., those taken from studies) and may therefore not accurately represent
the actual location-specific rates.
-Current baseline incidence rates may not approximate well baseline incidence rates in 2010 or 2040.
-Projected population and demographics may not represent well future-year population and demographics.
6. Uncertainties Associated With Economic Valuation
-Unit dollar values associated with health and welfare endpoints are only estimates of mean WTP and therefore have uncertainty
surrounding them.
-Mean WTP (in constant dollars) for each type of risk reduction may differ from current estimates due to differences in income or
other factors.
7. Uncertainties Associated With Aggregation of Monetized Benefits
-Health benefit estimates are limited to the available C-R functions. Thus, unquantified or unmonetized benefits are not included.
-Health benefit estimates are limited to the available air quality data. Though only annual mean PM2.5 is considered in this
analysis, the SAM1 control scenarios will also bring reductions in other pollutants, such as PM10, ozone, NOx and S02. The
avoided health effects associated with these reductions, though potentially_signifjcanL_aiQgt_guannfied^^
3.4.1 Alternative Calculations
The alternative calculations included in this analysis are based on relatively plausible alternatives
to the assumptions used in deriving the primary benefit estimates. We do not attempt to assign
probabilities to these alternative calculations, as we believe this would only add to the uncertainty of the
Abt Associates Inc.
3-6
April 2002
-------
analysis or present a false picture about the precision of the results.7 Instead, the reader is invited to
examine the impact of applying the different assumptions on the estimate of total benefits. While it is
possible to combine all of the alternative calculations with a positive impact on benefits to form a "high"
estimate or all of the alternative calculations with a negative impact on benefits to form a "low" estimate,
we do not recommend this because the probability of all of these alternative assumptions occurring
simultaneously is likely to be very low. Instead, the alternative calculations are intended to demonstrate
the sensitivity of our benefits results to key parameters which may be uncertain. Exhibit 3-5 summarizes
the alternative calculations included in this analysis.
Studies Used for Alternative Calculations
A number of studies that estimate plausible alternative relationships between PM exposure and
premature mortality are presented as alternative calculations to the mortality study included in the primary
analysis (Krewski et al., 2000, mean all-cause mortality). These alternative mortality functions are
discussed in more detail in Section 4.
The value of statistical life years alternative calculation recognizes that individuals who die from
air pollution related causes tend to be older than the average age of individuals in the VSL studies used to
develop the $5.9 million value. To employ the value of statistical life-year (VSLY) approach, we first
estimated the age distribution of those lives projected to be saved by reducing air pollution. Based on life
expectancy tables, we calculate the life-years saved from each statistical life saved within each age and
gender cohort. To value these statistical life-years, we hypothesized a conceptual model which depicted
the relationship between the value of life and the value of life-years. The average number of life-years
saved across all age groups for which data were available is 14 for PM-related mortality. The average for
PM, in particular, differs from the 35-year expected remaining lifespan derived from existing wage-risk
studies. Using the same distribution of value of life estimates used above, we estimated a distribution for
the value of a life-year and combined it with the total number of estimated life-years lost.
An alternative to the calculation of life-years lost is age-based adjustments to the value of a
statistical life lost based on empirical estimates of WTP by age. Several studies conducted by Jones-Lee,
et al. (1985; 1989; 1993) found a significant effect of age on the value of mortality risk reductions
expressed by citizens in the United Kingdom. We used the results of the Jones-Lee et al. analysis to
calculate age-specific values of a statistical life. As described below, we started with the value of a
statistical life lost by an individual of about age 40, and then adjusted it with age-specific factors. We use
40 as the base because we use wage risk studies in developing the value of a statistical life, and the
average age in the wage-risk studies is about 40.
We apportioned the number of lives saved in each of the age groups used in the statistical life-
years-lost alternative calculation to the age groups used by Jones-Lee et al. (1989; 1993). We then
multiplied the number of lives saved in an age group by the age-adjusted value of a statistical life saved
for that age group. To calculate the value of a statistical life saved in an age group, we multiplied $6.12
million by the ratio of the WTP for mortality risk reduction in that age group to the WTP for mortality
risk reduction in the age 40-59 group, as reported by Jones-Lee et al. (1989; 1993). The five-year lag
structure used in the primary method is also applied under two alternative discount rate assumptions of
1 Some recent benefit-cost analyses in Canada and Europe (Lang et al., 1995; Holland et al.. 1999) have estimated ranges
of benefits by assigning ad hoc probabilities to ranges of parameter values for different endpoints. Although this does generate a
quantitative estimate of an uncertainty range, the estimated points on these distributions are themselves highly uncertain and very
sensitive to the subjective judgements of the analyst. To avoid these subjective judgements, we choose to allow the reader to
determine the weights they would assign to alternative estimates.
Abl Associates Inc.
3-7
April 2002
-------
three percent and seven percent. Because the two Jones-Lee studies reported different ratios, this
alternative calculation is carried out separately using each of the two Jones-Lee studies.
Reversals in chronic bronchitis incidences are defined as those cases where an individual reported
having chronic bronchitis at the beginning of the study period but reported hot having chronic bronchitis
in follow-up interviews at a later point in the study period. Since, by definition, chronic diseases are
long-lasting or permanent, if the disease goes away it is not chronic. In the primary analysis, these
reversals are given a value of zero. As an alternative calculation, we estimate reversals and value each as
a case of the mildest form of chronic bronchitis.
Exhibit 3-5 Alternative Benefits Calculations for the SAMI Benefit Analyses
Alternative Calculations
' Description
PM-related premature mortality
A number of studies provide an alternative estimate of the relationship between chronic
PM exposure and mortality.
Value of avoided premature
mortality incidences based on
statistical life years
Calculate the incremental number of life-years lost from exposure to changes in ambient
PM and use the value of a statistical life year based on a S5.9 million value of a statistical
life.
Age-based adjustments to the value
of a statistical life lost
Results of the Jones-Lee et al. (1985: 1989; 1993) analysis were used to calculate age-
based adjustment factors to adjust the value of a statistical life lost by an individual of
about age 40 to age-specific values.
Reversals in chronic bronchitis
treated as lowest severity cases
Instead ot omitting those cases of chronic bronchitis that reverse after a period of time,
they are treated as being cases with the lowest severity rating.
3.4.2 Sensitivity Analyses
In addition to alternative calculations, we perform sensitivity analyses, briefly described in
Exhibit 3-6. Sensitivity analyses, as opposed to alternative calculations, examine the sensitivity of
estimated benefits results to less plausible alternatives to the assumptions used in the primary analysis.
Sensitivity calculations also demonstrate the sensitivity of our benefits results to key analytical
parameters. The sensitivity analyses calculated for this analysis includes an examination of how a PM
concentration threshold could influence mortality incidence estimates, and alternative lag structures when
valuing mortality. Results from the sensitivity analyses are presented in Appendix A.
Exhibit 3-6 Sensitivity Analyses for the SAMI Benefit Analyses
Sensitivity Analysis
Description
Threshold assumptions
Calculate the impact varying threshold assumptions have on the estimation of mortality
incidence based on the Krewski et al. (2000) study.
Alternative mortality lag structures
Calculate the impact different lag structures have on the estimation of benefits associated
with avoided mortalitv incidence |
3.4.3 Statistical Uncertainty Bounds
Although there are several sources of uncertainty affecting estimates of endpoint-specific
benefits, the sources of uncertainty that are most readily quantifiable in this analysis are the C-R
Abt Associates Inc.
3-8
April 2002
-------
relationships and uncertainty about unit dollar values. The total dollar benefit associated with a given
endpoint depends on how much the endpoint will change due to the SAMI emission control scenarios
(e.g., how many premature deaths will be avoided) and how much each unit of change is worth (e.g., how
much a premature death avoided is worth).8 Based on these distributions, we provide estimates of the 5'b
and 95th percentile values of the distribution of estimated benefits. However, we hasten to add that this
omits important sources of uncertainty, such as the contribution of air quality changes, baseline
population incidences, projected populations exposed, transferability of the C-R function to diverse
locations, and uncertainty about premature mortality. Thus, a confidence interval based on the standard
error would provide a misleading picture about the overall uncertainty in the estimates. The empirical
evidence about uncertainty is presented where it is available.
Both the uncertainty about the incidence changes and uncertainty about unit dollar values can be
characterized by distributions. Each "uncertainty distribution" characterizes our beliefs about what the
true value of an unknown (e.g., the true change in incidence of a given health effect) is likely to be, based
on the available information from relevant studies.9 Unlike a sampling distribution (which describes the
possible values that an estimator of an unknown value might take on), this uncertainty distribution
describes our beliefs about what values the unknown value itself might be. Such uncertainty distributions
can be constructed for each underlying unknown (such as a particular pollutant coefficient for a particular
location) or for a function of several underlying unknowns (such as the total dollar benefit of a
regulation). In either case, an uncertainty distribution is a characterization of our beliefs about what the
unknown (or the function of unknowns) is likely to be. based on all the available relevant information.
Uncertainty statements based on such distributions are typically expressed as 90 percent credible
intervals. This is the interval from the fifth percentile point of the uncertainty distribution to the ninety-
fifth percentile point. The 90 percent credible interval is a "credible range" within which, according to
the available information (embodied in the uncertainty distribution of possible values), we believe the true
value to lie with 90 percent probability.
The uncertainty about the total dollar benefit associated with any single endpoint combines the
uncertainties from these two sources, and is estimated with a Monte Carlo method. In each iteration of
the Monte Carlo procedure, a value is randomly drawn from the incidence distribution and a value is
randomly drawn from the unit dollar value distribution, and the total dollar benefit for that iteration is the
product of the two.10 If this is repeated for many (e.g., thousands of) iterations, the distribution of total
dollar benefits associated with the endpoint is generated.
Using this Monte Carlo procedure, a distribution of dollar benefits may be generated for each
endpoint. The mean and median of this Monte Carlo-generated distribution are good candidates for a
point estimate of total monetary benefits for the endpoint. As the number of Monte Carlo draws gets
larger and larger, the Monte Carlo-generated distribution becomes a better and better approximation to the
underlying uncertainty distribution of total monetary benefits for the endpoint. In the limit, it is identical
to the underlying distribution.
' Because this is a regional analysis in which, for each endpoint. a single C-R function is applied everywhere, there are
tw o sources of uncertainty about incidence: (1) statistical uncertainty (due to sampling error) about the true value of the pollutant
coefficient in the location where the C-R function was estimated, and (2) uncertainty about how well any given pollutant coefficient
approximates incidence in areas beyond where the C-R function was estimated.
* Although such an "uncertainly distribution" is not formally a Bavesian posterior distribution, it is very similar in concept
and function (see, for example, the discussion of the Bayesian approach in Kennedy 1990, pp. 168-172).
111 This method assumes that the incidence change and the unit dollar value for an endpoint are stochastically independent.
Abr Associates Inc.
3-9
April 2002
-------
3.4.4 Unquantified Benefits
In considering the monetized benefits estimates, the reader should remain aware of the
limitations. One significant limitation of the SAMI health benefits analyses is the inability to quantify
many of the adverse effects associated with exposures to pollutants other than PM2.5. Section 6
discusses these unquantified benefits in detail. Another limitation of the SAMI analyses is that for many
additional health and welfare effects associated with exposures to PM, such as PM-related materials
damage, reliable C-R functions and/or valuation functions are not currently available. In general, if it
were possible to monetize these benefits categories, the benefits estimates presented in this RIA would
increase. In addition to unquantified benefits, there may also be environmental costs that we are unable
to quantify. Several of these environmental cost categories are related to nitrogen deposition, while one
category is related to the issue of ultraviolet light. The net effect of excluding benefit and disbenefit
categories from the estimate of total benefits depends on the relative magnitude of the effects.
Abt Associates Inc.
3-10
April 2002
-------
4. Health Benefits
Typically, in benefit analyses of air policies, the most significant monetized benefits of reducing
ambient concentrations of PM are associated with reductions in health risks associated with the fine
particulate portion of PM, PM,5. This Section describes individual effects and the methods used to
quantify and monetize changes in the expected number of incidences of various health effects related to
changes in PM, 5. Though only effects associated with annual mean PM, 5 are valued in this analysis,
many additional benefits are associated with incidence changes associated with reductions in other
pollutants like ozone, PM10, NOx, and SO,. Yet, because we only have data for PM, 5, these other
benefits can not be quantified. We discuss these endpoints qualitatively in Section 6.
We estimate the incidence of adverse health effects using C-R functions based on PM,5. The
changes in incidence of PM-related adverse health effects and corresponding monetized benefits
associated with these changes are estimated separately. The PM-related health endpoints for which C-R
functions are estimated are shown in Exhibit 4-1. The unit monetary values for each of these endpoints,
and associated uncertainty distributions, are presented in Exhibit 4-2. Issues relating to the calculation of
changes in incidence and the monetization of these changes are discussed below for each endpoint.
Note also that in some cases there are alternative endpoints, studies, or unit dollar values that
could be used in calculating the benefits of a change in pollution. Again, this analysis follows the
methodology the EPA's Heavy Duty Diesel Analysis and the Section 812 Analysis, which have
undergone extensive review and critique. In following this methodology, these alternatives are presented
where appropriate in Exhibits 4-1 and 4-2 in italics to indicate that they are not used in the primary
analysis but may be used in alternative analyses. Appendix B presents the functional forms for each C-R
function and how they were derived.
Exhibit 4-1 PM-Related Health Endpoints
Endpoint 1
Population !
PM |
Study
Mortality
Associated with long-term exposure
Ages 30+
PM.,
Krewski et al. (2000), reanalysis of Pope
et al. (1995) using the annual mean and
all-cause mortality
Associated with long-term exposure'
Ages 30+
PM:,
Krewski et al. (2000), reanalysis of
Dockeiy et al. (1993)
Associated with long-term exposure
Ages 27+
PM:s
Docker}- et al. (1993)
Chronic Illness
Chronic Bronchitis
>26
PM;,
Abbey et al. (1995b)
Respiratory Symptoms/Illnesses Not Requir
ing Hospitalization
Acute bronchitis
Ages 8-12
r pm,.
1 Dockery et al. (1989)
" Italicized entries are either alternative or supplemental calculations to the endpoints and'or studies used in the primary analysis.
Abl Associates Inc.
4-1
April 2002
-------
Exhibit 4-2 Unit Values for Economic Valuation of Health Endpoints (2000 $)
Health Endpoint
Mean Estimate"
Assumed Uncertainty Distribution *
Mortality
Value of a statistical life
$6,324 million per statistical life
Weibull distribution, mean = $6,324 million;
std. dev. = 4.27 million.
Chronic Bronchitis
WTP approach
$340,568 per case
A Monte Carlo-generated distribution, based on three underlying
distributions.
Respiratory Ailments N<
>t Requiring Hospitalization
Acute bronchitis
j $59.29 per case
Continuous uniform distribution over f$17.13, $101,451.
"The derivation of each of the estimates is discussed in the text. All WTP-based dollar values were obtained by multiplying
rounded 1990 S values used in the §812 Prospective Analysis by 1.318 to adjust to 2000 S. Entries in italics are not used in the
primary benefits analysis.
4.1 Premature Mortality
Health researchers have consistently linked air pollution, especially PM, with increases in
premature mortality. A substantial body of published scientific literature recognizes a correlation
between elevated PM concentrations and increased mortality rates. For instance, studies have found
associations between day-to-day particulate air pollution and increased risk of various adverse health
outcomes, including cardiopulmonary mortality (Pope et al., 2000). Fine particles, PM; 5, are the largest
health concern because they can be breathed most deeply into the lung. Much of this literature is
summarized in 1996 PM Criteria Document (US EPA, 1996) and the Heavy Duty Diesel Regulatory
Impact Analysis (US EPA, 2000).
These epidemiological studies typically estimate the relationship between air quality changes and
the relative risk of a health effect, rather than estimate the absolute number of avoided cases of premature
mortality. For example, the risk of mortality at ambient PM level x0 relative to the risk of mortality at
ambient PM level x may be characterized by the ratio of the two mortality rates: the mortality rate among
individuals when the ambient PM level is x0 and the mortality rate among (otherwise identical)
individuals when the ambient PM level is x. We incorporate the relative risk into a concentration
response function so that the effects of changes in PM concentrations on mortality can be estimated by a
count of the expected number of deaths avoided due to a given reduction in PM concentrations. An
alternative measure is to infer the number of years of life that are saved by a given reduction in PM
concentrations: years of life that each individual was expected to live and that would have been lost had
the reduction in PM concentrations not occurred. To provide a range of the possible cost of premature
mortality, we estimate both measures of mortality in this analysis.
There are two types of exposure to elevated levels of air pollution that may result in premature
mortality. Acute (short-term) exposure (e.g., exposure on a given day) to peak pollutant concentrations
may result in excess mortality on the same day or within a few days of the elevated exposure. Chronic
(long-term) exposure (e.g., exposure over a period of a year or more) to levels of pollution that are
generally higher may result in mortality in excess of what it would be if pollution levels were generally
lower. The excess mortality that occurs will not necessarily be associated with any particular episode of
elevated air pollution levels.
4.1.1 Short-Term Versus Long-Term Studies
Abi Associates Inc. 4-2 April 2002
-------
There are two types of epidemiological studies that examine the relationship between mortality
and exposure. Long-term studies (e.g., Pope et al., 1995) estimate the association between long-term
(chronic) exposure to air pollution and the survival of members of a large study population over an
extended period of time. Such studies examine the health endpoint of concern in relation to the general
long-term level of the pollutant of concern, for example, relating annual mortality to some measure of
annual pollutant level. Daily peak concentrations would impact the results only insofar as they affect the
measure of long-term (e.g., annual) pollutant concentration. In contrast, short-term studies relate daily
levels of the pollutant to daily mortality. By their basic design, daily studies can detect acute effects but
cannot detect the effects of long-term exposures. A chronic exposure study design (a prospective cohort
study, such as the Pope study) is best able to identify the long-term exposure effects, and may detect some
of the short-term exposure effects as well. Because a long-term exposure study may detect some of the
same short-term exposure effects detected by short-term studies, including both types of study in a benefit
analysis would likely result in some degree of double counting of benefits. While the long-term study
design is preferred, these types of studies are expensive to conduct and consequently there are relatively
few well designed long-term studies.
4.1.2 Degree of Prematurity of Mortality
It is possible that the short-term studies are detecting an association between PM and mortality
that is primarily occurring among terminally ill people. Critics of the use of short-term studies for policy
analysis purposes correctly point out that an added risk factor that results in terminally ill people dying a
few days or weeks earlier than they otherwise would have (referred to as "short-term harvesting") is
potentially included in the measured PM mortality "signal" detected in such a study. While some of the
detected excess deaths may have resulted in a substantial reduction in lifespan, others may have resulted
in a relatively small decrease in lifespan. Studies by Spix et al (1993) and Pope et al. (1992) yield
conflicting evidence, suggesting that harvesting may represent anywhere from zero to 50 percent of the
deaths estimated in short-term studies. However, recent work by Zeger et al. (1999) and Schwartz (2000)
that focused exclusively on this issue, reported that short-term harvesting does not play a major role in the
PM-mortality relationship."
It is not likely, however, that the excess mortality reported in a long-term prospective cohort
study like Pope et al. (1995) contains any significant amount of this short-term harvesting. The Cox
proportional hazard statistical model used in the Pope study examines the question of survivability
throughout the study period (ten years). Deaths that are premature by only a few days or weeks within
the ten-year study period (for example, the deaths of terminally ill patients, triggered by a short duration
PM episode) are likely to have little impact on the calculation of the average probability of surviving the
entire ten-year interval.
4.1.3 Estimating PM-Related Premature Mortality
The benefits analysis estimates PM:, -related mortality using the C-R function estimated by
Krewski et al. (2000). This study is a reanaiysis of Pope et al. (1995), which estimated the association
between long-term (chronic) exposure to PM, 5 and the survival of members of a large study population.
Our decision to use Pope et al. in previous benefits analyses reflected the EPA Science Advisory
1 'Zeger et al. (1999. p. 171) reported that: "The TSP-mortality association in Philadelphia is inconsistent with the
harvesting-only hypothesis, and the harvesting-resistant estimates of the TSP relative risk are actually larger - not smaller - than the
ordinary estimates."
Abt Assoc iates Inc. 4-3 April 2002
-------
Board's1" explicit recommendation for modeling the mortality effects of PM in both the Clean Air Act
§812 Retrospective Report to Congress and the Clean Air Act §812 Prospective study (U.S. EPA, 1999a,
p. 12). An advantage of Krewski et al. over Pope et al. is that Krewski et al.'s reanalysis uses the annual
mean PM2.5 concentration rather than the annual median. Because the mean is more readily affected by
high PM values than is the median, if high PM days are actually important Tn causing premature
mortality, the annual mean may be a preferable measure of long-term exposure than the median.
The Krewski et al. (2000) long-term study is selected for use in the benefits analysis instead of
short-term (daily pollution) studies for a number of reasons. It is used alone- rather than considering the
total effect to be the sum of estimated short-term and long-term effects— because summing creates the
possibility of double-counting a portion of PM-related mortality. The Krewski et al. study and the Pope
study it reanalyzes are considered preferable to other available long-term studies because they use better
statistical methods, have a much larger sample size, the longest exposure interval, and more locations (51
cities) in the United States, than other studies.
It is currently unknown whether there is a time lag (a delay between changes in PM exposures
and changes in mortality rates) in the chronic PM/premature mortality relationship. The existence of such
a lag is important for the valuation of premature mortality incidences because economic theory suggests
that benefits occurring in the future should be discounted. Although there is no specific scientific
evidence of the existence or structure of a PM effects lag, current scientific literature on adverse health
effects, such as those associated with PM (e.g., smoking related disease) and the difference in the effect
size between chronic exposure studies and daily mortality studies suggest that it is likely that not all
incidences of premature mortality reduction associated with a given incremental change in PM exposure
would occur in the same year as the exposure reduction. This same smoking-related literature implies that
lags of up to a few years are plausible. Following explicit advice from the SAB, we assume a five-year
lag structure, with 25 percent of premature deaths occurring in the first year, another 25 percent in the
second year, and 16.7 percent in each of the remaining three years (U.S. EPA, 1999c, p. 9). It should be
noted that the selection of a five-year lag structure is not directly supported by any PM-specific literature.
Rather, it is intended to be a best guess at the appropriate time distribution of avoided incidences of PM-
related mortality.
Alternative Calculations: PM-Related Premature Mortality
Although we use the Krewski, et al. (2000) mean-based ("PM2.5(DC), All Causes") model
exclusively to derive our primary estimates of avoided premature mortality, we also examine the impacts
of selecting alternative C-R functions for premature mortality. There are several candidates for
alternative C-R functions, some from the Krewski, et al. study, and others from the original ACS study by
Pope et al. (1995) or from the "Harvard Six-City Study" by Dockery et al. (1993), however, for this
analysis, we are limited to examining those based on annual mean PM, 5.
The Krewski et al. (2000) reanalysis provides results for several models that control for spatial
correlations in the data. Krewski et al. pointed out that "if not identified and modeled correctly, spatial
correlation could cause substantial errors in both the regression coefficients and their standard errors,"
Science Advisory Board was established by Congress to provide independent scientific and engineering
advice to the EPA Administrator on the technical basis for EPA regulations. Expressed in terms of the current parlance of the risk
assessment/risk management parad.gm of decision making (National Research Council, Managing Risk in the Federal Government.
), the SAB deals with risk assessment issues (hazard identification, dose-response assessment, exposure assessment and risk
characterization) and only that portion of risk management that deals strictly with the technical issues associated with various
control options. Source: http: '/www .ena.gov/sab/overview him April 5, 2002
Abt Associates Inc.
4-4
April 2002
-------
^SiTn PM These modc- s are based on the original ACS air quality dataset. which contained
? £ COnc™°nS' "f "y- °Ur C* for premature mortality would be
both based on the mean and adjusted for regional variability. Unfortunately, Krewski et al. do not
provtde such an earnrne As such, we have chosen to use the mean-based relative risk in our primary
analysis, because the SAMI analysis only generated annual mean PM, „ we" are unable to use the median-
based regionally adjusted relative risks to provide alternative estimates exploring the impact of
adjustments for spatial correlations.
thoughwe are unable to estimate median-based mortality incidence based on relative risks
adjusted for spatial correlation, we can compare the relative risks between our primary mortality model
and the alternative Krewski model that is based on the annual median and adjusted for regional
variability The pnmaiy model we use has a reported relative risk of 1.12; the alternative model has a
reported relati ve risk of 1 16. If the two models shared the same PM metric (average or median) and all
e se was equa , t e model with the adjustment for spatial correlation would yield the larger estimate of
avoided mortality incidence. However, the two models are not based on the same PM metric. As we
exp aine a ove, a mean-based estimate is preferred to the median because changes in the mean more
accurate y re ect c anges in peak values than do changes in the median. For emission control scenarios
that affect peak PM days more than average PM days, larger changes in the mean will be observed
compared to median changes. Because of the competing effects of the spatial correlation adjustment and
the different PM metrics, it is difficult to say how the adjustment for spatial correlation impacts the
overall estimate of avoided incidence in a given policy analysis. We can, however, compare the impact
using each relative risk had on the estimate of mortality incidence in the Heavy Duty Diesel Analysis.
The median-based, spatial correlation adjusted estimate of avoided mortality incidence was approximately
13% greater than the primary estimate of avoided mortality incidence (Abt Associates. 2000).
Krewski, et al. (2000) also reanalyzed the data from another prospective cohort study (the
Harvard Six Cities Study ) authored by Dockery et al. (1993). The Dockery et al. study used a smaller
sample of individuals from fewer cities than the study by Pope et al. (1995); however, it features
improved exposure estimates, a slightly broader study population (adults aged 25 and older), and a
follow-up period nearly twice as long as that of Pope et al. The SAB has noted that "the [Harvard Six
Cities] study had better monitoring with less measurement error than did most other studies" (U.S. EPA.
1999d, p. 10).
/, nnc\ ery a'- (1993) study finds a larger effect of PM on premature mortality relative to the
ope et a . ( ) stu y. To provide a more complete picture of the range of possible premature mortality
ris s t at may e associated with long-term exposures to fine particles, we also present alternative
estimates ase on t e Krewski et al. (2000) reanalysis of the Dockery et al. data and the original study
estimates. The Health Review Committee (2000, p. 270) commentary noted the "inherent limitations of
using only six cities, understood by the original investigators, should be taken into account when
interpreting the results of the Six Cities Study." We emphasize, that based on our understanding of the
relative merits of the two datasets, the Krewski et al. ACS model based on mean PM„ levels in 63 cities
is the most appropriate model for analyzing the premature mortality impacts of the HD Engine/Diesel
Fuel rule. It is thus used for our primary estimate of this important health effect.
Some of the functions are based on changes in mean PM25 concentrations while others are based
on median PM; 5 concentrations. Given the available SAMI air quality data, we have only considered
those based on annual mean changes. Estimated reductions in premature mortality will depend on both
the size of the C-R coefficient and the change in the annual mean PM,, metric. We also estimated
alternative premature mortality incidence using both non-accidental and all-cause mortality rates. In
previous benefit analyses conducted for the EPA, premature mortality was calculated using non-
accidental mortality rates. For the sake of comparability to previous analyses, we included estimates of
premature mortality based on both rates.
Abl Associates Inc.
4-5
April 2002
-------
Sensitivity Calculation: Mortality Lag Structure
Just when PM-related mortality occurs in relation to exposure to PM is uncertain. We do not
know what percentage of PM-related mortality occurs in the same year as exposure, in the following year,
and so forth. To account for the uncertainty about possible lags in PM-related mortality, we examine the
sensitivity of mortality-related benefits to alternative lag structures. Exhibit 4-3 presents the lags that are
used in these sensitivity calculations. As stated earlier, the primary analysis uses a five-year lag structure
in the valuation of mortality and chronic bronchitis, with incidence apportioned as follows: 25 percent in
the first year, 25 percent in the second year, and 16.67 percent in each of the last three years.
To examine the effect alternate lag-structures have on the estimation of both mortality and
chronic bronchitis valuation, the mortality benefits will be calculated using five alternative lag structures.
Lag 1 will apportion the occurrence of all incidence to the first year. Valuation of these cases will not be
discounted. In lag 2, based on the length of the study period for the Dockery et al. (1993) study, 100
percent of mortality incidence occurs in fifteen years from the modeled future-year. Lag 3, based on the
length of the study period for the Pope et al. (1995) study, assigns 100 percent of the occurrence of
mortality incidence to the eighth year out from the modeled future-year. Lag 4 front loads the occurrence
of mortality incidence. Incidence is apportioned in decreasing amounts out to fifteen years. Lag 5
apportions incidence over fifteen years, assigning a lesser percentage of incidence in the beginning years,
with the percentage of incidence increasing over time out to fifteen years." The latter two lag structures
are intended to show how the distribution of incidences within a lag period affects benefit estimates.
Sensitivity Calculation: Threshold Analysis
To examine the effect an implied PM threshold has on the estimation of health effects in this
analysis, we applied an increasingly stringent threshold to the Krewski et al. (2000) mortality function in
one ug/m3 increments. The results of this sensitivity analysis can be found in Appendix A.
Abl Associates Inc.
4-6
April 2002
-------
Exhibit 4-3 Mortality Lag Structures Examined in Sensitivity Analyses
Year
Primary
Sensitivity 1
Sensitivity- 2
Sensitivity 3
Sensitivity 4
Sensitivity 5
1
25
100
0
0
30
1
2
25
0
0
0
25
1
3
16.67
0
0
0
15
]
4
16,67
0
0
0
6
2
5
16.67
0
0
0
4
2
6
0
0
0
0
3
2
7
0
0
0
0
3
2
8
0
0
0
100
3
3
9
0
0
0
0
2
3
10
0
0
0
0
2
3
11
0
0
0
0
2
4
12
0
0
0
0
2
6
13
0
0
0
0
1
15
14
0
0
0
0
1
25
15
0
0
100
0
1
30
See Appendix A for results of the sensitivity analys.s and its implications on the avoided mortality incidence benefits total.
4.1.4 Valuing Premature Mortality
Three methods for valuing avoided prematire mortality are presented in this analysis. The firs,
and primary one ,s the statical lives lost approach, which derives the value of a "statistical life" lost
from '"formation about what people are wtlltng to pay for mortal risk reduction. In contrast to the
statist,ca hves lost approach, the second and third valuation approaches Uy to take into account that an
individual s willingness to pay for mortal risk reduction may depend on his age. Using these approaches,
¦he value of an avoided premature death depends on the age a, which the individual dies. In allLe
methods we assume for this analysis that PM-related premature mortality is distributed over the five
years following exposure (the five-year mortality lag). To take this into account in Ihe valuation of
reductions in premature deaths^ we apply an annual three percent discount rate to the value of avoided
premature deaths occurring in future years.
Statistical Lives Lost
We estimate the monetary benefit of reducing premature mortality risk using the value of a
"statistical life lost" approach, even though the actual valuation is of small changes in mortality risk
experienced by a large number of people. The estimated value of a "statistical life lost" is an intermediate
value from a variety of estimates m the economics literature, and is a value that EPA has frequently used
in RIAs for other rules. This estimate is the mean of a distribution fitted to the estimates from 26 value-
of-life studies identified in the §812 study as "applicable to policy analysis." The approach and set of
selected studies mirrors that of Viscusi (1992) (with the addition of two studies), and uses the same
criteria used by Viscusi in his review of value-of-life studies. The estimate is consistent with Viscusi's
conclusion (updated to 2000 $) that most of the reasonable estimates of the value of life are clustered in
the $4 to S9.2 million range. Uncertainty associated with the valuation of premature mortality avoided is
expressed through a Weibull distribution (see Exhibit 4-3) (IEc 1992, p. 2).
Abt Associates Inc.
4-7
April 2002
-------
Five of the 26 studies are contingent valuation (CV) studies, which directly solicit WTP
information from subjects; the rest are wage-risk studies, which base WTP estimates on estimates of the
additional compensation demanded in the labor market for riskier jobs. The 26 studies are listed in
Exhibit 4-4. The references for all but Gegax et al. (1985) and V.K. Smith (1983) may be found in
Viscusi (1992). Although each of the studies estimated the mean WTP (MWTP) for a given reduction in
mortality risk, the amounts of reduction in risk being valued were not necessarily the same across studies,
nor were they necessarily the same as the amounts of reduction in mortality risk that would actually be
conferred by a given reduction in ambient pollutant concentrations.
The transferability of estimates of the value of a statistical life from the 26 studies to this analysis
rests on the assumption that, within a reasonable range, WTP for reductions in mortality risk is linear in
risk reduction, or equivalently, that the marginal willingness to pay curve is horizontal within a
reasonable range. For example, suppose a study estimates that the average WTP for a reduction in
mortality risk of 1/100,000 is $30. Suppose, however, that the actual mortality risk reduction resulting
from a given air quality improvement is 1/10,000. If WTP for reductions in mortality risk is linear in risk
reduction, then a WTP of $30 for a reduction of 1/100,000 implies a WTP of $300 for a risk reduction of
1/10,000 (which is ten times the risk reduction valued in the study). Under the assumption of linearity,
the estimate of the value of a statistical life does not depend on the particular amount of risk reduction
being valued.
Abl Associates Inc.
4-8
April 2002
-------
Exhibit 4-4 Summary of Mortality Valuation Estimates
Study
Kneisner and Leeth (1991) (US)
Smith and Gilbert (1984)
Dillingham (1985)
Butler (1983)
Miller and Guria(1991)
Moore and Viscusi (1988)
Viscusi et al. (1991)
Gegax et al. (1985; 1991)
Type of Estimate
Labor Market
Labor Market
Labor Market
contingent Valuation
Labor Market
Contingent Valuation
contingent Valuation
Valuation (millions 2000 $)
0.8
0.9
1.2
1.5
1.6
3.3
3.6
4 3
Marin and Psacharopoulos (1982)
Kneisner and Leeth (1991) (Australia)
Labor Markel
Labor Market
3.7
4 3
Gerking et al. (1988)
contingent Valuation
4.5
Cousineau et al. (1988; 1992)
Labor Market
4.7
Jones-Lee (1989)
contingent Valuation
5.0
Dillingham (1985)
Labor Markel
5 1
Viscusi (1978; 1979)
Labor Market
5.4
R.S. Smith (1976)
Labor Market
6.1
V.K. Smith (1983)
Labor Market
6.2
Olson (1981)
Labor Market
6.9
Viscusi (1981)
Labor Market
8.6
R.S. Smith (1974)
Labor Market
9.5
Moore and Viscusi (1988)
Labor Market
9.6
Kneisner and Leeth (1991) (Japan)
Labor Market
10.0
Herzog and Schlottman (1987; 1990)
Labor Market
12.0
Leigh and Folson (1984)
Labor Market
12.8
Leigh (1987)
Labor Market
13.7
Garen (1988)
Labor Market
17.8
Although the particular amount of mortality risk reduction being valued in a study may not affect
the transferability of the WTP estimate from the study to this analysis, the characteristics of the study
subjects and the nature of the mortality risk being valued in the study could be important. Certain
Cfar^0th!, P°Pulatl,on5ecled and the mortality risk facing that population are believed to
affect the M WTP to reduce the risk. The appropriateness of the MWTP estimates from the 26 studies for
valuing the mortality-related benefits of reductions in ambient air concentrations therefore depends not
only on the quality of the studies (i.e., how well they measure what they are tiying to measure), but also
on (1) the extent to which the subjects in the studies are similar to the population affected by changes in
ambient a.r concentrations and (2) the extent to which the risks being valued are similar.
Focusing on the wage-risk studies, which make up the substantial majority of the 26 studies relied
upon, the likely differences between (1) the subjects in these studies and the population affected by
changes m air concentrations and (2) the nature of the mortality risks being valued in these studies and the
Abl Associates Inc. 4.9 April 2002
-------
nature of air pollution-related mortality risk are considered The direction of bias in which each difference
is likely to result is also considered.
Compared with the subjects in wage-risk studies, the population believed to be most affected by
air pollution (i.e., the population that would receive the greatest mortality risk reduction associated with a
given reduction in air concentrations) is, on average, older and probably more risk averse. For example,
citing Schwartz and Dockery (1992) and Ostro et al. (1996), Chestnut (1995) estimated that
approximately 85 percent of those who die prematurely from ambient air pollution-related causes are over
65. The average age of subjects in wage-risk studies, in contrast, is well under 65.
There is also reason to believe that those over 65 are, in general, more risk averse than the general
population while workers in wage-risk studies are likely to be.less risk averse than the general population.
Although Viscusi's (1992) list of recommended studies excludes studies that consider only much-higher-
than-average occupational risks, there is nevertheless likely to be some selection bias in the remaining
studies — that is, these studies are likely to be based on samples of workers who are, on average, more
risk-loving than the general population. In contrast, older people as a group exhibit more risk averse
behavior.
In addition, it might be argued that because the elderly have greater average wealth than those
younger, the affected population is also wealthier, on average, than wage-risk study subjects, who tend to
be blue collar workers. It is possible, however, that among the elderly it is largely the poor elderly who
are most vulnerable to air pollution-related mortality risk (e.g., because of generally poorer health care).
If this is the case, the average wealth of those affected by a reduction in air concentrations relative to that
of subjects in wage-risk studies is uncertain.
The direction of bias resulting from the age difference is unclear, particularly because age is
confounded by risk aversion (relative to the general population). It could be argued that, because an older
person has fewer expected years left to lose, his WTP to reduce mortality risk would be less than that of a
younger person. This hypothesis is supported by one empirical study, Jones-Lee et al.(1985), that found
the value of a statistical life at age 65 to be about 90 percent of what it is at age 40. Citing the evidence
provided by Jones-Lee et al.. Chestnut (1995) assumed that the value of a statistical life for those 65 and
over is 75 percent of what it is for those under 65.
The greater risk aversion of older people, however, implies just the opposite. Citing Ehrlich and
Chuma (1990), Industrial Economics Inc. (1992) noted that "older persons, who as a group tend to avoid
health risks associated with drinking, smoking, and reckless driving, reveal a greater demand for reducing
mortality risks and hence have a greater implicit value of a life year." That is, the more risk averse
behavior of older individuals suggests a greater WTP to reduce mortality risk.
There is substantial evidence that the income elasticity of WTP for health risk reductions is
positive (Loehman and De, 1982; Jones-Lee et al., 1985; Mitchell and Carson, 1986; Gerking et al., 1988;
Alberini et al., 1997). However, there is uncertainty about the exact value of this elasticity. Individuals
with higher incomes (or greater wealth) should, then, be willing to pay more to reduce risk, all else equal,
than individuals with lower incomes or wealth. Whether the average income or level of wealth of the
population affected by ambient air pollution reductions is likely to be significantly different from that of
subjects in wage-risk studies, however, is unclear.
Finally, although there may be several ways in which job-related mortality risks differ from air
pollution-related mortality risks, the most important difference may be that job-related risks are incurred
voluntarily whereas air pollution-related risks are incurred involuntarily.
Abt Associates Inc.
4-10
April 2002
-------
There is some evidence that people will pay more to reduce involuntarily incurred risks than risks
incurred voluntarily (e.g., Violette and Chestnut, 1983). Job-related risks are incurred voluntarily
whereas air pollution-related risks are incurred involuntarily. If this is the case, WTP estimates based on
wage-risk studies may be downward biased estimates of WTP to reduce involuntarily incurred ambient air
pollution-related mortality risks.
The potential sources of bias in an estimate of MWTP to reduce the risk of air pollution related
mortality based on wage-risk studies are summarized in Exhibit 4-5. Although most of the individual
factors tend to have a downward bias (i.e. the given WTP estimate is understated), the overall effect of
these biases is unclear.
Exhibit 4-5 Potential Sources of Bias in Estimates of Mean WTP to Reduce the Risk of PM Related
Mortality Based on Wage-Risk Studies
Factor
Likely Direction of Bias in Mean WTP Estimate
Age
Uncertain
Degree of Risk Aversion
Downward
Income
Downward, if the elderly affected are a random sample of the elderly. It is
unclear, if the elderly affected are the poor elderly.
Risk Perception: Voluntary vs. Involuntary risk
Downward
Alternative Calculation: Statistical Life-Years Lost
In an alternative calculation, we value statistical life-years, rather than valuing statistical lives.
Moore and Viscusi (1988) value a statistical life-year lost, by assuming that the WTP to save a statistical
life is the value of a single year of life times the expected number of years of life remaining for an
individual. They suggest that a typical respondent in a mortal risk study has a life expectancy of an
additional 35 years. Using a mean estimate of $4.8 million (1990 $) to save a statistical life, their
approach yields an estimate of $137,000 per life-year lost or saved, assuming no discounting. If an
individual discounts future additional years using a standard discounting procedure, the value of each life-
year lost must be greater than the value assuming no discounting. Using a 35 year life expectancy, a
$6,324 million value of a statistical life, and a three percent discount rate, the implied value of each life-
year lost is $293,807 in 2000 dollars.
In addition, the "statistical life-years lost" analysis must accommodate the five-year lag structure.
For each person dying at a given age, using the expected number of years remaining for that age, based on
1997 life expectancy tables (National Center for Health Statistics, 1999, Table 5), and a VSLY of
$293,807, we calculate the present discounted value (discounted back to the beginning of the year of
death) for that person. All values are then discounted back to the beginning of the scenario year (2010 or
2040), whether the individual dies in that year or in a subsequent year. The present discounted value
(discounted back to the beginning of the scenario year) of an avoided premature mortality will vary from
one individual to another, depending on the age of the individual at death and on the extent of lag
between exposure and death. The age at death determines the expected number of life-years lost, while
the extent of lag between exposure and death determines the amount of discounting-needed.
Alternativ e Calculation: Age-Based Adjustments of the Value of a Statistical Life Lost
Abt Associates Inc. 4-11 Am-
-------
There are drawbacks to the "statistical life-years lost" annrn h u
Scientific Advisoiy Board (SAB) notes*that "inferring the value of however. In a recent report, the
assumptions about the discount rate and about the time path of e* 3 St^tlstlcal life year... requires
EPA, 2000a, p. 8). In considering the merits of age-based adjustme t ^ °f consumPtion" (U.S.
theoretically appropriate method is to calculate WTp for individual "uB also notes that "the
the affected population, and that it is preferable to base these calculi 3gCS corresPond to those of
by age." Several studies conducted by Jones-Lee, et al. (1985- 1989-Tqq °n empirical estimates of WTP
age on the value of mortality risk reductions expressed by citizens in'th n f0U"d 3 significant effect of
Lee-based analysis suggests a U-shaped relationship between age and vcr Kingdom- The Jones-
declining to between 60 and 90 percent of the mean VSL value for i H h peaking around age 40, and
declining further as individuals age. This finding has been suDDorfpH , Uals over the a8e of 70, and
by Krupnick, et al. (2000; 2000), which asked samples of Canadian a riVne° recent ana'yses conducted
reductions in mortality risk. -S- residents their values for
The results of the Jones-Lee et al. analysis were used to calculate a u
adjust the value of a statistical life lost by an individual of about age 40 h ed adJustment factors to
studies on which the value of a statistical life is based in the primarv h 3Verage a8e in the wage-risk
example, Jones-Lee et al. (1989) found that people ages 30-39 were ""lT" ^ l° age"specific vaIues- For
people ages 40-59 for the same mortality risk reduction. If the valued ^ l° ^ 89 percent as much as
40 years old is $6.12 million, then the value of a statistical life saved °f 3 statlstlcal saved of someone
percent of that, or S5.45 million. Numbers of lives saved in each of th S°meone age 30-39 would be 89
life-years-lost alternative calculation were apportioned to the aee ?r C ^ groups used in the statistical
1993). The number of lives saved in an age group was then multinliedh "T* by Jones"Lee et al- (1989;
statistical life saved for that age group. The value of a statistical life age"adjusted value of a
as $6.12 million times the ratio of the WTP for mortality risk redurti !n an age grouP was calculated
mortality risk reduction in the age 40-59 group, as reported bv Jonec.TV" , Sge group t0 the WTP for
year lag structure used in the primary method was applied under two alJm ? (1.989' 1993)' The flve"
of three percent and seven percent. Because the two Jones-Lee studies 3 dlscount rate assumptions
alternative calculation was carried out separately using each of the two7o ^L different ratios'this
4.2 Chronic Illness
The only benefit category evaluated in this analysis associated with j, •
bronchitis. Onset of bronchitis has been associated with exposure to " 11omc (,ong-term) illness is
fact, linked the onset of chronic bronchitis in adults to particulate mau^ P^! ants' Three studies have, in
with research that has found chronic exposure to pollutants leads rf ^ ese resu*ts are consistent
(Detels et al., 1991; Ackermann-Liebrich et al., 1997; Abbey et al ° 1 pulmonary functioning
Heavy Duty Diesel Technical Support Document (Abt Associates 2000) F°rmore information, see the
4.2.1 Chronic Bronchitis
In past analyses, we have estimated the changes in the number nfn
chronic bronchitis using studies by Schwartz (1993) and Abbey et al n99shT °tuS of PM"related
examined the relationship between exposure to PMI0 and prevalence of ch • ^**Sc|lwartz study
al. study examined the relationship between PM,5 and new incidences nf °nic "ronc^'tis- The Abbey et
have strengths and weaknesses which suggest that, if both measures of PM WmC bronchit,s- Both studies
effect estimates from each study may provide a better estimate of the evne^^i3^11^16'130011118 the
chronic bronchitis-than using either study alone. However since the a mt c"ange in incidences of
nuwever, since the SAMI analysis is based solely upon
Abt Associates Inc.
4-12
April 2002
-------
changes in annual mean PM2.5, we use the Abbey et al. study to predict changes in chronic bronchitis
incidence.
It should be noted, however, that reliance on the Abbey et al. (1995b) estimate will result in an
underestimate of the change in chronic bronchitis incidences if both the fine and coarse fractions of PM,0
are associated with chronic bronchitis. The SAMI control scenarios would likely result in reductions in
both the fine and coarse fractions of PM,0.
Alternative Calculation: Chronic Bronchitis Reversals
In developing the C-R functions for chronic bronchitis, it is necessary to estimate its annual
incidence rate. The annual incidence rate is derived by taking the number of new cases (234), dividing by
the number of individuals in the sample (3,310), as reported by Abbey et al.(1993, Table 3), dividing by
the ten years covered in the sample, and then multiplying by one minus the reversal rate.13 Reversals refer
to those cases of chronic bronchitis that were reported at the start of the Abbey et al. survey, but were
subsequently not reported at the end of the survey. Since we assume that chronic bronchitis is a
permanent condition, we subtract these reversals. Nevertheless, reversals may likely represent a real
effect that should be included in the analysis. To allow for this possibility, we present an estimate of
reversals in an alternative calculation in which reversals are considered to be chronic bronchitis cases of
the lowest severity level, as described below.
Valuing Chronic Bronchitis
, PMrehted chronic bronchitis is expected to last from the initial onset of the illness throughout
the rest of the individual s life. WTP to avoid chronic bronchitis would therefore be expected to
!scou"te<* Ja'ue °f a potentially long stream of costs (e.g., medical expenditures
and lost earnings) and pain and suffering associated with the illness. Two studies, Viscusi et al. (1991)
and Krupnick and Cropper (1992), provide estimates of WTP to avoid a case of chronic bronchitis.
. ^>1SCUS' Ct a'' anc* ^ruPnick and Cropper (1992) studies were experimental studies
mtended to examme new methodolog.es for eliciting values for morbidity endpoints. Although these
studies were not specifically designed for policy analysis, we believe the studies provide reasonable
fi! 'wtp °«t' Cf a °r j r°mf bronchitis. ot'ier contingent valuation studies, the reliability of
the WTP estimates depends on the methods used to obtain the WTP values. The Viscusi et al. and the
Krupnick and Cropper stud.es are broadly consistent with current contingent valuation practices, although
specific attributes of the studies may not be.
The study by Viscusi et al. (1991) uses a sample that is larger and more representative of the
genera popu a ion t an the study by Krupnick and Cropper (1992), which selects people who have a
relative with the disease. Thus, the valuation for the high-end estimate is based on the distribution of
• a r{T,nSeS r°m ,CUS'et a'' WTP to avoid a case of pollution-related chronic bronchitis (CB)
1S wlnm \ ST"g W' t0 avo^ a severe case of chronic bronchitis, as described by Viscusi
e\a ' i r-D ,an. 3 Justin8 downward to reflect (1) the decrease in severity of a case of pollution-
related CB relative to the severe case described in the Viscusi et al. study, and (2) the elasticity of WTP
wit respect to severity reported in the Krupnick and Cropper study. Because elasticity is a marginal
concept and because it is a function of severity (as estimated from Krupnick and Cropper), WTP
adjustments were made incrementally, in one percent steps. A severe case of CB was assigned a severity
"The percentage of reversals is estimated to be 46.6% based on Abbey et al. (1995a. Table 1).
Abt Associates Inc. 4-13
April 2002
-------
level of 13 (following Krupnick and Cropper). The WTP fx,
by: CWTPforaone Percent decrease in severity is given
= »TP„, ¦ (1-0.01 e) ,
where sev is the original severity level (which, at the start k i ^ ^
respect to severity. Based on the regression in Krupnick and Cropper'ri WW £7 ^
e is 0.18*sev. At the mean value of sev (6 47) e = 1 16 A PP (1992) the estimate c
decreases. Using the regression coefficient of 0 18 the abovfJ!"? decreaLses' however> the elasticity
' doove equation can be rewritten as:
For a given WTPSCV and a given coefficient of sev (0.18) the WTP f™ * ^ • •
can be obtained iteratively, starting with sev =13, as follows- percent reduction in severity
*^12.87 = WTPmn = WTPxi ¦ (1 - 0.01 • 0.18 13)
^7^12.74 = ^^0.99 12.87 = ^^2.87 ' 0 ~~ 0.0 1 • 0.1 8 • 1 2.87)
^2.6. = &770.W.I2.74 = *^2.74 ' (1 ~ 0.01 0.1 8 12.74)
and so forfc This iterative procedure eventually yields WTP(!, 0, WTP ,o avoid a case of chronic
bronchitis that is of average severity.
The derivation ofthe WTP to avoid a case of pollution rpiat^^i, • u u u a
three components. each of which ,s uncenain: „) ,he h
iv sc?r *199 11,6 seve?? 't' °f m ^2*£ s:
to that of the case described by Viscus, et al.), and (3) the elasticity of WTP with respect to severity of the
illnKS Because of these three sources of uncertainty, the WTP is uncertain. Based on assumption about
the distributions of each of the three uncertain components, a distribution of WTP to avoid a pollution-
related case of CB was derived by Monte Carlo methods. The mean of this distribution is taken as the
central tendency est,mate of WTP to avoid a pollution-related case of CB. Each of the three underlying
distributions is described briefly below.
1. The distribution of WTP to avoid a severe case of CB was based on the distribution of WTP
responses in the Viscusi et al. (1991) study. Viscusi et al. derived respondents' implicit WTP to avoid a
statistica case o c ironic ronchitis from their WTP for a specified reduction in risk. The mean response
implied a WTP of about S1,318,000 (2000 $) ; the median response implied a WTP of about $700,000
(2000 $). However, the extreme tails of distributions of WTP responses are usually considered
"There is an indication in the Viscusi et al. (199]) paper that the dollar values in the paper are in 1987 dollars. Under this
assumption, the dollar values were converted to 2000 dollars.
Abl Associates Inc. 4-14 April 2002
-------
unreliable. Because the mean is much more sensitive to extrw*.,, 1 ,
often used rather than the mean. Viscusi et al report not onlv th 3 median of WTP responses is
of WTP responses, however, but the decile points as well Th 7 a"d media" °f their distribution
from the Viscusi et al. study could therefore be approximated h 1°" of'ehab,e WTP ^ponses
probability of 1/9 to each of the first nine decile points. This omit« g'Vmg 3
of the responses (the extreme tails, considered unreliable) Thic tr ^ f 30 ® fSt Percent
from the Viscusi et al. study was assumed to be the distribution ofWTP ^tl0n of WTP resP°n®es
oution oi WTPs to avoid a severe case of CB.
2. The distribution of the severity level of an „ „ .
as a triangular distribution centered at 6.5, with endpoints at 1 1 £° "°n~reated5® wa® mod^ed
on the severity levels used in Krupnick and Cropper (1992) which P<=f se1venty /eV6'S are
ln(WTP) and severity level, from which the elasticity is derived T IT, relationship between
study is assigned a severity level of 13. The mean of the triano i a severe case °f CB in that
50 (ircen, reduction in Verity from a severe «e 8Ul" d,St"bU"°n 'S 6 5 This repreSeMS '
,m,L^ethia!S^W1iP T°taV°id a,°aSe °f CB Wi'h re$PKt <° "* °f »>»• «se of CB is a
constant t,me the seventy level. This constant was estimated by Krupnick and Cropper (1992) in the
regres on of ln(WTP) on sevens discussed above. This estimated constant (regression coefficient) is
CrZir) " 18 and S,Md"d devia,i°" * 0 0669 (obtained from Krupnick and
h A ^I]C ^ '° avoid a case of pollution-related CB was generated by Monte Carlo
methods, drawing ftom the three distnbutions described above. On each of 16,000 iterations (1) a value
was se ected from each distribution, and (2) a value for WTP ,1,00 .,, , . ,
7 .. . . u- u u • , , avaiuclor w 1 r was generated by the iterative procedure
described above in wh.ch the seventy level was decreased by one percent on each iteration, and the
corresponding WTP was derive* The mean of the resulting distribution ofWTP to avoid a case of
pollution-related CB was $340,568.
This WTP estimate is reasonably consistent with full COI estimates derived for chronic
bronchitis, using average annua lost earnings and average annual medical expenditures reported by
Cropper and Krupnick ( 990) Using a 5 percent discount rate and assuming that (1) lost earnings
continue until age 65, (2) medical expenditures are incurred until death, and (3) life expectancy is
unchanged by chronic bronchitis, the present discounted value of the stream of medical expenditures and
lost earnings associated with an average case of chronic bronchitis is estimated to be about $ 117,000 for a
30 year old, about $113,000 for a 40 year old, about $103,000 for a 50 year old, and about $59,000 for a
60 year old. A WTP estimate would be expected to be greater than a full COI estimate, reflecting the
willingness to pay to avoid the pain and suffering associated with the illness The WTP estimate of
$340,568 is from 2.9 times the full COI estimate (for 30 year olds) to 5.8 times the full CO I estimate (for
60 year olds).
Alternative Calculation: Valuing Chronic Bronchitis Reversals
In an alternative calculation, we estimate chronic bronchitis reversals and value them using the
same method used to value cases of chronic bronchitis. However, instead of allowing the severity level to
range from one to 13, we value all reversals at a severity level of one.
4.3 Acute Illnesses and Symptoms Not Requiring Hospitalization
We consider in this section the only acute symptom based upon a measure of annual mean
PM2.5, acute bronchitis. Note that there are several additional types of acute illnesses and symptoms
Abt Associates Inc.
4-15
April 2002
-------
associated with daily average PM2.5 and PM10, as well Th«- h u
previous analyses, such as the Heavy Duty Diesel analysis / ak considered in a number of
acute illness endpoints that are not evaluated in this anal,,*; j Ass°ciates, 2000). The extent of the
analysis is discussed in Section 6.
4.3.1 Acute Bronchitis
Dockery et al. (1996) examined the relationship between Pu ^ ,
rates of asthma, persistent wheeze, chronic cough and bro w 0 P°Uutants on the reported
living in 24 communities in the U.S. and Canada.' Health daf'tlS'3 StUdy of 13'369 children ages 8-12
pollutant models were used in the analysis to test a number / C°^ectec* 'n 1988-1991, and single-
Dockery et al. found that annual level of sulfates and particle m®.aSUres of particulate air pollution,
bronchitis, and PM15 and PM10 were marginally significantly related l°
Valuing Acute Bronchitis
Estimating WTP to avoid a case of acute bronchitis is difficult r
avoid acute bronchitis itself has not been estimated. EstimationS " ^ reasons' FirSt' WTP10
therefore must be based on estimates of WTP to avoid svmnt L l° aVOld thlS health endPoint
case of acute bronchitis may last more than one day whereasT?, °°Cf W"h this lHneSS' Sec0nd' a
typically valued. Finally, the C-R function used in the benefit analvs^f Sympt°mS that is
for children, whereas WT? estimates for those symptoms associated u ¦ u . ,
from adults. "ociated with acute bronchitis were obtained
With these caveats in mind, the values used for acute bronchi,;. • . • u •
adjusting the values used in the CAA §812 Prospective analyst n 7 *
multiplying by 1.318. WTP ,o avoid a case of acme brooch 2 wa™ '!? ™ , ,t? *
low estimate and a high estimate. The low estimate is the sum of he S "S P
lEc (1994) for two symptoms believed to be associated withTcut, tSlT rec0™"e"ded **
tightness. The high estimate was taken to be twice the value of a minor cou8hlnS and chest
The unit value is the midpoint between the low and high estimates "h f T"* *
estimates used in the §812 Prospective analysis were $ 3, $ 7 5? 7' TV"1,, SSTJL
corresponding values in 2000 dollars are $17.13. $101.45 and's^ l^Sly
Abl Associates Inc.
4-16
April 2002
-------
5. Results
This Section provides estimates of the magnitude anH ,
endpoints associated with each of the SAMI emission cont 1 ^ ° n^es in se,ected health
PM-,, concentrations. The total dollar benefits associated Scenano'related changes in annual average
endpoint will change (e.g., how many premature deaths Jill k 3 8IV? endpoint dePend on how much the
change is worth (e.g., how much a premature death avoided ^ ^ ^ °f
Before presenting the benefits associated with each nfth. c a
however, we provide a list of benefit categories that wer*» n ^ emission control scenarios,
Each of these unqualified endpoints are likely to be assnH* °r monetized in Exhlblt 5_1-
achieved under the emission control assumptions each Samt ambient pollution reductions
these endpoints have been quantified in previous analvs«7iemiSSI0? C°ntr0i SCena"° makeS' Many of
available. This table is meant to illustrate the extent of h^lTh * P°Uutant data W3S
pollution reductions beyond those considered in this analysis effects that are associated with
We also note that annual average PM2 5-relaf^ ~
been estimated in association with the SAMI control scenario, ^
estimated benefits associated with improved rKrado^Sli^ ""T
opportunities due to improvements in SAMI-region visual n' , improved recreational fishing
these benefits, as well as a symbolic representation of aH nth ^ a"d ^ qUality' We haVe added
displayed in the subsequent primary analysis tables to clt,heT benef,ts' "B" t0 the beneflts
with the SAMI emission control strategies. P C magnitude of total beneflts associated
To place SAMI estimated incidence changes into context with nredicted baseline incidence
throughout the eight state SAMI region, Exhibit 5-2 disolavs the hac , • T ® I " 5
endpoints. Both the mean estimated incidence chanse and 1 tnc.dence figures for PM; s
control incidence reductions and the predicted incidence baseKn? 8 T2 T T~
incidences include a,i incidences, no,'jus, those
Exhibits 5-3 through 5-6 present results of our orimarv ^u,,- ¦ j j u .
primary analysis; incidence and benefit estimates
asso .a ed witeachof *ejce™,os m each taure year, Note that Scenarjo A2 is C(Jnsi(iered the
baseline - the expected state of the world ,n 2010 and 2040 before SAMI controls are applied. These
P'esent results for Scenano Bl and Scenario B3 in 2010 and 2040, respectively. T5* percentile,
mean, and 95 percent,le st,mate for both incidence and benefits is presented for each endpoint, as well
as the simple mean benefit (calculated by multiplying the mean esti£ate of incide„ce „
corresponding mean valua ion). Total benefits are also displayed, calculated by summing the mean of
benefit" P°"" eS"maKS °f reCreatioMl ™ibi% and recreational fishing
,T„hc ^ne^""s assoc'atG^ wi,h'!lf SAMI emission control scenarios are substantial. Under scenario
Bl, total benefits are approximately $12 b.ll.on + Bin 2010 and grow ,o over $36 billion + B by 2040.
These benefits are refiectec1 by the number of incidences avoided by each scenario; by 2040, premature
mortality ,s reduced by «»o cases, there are 3,300 fewer cases of chronic branch,US, and cases of acute
bronchms are reduced by 9,000. Under scenario B3, total benefits are approximately $45 billion + B in
2010 and grow to approximately $68 billion + B by 2040. By 2040, premature mortality is reduced by
8,000 cases, there are 6,000 fewer cases of chronic bronchitis, and cases of acute bronchitis are reduced
by 16,000.
We present total health-related benefits by SAMI state in Exhibit 5-7 for the 2040 Bl and B3
scenarios. State-level benefits only include benefits associated with annual average PM2.5-related
A bt Associates Inc. 5. / Apn, m2
-------
endpoints; other benefits, such as recreational visibility and
states included in the 8-state SAMI region, North Caroli ecreat,onal fishing are not included. Of the
health-related benefits. In 2040, North Carolina received 24?"? ^IargeSt Percenta8e share of total'
and B3. Georgia receives the next largest percentage cha ° a, ^0/° tota' benefits for scenarios B1
B1 and B3 in 2040. The state receiving the smallest sha ^ t0.2l% °f total benefits for scenarios
benefits for scenarios B1 and B3 in 2040. fC IS St Virg'nia with 2% and 3% of total
Exhibits 5-8 through 5-11 display the alternative ' H
included in the primary analysis. Where possible a 5lh n 100 anc^ benefit calculations to those
incidence and/or benefits is presented for each alternativ - mean'and 95'h percentile estimate for
the alternative mortality functions. A 5lh percentile mean^ d ov* ^^its 5-12 through 5-15 present
benefits is presented for each alternative mortality fiinct'"' Percentile estimate for incidence and
mortality functions uses a 3% discount rate in the anni;^"' ^ that the valuation of alternative
ion of a mortality lag adjustment factor.
We provide quantified estimates of the 90 percent confix ¦
based solely on the standard errors of the C-R coefficients Th °e lntervals around these estimates
uncertainties in the change in air quality, population proie^tinn!^1"^318 d° n0t account for any
uncertainties. We also provide estimates of the 90 percent r sellne incidence rates, or model
estimates for each endpoint, using Monte Carlo techniauec 1°" "ce intervals around the dollar value
effect estimates and the valuation estimates. Again these ' t C° , lne the distributions of the health
other factors. We do not calculate a confidence interval fn s f° not a«ount for any uncertainties in
outcomes, as this would imply a precision which is not waLnVH? ec,onomic va,ue of aI1 health
the impact of unqualified sources of uncertainty. SC °" the gaps in information about
Abt Associates Inc.
5-2
April 2002
-------
Exhibit 5-1 Unqualified Endpoints Associated with P«n „
SAMI Emission Control AsS°dated wkh the
Asthma Attacks (asthmatics, all ages)
Lower Respiratory Symptoms (children, 7-14)
PM-Related Endpoints - Welfare Effe
Hospital Admissions - Respiratory Causes (all ages)
Daily Average PM10
Daily 12-Hour Average Ozone
Hospital Admissions - Cardiac Dysrhythmias (all ages)
Daily Average Ozone
Emergency Room Visits for Asthma (all ages)
Varies
Asthma Attacks (asthmatics, all ages)
Minor Restricted Activity Days (adults, age 18-65)
Daily One-Hour Max Ozone
Daily One-Hour Max Ozone
Decreased Worker Productivity (adult working population)
Daily Average Ozone
Other Ozone-Related Health Effects
Ozone-Related Endpoints - Welfare Effects"
Commercial Agricultural Benefits (6 major crops)
Commercial Forestry Benefits
Other Ozone-Related Welfare Effects
Sum 06
Varies
CO-Related Health and Welfare Effects
SO^RelatedHealth^mdWel^ar^E/jkcts
Varies
NOx^Relatec^eailh^nc^WelJbre^ffec^
Varies
Hazardous Air Pollutant-Related Health and Welfare Effj^_
Varies
Total Unmonetized Health- and Welfare-Related Benefits
B k
" For a complete discussion of the nature, estimation, and valuation of PM n,„ ¦ . . , „ _ .
DtaH ^ ,„P„ Analysis ,U,. HP A.»»,
* "B represents the sum of all unmonetized benefit categories.
Abt Associates Inc.
5-3
April 2002
-------
Exhibit 5-2 Baseline Percentages
Baseline Scenarios
2010 B1 Scenario
2040 B1 Scenario
2010 B3 Scenario
2040 B3 Scenario
Endpoint
Reference
2010
2040
Mean
% of Baseline
Mean
% of Baseline
Mean
% of Baseline
Mean
% of Baseline
Ages 30+
Krcwski et al. (2000)
325,741
399,728
1,662
0.51%
4,273
1.07%
6,155
1.89%
8,007
2.00%
Chronic
Bronchitis
Abbey ct al. (1995b)
76,832
95,580
1.258
1.64%
3.303
3.46%
4,531
5.90%
6,051
6.33%
Acute Bronchitis
Dockcry et al. (1996)
120,528
149,809
3,464
2.87%
8,952
5.98%
12,192
10.12%
16,177
10.80%
Exhibit 5-3 Total Quantified Benefits of SAM I: B1 Scenario in 20 fO
Endpoint
Reference
Avoided Incidence (cases/year)
5'h %ile Mean 95"' %ile
Monetary Benefits (millions 2000$)
5"" %ilc Mean 95"' %ile
Mortality
Ages 30+, 3% Discount Rate
Krcwski ct al. (2000), Tabic 31
931
1,662
2,355
$1,568
$11,114
$26,902
Ages 30+, 7% Discount Rate
Krcwski ct al. (2000)
931
1,662
2,355
$1,473
$10,439
$25,269
Chronic Illness
Chronic Bronchitis
Abbey et al. (1995b)
222
1,258
2,269
$25
$483
$1,569
Minor Illness
Acute Bronchitis
Dockery et al. (1996)
0
3,464
6,805
$0.0
$0.2
$0.5
Welfare Effects
I Recreational Visibility Benefits
Abt Associates Inc. (2002a)
Direct P.conomic Valuation
--
$796
--
Recreational Fishing Benefits
Abt Associates Inc. (2002b)
Direct P.conoinic Valuation
-
$0.6
--
Total Primary Benefits (3% discount rate)
-
•
--
--
$12394 +B
-
Total Primary Benefits (7% discount rate)
-
-
-
-
$11,719 + B
-
A hi Associates Inc.
5-4
April 2002
-------
Exhibit 5-4 Total Quantified Benefits of SAMI: B1 Scenario in 2040
Endpoint
Reference
Avoided Incidence (cases/year)
5"1 %ilc Mean 95"' %ile
Monetary Benefits (millions 2000$)
5'h Voile Mean 95"1 %ile
Mortality
Ages 30+, 3% Discount Rate
Krewski ct al. (2000), Table 31
2,399
4,273
6,049
$4,710
$33,332
$80,666
Ages 30+, 7% Discount Rate
Krewski ct al. (2000)
2,399
4,273
6,049
$4,424
$31,308
$75,767
Chronic Illness
Chronic Bronchitis
Abbey ct al. (1995b)
590
3,303
5,924
$94
$1,508
$5,456
Minor Illness
Acute Bronchitis
Dockcry ct al. (1996)
0
K,952
17,362
$0.0
$0.6
$1.4
Welfare EITccIs
Recreational Visibility Benefits
Abt Associates Inc. (2002a)
Direct Economic Valuation
--
$1,474
. 1
Recreational Fishing Benefits
Abt Associates Inc. (2002b)
Direct Economic Valuation
-
$1.4
Total Primary Benefits (3% discount rate)
--
--
"
-
$36,316+ B
--
Total Primary Benefits (7% discount rate)
--
~
- 1
--
534,292 + B
-
A hi Associates Inc.
5-5
April 2002
-------
Exhibit 5-5 Total Quantified Benefits of SAMI: B3 Scenario in 2010
Endpoint
Reference
Avoided Incidence (cases/year)
5'1" %ilc Mean 95"" %ilc
Monetary Benefits (millions 2000$)
5" %ile Mean 95"' %ile
Mortality
Ages 30+, 3% Discount Rate
Krcwski ct al. (2000), Table 31
3.463
6,155
8,700
$5,828
$41,163
$99,672
Ages 30+, 7% Discount Rate
Krcwski et al. (2000)
3,463
6,155
8,700
$5,474
$38,664
$93,619
Chronic Illness
-
Chronic Bronchitis
Abbey ct al. (1995b)
X20
4,531
8,072
$92
$1,740
$5,630
Minor Illness
Acute Bronchitis
Dockery et al. (19%)
0
12,192
23,340
$0.0
$0.8
$1.8
Welfare Effects
Recreational Visibility Benefits
Abt Associates Inc. (2002a)
Direct Fxonomic Valuation
-
$2,502
--
Recreational Fishing Benefits
Abt Associates Inc. (2002b)
Direct Economic Valuation
--
$1.4
--
Total Primary Benefits (3% discount rate)
--
--
--
--
$45,407 + B
--
Total Primary Benefits (7% discount rate)
-
-
-
--
$42,907 + B
- .
Ah! Associates Inc.
5-6
April 2002
-------
Exhibit 5-6 Total Quantified Benefits of SAMI: B3 Scenario in 2040
Endpoint
Reference
Avoided Incidence (cases/year)
5" %ilc Mean 95"1 %ilc
Monetary Benefits (millions 2000$)
5'" %ile Mean 95" %ile
Mortality
Ages .30+, 3% Discount Rate
Krewski et al. (2000), Table 31
4,507
8,007
11,314
$8,845
$62,457
$151,219
Ages .10+. 7% Discount Rate
Krcwski et al. (2000)
4,507
8,007
11,314
$8,307
$58,665
$142,037
Chronic Illness
Chronic Bronchitis
Abbey ctal. (1995b)
1,099
6,051
10,759
$174
$2,763
$10,019
Minor Illness
Acute Bronchitis
Dockcry et al. (1996)
0
16,177
30.844
$0.0
$1.1
$2.5
Welfare Effects
Recreational Visibility Benefits
Abt Associates Inc. (2002a)
Direct Economic Valuation
--
$2,705
-
Recreational Fishing Benefits
Abt Associates Inc. (2002b)
Direct Hconomic Valuation
--
$4.9
--
Total Primary Benefits (3% discount rate)
--
--
-
--
$67,931+ B
--
Total Primary Benefits (7% discount rate)
-
-
-
-
$64,138 + B
-
Aht Associates Inc.
5-7
April 2002
-------
Exhibit 5-7 Total Health Related Benefits in 2040 by State
State
2040 Bl
2040 B3
Monetary Benefits (millions
2000$)
Percent of Total
Monetary Benefits (millions
2000$)
Percent of Total
SAMI-Rcgion Total
S34,S41
100%
$65,221
100%
Alabama
$4,149
12%
$7,209
11%
Georgia
$7,285
21%
$13,948
21%
Kentucky
$1,196
3%
$3,555
5%
North Carolina
$8,404
24%
$14,992
23%
South C'aiolina
$2,827
8%
$5,986
9%
Tennessee
$5,772
17%
$10,251
16%
Virginia
$4,345
12%
$7,469
11%
West Virginia
$863
2%
$1,8 II
3%
Abt Associates Inc.
5-
-------
Exhibit 5-8 Alternative Benefit Calculations for the 2010 SAMI "Bl" Scenario
Avoided Incidence (cases/year)
Monetary Benefits (millions 2000$)
Endpoint
Reference/Alternative Valuation
5" %ilc
Mean
•)5,k %ile
S" %ile
Mean
95"' %ile
PM-related Alternative Calculations
Life Years Lost, by age:
Krewski c! al. (2000), Table 31
30-34
707
1,268
1,798
-
-
-
35-44
1,703
3,055
4,331
-
-
-
45-54
1.817
3.260
4,622
-
-
-
55-64
2,604
4,671
6,623
-
-
-
65-74
3.104
5.568
7,895
-
-
-
75-84
2,343
4,202
5,957
-
-
-
85+
1.166
2,091
2.965
-
-
-
Life years lost
3% discount rate
-
-
-
$790
$5,346
$12,398
Life years lost
7% discount rate
-
-
-
$914
$6,207
$14,337
Age-Adjusted Value of
Jones-Lee et al. (19X9) 3% discount rale
-
-
-
$3,546
$6,361
$9,018
Statistical Lives Lost
Jones-Lee et al. (1989) 7% discount rate
-
-
-
$3,331
$5,974
$8,470
Jones-Lee el al. (1993) 3% discount rale
-
-
-
$5,617
$10,075
$14,284
Jones-Lee et al. (1993) 7% discount rate
-
-
-
$5,276
$9,463
$13,417
Chronic Broachitis
Reversals
194
1.095
1.980
$8
$183
$675
Alternative benefit estimates are presented to display the impact alternative assumptions have on total benefits. We caution, however, that multiple alternative calculations cannot be
added together and substituted into the primary benefit estimate because the likelihood of all alternative assumptions occurring simultaneously is low. See the Heavy Duty Diesel
Regulatory Impact Analysis for a more detailed discussion on the use of alternative estimates (U.S. F.PA, 2000b).
Abt Associates Inc.
5-9
April 2002
-------
Exhibit 5-9 Alternative Benefit Calculations for the 2040 SAMI "Bl" Scenario
Endpoint
Reference/Alternative Valuation
Avoided Incidence (cascs/ycar)
5" %ilc Mean 95"' %ile
Monetary Benefits (millions 2000$)
5* %ile Mean 95,t %ile
PVf-rclated Alternative Calculations
Life Years Lost, by age:
Krcwski ct al. (2000), Table 31
30-34
1,924
3,458
4.910
-
-
-
35-44
4.611
8,288
11,767
-
-
-
45-54
4,795
8,619
12,237
-
-
-
55-64
6,740
12,113
17,198
-
-
-
65-74
7,983
14,347
20,368
-
-
-
75-84
6.012
10,805
15,340
-
-
-
85+
3,007
5.403
7,671
-
-
-
Life years lost
3% discount rate
-
-
-
$2,396
$16,253
$37,741
Life years lost
7% discount rate
-
-
-
$2,775
$18,830
$43,514
Age-Adjusted Value of
Statistical Lives Lost
Jones-Lee et al. (1989) 3% discount rate
Jones-Lee ct al. (1989) 7% discount rate
-
$10,736
$10,084
$19,295
$18,124
$27,395
$25,731
Jones-Lee et al. J1993) 3% discount rate
-
-
$16,955
$30,472
$43,262
Jones-Lee et al. (1993) 7% discount rate
-
-
$15,925
$28,621
$40,634
Chrome Bronchitis
Reversals
515
2.877
5.169 I
$24
$573
$2,105
A hi Associates Inc.
5-10
April 2002
-------
Exhibit 5-10 Alternative Benefit Calculations for the 2010 SA1M1 "B3" Scenario
End point
Reference/Alternative Valuation
Avoided Incidence (cases/year)
S" %ile Mean 95" Voile
Monetary Benefits (millions 2000$)
5'h %ile Mean 95"" %ile
PM-relafed Alternative Calculations
Life Years Lost, by age:
Krcwski et al. (2000), Table 31
30-34
2,632
4.742
6,743
-
-
-
35-44
6,343
11,426
16.247
-
-
-
45-54
6,805
12,257
17.428
-
-
-
55-64
9,779
17,612
25,041
-
-
-
65-74
11,678
21,032
29.902
-
-
-
75-84
8,799
15,846
22,529
-
-
-
85+
4,367
7,864
11.181
-
. -
-
LiTe years lost
3% discount rate
-
-
-
$2,979
$20,131
$46,724
Life years lost
7% discount rate
-
-
-
$3,439
$23,379
$53,960
Age-Adjusted Value of
Statistical Lives Lost
Joncs-Lec et al. (19X9) 3% discount rate
Jones-Lee et al.'(1989) 7% discount rate
-
$13,307
$12,500
$23,966
$22,511
$34,075
$32,005
Joncs-Lec et al. (1993) 3% discount rate
-
-
-
$21,078
$37,960
$53,970
Jones-Lee et al. (1993) 7% discount rale
-
-
-
$19,798
$35,654
$50,693
Chronic Bronchitis
Reversals
716
3,946
7.044
$28
$661
$2,423 II
Abl Associates Inc.
5-11
April 2002
-------
Exhibit 5-11 Alternative Benefit Calculations for the 2040 SAMI "B3" Scenario
Endpoint
Reference/Alternative Valuation
Avoided Incidence (cases/year)
5'" Voile Mean 95,h %ile
Monetary Benefits (millions 2000$)
S" %ile Mean 95"' %ile
PM-rclatcd Alternative Calculations
Life Years Lost, by age:
Krewski et al. (2000), Table 31
30-34
3,600
6,489
9,233
-
-
-
35-44
8,630
15,556
22,131
-
-
-
45-54
9,030
16,274
23,150
-
-
-
55-64
12,741
22,961
32,658
-
-
-
65-74
15,133
27,268
38,784
-
-
-
75-84
11,389
20,522
29,188
-
-
-
85+
5,682
10.238
14,561
-
-
-
Life years lost
3% discount rate
-
-
-
$4,542
$30,781
$71,480
Life years lost
7% discount rate
-
-
-
$5,255
$35,677
$82,481
Age-Adjusted Value of
Statistical Lives Lost
Jones-Lee et al. (1989) 3% discount rate
Jones-Lee et al. (1989) 7% discount rate
-
$20,291
$19,059
$36,566
$34,345
$52,010
$48,851
Jones-Lee et al. (1993) 3% discount rate
-
-
-
$32,057
$57,767
$82,165
Jones-Lee ct al. (1993) 7% discount rate
-
-
-
$30,110
$54,260
$77,176
Chronic Bronchitis
Reversals
959
5,270
9,389
$44
$ 1.050
$3,845
Exhibit 5-12 Alternative Mortality Calculations for the 2010 SAMI "Bl'
Scenario
Age Group
Statistic
Mortality
Reference
Mortality Incidence (eases/year)
5"1 %ilc Mean 95"' %ile
Monetary Benefits (millions 2000$)
J'" %ile Mean 95"' %ilc
Age 30+
Mean
Non-Accidental
Krewski et al. (2000), Table 31
922
1,583
2,271
$1,609
$10,805
$26,256
Age 25+
Mean
Non-Accidental
Dockery ct al. (1993)
1,821
4,280
6,615
$3,751
$28,628
$70,855
Age 25+
Mean
All-Cause
Dockcry ct al. (1993)
2,086
4,547
7,175
$3,768
$30,260
$77,049
Age 25+
Mean
Non-Accidental
Krewski et al. (2000)
2,355
4,5X4
6,960
$4,129
$31,255
$78,750
Arc 25+
Mean
All-Cause
Krewski ct al. (2000)
2,356
4,864
7,249
$4,430
$32,382
$82,749
A hi Associates Inc.
5-12
April 2002
-------
Exhibit 5-13 Alternative Mortality Calculations for the 2040 SAMI "B1" Scenario
Age Croup
Statistic
Mortality
Reference
Mortality Incidence (cases/year)
5,h %ilc Mean 95"1 %ile
Monetary Benefits (millions 2000$)
5"1 %ile Mean 95"' %ile
Age 30+
Mean
Non-Accidental
Krcwski ct al. (2000). Table 31
2,374
4,068
5,829
$4,825
$32,383
$78,642
Age 25+
Mean
Non-Accidental
Dockcry ct al. (1993)
4,685
10,932
16,823
$11,225
$85,270
$211,237
Age 25+
Mean
All-Cause
Dockcry et al. (1993)
5,368
11,622
18,252
$11,303
$90,186
$229,584
Age 25+
Mean
Non-Accidental
Krewski ct al. (2000)
6,050
11,701
17,685
$12,306
$93,037
$234,447
Age 25+
Mean
All-Cause
Krcwski ct al. (2000)
6,058
12,424
18,438
$13,224
$96,451
$245,821
Exhibit 5-14 Alternative Mortality Calculations for the 2010 SAMI "B3" Scenario
H Age Group
Statistic
Mortality
Reference
Mortality Incidence (cases/year)
5,h %ilc Mean 95" %ile
Monetary Benefits (millions 2000$)
5,b %ilc Mean 9S'h %ile
| Age 30+
Mean
Non-Accidental
Krcwski et al. (2000). Table 31
3,429 5,862
8,388
$5,968
$40,021
$97,280
Age 25+
Mean
Non-Accidental
Dockcry ct al. (1993)
6,742 15,617
23,919
$13,81$
$104,460
$258,452
Age 25+
Mean
All-Cause
Dockery et al. (1993)
7.716 16.589
25,923
$13,895
$110,394
$281,087
Age 25+
Mean
Non-Accidental
Krcwski ct al. (2000)
8,696 16,704
25,123
$15,140
$113,900
$286,111
Age 25+
Mean
All-Cause
Krcwski ct al. (2000)
8,703 17,724
26,181
$16,183
$117,987
$300,270
Abt AssiK-iales Inc.
5-13
April 2002
-------
Exhibit 5-15 Alternative Mortality Calculations for the 2040 SA1MI "B3" Scenario
Age Group
Statistic
Mortality
Reference
Mortality Incidence (cases/year)
5'" %ilc Mean 95"' %ilc
Monetary Benefits (millions 2000$)
5" Voile Mean 95"' %ilc
Age 30+
Mean
Non-Accidental
Krcwski ct al. (2000). Table 31
4.461 7,623
10,904
$9,052
$60,688
$147,523
Age 25 +
Mean
Non-Accidental
Dockety et al. (1993)
8,776 20,2X6
31,033
$20,959
$I5X,237
$391,388
Age 25+
Mean
All-Cause
Dockcry ct al. (1993)
10,048 21,561
33,647
$21,072
$167,317
$426,021
Age 25+
Mean
Non-Accidental
Krcwski ct al. (2000)
11.314 21,6%
32,5X8
$22,992
$172,507
$432,478
Age 25+
Mean
All-Cause
Krcwski ct al. (2000)
11,331 '23,033
33,9X2
$24,531
$178,794
$454,594
Aht Associates Inc.
5-14
April 2002
-------
6. Unqualified Benefits From Other Pollutant Reductions
The SAMI integrated assessment generated ~
with future year scenarios that control emissions from uE h , * concentrations associated
engines, and area sources. Using these highway Vehide' n°n"r0ad
premature mortality, chronic bronchitis, and other measlesWhile th^ ^ ,1" f
the economic benefits associated with reducing these emission^ have captured the bulk of
benefits. In addition to reductions in PM,, the emissi™ Sfc ! ^ 7* mlSS 3 Va"ety Potential
the SAMI emission control srpnnrinc or» a reductions likely to be achieved under each of
me oAivii emission control scenarios are expected to also reHi.r.® r
PMI0 and particulate matter between 2.5 and 10 microns^ ,T concen5atIons °f
hazardous air pollutants (HAPs) and ozone. 10 as We as N0-' S°2' '
Previous analyses have found that these nollntnnte , -
well as cron and forest™ ln« P utants are associated with adverse health effects, as
-d0,he' eff?r <**•"*
pollutants, there is no direct method to estimate their elS fcSl! conc"\'r"1°"s °flhese °"°Plankt°'> srowh, and low or no dissolved oxygen in
S f 8 S° SUnligh'• cau,i"8losSK submerged aquatic vegetation
terrc«rialLk i ecosJs[e'lls Deposition of nitrogen-containing compounds also affects
estna ecosystems. Nitrogen fertilization can alter growth patterns and change the balance of species
Abt Associates Inc.
6-1
April 2002
-------
in an ecosystem. Nitrogen dioxide and airborne nitrate also contribute to pollutant haze which impairs
visibility and can reduce residential property values and the value placed on scenic views.
N,°\in ,COmbr;i0n VOlatile 0rganic COmP°unds (V°C) also serve as precursors to ozone.
PM can also be formed from NOx emissions; secondary PM is formed in the atmosphere through a
number of physical and chemical processes that transform gases, such as NOx into particles The effects
of secondary PM and ozone exposures due to NOx emissions are the same as those of directly emitted PM
and ozone.
Benefits of Sulfur Dioxide Reductions
High concentrations of sulfur dioxide (S02) affect breathing and may aggravate existing
respiratory and cardiovascular disease. Sensitive populations include asthmatics individuals with
bronchitis or emphysema, children and the elderly. S02 is also a primaiy contributor to acid deposition,
or acid rain, which causes acidification of lakes and streams and can damage trees crops historic
buildings and statues. In addition, sulfur compounds in the air contribute to visibility impairment in large
parts of the country. This is especially noticeable in national parks.
PM can also be formed from S02 emissions. Secondary PM is fomied in the atmosphere through
a number of physical and chemical processes that transform gases, such as S02 into particles The
effects of secondary PM exposures due to S02 emissions are the same as those of directly emitted PM.
Benefits of Reduction in Carbon Monoxide Emissions
Human health effects associated with exposure to CO include cardiovascular system and central
nervous system (CNS) effects. Cardiovascular effects of CO are directly related to reduced oxygen
content of blood, resulting in tissue hypoxia (i.e., oxygen starvation). Most healthy individuals have
mechanisms (e.g. increased blood flow, blood vessel dilation) which compensate for this reduction in
tissue oxygen, although the effect of reduced maximal exercise capacity has been reported in some
healthy persons. Several other medical conditions such as occlusive vascular disease, chronic obstructive
lung disease, and anemia can increase susceptibility to potential adverse effects of CO during exercise.
Effects of CO on the CNS involve both behavioral and physiological changes. These include
modification of visual perception, hearing, motor and sensorimotor performance, vigilance, and cognitive
ability.
Although acute poisoning induced by CO can be lethal and is probably the best known health
endpoint of CO, this only occurs at very high concentrations of CO (greater than 100 ppm, hourly
average). In the ambient air, exposures to lower-levels of CO predominate and at these leCels the best
documented adverse health endpoint in human subjects is the decrease in time to onset of reproducible
exercise-induced chest pain. Results of some human exposure studies and reports of workers routinely
exposed to combustion products provide support for recent epidemiology research suggesting day-to-day
variations in ambient CO concentrations are related to cardiovascular hospital admissions and daily
mortality, especially for individuals over 65 years of age. Uncertainties about the association between
these health endpoints and ambient CO and the relative influence of indoor vs. outdoor CO have not been
resolved and will require further research.
There are certain people who are more at risk" to CO exposures. Individuals with preexisting
illness or cardiovascular diseases which limit oxygen absorption or oxygen transport to body tissues
would be somewhat more susceptible to the effects of CO. Very little data are available demonstrating
human health effects in healthy individuals caused by or associated with exposures to low CO
concentrations. Decrements in maximal exercise duration and performance in healthy individuals have
been reported, however, these decrements are small and likely to affect only athletes in competition. No
Abt Associates Inc.
6-2
April 2002
-------
effects were seen m healthy individuals during submaximal exercise, representing more typical daily
activities. Most recent evidence of CNS effects induced by exposure to CO indicates that behavioral
impairments in healthy individuals should not be expected until CO levels are well above what would be
Can u T5 nt ai:JeVe,:,0fCa AIS0J evidence 0f 1fetal toxicity or Of interactions
with high altitudes, drugs other pollutants, or other environmental stresses remains uncertain or suggests
that effects of concern will occur in healthy individuals only with exposure to very high levels of CO.
Benefits from Reductions in Hazardous Air Pollutants (HAPs)
Human exposure to HAPs may occur directly through inhalation or indirectly through ingestion
of food or water contaminated by HAPs or through exposure to the skin. HAPs may also enler terrestrial
and aquatic ecosystems through atmosphenc deposition. HAPs can be deposited on vegetation and soil
through wet or dry deposition. HAPs may also enter the aquatic environment from the atmosphere via
gas exchange tetween surface water and the ambient air, wet or dry deposition of particulate HAPs and
pamdes to which HAPs atab and wet or dty deposition to watersheds with subsequent leaching or
runoffto bodies of water (EPA,1992a). This analysis is focused only on the air quality benefits of HAP
reduction.
Health Benefits of HAP Reductions
The HAP emission reductions likely to be achieved under each of the SAMI emission control
scenarios are expected to reduce exposure to ambient concentrations of arsenic, cadmium, chromium,
hydrogen chloride hydrogen flounde lead, manganese, mercury, and nickel, which will reduce a variety
of adverse health effects considering both cancer and noncancer endpoints. These adverse health effects
include chronic health disorders (e.g. irritation of the lung, skin, and mucus membranes and effects on
the blood, digestive tract, kidneys, and central nervous system), and acute health disorders (e.g., lung
irritation and congestion, alimentary effects such as nausea and vomiting and effects on the central
nervous system). EPA has classified several of these HAPs as known of p^SlncaSnogL.
Noncancer health effects can be generally grouped into the following broad categories:
genotoxicity, developmental toxicity, reproductive toxicity, systemic toxicity and irritation
Genotoxicity is a broad term that usually refers to a chemical that has the ability to damage DNA or the
chromosomes. Developmental toxicity refers to adverse effects on a developing organism that may result
from exposure prior to conception, during prenatal development, or postnatally to the time of sexual
maturation. Adverse developmental effects may be detected at any point in the life span of the organism
Reproductive toxicity refers to the harmful effects of HAP exposure on fertility, gestation, or offspring,
caused by exposure of either parent to a substance. Systemic toxicity affects a portion of the body other
than the site of entry. Irritation, for the purpose of this document, refers to any effect which results in
irritation of the eyes, skin, and respiratory tract. Though we do not know the extent to which the adverse
health effects described above occur in the SAMI region population, to the extent the adverse effects do
occur, the emission control scenarios will likely reduce emissions and subsequent exposures.
Welfare Benefits of HAP Reductions
The welfare effects of exposure to HAPs have received less attention from analysts than the
health effects. However, this situation is changing, especially with respect to the effects of toxic
substances on ecosystems. Over the past ten years, ecotoxicologists have started to build models of
ecological systems which focus on interrelationships in function, the dynamics of stress, and the adaptive
potential for recovery. Chronic sub-lethal exposures may affect the normal functioning of individual
species in ways that make it less than competitive and therefore more susceptible to a variety of factors
including disease, insect attack, and decreases in habitat quality. All of these factors may contribute to an
overall change in the structure (i.e., composition) and function of the ecosystem.
Abt Associates Inc.
6-3
April 2002
-------
The adverse, non-human biological effects of HAP m'
and commercial fishery impacts. Atmospheric deposition ofHA?0? inC|Ude ecosystem and recreational
ecosystems. Atmospheric deposition of HAPs also contributed L * l° la"d may affeCt terrestrial
not only has adverse implications for individual wildlife sJhL ^ 3qUat,C ecosystem effects- This
humans who may ingest contaminated fish and waterfowlIn™ * Wh°le' bUt al'° ,
through the Industrial Boilers and Process Heaters NESHAP? emiss,on reductions achieved
environmental impacts. should reduce the associated adverse
6.2 Extrapolation of Benefits from Other Studies
monetary benefits associaTe^khfeduchj^PM°'
g extrapolate past results to the present study.
Two ways might be considered:
• Estimate the dollar benefits per ton of pmi«sir»«
the dollar per ton estimates is calculated by rividinelT T
emissions associated with these benefits. g tCS of benefits the
• Estimate {he relative damages of PM to other nnlliitant* f ,¦ ,
ac«nmf> that -=i 1 , r Pollutants from previous studies, and
assume that these relative damages apolv to the m,«„, i • r- , r u
lipalth anH „ u <•> . current analysis. For example, if the
nealtn and weliare benefits of PM associated with P\/r ~ j > * ¦ i i_
and welfare benefits associated with ocJmZLTZT T ""
similar ratio holds for the present studjT °°e m'8 "
With the exception ofusing impacts associated with annual PM,, to estimate other impacts
associated wtth pm.cul.it matter, these two approaches fail to provide reliabilities for ofher
pollutants. Prevtons analyses deviate from the current analysis in one or more thaI significallt|y
reduce their comparab.l.ly w,,h the present study. Tbey differ in terns of; basehne assumptions
77„XaXS„™8n,P f T Ch°iCe fheakh Wdfare effKK- '"—¦»> of ^cts, and so
vll o ea2 Lrrvn?, a C0mp X' ,e.mPera,u^P»den, interaction between NO, and
volatile organic carbon (VOC); in some cases ozone levels will increase when the concentrations of one
fZZZeZt7aSeS (UuS- EPA' 1 "6' FigUrC 3"25)- U iS difflCuh ~ without an air quality model -
presentTtlidy " PFeV1°US W°Uld be Similar t0 the relative changeS in the
,, •. vJXhib!l SU!"marizes d°llar beneflts associated with the annual PM, < C-R functions versus
daily Pmi5 and PMi0 functions. The dollar benefits of these other PM measures are no more than five
percent of the dollar benefits associated with annual PM,5. However, these other measures of PM capture
a range o e ec: s ca egories that we have not captured thus far in the analysis, such as hospital admissions
Samet et al 2000), emergency room visits (Schwartz et al„ 1993), lower respiratory symptoms
i^D Arwrw ',Upff,resPiratory symptoms (Pope et al., 1991), work loss day (Ostro, 1987),
NlRAD,(°Strv0 and Rol,;schild, 1989), and recreational visibility reduction. The number of persons
affected by these excluded categories greatly exceeds those affected by the C-R functions linked to annual
''Lead is a criteria pollutant, but it is discussed in the next section on toxic pollutants
Abt Associates Inc. a.a
April 2002
-------
Exhibit 6-1 Comparison of Annual PM and rwk„.. n»y.
annual and Other PM-related Benefits From Previous Analyses
Studg
Heavy Duty Diesel (Abt Associates Inc.,
2000, Exhibits 6-2 & 6-3).
Location: 48 contiguous U.S. States for PM,
and 37 Eastern U.S. States for ozone.
Ben-fits by Catepnrv mi||in„, ^
m°riy (s48j29)-cha>nic
Bronchitis ($1,803), acute bronchitis (Si)
Total Benefits ($ million)
$49,933
Other PM effects: hospital admissions ($82)
emergency room visits (SI), lower respiratory
symptoms (S3), upper respiratory symJt0mS?S5)
work oss day (S178), MRAD (S391), recreational
visibility reduction (SI,789)
2,449
°'hel PM effects as percent of annual PM.. .fr*,,*
5%
Tier 2 (Abt Associates Inc., 1999, Exhibit 6-
2 & 6-3).
Location: 48 contiguous U.S. States.
Annua! PM,, effects: mortality (523,370), chronic
bronchitis (S727). acute bronchitis (SO 4'
S24.097
utner rM effects: hospital admissions (Si 8)
emergency room visits (S0.3), lower respiratory
symptoms (Si), upper respirator symptoms (S2)
work loss day (S70), MRAD (S173) recreation!'
visibility reduction (S371)
635
Other PM effects as percent of annual PM..
3%
"NOs SIP Call" (Abt Associates Inc.. 1998a.
Exhibits 4.b. 1 & 4.b.2: Abt Associates Inc..
1998b. Exhibit 13).
Location: 37 Eastern U.S. States.
Annual PM, < effects: mortality (SI.123), chronic
bronchitis (SI 87). acute bronchitis (< SO H
SI.309
Othe, PM effects: hospital admissions (S2) lower
respiratory symptoms (SO. 1), upper respiratory
symptoms (< S0.1). work loss day (S6) MRAD
(S22), household soiling (S9). recreational visibility
reduction (S30) '
68
Other PM effects as percent of annua! P u
5%
^7dT.oroRundTgand S"R'reSUhS f°r" EaStem US S,ateS ^ ,hC 015 Tradin« Alternate. Numbers may not sum to the
Exhibit 6-2 compares the estimated benefits associated with annual PM, < and ozone benefits from
previous studies. For the Heavy Duty Diesel and Tier 2 analyses, both transportation-related analyses, the
annual PM, 5 benefits dominate the benefits associated with ozone, with ozone representing three percent
or less of annual PM, 5 benefits. For the NOx SIP Call, an analysis of NOx power plant emission
reductions, the ozone benefits are about 24 percent of the annual PM,, benefits, due in large part to
ozone-related agricultural crop losses. Since the SAMI program affects both automotive and power plant
emission sources, the relative contribution of ozone is likely to be somewhere between these two prior
estimates. However, we guess that it is likely to tend toward the lower end of this range, because the
annual PM, ? benefits in the current analysis are an order of magnitude greater than that found in the NO,
SIP Call, and the impact of ozone on agricultural sources in the eight SAMI states is unlikely to be as
large as that found in the 37 state NOx SIP Call analysis.
Including the estimated benefits of NO:, SO,, and CO probably would not have changed the
results in Exhibit 6-2 greatly. Changes in NO:, SO,, and CO are generally related to a relatively small
subset of effects; the most serious of which is perhaps hospitalization for heart-related problems (e.g.,
Schwartz and Morris, 1995). There have been studies finding some evidence that NO, and CO are linked
to mortality but it is difficult to determine if these effects are in addition to any effects"associated with PM
Abt Associates Inc.
6-5
April 2002
-------
and ozone (e.g., Kinney and Ozkayank, 1991; Kinney et al 199^ '<¦ t„
acting as indicators for NO,, SO,, and CO, then it is reasonable tr. ¦ i . extent that PM and ozone afe
PM and ozone. " " reasonable to simply look at the relative effects of
Exhibit 6-2 Comparison of Annual PM25 and Ozone Benefits From Previous Analyses
Study
Benefits bv Category ($ million)
Total Benefits ($ million)
Heavy Duty Diesel (Abt Associates Inc.,
2000, Exhibits 6-2 & 6-3).
Location: 48 contiguous U.S. States for PM,
and 37 Eastern U.S. States for ozone.
Annual PM,} effects: mortality ($48,129), chronic
bronchitis ($1,803), acute bronchitis (SI)
$49,933
Ozone effects: hospital admissions ($21),
emergency room visits ($0.1), MRAD ($90).
decreased worker productivity ($142), agricultural
crop losses ($1,078)
$1,331
Ozone effects as percent of annual PM,, effects
3%
Tier 2 (Abt Associates Inc.. 1999. Exhibit 6-
2 & 6-3).
Location: 48 contiguous U.S. States.
Annual PM: , effects: mortality (S23.370), chronic
bronchitis ($727), acute bronchitis ($0.4)
$24,097
Ozone effects: hospital admissions (S13).
emergency room visits (S0.1), MRAD (S I 01).
decreased worker productivity ($142). agricultural
crop losses ($217)
S473
Ozone effects as percent of annual PM,, effects
2%
"NO/SIP Call" (Abt Associates Inc., 1998b,
Exhibit 13).
Location: 37 Eastern U.S. States.
Annual PM:} effects: mortality ($1,123). chronic
bronchitis (SI87). acute bronchitis (< S0.1)
$1,309"
Ozone effects: hospital admissions ($4), any-of-19
symptoms ($1), decreased worker productivity
(S22), agricultural crop losses ($283 million)
S310
Ozone effects as percent of annual PM., effects
24%
"Average of RADM and S-R. results for 37 Eastern U.S. states from the 0.15 Trading Alternative.
"¦Some fraction of persons hospitalized subsequently die at the hospital,
evidence for this relationship is not well developed.
Abt Associates Inc. 6-6
CO may thus be related to mortality, although the
April 2002
-------
7. References
Abbey, D.E., F. Petersen, P.K. Mills and W.L. Beeson 1993 inn(lT ,
Total Suspended Particulates, Ozone, and Sulfur Dioxide1^™ ,Ambient Concentrations of
Nonsmoking Population. Archives of Environmental Heafth * '
Abbey, D.E., B.L. Hwang, R.J. Burchette, T. Vancuren and P K vr.i ,™
Ambient Concentrations of Pm(10) and Development of Estimated Long-Term
Population. Archives of Environmental Health Vol. 50(2)-T39 Sympt°mS 3 Nonsmoking
Abbey, D.E., B.E. Ostro, F. Petersen and R.J. Burchette !99Sh r>u • „
Associated with Estimated Long-Term Ambient Cn™ ° ResPirat0^ Symptoms
2.5 Microns in Aerodynamic Diameter (PM2 5) and OtherTTi^'"6 Particulates Less Than
Epidemiol. Vol. 5(2): 137-159 Other Air Pollutants. J Expo Anal Environ
Abbey, D.E., R.J. Burchette, S.F. Knutsen, W.F. McDonnell M n r u •
Long-tern, particulate and other air pollutants fu„c«„ °W'C """ tS~ Enri8h" '"8
Journal of Respiratory and Critical Care Medicine. Vol. S i , 289.°
Abt Associates Inc. 1998a. Agricultural Benefits U^ino ** >
EPA, Office of Air Quality Planning«d the N°X SIP CaH' Prepared f°r US'
Triangle Park, NC. September. ' er contract68-D-98-001. Research
Abt Associates Inc. 1998b. Benefit Analysis Results of Selected Health a aw if c a ¦ » r *1
NOx SIP Call RIA. Prepared for U S EPA Office of Air n ! elfarC EndPomts for the
contract no. 6S.D-98-001. Research Triangle Park, NC s^pTembe' 8 S'a° '
Abt Associates I:nc 1999. Final Tier 2 Rule: Air Quality Est,ma,ion, Selected Health and We.fare
Eseneuts Metnods, and Benefit Analysis Results. Prepared for i t c cm ^ a ¦ /-* 1
Planning and Standards, Research Triangle Park, NC. Bethesda, MD. December"
Ab' AS feath and wZ,eFRei?,MM K E"8jn=®ie,sel ™ *»>e: Air Quality Estimation, Slelected
Sffi e o?l ^0™ L ^ ?Tf" A"a'ysis Res»ta ^VS. EPA,
December 8 ' RKe"rCh TrianSle P«l<, NC, Bethesda, MD.
Abt Associates Inc. 2002a. The Value of Visibility Improvements in e .u a , u-
srM^ureparedforSou,heraAp~
A"' ASSinCihe Solerf' """""T" ReCrea,i0haI Fishi»8 Reduced Acid Deposition
"'^"Sch.^ f°' S°*m A^laChia" M°Um*inS
Ackermann-Liebrich U., P. Leuenberger, J. Schwartz, C. Schindler, C. Monn C Bolognini JP
Bongard, O. Brandli, G. Domenighetti, S. Elsasser, L. Grize, W Karrer R Keller H
KellerWossidlo N. Kunzli, B.W. Martin, T.C. Medici, A.P. Perruchoud', M.H. Schoni, J.M.
Tschopp, B. Vilhger, B Wuthrich, J.P. Zellweger and E. Zemp. 1997. Lung function and long
TZr^TJI n?! f? antS in Switzer,and Study on Ai' Pollution and Lung Diseases in
Adults (SAPALDIA) Team. Am J Respir Crit Care Med. Vol. 155(1); 122-129.
Abt Associates Inc, 7/ ,
April 2002
-------
Adams, RR and M A, Marano .1995. Cmrent Estimates fr„„, the Nationl| Hea|th Inte„iew Survey,
vi Center for Health Statistics. Hyattsviile, MD. Vital Health Statistics, Series 10,
No. 193. December.
Alberini A M. Cropper, T.T. Fu, A. Krupnick, J.T. Liu, D. Shaw and W. Harrington 1997. Valuing
health effects of air pollution in developing countries: The case of Taiwan. Journal of
Environmental Economics and Management. Vol. 34(2): 107-126
Butler, R. J 1983. Wage and Injury Rate Responses to Shifting Levels of Workers'Compensation. In:
Safety and the Work Force. Worrall, J.D., Ed. Cornell University, ILR Press: Ithaca, NY.
Chestnut, L.G. 1995. Dollars and Cents: The Economic and Health Benefits of Potential Particulate
Matter Reductions in the United States. Prepared for American Lung Association.
Collins, J.G. 1997. Prevalence of Selected Chronic Conditions: United States 1990-1992. National
Center for Health Statistics. Hyattsviile, MD. Vital Health Statistics, Series 10. No. 194.
Council of Economic Advisers. 2000. The Annual Report of the Council of Economic Advisers. In:
Economic Report of the President. U.S. Government Printing Office: Washington, DC.
Cousineau, J., R. Lacroix and A. Girard. 1988. Occupational Hazard and Wage Compensating
Differentials. University of Montreal Working Paper.
Cousineau. J.. R. Lacroix and A. Girard. 1992. Occupational Hazard and Wage Compensating
Differentials. The Review of Economics and Statistics. Vol. 74: 166-169
Cropper, M.L. and A.J. Krupnick. 1990. The Social Costs of Chronic Heart and Lung Disease.
Resources for the Future. Washington, DC. Discussion Paper QE 89-16-REV.
Detels, R., D.P. Tashkin, J.W. Sayre, S.N. Rokaw, F.J. Massey, A.H. Coulson and D.H. Wegman. 1991.
The Ucla Population Studies of Cord .10. a Cohort Study of Changes in Respiratory Function
Associated With Chronic Exposure to Sox, Nox, and Hydrocarbons American Journal of Public
Health. Vol. 81(3): 350-359.
Dillingham, A. 1985. The Influence of Risk Variable Definition on Value of Life Estimates. Economic
Inquiry. Vol. 24: 277-294.
Dockery, D.W., F.E. Speizer, D.O. Stram, J.H. Ware, J.D. Spengler and B.G. Ferris, Jr. 1989. Effects of
Inhalable Particles on Respiratory Health of Children. Am Rev Respir Dis. Vol. 139: 587-594.
Dockery, D.W., C.A. Pope, X.P. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, B.G. Ferris and F.E. Speizer.
1993. An association between air pollution and mortality in six U.S cities N Engl J Med. Vol.
329(24): 1753-1759.
Dockery, D.W., J. Cunningham, A.I. Damokosh, L.M. Neas, J.D. Spengler, P. Koutrakis, J.H. Ware, M.
Raizenne and F.E. Speizer. 1996. Health Effects of Acid Aerosols On North American Children
- Respiratory Symptoms. Environmental Health Perspectives. Vol. 104(5): 500-505.
Ehrlich, I. and H. Chuma. 1990. A Model of the Demand For Longevity and the Value of Life
Extension. Journal of Political Economy. Vol. 98(4): 761-782.
Abt Associates Inc.
7-2
April 2002
-------
Garen, J. 1988. Compensating Wage Differential.; anH c j
Economics and Statistics. Vol 70(1 )• 9-16 d°geneity °f Job Riskiness. The Review of
Gegax, D„ S. Gerking and W. Shulze 1985 Perceive p • v a,
paper prepared for the U. S. Environmental ProteaLn^ency^8'"3' ValUC °f Working
Gegax, D., S. Gerking and W. Shulze. 1991. Perceived Risk and tk \*
Review of Economics and Statistics. Vol. 73(4); 589-5% arginal Value of Safety- The
Gerking, S., M. DeHaan and W. Schulze. 1988. The Marginal Vol »
Valuation Study. Journal of Risk and Uncertainty VoU ^99 A Contingent
Health Review Committee. 2000. Commentary: Reanalvsis of thP u , •
American Cancer Society Study of Particulate Mr Pollution^ CitieS StUdy and the
Institute. Cambridge. July. Pollution and Mortality. Health Effects
Herzog, H.W., Jr., and A.M. Schlottmann. 1987. Valuing Risk in \\r 11 1 ^ •
Willingness to Pay, and the Optimal Provision of Safetv i i WorkP,ace: Market Pnce'
Paper. University of Tennessee Working
Herzog, H.W., Jr., and A.M. Schlottmann. 1990. Valuing Risk in th* m i 1 < n
Willingness to Pay. and the Optimal Provision of Safe^ Tte Marke• °®CC °f**
March 31 °W' y mam'M T Dickie.and R E t,nswonh
Jones-Lee, MW M^Hammerton and P R Philips. 1985. The Value of Safety: Result of a National
Sample Survey. Economic Journal. Vol. 95(March): 49-72
Jones-Lee, M.W. 1989. The Economics of Safety and Physical Risk. Basil Blackwell: Oxford.
Jones-Lee. M.W., G. Loomes. D. O'Reilly and P.R. Phillips. 1993. The Value of Preventing Non-fatal
WP SRa16^ S 3 Wlllingness"t°-Pa>' Nation£>l Sample Survey. TRY Working Paper.
Kennedy. 1990. A Guide to Econometrics. 2nd ed. MIT Press: Cambridge, MA.
Abt Associates Inc.
7-3
April 2002
-------
Kinney, P.L. and H. Ozkayank. 1991. Associating n.-i **
County. Environmental Research. Vol. 54: 99-120 V nahty and Air Pollution in Los Angeles
Kinney, P.L., K. Ito and G.D. Thurston. 1995. A Sensitivitv And r*
in Los Angeles. Inhalation Toxicology. Vol. 7(l): 59^9 of Mortality Pm"10 Associations
Japan, and tfe Unted SmL 4^. *£* ** ta AUStra"a'
White. 20006 Re^natylifo'f1^hfHa^ar^Six cSmdy L^a' M\Abrahamowicz and M'
Study of Particulate Air Pollution and Mortality Heakh Fffl t Amencan Cancer Soc]e?
y' " Effects Institute. Cambridge. July.
Krupnick, A.J. and M.L. Cropper. 1992. The Effect nf „
of Risk and Uncertainty. Vol. 5(1): 29-48 n 0n Health Risk Valuations. Journal
Krupnick, A.J. 2000. Resources for the Future. Personal rnmm,. •
5. nwiunication with Lisa Akeson. December
Krupnick, A.J., A. Alberini, M. Cropper, N. Simon B O'Rripn d n
Age, Health, and the Willingness to Pay MortalityR^ZTa MA "ein!zel,na^
Survey of Ontario Residents. Resources for .he FumeWa^nZ"^ Cn0"'",ge"' pa'Ua"™
37. September. was"mgton, DC. Discussion Paper 00-
Lang, C., G. Yarwood, F. Lalonde and R Bloxam iqo<; p„,,:
Vehicles and Fuek prP„orA^i f r A n' mental an^ Health Benefits of Cleaner
verncies ana rueJs. Prepared for: Canadian Council of Mint'et^c r- ~ t 1 c
on fleanpr Vohiri^o c 1 u/- • or Ministers of the Environment Task Force
on v_leaner Vehicles and Fuels. Winnipeg, Manitoba. October.
Leigh, J.P. and R.N. Folsom. 1984. Estimates of the Value of Accident a a f, , . n .
Conravitv r>f fho Accident Avoidance at the Job Depend on
concavity ot the Equalizing Differences Curve. The Ouarff>rK,u^ • fr • a
Business. Vol. 24( 1): 56-66 Quarterly Review of Economics and
Leigh, J.P. 1987. Gender, Firm Size, Industry and Estimates of the 1 1 rv 1*1
Economics. Vol. 6: 255-273. value-of-Life. Journal of Health
Loehman, E.T. and V.H. De. 1-982. Application of Stochastic Chnire r n r a i r
Public Goods. The Review of Economics and sS£ Vol 474.4°80 ' "
Man"' Un^KtatrT10^ 1982t 71,5 ^ fM R'Sk in ,h= Lab°< ¦*<*«*
90(4) 8^7-853 " ReCOnC,"a,U>n WUh 0,her Smdies Journal of Political Economy. Vol.
MllkT- TM*"d / G"™ 1991 The Value of Statistical Life in New Zealand. Report to the New Zealand
Ministry of Transport, Land Transport Division.
Mitchell, R^C. and R.T. Carson. 1986. The Use of Contingent Valuation Data for Benefit/Cost Analysis
Pr 3fer °A Utl°n °"tr°L rePorl submitted by Resources for the Future to Environmental
Protection Agency, Office of Policy Analysis. Washington, DC. September.
Moore, M.J. and WX Viscusi. 1988. Doubling the Estimated Value.of Life: Results Using New
Occupational Fatality Data. Journal of Policy Analysis and Management. Vol. 7(3): 476-490.
Abt Associates Inc.
7-4
April 2002
-------
National Center for Health Statistics. 1999. National Vital Statistics Reports. U.S. Department of Health
and Human Services, Centers for Disease Control and Prevention, National Center for Health
Statistics. Washington, DC. Volume 47, Number 19. June 30.
Olson, C. A. 1981. An Analysis of Wage Differentials Received by Workers on Dangerous Jobs. Journal
of Human Resources. Vol. 16: 167-185.
Ostro, B.D. 1987. Air Pollution and Morbidity Revisited: A Specification Test. Journal of
Environmental Economics and Management. Vol. 14: 87-98.
Ostro, B.D. and S. Rothschild. 1989. Air Pollution and Acute Respiratory Morbidity - an Observational
Study of Multiple Pollutants. Environ Res. Vol. 50(2): 238-247.
Ostro, B.D., J.M. Sanchez, C. Aranda and G.S. Eskeland. 1996. Air Pollution and Mortality - Results
From a Study of Santiago, Chile. J Expo Anal Environ Epidemiol. Vol. 6(1): 97-114.
Pope, C.A., D.W. Dockery, J.D. Spengler and M.E. Raizenne. 1991. Respiratory Health and PmlO
Pollution - a Daily Time Series Analysis. American Review of Respiratory Disease. Vol. 144(3):
668-674.
Pope, C.A., J. Schwartz and M.R. Ransom. 1992. Daily Mortality and PM10 Pollution in Utah Valley.
Archives of Environmental Health. Vol. 47(3): 211-217.
Pope, C.A., M.J. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer and C.W. Heath.
1995. Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults.
Am J Respir Crit Care Med. Vol. 151(3): 669-674.
Samet, J., S. Zeger, F. Dominici, F. Curriero, I. Coursac, D. Dockery, J. Schwartz and A. Zanobetti.
2000. The National Morbidity, Mortality, and Air Pollution Study. Health Effects Institute.
Cambridge, MA. Report No. 94. May.
Schwartz, J. and D.W. Dockery. 1992. Particulate Air Pollution and Daily Mortality in Steubenville,
Ohio. American Journal of Epidemiology. Vol. 135(1): 12-19.
Schwartz, J. 1993. Particulate Air Pollution and Chronic Respiratory Disease. Environ Res. Vol. 62: 7-
13.
Schwartz, J., D. Slater, T.V. Larson, W.E. Pierson and J.Q. Koenig. 1993. Particulate air pollution and
hospital emergency room visits for asthma in Seattle. Am Rev Respir Dis. Vol. 147(4): 826-31.
Schwartz, J., D.W. Dockery, L.M. Neas, D. Wypij, J.H. Ware, J.D. Spengler, P. Koutrakis, F.E. Speizer
and B.G. Ferris. 1994. Acute Effects of Summer Air Pollution On Respiratory Symptom
Reporting in Children. Am J Respir Crit Care Med. Vol. 150(5): 1234-1242.
Schwartz, J. and R. Morris. 1995. Air Pollution and Hospital Admissions For Cardiovascular Disease in
Detroit, Michigan. American Journal of Epidemiology. Vol. 142(1): 23-35.
Schwartz, J. 2000. Harvesting and long term exposure effects in the relation between air pollution and
mortality [see comments]. Am J Epidemiol. Vol. 151(5): 440-8.
Smith, R.S. 1974. The Feasibility of an 'Injury Tax' Approach to Occupational Safety. Law and
Contemporary Problems. Vol. 38(4): 730-744.
Abt Associates Inc.
7-5
April 2002
-------
Smith, R.S. 1976. The Occupational Safety and Health Act: Its Goals and Achievements. American
Enterprise Institute. Washington, DC.
Smith, V.K. 1983. The Role of Site and Job Characteristics in Hedonic Wage Models. Journal of Urban
Economics. Vol. 13: 296-321.
Smith, V.K. and C. Gilbert. 1984. The Implicit Risks to Life: A Comparative Analysis. Economics
Letters. Vol. 16: 393-399.
Southern Appalachian Mountains Initiative. 2001. 2001 Interim Report. Asheville, North Carolina.
May.
Spix, C., J. Heinrich, D. Dockery, J. Schwartz, G. Volksch, K. Schwinkowski, C. Collen and H.E.
Wichmann. 1993. Air pollution and daily mortality in Erfurt, east Germany, 1980-1989.
Environ Health Perspect. Vol. 101(6): 518-26.
U.S. Bureau of Economic Analysis. 1995. BEA Regional Projections to 2045: Volume 1, States. U.S.
Department of Commerce. Washington, DC. July.
U.S. Bureau of the Census. 1998. Statistical Abstract of the United States: 1998. 118 ed. Washington,
DC.
U.S. Centers for Disease Control. 1999. CDC Wonder.
U.S. EPA. 1996. Review of National Ambient Air Quality Standards for Ozone: Assessment of
Scientific and Technical Information. OAQPS Staff Paper. U.S. EPA, Office of Air Quality
Planning and Standards. Research Triangle Park, NC. EPA-452\R-96-007. June.
U.S. EPA. 1999a. An SAB Advisory: The Clean Air Act Section 812 Prospective Study Health and
Ecological Initial Studies. Prepared by the Health and Ecological Effects Subcommittee (HEES)
of the Advisory Council on the Clean Air Compliance Analysis, Science Advisory Board, U.S.
Environmental Protection Agency. Washington, DC. EPA-SAB-Council-ADV-99-005.
February.
U.S. EPA. 1999b. The Clean Air Act Amendments (CAAA) Section 812 Prospective Study of Costs and
Benefits (1999): Advisory by the Advisory Council on Clean Air Compliance Analysis: Costs
and Benefits of the CAAA. Prepared by the Advisory Council on the Clean Air Compliance
Analysis, Science Advisory Board, U.S. Environmental Protection Agency. Washington, DC.
EPA-SAB-COUNCIL-ADV-00-002. October 29.
U.S. EPA. 1999c. The Clean Air Act Amendments (CAAA) Section 812 Prospective Study of Costs and
Benefits (1999): Advisory by the Health and Ecological Effects Subcommittee on initial
assessments of Health and Ecological Effects; Part 2. Prepared by the Health and Ecological
Effects Subcommittee, Science Advisory Board, U.S. Environmental Protection Agency.
Washington, DC. EPA-SAB-COUNCIL-ADV-00-001. October 29.
U.S. EPA. 1999d. The Clean Air Act Amendments (CAAA) Section 812 Prospective Study of Costs and
Benefits (1999): Advisory by the Health and Ecological Effects Subcommittee on Initial
Assessments of Health and Ecological Effects; Part 1. Prepared by the Health and Ecological
Effects Subcommittee (HEES) of the Advisory Council on the Clean Air Compliance Analysis,
Science Advisory Board, U.S. Environmental Protection Agency. Washington, DC. EPA-SAB-
Council-ADV-99-012. July 28.
Abt Associates Inc.
7-6
April 2002
-------
U.S. EPA. 2000a. An SAB Report on EPA's White Paper Valuing the Benefits of Fatal Cancer Risk
Reduction. Prepared by the Environmental Economics Advisory Committee (EEAC) of the
Science Advisory Board, U.S. Environmental Protection Agency. Washington, DC. EPA-SAB-
EEAC-00-013. July 27,.
U.S. EPA. 2000b. Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and
Highway Diesel Fuel Sulfur Control Requirements. U.S. EPA, Office of Air and Radiation.
Washington, DC. EPA 420-R-00-026. December.
Violette, D.M. and L.G. Chestnut. 1983. Valuing Reductions in Risks: A Review of the Empirical
Estimates. Prepared for U.S. Environmental Protection Agency. Washington DC. EPA-230-05-
83-002.
Viscusi, W.K. 1978. Labor Market Valuations of Life and Limb: Empirical Estimates and Policy
Implications. Public Policy. Vol. 26(3): 359-386.
Viscusi, W.K. 1979. Employment Hazards: An Investigation of Market Performance. Harvard
University Press: Cambridge.
Viscusi, W.K. 1981. Occupational Safety and Health Regulation: Its Impact and Policy Alternatives. In:
Research in Public Policy Analysis and Management. Crecine, J., Ed. JAI Press: Greenwich,
CT. Vol: 2. p. 281-299.
Viscusi, W.K., W.A. Magat and J. Huber. 1991. Pricing Environmental Health Risks - Survey
Assessments of Risk - Risk and Risk - Dollar Trade-Offs For Chronic Bronchitis. Journal of
Environmental Economics and Management. Vol. 21(1): 32-51.
Viscusi. W.K. 1992. Fatal Tradeoffs: Public and Private Responsibilities for Risk. Oxford University
Press: New York.
Wessex, I. 1994. PRO/FILER, US Demographics and ZIPS: Summary Tape File 1 A. Wessex, Inc.
Zeger, S.L., F. Dominici and J. Samet. 1999. Harvesting-resistant estimates of air pollution effects on
mortality. Epidemiology. Vol. 10(2): 171-5.
Abt Associates Inc.
7-7
April 2002
-------
Appendix A: Results for Sensitivity Analyses
Exhibit A-l Sensitivity Analysis Results for the 2010 SA1M1 "Bl" Scenario
Avoided Incidence (cases/year)
Monetary Benefits (millions 2000$)
Mortality Lag
Rcfcrencc/Altcrnativc Valuation
5,h %ilc
Mean 95'" %ile
II
<*>
r = 5%
r = 7%
No Lag
-
1.662
$11.690
$1 1,690
$11,690
5 Year
25%. 25%. 17%, 17%, 16%
-
1,662
$11,114
$10,764
$10,439
8 Year
Incidence Occurs X"' Year
-
1,662
$9,505
$8,308
$7,280
15 Year
Incidence Occurs 15"' Year
-
1,662
$7,729
$5,904
$4,534
15 Year
Incidence Skewed F.arly
-
1.662
$10,882
$10,433
$10,041
15 Year
Incidence Skewed Late
-
1,662
$8,384
$6,794
$5,556
Exhibit A-2 Sensitivity Analysis Results for the 2010 SAMI "B3" Scenario
Avoided Incidence (cases/year)
Monetary Benefits (millions 2000$)
Mortality Lag
Reference/Alternative Valuation
5"1 %ilc
Mean 95'" %ilc
r = 3%
r = 5%
r = 7%
No Lag
-
6.155
$43,298
$43,298
$43,298
5 Year
25%, 25%, 17%, 17%, 16%
-
6,155
$41,163
$39,868
$38,664
8 Year
Incidence Occurs 8"' Year
-
6.155
$35,205
$30,771
$26,964
15 Year
Incidence Occurs 15"' Year
-
6.155
$28,625
$21,868
$16,792
15 Year
Incidence Skewed Rarly
-
ft. 155
$40,303
$38,640
$37,188
15 Year
Incidence Skewed Late
-
6,155
$31,052
$25,163
$20,577
Abt Associates Inc.
A-l
April 2002
-------
Exhibit A-3 Sensitivity Analysis Results for the 2040 SAM1 "Bl" Scenario
Avoided Incidence (cases/year)
Monetary Benefits (millions 2000$)
Mortality Lag
Reference/Alternative Valuation
5,h %ilc
Mean 95"1 %ile
r = 3%
r = 5%
r = 7%
No Lag
-
4,273
$35,054
$35,054
$35,054
5 Year
25%, 25%, 17%, 17%, 16%
-
4,273
$33,326
$32,277
$31,302
8 Year
Incidence Occurs 8"' Year
-
4,273
$28,502
$24,912
$21,830
15 Year
Incidence Occurs 15"' Year
-
4.273
$23,175
$17,705
$13,595
15 Year
Incidence Skewed F.arly
4,273
$32,630
$31,283
$30,107
15 Year
Incidence Skewed Late
-
4,273
$25,140
$20,372
$16,659
Exhibit A-4 Sensitivity Analysis Results for the 2040 SAMI "B3" Sc
enario
Avoided Incidence (eases/year)
Monetary Benefits (millions 2000$)
Mortality Lag
Reference/Alternative Valuation
5,h %ilc
Mean 95"1 %ile
r = 3%
r = 5%
r = 7%
No Lag
-
8,007
$65,685
$65,685
$65,685
5 Year
25%, 25%, 17%, 17%, 16%
-
8,007
$62,446
$60,481
$58,654
X Year
Incidence Occurs 8'1' Year
-
8,007
$53,408
$46,681
$40,905
15 Year
Incidence Occurs 15"'Year
-
8,007
$43,425
$33,175
$25,474
15 Year
Incidence Skewed Early
-
8,007
$61,141
$58,618
$56,415
J 5 Year
Incidence Skewed Lale
-
8,007
$47,107
$38,172
$31,216
Abl Associates Inc.
A-2
April 2002
-------
Exhibit A-5 Sensitivity Analysis: Effect of Thresholds on Estimated PM-Related Mortality Based on Krewski et al. (2000) - Mean, All-
Cause for the 2010 SAM1 "Bl" Scenario
2,000
N
«
w
e
«
2
"3
c
M
¦o
4»
2
'©
>
<
0
10 15 20
Assumed Effect threshold (Annual [Mean PM2.5 (ug/m3))
25
30
A hi Associates Inc.
A-5
April 2002
-------
Exhibit A-6 Sensitivity Analysis: Effect of Thresholds on Estimated PM-Related Mortality Based on Krewski et al. (2000) - Mean, All-
Cause for the 2010 SAMI "B3" Scenario
o
W
C
4>
!2
"S
a
i—
"O
2
"o
>
<
5 10 15 20
Assumed EITect Threshold (Annual Mean PM2.5 (ug/m3))
Abt Associates Inc.
A-4
April 2002
-------
Exhibit A-7 Sensitivity Analysis: Effect of Thresholds on Estimated PM-Related Mortality Based on Krewski et al. (2000) - Mean, All-
Cause for the 2040 SAMI "I}1" Scenario
Assumed Effect Threshold (Annual Mean PM2.5 (ug/m3))
A hi Associates Inc.
A-5
April 2002
-------
Exhibit A-8 Sensitivity Analysis: Effect of Thresholds on Estimated PM-Related Mortality Based on Krewski et al. (2000) - Mean, All-
Cause for the 2040 SAMI "B3" Scenario
T
o
N
OJ
u
e
12
"3
<
0
10 15 20 25
Assumed Fffect Threshold (Annual Mean PIM2.5 (ug/m3))
30
35
Aht Associates Inc.
A-6
April 2002
-------
Appendix B: Particulate Matter C-R Functions
Note that APM is defined -- for all of the concentration-response (C-R) functions - as PMbascljnc -
PMcomrol, and that the change is defined to be: - (incidenceconlrol - incidencebJcMnc).
B .1 Mortality
There are two types of exposure to PM that may result in premature mortality. Short-term
exposure may result in excess mortality on the same day or within a few days of exposure. Long-term
exposure over, say, a year or more, may result in mortality in excess of what it would be if PM levels
were generally lower, although the excess mortality that occurs will not necessarily be associated with
any particular episode of elevated air pollution levels. In other words, long-term exposure may capture a
facet of the association between PM and mortality that is not captured by short-term exposure.
B .1.1 Mortality (Krewski et al., 2000) Based on ACS Cohort: Mean PM25
The C-R function to estimate the change in long-term mortality is:
A Mortality = ~[y0 • - l)j. pop,
where:
y0 = county-level all-cause annual death rate per person ages 30 and older
P = PM; 5 coefficient = 0.0046257
APM; 5 = change in annual mean PM;, concentration
pop = population of ages 30 and older
op = standard error of (5 = 0.0012046
Incidence Rate. To estimate county-specific baseline mortality incidence among individuals ages 30 and
over, this analysis used the average annual all-cause county mortality rate from 1994 through 1996 (U.S.
Centers for Disease Control, 1999). Note that the Krewski et al. (2000) replication of Pope et al. (1995)
used the same all-cause mortality when estimating the impact of PM.
Coefficient Estimate (P). The coefficient (P) is estimated from the relative risk (1.12) associated with a
change in mean exposure of 24.5 ng/m3 (based on the range from the original ACS study) (Krewski et al.,
2000, Part II - Table 31).
„ ln(1.12)
00046257¦
Abt Associates Inc.
B-l
April 2002
-------
Standard Error (ap). The standard error (op) was calculated as the average of the standard errors implied
by the reported lower and upper bounds of the relative risk (Krewski et al., 2000, Part II - Table 31).
f ln(1.19) ln(1.12)^
A,g/, - P [ 245 245 J
®p.'ugh j ]~96 - 0.0012625
(ln(1.12) ln(1.06)^|
P~P,m I 245 24.5 J
= 1.96 = L96 = 00011466
<7hil!h+CJ'h»
G _ —£_ 0.0012046
B .1.2 Mortality (Krewski et al., 2000), Based on Six-City Cohort: Mean PM25
The C-R function to estimate the change in long-term mortality is:
A Mortality = ~[y0 ¦ (e'piPh'-' - i)]. popj
where:
y0 = county-level all-cause annual death rate per person ages 25 and older
P = PM; 5 coefficient = 0.013272
APM;, = change in annual mean PM: 5 concentration
pop = population of ages 25 and older
op = standard error of p = 0.004070
Incidence Rate. To estimate county-specific baseline mortality incidence among individuals ages 25 and
over, this analysis used the average annual county mortality rate from 1994 through 1996 (U.S. Centers
for Disease Control, 1999). The Krewski et al. (2000) reanalysis of Dockery et al. (1993, p. 1754)
appears to have used all-cause mortality when estimating the impact of PM.
Coefficient Estimate (P). The coefficient (P) is estimated from the relative risk (1.28) associated with a
change in mean exposure going from 11.0 ng/m3 to 29.6 ng/m3 (Krewski et al., 2000, Part I - Table 19c):
„ ln(1.28)
(29.6- 11) = 0013272 ¦
Standard Error (op). The standard error (ap) was calculated as the average of the standard errors implied
by the reported lower and upper bounds of the relative risk (Krewski et al., 2000, Part I - Table 19c):
Abt Associates Inc.
B-2
April 2002
-------
(ln(1.48) ln(1.28)
frigh-P V 18.6 ~ 18.6
& P. high ~ J 96 ~ \96 ~ ®-^3982
(ln(1.28) ln(1.10)
P~ P,m v 18.6 " 18.6
j 95 196 - 0.004157
o> = °*** * g/''" = 0.004070
B .1.3 Mortality (Dockerv et al., 1993), Based on Six-City Cohort: Mean PMj5
Dockery et al. (1993) examined the relationship between PM exposure and mortality in a cohort
of 8,111 individuals aged 25 and older, living in six U.S. cities. They surveyed these individuals in 1974-
1977 and followed their health status until 1991. While they used a smaller sample of individuals from
fewer cities than the study by Pope et al., they used improved exposure estimates, a slightly broader study
population (adults aged 25 and older), and a follow-up period nearly twice as long as that of Pope et al.
(1995). Perhaps because of these differences, Dockery et al. study found a larger effect of PM on
premature mortality than that found by Pope et al.
The C-R function to estimate the change in long-term mortality is:
A Mortality = -[y0 • - 1)]- pop,
where:
y0 = county-level all-cause annual death rate per person ages 25 and older
P = PM, 5 coefficient = 0.0124
APM2 5 = change in annual mean PM, 5 concentration
pop = population of ages 25 and older
op = standard error of P = 0.00423
Incidence Rate. Dockery et al. (1993, p. 1754) appear to have used all-cause mortality when estimating
the impact of PM. To estimate county-specific baseline mortality incidence among individuals ages 25
and over, this analysis used the average all-cause annual county mortality rate from 1994 through 1996
(U.S. Centers for Disease Control, 1999).
Coefficient Estimate (P). The coefficient (P) is estimated from the relative risk (1.26) associated with a
change in mean exposure going from 11.0 |ig/m3 to 29.6 ng/m3 (Dockery et al., 1993, Tables 1 and 5):
„ ln(1.26)
0.0124.
(29.6- 11)
Standard Error (op). The standard error (op) was calculated as the average of the standard errors implied
by the reported lower and upper bounds of the relative risk (Dockery et al., 1993, Table 5):
Abt Associates Inc.
B-3
April 2002
-------
(ln(1.47) ln(1.26)^
Pugh - P y 18.6 18.6 J
Gp.high ~ ~ T96 = 0.00423
(ln(1.26) 111(1.08)^
1-/L v 18.6 18.6 J
&fi.iow ~ j j 96 - 0.00423
+ °"/oh'
a = — 0.00423.
B .2 Chronic Morbidity
Onset of bronchitis has been associated with exposure to air pollutants. Three studies have, in
fact, linked the onset of chronic bronchitis in adults to particulate matter. These results are consistent
with research that has found chronic exposure to pollutants leads to declining pulmonary functioning
(Deteis et al., 1991; Ackermann-Liebrich et al., 1997; Abbey et al., 1998).
In past analyses, we have estimated the changes in the number of new cases of PM-related
chronic bronchitis using studies by Schwartz (1993) and Abbey et al. (1995b). The Schwartz study
examined the relationship between exposure to PM|0 and prevalence of chronic bronchitis. The Abbey et
al. study examined the relationship between PM: 5 and new incidences of chronic bronchitis. Both studies
have strengths and weaknesses which suggest that, if both measures of PM were available, pooling the
effect estimates from each study may provide a better estimate of the expected change in incidences of
chronic bronchitis than using either study alone. However, since the SAMI analysis is based solely upon
changes in annual mean PM2.5. we use the Abbey et al. study to predict changes in chronic bronchitis
incidence.
B .2.1 Chronic Bronchitis (Abbey et al., 1995b, California)
Abbey et al. (1995b) examined the relationship between estimated PM,, (annual mean from 1966
to 1977), PM,0 (annual mean from 1973 to 1977) and TSP (annual mean from 1973 to 1977) and the
same chronic respiratory symptoms in a sample population of 1,868 Californian Seventh Day Adventists.
The initial survey was conducted in 1977 and the final survey in 1987. To ensure a better estimate of
exposure, the study participants had to have been living in the same area for an extended period of time.
In single-pollutant models, there was a statistically significant PM, ? relationship with development of
chronic bronchitis, but not for AOD or asthma; PM,0 was significantly associated with chronic bronchitis
and AOD; and TSP was significantly associated with all cases of all three chronic symptoms. Other
pollutants were not examined.
Abt Associates Inc.
B-4
April 2002
-------
The C-R function to estimate the change in chronic bronchitis is:
A Chronic Bronchitis = -[y0 • - l)j. pop,
where:
y0 = annual bronchitis incidence rate per person (Abbey et al., 1993, Table 3) = 0.00378
P = estimated PM2 < logistic regression coefficient = 0..0132
APM; 5 = change in annual average PM15 concentration
pop = population of ages 27 and older without chronic bronchitis17 = 0.9465^population 27+
cp = standard error of P = 0.00680
Incidence Rate. The annual incidence rate is derived by taking the number of new cases (234), dividing
by the number of individuals in the sample (3,310), as reported by Abbey et al.(1993, Table 3), dividing
by the ten years covered in the sample, and then multiplying by one minus the reversal rate (estimated to
be 46.6% based on Abbey et al. (1995a, Table 1)). Using the same data base, Abbey et al. (1995a, Table
1) reported the incidences by three age groups (25-54, 55-74, and 75+) for "cough type" and "sputum
type" bronchitis, but they did not report an overall incidence rate for bronchitis.
Coefficient Estimate (P). The estimated coefficient (p) is based on the relative risk (=1.81) associated
with 45 yug/m'1 change in PM; 5 (Abbey et al., 1995b, Table 2). The coefficient is calculated as follows:
„ ln( 1.81)
P = = 0.0132.
45
Standard Error (op). The standard error for the coefficient (op) is calculated from the reported lower and
upper bounds of the relative risk (0.98 to 3.25) (Abbey et al., 1995b, Table 2):
(ln(3.25) ln(1.81)Ni
PM,h - P V 45 ~ 45 i ^
a!!¦!'¥' - 196 ~ 1.96 ~ 0.00664
(ln(1.81) ln(0.98)
P- /L I 45 " 45
" 196 196 - 0.00696
op = °h¥' * Cr'"" = 0.00680.
B .3 Acute Morbidity
There is a considerable body of scientific research that has estimated significant relationships
between elevated air pollution levels and other morbidity health effects. Chamber study research has
' Using the same data set. Abbey et al. (1995a, p. 140) reported that the respondents in 1977 ranged in age from 27 to 95.
Chronic bronchitis prevalence from Adams and Marano (1995, Tables 62 and 78).
Abt Associates Inc. B-5 April 2002
-------
established relationships between specific air pollution chemicals and symptoms such as coughing, pain
on deep inspiration, wheezing, eye irritation and headaches. In addition, epidemiological research has
found air pollution relationships with aeute infectious diseases (e.g., bronchitis, sinusitis) and a variety of
"symptom-day" categories. Some "symptom-day" studies examine excess incidences of days with
identified symptoms such as wheezing, coughing, or other specific upper or lower respiratory symptoms.
Other studies estimate relationships for days with a more general description of days with adverse health
impacts, such as "respiratory restricted activity days" or work loss days.
A challenge in preparing an analysis of the minor morbidity effects is identifying a set of effect
estimates that reflects the ftill range of identified adverse health effects but avoids double counting.
However, because only annual mean PM2.5 was available for the analysis, only one acute morbidity
endpoint was available for evaluation, acute bronchitis.
B .3.1 Acute Bronchitis C-R Function (Dockery et al., 1996)
Dockery et al. (1996) examined the relationship between PM and other pollutants on the reported
rates of asthma, persistent wheeze, chronic cough, and bronchitis, in a study of 13,369 children ages 8-12
living in 24 communities in U.S. and Canada. Health data were collected in 1988-1991, and single-
pollutant models were used in the analysis to test a number of measures of particulate air pollution.
Dockery et al. found that annual level of sulfates and particle acidity were significantly related to
bronchitis, and PM:, and PMI0 were marginally significantly related to bronchitis.18 They also found
nitrates were linked to asthma, and sulfates linked to chronic phlegm. It is important to note that the
study examined annual pollution exposures, and the authors did not rule out that acute (daily) exposures
could be related to asthma attacks and other acute episodes.
Earlier work, by Dockery et al. (1989), based on six U.S. cities, found acute bronchitis and
chronic cough significantly related to PM,5. Because it is based on a larger sample, the Dockery et al.
(1996) study is the better study to develop a C-R function linking PM15 with bronchitis. The C-R
function to estimate the change in acute bronchitis is:
k Acute Bronchitis = -
y0
(l->'0) v™^ + >>0
Jo
pop,
where:
y0 = annual bronchitis incidence rate per person = 0.044
P = estimated PM:; logistic regression coefficient = 0.0272
APM: 5 = change in annual average PM, 5 concentration
pop = population of ages 8-12
op = standard error of p = 0.0171
Incidence Rate. Bronchitis was counted in the study only if there were "reports of symptoms in the past
12 months" (Dockery et al., 1996, p. 501). It is unclear, however, if the cases of bronchitis are acute and
temporary, or if the bronchitis is a chronic condition. Dockery et al. found no relationship between PM
The original study measured PM;,, however when using the study's results we use PM;_«. This makes only a negligible
difference, assuming that the adverse effects of PM;, and PM;, are comparable.
Abt Associates Inc.
B-6
April 2002
-------
and chronic cough and chronic phlegm, which are important indicators of chronic bronchitis. For this
analysis, we assumed that the C-R function based on Dockery et al. is measuring acute bronchitis.
In 1994, 2,115,000 children ages 5-17 experienced acute conditions (Adams and Marano, 1995,
Table 6) out of population of 48.110 million children ages 5-17 (U.S. Bureau of the Census, 1998, Table
14), or 4.4 percent of this population. This figure is somewhat lower than the 5.34 percent of children
under the age of 18 reported to have chronic bronchitis in 1990-1992 (Collins, 1997, Table 8). Dockery
et al. (1996, p. 503) reported that in the 24 study cities the bronchitis rate varied from three to ten percent.
Finally a weighted average of the incidence rates in the six cities in the Dockery et al. (1989) study is 6.34
percent, where the sample size from each city is used to weight the respective incidence rate (Dockery et
al., 1989, Tables 1 and 4)." This analysis assumes a 4.4 percent prevalence rate is the most
representative of the national population. Note that this measure reflects the fraction of children that have
a chest ailment diagnosed as bronchitis in the past year, not the number of days that children are adversely
affected by acute bronchitis.20
Coefficient Estimate (|3). The estimated logistic coefficient (P) is based on the odds ratio (= 1.50)
associated with being in the most polluted city (PM2, = 20.7 ng/m3) versus the least polluted city (PM;, =
5.8 (ig/m ) (Dockery et al., 1996, Tables 1 and 4). The original study used PM,,, however, we use the
PM-,, coefficient and apply it to PM2 ? data.
ln(1.50)
(20.7 - 5.8)
&¦>;-, = MnVl = 0-0272.
Standard Error (op). The standard error of the coefficient (a„) is calculated from the reported lower and
upper bounds of the odds ratio (Dockery et al., 1996, Table 4):
ln(2.47) ln( 1.50)^i
PW,~P V 14.9 14.9
= ~i^~= m = 00171
ln(1.50) ln(0.91)
P~ P,.m ^ 14.9 ~ 14.9
al>-" 1.96 " 1.96 " 00171
Gp.hith +
= —= 0.0171.
'The unweighted average of the six city rates is 0.0647.
:"ln 1994. there were 13,707.000 restricted activity days associated with acute bronchitis, and 2,115,000 children (ages 5-
17) experienced acute conditions (Adams and Marano. 1995. Tables 6 and 21). On average, then, each child with acute bronchitis
suffered 6.48 days.
Abt Associates Inc.
B-7
April 2002
-------
£$2? .3-
. 53
016543
/-oi
------- |