United States Office of Water EPA-821-R-13-004
Environmental Protection Washington, DC 20460 April 2013
Agency
4>EPA Benefit and Cost Analysis
for the Proposed Effluent
Limitations Guidelines and
Standards for the Steam
Electric Power Generating
Point Source Category
-------
Benefit and Cost Analysis for Proposed ELGs Table of Contents
Table of Contents
1 INTRODUCTION 1-1
1.1 STEAM ELECTRIC PLANTS
1.2 REGULATORY OPTIONS CONSIDERED FOR THE PROPOSED ELGs.
1.3 ANALYTIC FRAMEWORK
1.3.1 Constant Prices
1.3.2 Discount Rate and Year
1.3.3 Period of Analysis
1.3.4 Population and Income Growth..
-1
-2
-1
-2
-2
-3
-3
1.4 ORGANIZATION OF THE BENEFIT AND COST ANALYSIS REPORT -3
2 BENEFITS OVERVIEW 2-1
2.1 HUMAN HEALTH BENEFITS ASSOCIATED WITH IMPROVEMENTS IN SURFACE WATER QUALITY 2-3
2.1.1 Fish Consumption 2-3
2.1.2 Drinking Water Consumption 2-4
2.1.3 Complementary Measure of Human Health Benefits 2-5
2.2 ECOLOGICAL BENEFITS ASSOCIATED WITH IMPROVEMENTS IN SURFACE WATER QUALITY 2-5
2.2.1 Improved Surface Water Quality 2-5
2.2.2 Benefits to Threatened and Endangered Species 2-7
2.2.3 Reduced Sediment Contamination 2-8
2.3 BENEFITS ASSOCIATED WITH IMPROVEMENTS IN GROUNDWATER QUALITY 2-8
2.4 ECONOMIC PRODUCTIVITY BENEFITS 2-9
2.4.1 Reduced Impoundment Failures 2-9
2.4.2 Water Supply and Use 2-9
2.4.3 Commercial Fisheries 2-10
2.4.4 Tourism 2-10
2.4.5 Property Values 2-10
2.5 REDUCED AIR POLLUTION 2-11
2.6 REDUCED WATER WITHDRAWALS 2-11
2.7 SUMMARY OF BENEFITS CATEGORIES 2-11
3 HUMAN HEALTH BENEFITS 3-1
3.1 REDUCED CANCER CASES FROM CONSUMPTION OF FISH CONTAMINATED WITH ARSENIC 3-2
3.1.1 Methodology and Data 3-2
3.1.2 Results 3-5
3.1.3 Revised Cancer Slope Factor 3-6
3.2 BENEFITS TO CHILDREN FROM REDUCED LEAD EXPOSURE VIA FISH CONSUMPTION 3-6
3.2.1 Methodology and Data 3-6
3.2.2 Results 3-10
3.3 BENEFITS TO INFANTS FROM REDUCED EXPOSURE TO MERCURY 3-11
3.3.1 Methodology and Data 3-11
3.3.2 Results 3-13
3.4 BENEFITS TO SUBSISTENCE FISHERS 3-13
3.5 POTENTIAL ADDITIONAL HEALTH BENEFITS 3-16
3.6 LIMITATIONS AND UNCERTAINTIES 3-17
4 NON-MARKET BENEFITS FROM WATER QUALITY IMPROVEMENTS 4-1
4.1 WATER QUALITY 4-1
4.1.1 WQI Calculation 4-2
4.1.2 Sources of Data on Ambient Water Quality 4-5
4.1.3 Baseline WQI 4-8
April 19, 2013 T
-------
Benefit and Cost Analysis for Proposed ELGs Table of Contents
4.1.4 Estimated Changes in Water Quality (AWQI) from the Regulation 4-9
4.2 WILLINGNESS TO PAY FOR WATER QUALITY IMPROVEMENTS 4-11
4.3 TOTAL WTP FOR WATER QUALITY IMPROVEMENTS 4-17
4.4 LIMITATIONS AND UNCERTAINTIES 4-19
5 IMPACTS AND BENEFITS TO THREATENED AND ENDANGERED SPECIES 5-1
5.1 INTRODUCTION 5-1
5.2 BASELINE STATUS OF FRESHWATER FISH SPECIES 5-1
5.3 T&E SPECIES AFFECTED BY THE PROPOSED ELGs 5-2
5.3.1 Identifying T&E Species Potentially Affected by the Proposed ELGs 5-2
5.3.2 Assessing Benefits of T&E Species Improvements from the Proposed ELGs 5-4
5.4 ESTIMATING WTP FOR T&E SPECIES POPULATION INCREASES 5-5
5.4.1 Economic Valuation Methods 5-5
5.4.2 Estimating WTP for Improved Protection of T&E Species 5-6
5.5 RESULTS 5-7
5.6 LIMITATIONS AND UNCERTAINTIES 5-11
6 BENEFITS FROM GROUNDWATER QUALITY IMPROVEMENTS 6-1
6.1 METHODOLOGY AND DATA 6-1
6.1.1 Baseline Water Quality 6-1
6.1.2 Water Quality Improvements 6-1
6.1.3 Affected Households 6-2
6.1.4 Monetary Values of Groundwater Quality Improvements 6-3
6.2 RESULTS 6-5
6.3 LIMITATIONS AND UNCERTAINTIES 6-5
7 BENEFITS FROM AVOIDED IMPOUNDMENT FAILURES 7-1
7.1 METHODS AND DATA 7-1
7.1.1 Failure Probability 7-1
7.1.2 Capacity Factor 7-2
7.1.3 Failure Costs 7-2
7.2 RESULTS 7-5
7.3 LIMITATIONS AND UNCERTAINTIES 7-6
8 AIR-RELATED BENEFITS 8-1
8.1 DATA AND METHODOLOGY 8-1
8.1.1 Changes in Air Emissions 8-1
8.1.2 NOxandSO2 8-3
8.1.3 CO2 8-6
8.1.4 Estimating Total Air-Related Benefits 8-7
8.2 RESULTS 8-8
8.3 LIMITATIONS AND UNCERTAINTIES 8-9
9 BENEFITS FROM REDUCED WATER WITHDRAWALS 9-1
9.1 REDUCED GROUNDWATER WITHDRAWALS 9-1
9.1.1 Methods 9-1
9.1.2 Results 9-2
9.1.3 Limitations and Uncertainties 9-2
10 SUMMARY OF TOTAL BENEFITS 10-1
10.1 TOTAL ANNUALIZED BENEFITS 10-1
10.2 TIME PROFILE OF BENEFITS 10-0
10.3 INFERRED BENEFITS FOR REGULATORY OPTIONS NOT ANALYZED EXPLICITLY 10-1
April 19, 2013
-------
Benefit and Cost Analysis for Proposed ELGs Table of Contents
11 SUMMARY OF TOTAL SOCIAL COSTS 11-4
11.1 OVERVIEW OF SOCIAL COSTS ANALYSIS FRAMEWORK 11-4
11.2 KEY FINDINGS FOR REGULATORY OPTIONS 11-5
12 BENEFITS AND SOCIAL COSTS 12-2
12.1 COMPARISON OF BENEFITS AND SOCIAL COSTS BY OPTION 12-2
12.2 ANALYSIS OF INCREMENTAL BENEFITS AND SOCIAL COSTS 12-3
13 REFERENCES 13-1
APPENDIX A: DETAIL ON ESTIMATING AFFECTED POPULATION 1
APPENDIX B: IEUBK MODEL DESCRIPTION AND APPLICATION 1
APPENDIX C: HUMAN HEALTH BENEFITS INCLUDING DOWNSTREAM REACHES 1
APPENDIX D: TSS, TN, AND TP ECOREGION-SPECIFIC SUBINDEX CURVES 1
APPENDIX E: META-ANALYSIS RESULTS 1
E.I LITERATURE REVIEW OF WATER RESOURCE VALUATION STUDIES 1
E.2 TOTAL WTP META-ANALYSIS REGRESSION MODEL AND RESULTS 6
E.3 MODEL LIMITATIONS 18
APPENDIX F: IMPACTS OF STEAM ELECTRIC POLLUTANTS ON AQUATIC SPECIES.
APPENDIX G: SENSITIVITY OF THE ESTIMATED BENEFITS FROM AVOIDED IMPOUNDMENT
FAILURES TO FAILURE PROBABILITY 1
APPENDIX H: CONCENTRATION RESPONSE FUNCTION USED IN THE ANALYSIS OF AIR-RELATED
BENEFITS 1
APPENDIX I: CO2-RELATED BENEFITS USING ALTERNATE SOCIAL COST OF CARBON VALUES 1
April 19, 2013 Mi
-------
Benefit and Cost Analysis for Proposed ELGs
List of Abbreviations
List of Abbreviations
BAT Best available technology economically achievable
BMP Best management practice
BOD Biochemical oxygen demand
BPT Best practicable control technology currently available
CAA Clean Air Act
CCR Coal combustion residuals
CSF Cancer slope factor
CPI Consumer Price Index
CWA Clean Water Act
E&T Endangered and threatened
EA Environmental Assessment
EJ Environmental justice
ELGs Effluent limitations guidelines and standards
EO Executive Order
EPA U.S. Environmental Protection Agency
FC Fecal coliform
FCA Fish consumption advisories
FGD Flue gas desulfurization
FGMC Flue gas mercury control
GDP Gross domestic product
HUC Hydrologic unit code
IEUBK Integrated Exposure, Uptake, and Biokinetics
IPM Integrated Planning Model
IRIS Integrated Risk Information System
IQ Intelligence quotient
LADD Lifetime average daily dose
MATS Mercury and Air Toxics Standards
MCL Maximum contaminant level
NAICS North American Industry Classification System
NERC North American Electric Reliability Corporation
NHD National hydrography dataset
NPDES National Pollutant Discharge Elimination System
O&M Operation and maintenance
OMB Office of Management and Budget
PbB Blood lead concentration
POTW Publicly owned treatment works
PSES Pretreatment Standards for Existing Sources
PSNS Pretreatment Standards for New Sources
RSEI Risk-Screening Environmental Indicators
SCC Social cost of carbon
SPARROW SPAtially Referenced Regressions On Watershed attributes
TDD Technical Development Document
TN Total nitrogen
TP Total phosphorus
TRI Toxic Release Inventory
TSS Total suspended solids
April 19, 2013 iv~
-------
Benefit and Cost Analysis for Proposed ELGs List of Abbreviations
USGS U.S. Geological Service
WQI Water quality index
WQL Water quality ladder
WTP Willingness to pay
April 19, 2013
-------
Benefit and Cost Analysis for Proposed ELGs
1: Introduction
1 Introduction
EPA is proposing a regulation that would strengthen the existing controls on discharges from steam electric
power plants by revising technology-based effluent limitations guidelines and standards (ELGs) for the steam
electric power generating point source category, 40 CFR part 423.
The proposed effluent limitation guidelines and standards for the Steam Electric Power Generating Point
Source Category are based on data generated or obtained in accordance with EPA's Quality Policy and
Information Quality Guidelines. EPA's quality assurance (QA) and quality control (QC) activities for this
rulemaking include the development, approval and implementation of Quality Assurance Project Plans for the
use of environmental data generated or collected from all sampling and analyses, existing databases and
literature searches, and for the development of any models which used environmental data. Unless otherwise
stated within this document, the data used and associated data analyses were evaluated as described in these
quality assurance documents to ensure they are of known and documented quality, meet EPA's requirements
for objectivity, integrity and utility, and are appropriate for the intended use.
This document presents an analysis of the benefits and social costs of the proposed ELGs and complements
other analyses EPA conducted in support of the proposed ELGs, described in separate documents:
> Environmental Assessment for the Proposed Effluent Limitations Guidelines and Standards for the
Steam Electric Power Generating Point Source Category (EA) (U.S. EPA, 2013a; DCN SE01995).
The EA summarizes the environmental and human health improvements that are expected to result
from implementation of the proposed ELGs.
> Technical Support Document for the Proposed Effluent Limitations Guidelines and Standards for the
Steam Electric Power Generating Point Source Category (TDD) (U.S. EPA, 2013b; DCN SE01964).
The TDD provides background on the proposed ELGs; applicability and summary of the proposed
ELGs; industry description; wastewater characterization and identification of pollutants of concern;
and treatment technologies and pollution prevention techniques. It also documents EPA's
engineering analyses to support the proposed ELGs including facility specific compliance cost
estimates, pollutant loadings, and non-water quality impact assessment.
> Regulatory Impact Analysis for the Proposed Effluent Limitations Guidelines and Standards for the
Steam Electric Power Generating Point Source Category (PJA) (U.S. EPA, 2013c; DCN SE03170).
The RIA describes EPA's analysis of the costs and economic impacts of the proposed ELGs. In
particular, it provides the basis for social cost estimates presented in this document. It also provides
information pertinent to meeting several legislative and administrative requirements.
The rest of this chapter discusses aspects of the proposed ELGs that are salient to EPA's analysis of the
benefits and social costs of the regulation and summarizes key analytic assumptions used throughout this
document.
1,
Steam Electric Plants
As defined in 40 CFR part 423, the steam electric industry covers establishments "primarily engaged in the
generation of electricity for distribution and/or sale, which results primarily from a process utilizing fossil-
type fuels (coal, petroleum coke, oil, gas) or nuclear fuel in conjunction with a thermal cycle employing the
steam water system as the thermodynamic medium." (40 CFR part 423.10) EPA intends the industry
definition to apply to operations where the generation of electricity is the predominant source of revenue
and/or principal reason for operation. EPA considers steam electric processes to be those with at least one
April 19, 2013
1-1
-------
Benefits and Cost Analysis for Proposed ELGs 1: Introduction
prime mover that utilizes steam. For example, combined cycle systems are composed of one or more
combustion turbines operating in conjunction with one or more steam turbines, and thus would be subject to
the proposed ELGs.
EPA estimated that 1,079 steam electric plants are subject to the proposed ELGs; a subset of these plants
would need to modify their operations to meet new effluent limits and standards (refer to the TDD and RIA
for details).
1.2 Regulatory Options Considered for the Proposed ELGs
EPA considered eight regulatory options for the proposed ELGs (see Table 1-1). These options differ in the
wastestreams controlled by the regulation, the size of the units controlled, and the stringency of controls (see
TDD for a detailed discussion of the technology bases for the options). Thus, EPA is proposing to revise or
establish Best Available Technology Economically Achievable (BAT), New Source Performance Standards
(NSPS), Pretreatment Standards for Existing Sources (PSES), and Pretreatment Standards for New Sources
(PSNS) that apply to discharges of up to seven wastestreams: flue gas desulfurization (FGD) wastewater, fly
ash transport water, bottom ash transport water, combustion residual leachate from landfills and surface
impoundments, wastewater from flue gas mercury control (FGMC) systems and gasification systems, and
nonchemical metal cleaning wastes.
April 19, 2013 1-2
-------
Benefit and Cost Analysis for Proposed ELGs
1: Introduction
Table 1-1. Steam Electric Regulatory Options
Technology Basis for BAT/NSPS/PSES/PSNS
Regulatory Options
Wastestreams
FGD
Wastewater
Fly Ash
Transport
Water
Bottom Ash
Transport
Water
Combustion
Residual
Leachate
FGMC
Wastewater
Gasification
Wastewater
Nonchemical
Metal Cleaning
Wastes
1
Chemical
Precipitation
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Evaporation
Chemical
Precipitation
3a
BPJ
Determination
Dry Handling
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Dry Handling
Evaporation
Chemical
Precipitation
2
Chemical
Precipitation +
Biological
Treatment
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Evaporation
Chemical
Precipitation
3b
Chemical
Precipitation +
Biological
Treatment
**
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Dry Handling
Evaporation
Chemical
Precipitation
3
Chemical
Precipitation +
Biological
Treatment
Dry Handling
Impoundment
(Equal to BPT)
Impoundment
(Equal to BPT)
Dry Handling
Evaporation
Chemical
Precipitation
4a
Chemical
Precipitation +
Biological
Treatment
Dry Handling
Dry Handling
/Closed Loop
**
Impoundment
(Equal to BPT)
Dry Handling
Evaporation
Chemical
Precipitation
4
Chemical
Precipitation +
Biological
Treatment
Dry Handling
Dry Handling
/Closed Loop
Chemical
Precipitation
Dry Handling
Evaporation
Chemical
Precipitation
5
Chemical
Precipitation +
Evaporation
Dry Handling
Dry Handling
/Closed Loop
Chemical
Precipitation
Dry Handling
Evaporation
Chemical
Precipitation
** Requirement is subject to applicability threshold. For Option 3b FGD wastewater: Chemical Precipitation + Biological Treatment for units at a facility with a total
wet-scrubbed capacity of 2,000 MW and more; BPJ determination for units at a facility with a total wet-scrubbed capacity <2,000 MW. For Option 4a bottom ash
transport water: Dry handling/Closed loop for units >400 MW; Impoundment (Equal to BPT) for units <400 MW.
BPT = Best Practicable Control Technology Currently Available.
BPJ = Best Professional Judgment.
Source: U.S. EPA Analysis, 2013
April 19, 2013
1-0
-------
Benefit and Cost Analysis for Proposed ELGs
1: Introduction
After considering these regulatory options, EPA identified Options 3a, 3b, 3 and 4a as the preferred options
for regulation of pollutant discharges from existing sources (BAT and PSES). For new sources, EPA
identified Option 4 as the preferred option for NSPS and PSNS. The preamble that accompanies the proposed
regulation explains the rationale for EPA's determination.
The load reductions that would be achieved by the proposed regulatory options vary depending on the
pollutant type, which are categorized as conventional (such as total suspended solids, biochemical oxygen
demand, and oil and grease), priority (such as mercury, arsenic, and selenium), and non-conventional (such as
phosphorus and total dissolved solids). Table 1-2 summarizes the estimated conventional, priority, non-
conventional, and toxic-weighted pound equivalent1 pollutant reductions under each of the eight regulatory
options. The table lists the options in increasing order of total toxic-weighted pollutant removals.
Table 1-2. Pollutant Removal for Proposed ELGs Regulatory Options
Pollutant Load Reduction
(million pounds per year)
Regulatory
Option
Option T
Option 3 a
Option 2
Option 3b
Option 3
Option W
Option 4
Option 5
Conventional
Pollutants3
2.8
16.0
2.8
17.1
19.0
28.0
35.0
36.0
Priority Pollutants
0.5
0.4
0.7
0.6
1.1
1.4
1.7
1.7
Nonconventional
Pollutants'1
c-418
468
1,155
914
1,623
2,612
3,328
5,287
Toxic- Weighted Pound
Equivalent
1.5
2.5
2.6
3.4
5.1
6.7
7.8
8.2
a. The loadings reduction for conventional pollutants includes BOD and TSS.
b. The loadings reduction for nonconventional pollutants excludes TDS and COD to avoid double-counting removals for certain
pollutants that would also be measured by these bulk parameters (e.g., sodium, magnesium).
c. Option 1 shows a negative removal for nonconventional pollutants because the mass of several pollutants (ammonia, chromium,
TKN, and BOD) are not quantified at baseline, and because some pollutant discharge concentrations are higher under Option 1.
d. EPA estimated the pollutant removals for Option 4a based on approximated plant-level bottom ash loadings for those plants that have
at least one generating unit with a nameplate capacity of 400 MW or less and at least one other generating unit with a nameplate
capacity of greater than 400 MW. For more details on how EPA estimated these plant-level bottom ash loadings, see the memorandum
entitled "Steam Electric ELG Regulatory Option 4a Estimation Methodologies" (DCN SE03834).
Additionally, the regulatory options would eliminate or reduce water withdrawals associated with wet ash
transport and wet FGD scrubbers. EPA estimates that power plants would reduce the use of water by 50
billion gallons per year (136 million gallons per day) under Option 3a, by 52 billion gallons per year (143
million gallons per day) under Option 3b, by 53 billion gallons per year (144 million gallons per day) under
Option 3, and by 103 billion gallons per year (282 million gallons per day) under Option 4a.
nalytic Framework
The analytic framework of this BCA includes four basic components used consistently throughout the
analysis of benefits and social costs of the proposed ELGs:
1. All values are presented in 2010 dollars;
2. Benefits and social costs are analyzed over a 24-year period (2017 to 2040);
For information on toxic weighted pound equivalents, see TDD (U.S. EPA, 2013b).
April 19, 2013
1-1
-------
Benefits and Cost Analysis for Proposed ELGs 1: Introduction
3. Future benefits and social costs are discounted using rates of 3 percent and 7 percent back to the
expected rule promulgation year of 2014; and
4. Future values account for annual U.S. population and income growth.
These components are discussed in the sections below.
Note that the benefits and social cost analyses presented in this document are generally performed at the level
of individual steam electric plants (for all plants that were surveyed), with plant-level estimates multiplied by
sample weights to represent all 1,079 steam electric plants and the affected resources.2'3
EPA estimated benefits for five of the eight regulatory options considered for the proposed ELGs (Options 1,
2, 3, 4, and 5). EPA did not estimate the benefits of Options 3a, 3b and 4a. However, EPA used its
understanding of the wastestreams and treatment technologies for these options, along with projections of
pollutant reductions for all eight options, to infer total monetized benefits for Options 3a, 3b, and 4a. This is
because the five options EPA analyzed can serve as upper and lower bounds for the benefits of Options 3a,
3b, and 4a. Specifically, monetized benefits for Options 3a and 3b are likely to be between those for Options
2 and 3. Similarly, monetized benefits for Option 4a are likely to be between those for Options 3 and 4.
However, EPA is less confident that the approach used to infer benefits for Options 3a, 3b, and 4a based on
the results of the other five options would yield reasonable estimates if applied to the individual categories of
benefits (e.g., water quality, air emissions, avoided impoundment failure cleanup costs) and therefore the
Agency did not infer values for individual benefit categories but presents category-specific estimates only for
Options 1, 2, 3, 4, and 5.
1.3.1 Constant Prices
This BCA applies a year 2010 constant price level to all future annual monetary values of costs and benefits.
Some monetary values of benefits and costs are based on actual past market price data (i.e., prior to 2010),
and in those instances, EPA has updated the prices to 2010 by multiplying them by appropriate indexes, or
specific sub-components of these general indexes (index-updated prices). However, not all dollar-monetized
benefits and costs in this BCA are based on actual market prices of goods or services. Several categories of
benefits presented in this report are estimated based on consumer or household willingness-to-pay (WTP)
surveys, such as WTP for surface water quality improvements applied in this BCA for monetizing ecological
benefits of the proposed ELGs. This BCA updates these non-market prices as needed using appropriate
indexes (e.g., Consumer Price Index (CPI)).
1.3.2 Discount Rate and Year
This BCA estimates the annualized value of future benefits using two discount rates: 3 percent and 7 percent.
The 3 percent discount rate reflects society's valuation of differences in the timing of consumption; the
7 percent discount rate reflects the opportunity cost of capital to society. In Circular A-4, the Office of
Management and Budget (OMB) recommends that 3 percent be used when a regulation affects private
consumption, and 7 percent in evaluating a regulation that will mainly displace or alter the use of capital in
the private sector (OMB, 2003; updated 2009). The same discount rates are used for both benefits and social
costs.
All future cost and benefit values are discounted back to the expected ELG promulgation year of 2014.
2 Refer to the TDD for a discussion of the 2010 Questionnaire for the Steam Electric Power Generating Effluent
Guidelines (industry survey) and development and application of sample weights.
3 Benefits incur only to waterbodies affected by plants which have a sample weight of one. Benefits associated with
waterbodies affected by other steam electric plants are equal to zero. Therefore, the use of sample weights does not
introduce additional uncertainty on benefit estimates.
April 19, 2013
-------
Benefits and Cost Analysis for Proposed ELGs 1: Introduction
1.3.3 Period of A nalysis
Benefits are expected to begin accruing when each plant implements the control technologies needed to
comply with any applicable new effluent limits or standards. As discussed in the RIA (in Chapter 3:
Compliance Costs), for the purpose of the economic impact and benefit analysis, EPA assumes that plants
would implement control technologies three years after their NPDES permit comes up for renewal after the
rule promulgation (2014). Assuming that NPDES permits are renewed every five years, technology
implementation across all steam electric plants would occur during the period of calendar years 2017 through
2021.4 This schedule recognizes that control technology implementation is likely to be staggered over time
across the universe of steam electric plants.
The period of analysis extends to 2040 to capture the life of the longest-lived compliance technology at any
steam electric plant (20 years), and the last year of technology implementation (2021).
1.3.4 Population and Income Growth
To account for future population growth or decline, EPA adjusted affected population estimates based on
growth projections contained in EPA's Environmental and Benefits Mapping and Analysis Program
(BenMap) for the years 2014 through 2030. These projections are based on 2000 Census data as well as
projections taken from the 2007 Woods and Poole projection estimates (Woods and Poole Economics, 2007).
Because BenMap (at the time this analysis was conducted) does not include population estimates beyond
2030, EPA used stepwise autoregressive forecasting to forecast population values at the state level between
2030 and 2040. This method combines linear and autoregressive models.5
Also, since willingness-to-pay (WTP) is expected to increase as income increases, EPA took into account
income growth for some analyses (e.g., WTP for water quality improvements and groundwater protection and
value of a statistical life (VSL)). To estimate median state level household income between 2017 and 2040,
EPA used historic state-specific median household income data from the U.S. Census Bureau's 2009
Community Population Survey (U.S. Census Bureau, 2010b) for the years 1984 to 2009. The Consumer Price
Index (CPI) was used to adjust all value to 2010 dollars (U.S. BLS, 2010) and a stepwise autoregressive
forecasting method was used to estimate future annual state level median household income. For the health
benefits analyses in Chapter 3, EPA applied the projected income data (described above) along with the
income elasticity to adjust the VSL.6
1
.4
This Benefits and Cost Analysis (BCA) report presents EPA's analysis of the benefits of the proposed ELGs,
assessment of the total social costs, and comparison of the social costs and monetized benefits.
4 As specified in the preamble, certain limitations and standards based on any of the five main regulatory options being
proposed for existing direct and indirect dischargers do not apply until July 1, 2017 (approximately three years from the
effective date of the rule). The implementation period analyzed corresponds roughly to the timing of implementation for
proposed BAT and PSES for existing direct and indirect dischargers.
5 We used PROC FORECAST in SAS, which provides a quick and automatic way to generate forecasts for many time
series in one step. Details about PROC FORECAST can be found at
http://www.dms.umontreal.ca/~duchesne/chapl2.pdf
6 This method for adjusting VSL differs slightly from what BenMAP uses. EPA evaluated the effects of using the
BenMAP income growth factors instead of the method used in this analysis and found negligible differences in the
results.
April 19, 2013 T^
-------
Benefits and Cost Analysis for Proposed ELGs 1: Introduction
The report is organized as follows:
> Chapter 2: Benefits Overview provides an overview of the main benefits expected to result from the
implementation of the proposed ELGs.
> Chapter 3: Human Health Benefits details the methods and results of EPA's analysis of the human
health benefits.
> Chapter 4: Non-Market Benefits from Water Quality Improvements discusses EPA's analysis of the
water quality improvements resulting from the proposed ELGs.
> Chapter 5: Impacts and Benefits to Threatened and Endangered Species discusses expected benefits
to threatened and endangered (T&E) species.
> Chapter 6: Benefits from Groundwater Quality analyzes the benefits of reductions in the risk of
groundwater contamination.
> Chapter 7: Benefits from Avoided Impoundment Failures assesses the benefits of reducing the
impacts of coal combustion residue (CCR) releases due to the failure of impoundments used by some
steam electric plants to manage their CCR waste.
> Chapter 8: Air-Related Benefits describes EPA's analysis of benefits associated with changes in
emissions of air pollutants due to increased electricity consumption, transportation, and changes in the
profile of electricity generation.
> Chapter 9: Benefits from Reduced Water Withdrawals discusses ancillary benefits arising from
reduced surface water intake and groundwater use.
> Chapter 10: Summary of Total Benefits summarizes results across benefit categories.
> Chapter 11: Summary of Total Social Costs summarizes social costs of the proposed ELGs.
> Chapter 12: Benefits and Social Costs addresses the requirements of Executive Orders that EPA is
required to satisfy for this proposal, notably Executive Order 12866, which requires EPA to compare
the benefits and social costs of its actions.
> Chapter 13: References provides references cited in the text.
Several appendices provide additional details on selected aspects of analyses described in the main text of the
report.
April 19, 2013 1-4
-------
Benefit and Cost Analysis for Proposed ELGs
2: Benefits Overview
Benefits Overview
This chapter provides an overview of the potential benefits to society resulting from implementation of the
proposed ELGs. EPA expects that benefits would accrue to society in several broad categories, including
enhanced surface quality, reduced health risks, and increased productivity in economic activities that are
adversely affected by steam electric discharges. These effects follow directly from changes in effluent limits
and standards, which would reduce pollutant loadings to receiving waters. Benefits of the proposed ELGs
would also include secondary effects of the implementation of control technologies or other changes in plant
operations, such as reduction in emissions of air pollutants (e.g., CO2, NOX, and SO2) which provide benefits
in the form of reduced mortality and CO2 impacts on environmental quality and economic activities; reduction
in water use, which provide benefits in the form of increased availability of surface water and groundwater;
and reduction in the use of surface impoundment to manage CCR wastes, with benefits in the form of avoided
cleanup and other costs associated with episodic impoundment failures.
This chapter also provides a brief discussion of the steam electric pollutants, their human health and
ecological effects, and a framework for understanding the benefits likely to be achieved by the proposed
steam electric ELGs. For a more detailed description of steam electric pollutants, their fate, transport, and
impacts on human health and environment, see the Environmental Assessment document (U.S. EPA, 2013a).
Figure 2-1 summarizes the potential effects of the proposed ELGs, the expected environmental changes, and
categories of benefits, as well as EPA's approach to analyzing those benefits. EPA was not able to bring the
same depth of analysis to all categories of benefits because of imperfect understanding of the link between
discharge reductions or other environmental effects of the proposed ELGs and benefit categories, and how
society values some of the benefits. EPA was able to quantify and monetize some benefits, quantify but not
monetize other benefits, and assess still other benefits only qualitatively.
April 19, 2013
2-1
-------
Benefit and Cost Analysis for Proposed ELGs
2: Benefits Overview
Effect
Proposed ELG
Environmental
Change
Benefit
Valuation
Reduced metal and
nutrient loadings to
surface waters
/
\
Reduced fish tissue
contamination
Improved surface
water quality
/
/
\
•il
Improved Human Health
• Avoided cancer cases from arsenic exposure
• Avoided IQ losses in children from mercury and lead exposure
• Reduced cases of other non-cancer health effects
Improved Ecological Conditions
• Improved recreational and non-use values
• Threatened and endangered (T&E) species protection
Improved Economic Productivity
• Improved tourism
• Increased commercial fishery yields
* Reduced need for water treatment
• Enhanced property values
• VSL
• Value of an IQ point
* Count of human health criteria
exceedances {non-monetized)
• WTP for use and non-use values of
surface water quality improvements
• WTP for T&E population increases
Qualitative discussion
Reduced reliance on
impoundments to
manage CCR
Change in:
* Auxiliary power
use
• Transportation
• Electricity
generation
Reduced water use
/
\
v
/
\
\
Reduced leachate to
groundwater
Reduced CCR waste
managed in
impoundments
Reduced air
emissions of COj,
NOxrand SOx
Reduced
groundwater
withdrawals
Reduced surface
water withdrawals
>
fc.
•fc
Reduced groundwater contamination
Fewer and less consequential impoundment failures
• Reduced human mortality
• Reduced C02 impacts
Increased groundwater availability
Reduced impingement and enlrainment mortality
— >
^
h
WTP for reduced risk of groundwater
contamination
Avoided cost of cleanup, natural resource
damage, and litigation
• VSL
• Social cost of carbon
Avoided cost of water purchase
Qualitative discussion
VSL = Value of Statistical Life; WTP = Willingness to Pay; CCR = Coal Combustion Residuals
Figure 2-1. Summary of Benefits Resulting from the Proposed ELGs.
April 19, 2013
2-2
-------
Benefit and Cost Analysis for Proposed ELGs 2: Benefits Overview
uman Health Benefits Associated with Improvements in Surface Water
Quality
Steam electric pollutants can cause a wide variety of adverse human health effects arising, for example, from
drinking contaminated water (surface or groundwater) or consuming contaminated fish tissue. Metals are of
particular concern because they do not volatilize, do not biodegrade, can be toxic to plants, invertebrates and
fish, adsorb to sediments, and bio-concentrate in fish tissues (U.S. EPA, 2000a). More details on the fate,
transport, and exposure risks of steam electric pollutants are provided in the EA (U.S. EPA, 2013a).
Reducing pollutant discharges to the nation's waterways provides human health benefits by several
mechanisms. The most important and readily analyzed benefits stem from reduced risk of illness associated
with the consumption of water, fish, shellfish, and other aquatic organisms that are taken from waterways
affected by steam electric discharges. Human health benefits are typically analyzed by estimating the change
in the expected number of adverse human health events in the exposed population resulting from a reduction
in effluent discharges. While some health effects such as cancer are relatively well understood and can be
quantified in a benefits analysis, others are less well characterized and cannot be assessed with the same rigor,
or at all.
The proposed ELGs have the potential to provide human health benefits by reducing exposure to pollutants in
water via two principal exposure pathways: (1) consumption offish and shellfish taken from waterways
affected by steam electric discharges, and (2) consumption of water from surface waters affected by steam
electric plant discharges.
2.1.1 Fish Consumption
Recreational anglers and subsistence fishers (and their household members) who consume fish caught in the
reaches receiving steam electric plant discharges are expected to benefit from reduced pollutant
concentrations in fish tissue. EPA analyzed the following four direct measures of change in risk to human
health from exposure to contaminated fish tissue:
1. Incidence of cancer from fish consumption;
2. Neurological effects to children ages 0 to 7 from exposure to lead;
3. Neurological effects to infants from in-utero exposure to mercury; and
4. Reduced risk of non-cancer toxic effects from fish consumption.
EPA was able to monetize only the first three of these four measures. Incidence of cancer was translated into
an expected number of avoided mortality events and, on that basis, monetized. Lead and mercury impacts to
children were evaluated in terms of potential intellectual impairment as measured by estimated changes in
intelligence quotient (IQ). Details on these analyses are provided in the following sections of this report:
> Reduced cancer cases from arsenic exposure (Section 3.1),
> Reduced IQ losses in children ages 0 to 7 resulting from exposure to lead (Section 3.2), and
> Reduced IQ losses among children exposed to mercury in-utero (Section 3.3).
The fourth effect (reduced risk of non-cancer toxic effects from fish consumption) is addressed indirectly in
EPA's assessment of changes in exceedances of ambient water quality criteria (see Section 1.1).
The value of health benefits is the monetary value that society is willing to pay to avoid the adverse health
effects. Willingness to pay (WTP) to avoid morbidity or mortality is generally considered to be a
April 19, 2013 24
-------
Benefit and Cost Analysis for Proposed ELGs 2: Benefits Overview
comprehensive measure of the costs of health care, losses in income, and pain and suffering of affected
individuals and their caregivers. For example, the value of a statistical life (VSL) (see Section 3.1.1) is based
on estimates of society's WTP to avoid the risk of premature mortality. Alternatively, the cost-of-illness
approach, which is not used in this analysis, is a less comprehensive measure of cost: it allows valuation of a
particular type of non-fatal illness by placing monetary values on metrics, such as lost productivity and the
cost of health care and medications that can be monetized.
Some health benefits of reduced exposure to steam electric pollutants, such as neurological effects to children
and infants exposed to lead and mercury, are measured based on avoided IQ losses. Changes in IQ cannot be
valued based on WTP approaches since available economic research provides little empirical data on
society's WTP to avoid IQ losses. Instead, EPA calculated monetary values for avoided neurological and
cognitive damages based on the impact of an additional IQ point on an individual's future earnings and the
cost of compensatory education for children with learning disabilities. These estimates represent only one
component of society's WTP to avoid adverse neurological effects and therefore produce a partial measure of
benefits from reduced exposure to lead and mercury. Employed alone, these monetized benefits will
underestimate society's WTP, and perhaps significantly so. See Sections 3.2.1 and 3.3.1 for applications of
this method to valuing benefits to children and infants from reduced exposure to lead and mercury.
EPA expects that there would also be material health benefits via the fish consumption pathway arising from
reduced discharges of other steam electric pollutants, such as cadmium, selenium, and zinc. Analyses of these
health benefits are not possible due to lack of data on a quantitative relationship between ingestion rate and
potential adverse health effects.
Despite numerous studies conducted by EPA and other researchers, dose-response functions are available
only for a handful of health endpoints associated with steam electric pollutants. In addition, the available
research does not always allow complete economic evaluation, even for quantifiable health effects. For
example, EPA's analysis of health benefits omits the following health effects: morbidity preceding cancer
mortality from exposure to arsenic; neonatal mortality from exposure to lead (U.S. EPA, 2009a); effects to
adults from exposure to lead (including increased incidence of hypertension, heart attack, strokes, and
premature mortality, nervous system disorders, anemia and blood disorders, and other effects; U.S. EPA,
2009a; 2013a); effects to adults from exposure to mercury, including vision defects, hand-eye coordination,
hearing loss, tremors, cerebellar changes, and others (Mergler, et al., 2007; CDC, 2009); and non-cancer
effects from exposure to other steam electric pollutants. Therefore, the total monetized human health benefits
included in this analysis represent only a subset of the potential health benefits that would result from the
proposed ELGs.
2.1.2 Drinking Water Consumption
Steam electric pollutants discharged to surface waters may affect the quality of water used for public drinking
supplies. However, public drinking water supplies are subject to legally enforceable maximum contaminant
levels (MCLs) established by EPA (U.S. EPA, 2012a). As the term implies, an MCL for drinking water
specifies the highest level of a contaminant that is allowed in drinking water. The MCL is based on the MCL
Goal (MCLG), which is the level of a contaminant in drinking water below which there is no known or
expected risk to human health. EPA sets the MCL as close to the MCLG as possible, with consideration for
the best available treatment technologies and costs.
Pursuant to MCLs, public drinking water supplies are already treated for pollutants that pose human health
risks. As such, EPA restricted the analysis of monetized health benefits from improved surface water quality
to benefits arising from the consumption of contaminated fish tissue. Although treatment may not remove all
contaminants from the drinking water supplies and there may be some incremental health-related benefits
April 19, 2013 Z4
-------
Benefit and Cost Analysis for Proposed ELGs 2: Benefits Overview
associated with reduced concentrations arising from the proposed ELGs, EPA expects that these benefits
would not be substantial.7
2.1.3 Complementary Measure of Human Health Benefits
EPA quantified but did not monetize the expected reduction of pollutant concentrations in excess of human
health-based aquatic water quality criteria (AWQC) limits. This benefit measure was obtained by comparing
in-waterway pollutant concentrations to toxic effect levels. This analysis provides a measure of the change in
cancer and non-cancer health risk by comparing the number of receiving reaches exceeding health-based
AWQC for steam electric pollutants in the baseline to the number exceeding AWQC under the proposed
ELGs (Section 1.1).
AWQC are set at levels to protect human health through ingestion of water and aquatic organisms.
Accordingly, reducing the frequency at which human health-based AWQC are exceeded should translate into
reduced risk to human health. This measure should be viewed as an indirect indicator of reduced risk to
human health because it does not reflect the size of the exposed population and does not quantify changes in
human health risk per se.
2.2 Ecological Benefits Associated with Improvements in Surface Water Quali
The composition of steam electric plant wastewater depends on a variety of factors, such as fuel composition,
air pollution control technologies used, and waste management techniques used; wastewater often contains
metals such as aluminum, arsenic, boron, cadmium, chromium, copper, iron, lead, manganese, mercury,
nickel, selenium, thallium, vanadium, and zinc (U.S. EPA, 2013a). Discharges of these pollutants to surface
water has a wide variety of environmental effects, including fish kills, reduction in the survival and growth of
aquatic organisms, behavioral and physiological effects in wildlife, and degradation of aquatic habitat in the
vicinity of steam electric plant discharges (U.S. EPA, 2013a). The adverse effects associated with releases of
steam electric pollutants depend on many factors such as the chemical-specific properties of the effluent, the
mechanism, medium, and timing of releases, and site-specific environmental conditions.
EPA expects that the ecological benefits from the proposed ELGs would include enhanced habitat for fresh-
and saltwater plants, invertebrates, fish, and amphibians, as well as terrestrial wildlife and birds that prey on
aquatic organisms exposed to steam electric pollutants. The reduction in pollutant loadings is expected to
reestablish productive ecosystems in damaged waterways and to protect resident species, including threatened
and endangered species. EPA expects that the regulation would enhance the general health offish and
invertebrate populations, increase their propagation to waters currently impaired, and expand fisheries for
both commercial and recreational purposes. Improvements in water quality would also favor recreational
activities such as swimming, boating, fishing, and water skiing. Finally, the Agency expects that the
regulation would augment nonuse values (e.g., option, existence, and bequest values) of the affected water
resources.
2.2.1 Improved Surface Water Quality
The proposed steam electric ELGs are expected to provide ecological benefits through improvements in the
habitats or ecosystems (aquatic and terrestrial) that are affected by steam electric plant discharges. Society
values such ecological improvements by a number of mechanisms, including increased frequency and value
of use of the improved habitat for recreational activities. In addition, individuals may also value the protection
7 There may also be market benefits associated with the decreased need for drinking water treatment, but EPA did not
estimate these benefits as part of its analysis of the proposed ELG.
April 19, 2013
-------
Benefit and Cost Analysis for Proposed ELGs 2: Benefits Overview
of habitats and species that are adversely affected by effluent discharges, even when those individuals do not
use or anticipate future use of the affected waterways for recreational or other purposes.
Recreational activities that may be enhanced by reducing steam electric discharges to surface waters include:
> Recreational Fishing. Degraded water can reduce fish populations by inhibiting reproduction, growth,
and survival of an aquatic species (Friedman et al. 1996; Niimi and Kissoon 1994; U.S. EPA, 2009a;
U.S. EPA, 2011) resulting in fewer and smaller fish and thereby reducing the value of a fishing trip.
Reducing pollutant loads in steam electric plant discharges is expected to improve aquatic habitat and
thus increase the number, size, diversity, and health of recreational fish species and, as a result, the
value of recreational fishing. Studies have shown that the value of water resources for recreational
fishing increases with declining level of toxic contamination in fish tissue (Phaneuf et al., 1998; and
Jakus et al., 1997). In addition, improved aesthetic qualities of the waterbody (e.g., from reduced
nutrient loadings) and knowledge that the water is cleaner and does not contain any or contains fewer
pollutants that harm humans and aquatic life, increases individuals' enjoyment of their recreational
experience.
> Outings. Participants in other recreational activities such as hiking, jogging, picnicking, and wildlife
viewing would also benefit from improved abundance and diversity of aquatic and terrestrial species.
For example, wildlife viewers may benefit from improved abundance of piscivorous birds (e.g.,
osprey, eagle) and waterfowl whose populations are likely to increase due to a reduction of mercury
and other heavy metals in the food web and an increase in the forage fish populations (Schoch et al.,
2011; U.S. EPA, 2011). In addition, improved aesthetic quality of surface waters (e.g., clarity and
odors) would enhance the recreational experience of wildlife viewers and other recreational users.
(Schoch et al., 2011; U.S. EPA, 2011).
> Boating. Boaters may benefit from enhanced opportunities for companion activities, such as fishing
and wildlife viewing (e.g., piscivorous birds), and from improved aesthetic quality.
> Swimming. Swimmers may benefit from improved aesthetic quality of surface waters including water
clarity and odor thereby enhancing swimmer's aesthetic enjoyment of a waterbody.
> Hunting. Waterfowl hunters may benefit from improved aesthetic enjoyment of a water resource, an
increase in the number and quality of game available, and the removal of waterfowl consumption
advisories. Reducing nutrient loadings from steam electric plants is likely to benefit diving ducks
populations by reducing eutrophication and turbidity in the affected waters and improving their food
sources. Diving ducks rely upon undisturbed and abundant plant and invertebrate sources to prepare
for migration. Excessive nutrient loadings can lead to eutrophic and turbid waters, with few plants
and invertebrates food sources (MDNR, 2010). Waterfowl populations are adversely affected by
consuming contaminated fish or invertebrates; zebra mussels are an attractive food source for ducks
and have been found to have high concentrations of methyl mercury (MDNR, 2010). High mercury
levels have led to duck consumption advisories (Utah DNR, 2005). Reduction in metal loading to
surface waters and of their presence in the food web may benefit waterfowl reproduction and lead to
removal of duck consumption advisories.
EPA quantified potential ecological impacts from the proposed ELGs by estimating in-waterway
concentrations of nutrients and toxic pollutants discharged by steam electric plants and translating water
quality measurements into a single numerical indicator (water quality index (WQI)). EPA used the expected
change in WQI as a quantitative measure of ecological benefit for this regulatory analysis. Section 4.1 of this
report provides detail on the parameters used in formulating the WQI and the WQI methodology and
calculations.
April 19, 2013 2-6
-------
Benefit and Cost Analysis for Proposed ELGs 2: Benefits Overview
A variety of primary methods exist for estimating recreational use values, including both revealed and stated
preference methods (Freeman, 2003). Where appropriate data are available or can be collected, revealed
preference methods can represent a preferred set of methods for estimating use values. These methods use
observed behavior to infer users' values for environmental goods and services. Examples of revealed
preference methods include travel cost, hedonic pricing, and random utility (or site choice) models.
In contrast to direct use values, nonuse values are considered more difficult to estimate. Stated preference
methods, or benefit transfer based on stated preference studies, are the generally accepted techniques for
estimating these values (U.S. EPA, 2010b; OMB, 2003). Stated preference methods rely on carefully
designed surveys, which either (1) ask people about their WTP for particular ecological improvements, such
as increased protection of aquatic species or habitats with particular attributes, or (2) ask people to choose
between competing hypothetical "packages" of ecological improvements and household cost (Bateman et al.,
2003). In either case, values are estimated by statistical analysis of survey responses.
Although the use of primary research to estimate values is generally preferred, the realities of the regulatory
process often dictate that benefit transfer is the only option for assessing certain types of non-market values
(Rosenberger and Johnston, 2007). Thus, EPA developed a benefit transfer approach based on a meta-analysis
of surface water valuation studies to evaluate the use and non-use benefits of improved surface water quality
resulting from the proposed ELGs. This analysis is presented in Chapter 4. Benefit transfer is described as the
"practice of taking and adapting value estimates from past research ... and using them ... to assess the value
of a similar, but separate, change in a different resource" (Smith et al. 2002, p. 134). It involves adapting
research conducted for another purpose to estimate values within a particular policy context (Bergstrom and
De Civita, 1999). In the benefit transfer used for analyzing non-market benefits associated with water quality
improvements, EPA used a regression-based meta-analysis of 115 estimates of total WTP (including both use
and nonuse values) for water quality improvements, provided by the 45 original studies.8 The estimated
econometric model allows calculation of total WTP for improvements in a variety of environmental services
affected by water quality and valued by humans, including enhanced recreational and commercial fishing
opportunities, water-based recreation, and existence services such as aquatic life, wildlife, and habitat
designated uses.
2.2.2 Benefits to Threatened and Endangered Species
For threatened and endangered (T&E) species vulnerable to future extinction, even minor changes to
reproductive rates and small levels of mortality may represent a substantial portion of annual population
growth. Consequently, steam electric plant discharges may either lengthen recovery time, or hasten the
demise of these species. By reducing the discharge of steam electric pollutants to aquatic habitats, the
proposed ELGs would enhance the survivability of some T&E species living in these habitats. These T&E
species may have both use and nonuse values. However, given the protected nature of T&E species and the
fact that the majority of T&E species do not have direct uses, the majority of the economic value for T&E
species comes from nonuse values.
Species-specific estimates of nonuse values held for the protection of T&E species can be derived only by
primary research using stated preference techniques. However, the cost, administrative burden, and time
required to develop primary research estimates to value effects of the proposed regulation on T&E species are
beyond the schedule and resources available to EPA for this rulemaking. As an alternative, EPA used a
8 Although the potential limitations and challenges of benefit transfer are well established (Desvousges et al., 1998),
benefit transfers are a nearly universal component of benefit cost analyses conducted by and for government agencies.
As noted by Smith et al. (2002; p. 134), "nearly all benefit cost analyses rely on benefit transfers, whether they
acknowledge it or not."
April 19, 2013
-------
Benefit and Cost Analysis for Proposed ELGs 2: Benefits Overview
benefit transfer approach that relies on information from existing studies (U.S. EPA, 2010b). This benefit
transfer approach is based on a meta-analysis of 31 stated preference studies valuing threatened, rare, or
endangered fish, bird or mammal species (Richardson and Loomis, 2009). The estimated WTP equation
provided in this meta-analysis was used to estimate the monetary value of the potential increases in T&E
populations resulting from the proposed ELGs. This analysis and results are presented in Chapter 5. WTP
values for improvements in water quality discussed in the preceding section may inherently include benefits
to T&E species. Although there may be some overlap between WTP estimates for T&E species and the WTP
estimates for improvements in water quality, this overlap is likely to be minimal, however, since none of the
studies in EPA's meta-analysis of WTP for water quality improvements specifically mentioned or otherwise
prompted respondents to include benefits to T&E species populations (see Chapter 4).
2.2.3 Reduced Sediment Contamination
Effluent discharges from steam electric plants can also contaminate waterbody sediments. For example,
adsorption of arsenic, selenium, and other pollutants found in steam electric plant discharges can result in
accumulation of contaminated sediment on stream and lake beds (Ruhl, et al., 2012), posing a particular threat
to benthic (i.e., bottom-dwelling) organisms. These pollutants can later be re-released into the water column
and enter organisms at different trophic levels; concentrations of selenium and other steam electric pollutants
in fish tissue of organisms of lower trophic levels can bio-magnify through higher trophic levels, posing a
threat to the food chain at large (Ruhl, et al., 2012).
By reducing discharges of pollutants to receiving reaches, the proposed ELGs would reduce the future
contamination of waterbody sediments, thereby mitigating impacts to benthic organisms and reducing the
probability that the pollutants would later be released into the water column and affect surface water quality
and the waterbody food chain. Due to data limitations, EPA did not quantify or monetize this benefit.
2.3 Benefits Associated with improvements in uroundwater Quality
Impoundments used by steam electric plants to manage their wastewater can leach pollutants into
groundwater aquifers, degrading water quality and potentially creating health hazards to households drawing
drinking water from affected aquifers. The operational changes prompted by the proposed ELGs are expected
to result in plants closing or significantly reducing their use of impoundments to manage coal combustion
residuals (CCR). The associated reduction in the risk of groundwater contamination represents benefits
attributable to the rule.
To evaluate benefits of reduced contamination of groundwater aquifers in the vicinity of steam electric
impoundments, EPA applied a benefit transfer approach based on the results of a meta-analysis of
groundwater valuation studies conducted by Poe et al. (2001). Poe et al. (2001) used data from 13 studies that
elicit household WTP values for reducing risk of groundwater contamination, to develop a meta-regression
function based on economic variables (including changes in contamination probability and income, whether
cancer was mentioned as a health concern, and the share of the population on public water supplies) and
methodological variables (including elicitation methods). EPA used the estimated meta-regression function to
estimate WTP for reducing risk of groundwater contamination for households that rely on private wells
drawing water from aquifers surrounding steam electric plants. This analysis and results are presented in
Chapter 6.
April 19, 2013 2-8
-------
Benefit and Cost Analysis for Proposed ELGs 2: Benefits Overview
2.4 Economic Productivity Benefits
The economic productivity benefits expected from the proposed ELGs include reduced impacts of
impoundment failures and the reduction in the costs associated with the cleanup, environmental damages, and
transaction costs from the resulting release of coal combustion residuals (CCR) into the environment. Other
economic productivity benefits may stem from reduced contamination of public drinking water supplies and
irrigation water; increased tourism; increased commercial fish harvests; and increased property values.
2.4.1 Reduced Impoundment Failures
Steam electric plants manage CCR such as fly ash and bottom ash through either wet or dry handling. For
plants that use wet handling, the waste is typically sluiced to one or more surface impoundments (e.g., settling
ponds), where the solids settle out of the water. Many plants also use surface impoundments to manage their
flue gas desulfurization (FGD) wastewater. In addition to solids associated with the ash and FGD wastes,
these impoundments typically contain water with high concentrations of steam electric pollutants, including
dissolved metals.
The operational changes prompted by the proposed ELGs are expected to cause some plant owners to reduce
their reliance on impoundments to handle CCR. These changes could affect the volume of CCR released in
the event of a failure and/or the future probability of impoundment failures. Benefits arising from the reduced
risk of impoundment failures include avoided cleanup costs, environmental damage, and transaction costs.
EPA quantified and monetized these benefits based on expected future impoundment structural failure rates,
the volumes of CCR that would be released in the event of a failure, and the costs of spill cleanup, natural
resource damages, and transaction costs. Chapter 7 describes this analysis.
2.4.2 Water Supply and Use
The proposed ELGs are expected to reduce loading of steam electric pollutants to surface waters and thus
enhance uses of these waters for drinking water supply and agriculture:
> Drinking water supply: The proposed ELGs are expected to reduce costs of drinking water treatment
(e.g., filtration and chemical treatment) by reducing metal concentrations and eutrophication in source
waters. Eutrophication is one of the main causes of taste and odor impairment in drinking water,
which has a major negative impact on public perceptions of drinking water safety. Additional
treatment to address foul tastes and odors can significantly increase the cost of public water supply.
Further, public drinking water sources do not always effectively remove bromides (a steam electric
pollutant) from raw surface waters. Elevated bromide concentrations in source waters result in
increased trihalomethanes (THMs) in drinking water (Pittsburgh Water and Sewer
Authority/University of Pittsburgh School of Engineering, 2012). Drinking water utilities downstream
of bromide sources are increasingly finding it difficult to meet drinking water standards for THMs. If
water treatment is not sufficient, an alternate water source needs to be substituted (if available). Long-
term solutions might require the development of new raw water supplies, which would involve costs
for the acquisition of land (if available), regulatory review and permitting, development of
infrastructure (dams, pumps, pipes), and watershed protection.
> Irrigation and other agricultural uses: Reducing steam electric pollutants discharges can improve
agricultural productivity by improving water quality used for irrigation and livestock watering (Clark
et al., 1985). Although elevated nutrient concentrations in irrigation water would not adversely affect
its usefulness for plants, concerns exist for potential residual effects due to steam electric pollutants
April 19, 2013 2-9
-------
Benefit and Cost Analysis for Proposed ELGs 2: Benefits Overview
entering the food chain. Further, eutrophication promotes cyanobacteria blooms that can kill livestock
and wildlife that drink the contaminated surface water.
EPA did not quantify or monetize benefits from enhanced quality of drinking and agricultural water sources
arising from the proposed ELGs due to data limitations.
2.4.3 Commercial Fisheries
Pollutants in steam electric plant discharges can reduce fish populations by inhibiting reproduction and
survival of aquatic species. These changes may negatively affect commercial fishing industries as well as
consumers offish, shellfish, and fish and seafood products. Estuaries are particularly important breeding and
nursery areas for commercial fish and shellfish species. In some cases, excessive pollutant loadings can lead
to the closures of shellfish beds, thereby reducing shellfish harvests. Improved water quality due to reduced
discharges of steam electric pollutants would enhance aquatic life habitat and, as a result, contribute to
reproduction and survival of commercially harvested species and larger fish and shellfish harvest, which in
turn lead to an increase in producer and consumer surplus.
EPA did not quantify or monetize benefits to commercial fisheries from the proposed ELGs. EPA's EA (see
U.S. EPA, 2013a) shows that a small number of steam electric plants discharge to estuaries or marine waters;
as a result, the benefits to commercial fisheries arising from the proposed ELGs are likely to be small.
2.4.4 Tourism
The proposed ELGs may also benefit local economies by contributing to the tourism industries (e.g., sales of
fishing equipment) in the areas surrounding affected waters due to improved recreational opportunities. The
effects of water quality on tourism are likely to be highly localized. Moreover, since substitute tourism
locations may be available, increased tourism in the vicinity of steam electric plants may lead to a reduction in
tourism in other locations. Due to these factors EPA believes that benefit from an increase in tourism would
be limited to communities in the vicinity of steam electric plants; although tourism revenue is potentially
important to these communities, the overall societal benefits are likely to be small. Therefore, EPA did not
quantify or monetize this benefit category.
2.4.5 Property Values
The proposed ELGs are expected to improve the aesthetic quality of land and water resources by reducing
pollutant discharges and thus enhancing water clarity, odor, and color in the receiving and downstream
reaches. Several studies (Boyle et al., 1999; Poor et al., 2001; Leggett and Bockstael, 2000) suggest that
waterfront property is more desirable when located near unpolluted water. Therefore, the value of properties
located in proximity to waters contaminated with steam electric pollutants may increase due to reduced steam
electric discharges. Although this benefit would accrue to the current property owners only, it represents an
overall increase in societal wealth.
Due to data limitations, EPA was not able to quantify or monetize the potential increase in property values
associated with the proposed ELGs. The magnitude of the potential increase depends on many factors,
including the number of housing units located in the vicinity of the affected waterbodies, community (e.g.,
residential density) and housing stock (e.g., single family or multiple family) and the effects of steam electric
pollutants on aesthetic quality of surface water. The expected changes in property values are likely to be
small. In addition, there may be some overlap between changes in property values and the estimated total
WTP for surface water quality improvements summarized in Section 2.2.1.
April 19, 2013 2-10
-------
Benefit and Cost Analysis for Proposed ELGs
2: Benefits Overview
2.5 Reduced Air Pollution
The proposed ELGs are expected to affect air pollution through three main mechanisms: 1) additional
auxiliary electricity use by steam electric plants to operate wastewater treatment, ash handling, and other
systems needed to comply with the new effluent limits and standards; 2) additional transportation-related
emissions due to the increased trucking of CCR waste to on-site or off-site landfills; and 3) the change in the
profile of electricity generation due to relatively higher cost to generate electricity at plants incurring
compliance costs for the proposed ELGs. The different profile of generation can result in lower or higher air
pollutant emissions due to differences in emission factors for coal or natural gas combustion, or nuclear or
hydroelectric power generation.
Of the three mechanisms above, the change in the emissions profile of electricity generation at the market
level is the most significant. Small reductions in coal-based electricity generation as a result of the proposed
ELGs are compensated by increases in generation using other fuels or energy sources - biomass, landfill gas,
natural gas, nuclear power, oil, and wind power. The changes in air emissions reflect the differences in
emissions factors for these other fuels, as compared to coal. Overall for the three mechanisms (auxiliary
services, transportation, and market-level generation), EPA estimates a net reduction in CO2 and SO2, and a
slight increase in NOX emissions. NOX, and SO2 are known precursors to PM2.5, a criteria air pollutant that
has been associated with a variety of adverse health effects - most notably, premature mortality. To estimate
benefits of reducing NOX, and SO2 emissions, EPA used estimates of national monetized benefits per ton of
emissions avoided. CO2is an important greenhouse gas that is linked to climate change effects including
global warming, sea level rise, increased frequency of extreme weather events, ocean acidification, etc. EPA
used estimates of the social cost of carbon (SCC) obtained from the Interagency Working Group on Social
Cost of Carbon (see IWGSCC 2010, p. 1) to derive benefits per ton for CO2 The SCC reflects abroad range
of climate change impacts, including changes in agricultural productivity, human health risks, property
damage from increased flood frequencies, the loss of ecosystem services, and others. Chapter 8 details this
analysis.
2
.6 Reduced Water Withdrawals
Steam electric plants use water wet ash transport and for operating wet FGD scrubbers. By eliminating or
reducing water used in sluicing operations or prompting the recycling of water in FGD wastewater treatment
systems, the proposed ELGs are expected to reduce demand on aquifers by plants that rely on groundwater
sources.
Reduced surface water intake would reduce impingement and entrainment mortality. Due to data limitations,
EPA did not quantify and monetize these benefits as part of this analysis.
Reduced water use from groundwater sources by steam electric plants would result in greater availability of
groundwater supplies for alternative uses. EPA used an avoided cost method (based on the cost of
desalination to replace groundwater supplies) to value the increased quantity of groundwater. This analysis is
presented in Chapter 9.
2.7 Summary of Benefits Categories
Table 2-1 summarizes the benefits of the proposed ELGs and the level of analysis applied to each category.
As indicated in the table, only a subset of anticipated benefits can be quantified and monetized (in which case
the table identifies the section of the report that discusses the analysis). The monetized benefits include
reductions in some human health risks, use and non-use values from improved surface water quality, benefits
to threatened and endangered species, improved groundwater quality, reduced impoundment failures, reduced
April 19, 2013 £TT
-------
Benefit and Cost Analysis for Proposed ELGs
2: Benefits Overview
air pollution, and reduced water withdrawals. Other benefit categories, including some human health risk
reductions, can be quantified but not monetized. Finally, EPA was not able to quantify or monetize other
benefits, including drinking water treatment costs and benefits to commercial fisheries; EPA evaluated these
benefits qualitatively as discussed above in Sections 2.1 through 2.6.
Tab|.e.2-1 . Benefits of Reduced Pollutant Discharges from Steam Electric Power Plants
Category
Effect of Proposed ELGs
Benefits Analysis
Quantified
Monetized
Methods (Report
Chapter or Section
where Analysis is
Detailed)
Human Health Benefits from Surface Water Quality Improvements
Reduced incidence of
cancer
Reduced IQ losses to
infants
Reduced IQ losses to
children ages 0 to 7
Reduced need for
specialized education
Reduced other adverse
health effects (cancer
and non-cancer)
Reduced adverse health
effects
Reduced exposure to arsenic from
fish consumption
Reduced in-utero mercury exposure
from maternal fish consumption
Reduced childhood exposure to lead
from fish consumption
Reduced childhood exposure to lead
from fish consumption
Reduced exposure to other
pollutants (arsenic, lead, etc.) via
fish consumption
Reduced exposure to pollutants from
recreational water uses
•/
^
^
•/
•/
•/
•/
•/
•/
VSL (Section 3.1)
IQ point valuation
(Section 3.3)
IQ point valuation
(Section 3.2)
Avoided cost
(Section 3.2)
Human health
criteria exceedances
(Section 3.5)
Qualitative
discussion
Ecological Conditions and Recreational use Benefits from Surface Water Quality Improvements
Improved aquatic and
wildlife habitat3
Water-based recreation3
Aesthetics3
Non-use values3
Aquatic and wildlife
Improved protection of
T&E species
Reduced sediment
contamination
Improved ambient water quality in
receiving reaches
Enhanced swimming, fishing,
boating, and near-water activities
from improved water quality
Increased aesthetics from improved
water clarity, color, odor, including
nearby site amenities (residing,
working, traveling)
Enhanced existence, option, and
bequest values from improved
ecosystem health
Reduced risks to aquatic life from
exposure to steam electric pollutants
Improved T&E habitat and thus
potential increase in T&E population
Reduced deposition of toxic
pollutants to sediment
•/
^
•/
^
•/
•/
•/
•/
•/
•/
•/
•/
Benefit transfer
(Chapter 4)
Benefit transfer
(Chapter 4)
Benefit transfer
(Chapter 4)
Benefit transfer
(Chapter 4)
Benefit transfer
(Chapter 4)
Benefit transfer
(Chapter 4)
Qualitative
discussion
Groundwater Quality Benefits
Groundwater quality
Reduced groundwater contamination
Benefit transfer
(Chapter 6)
April 19, 2013
2-12
-------
Benefit and Cost Analysis for Proposed ELGs 2: Benefits Overview
Table 2-1. Benefits of Reduced Pollutant Discharges from Steam Electric Power Plants
Category
Effect of Proposed ELGs
Benefits Analysis
Quantified
Monetized
Methods (Report
Chapter or Section
where Analysis is
Detailed)
Market and Productivity Benefits
Impoundment failures
Reduced water treatment
costs for drinking water
and irrigation water
Commercial fisheries
Benefits to tourism
industries
Property values
Reduced risk of impoundment
failures due to changes in the use of
impoundment
Improved quality of source water
used for drinking and irrigation
Improved fisheries yield and harvest
quality due to aquatic habitat
improvement
Increased participation in water-
based recreation
Increased property values from
water quality improvements
•/
•/
Avoided cost of
cleanup, natural
resource damages,
and transaction
costs (Chapter 7)
Qualitative
discussion
Qualitative
discussion
Qualitative
discussion
Qualitative
discussion
Air-Related Benefits
Reduced air emissions of
NOX, S02
Reduced air emissions of
C02
Reduced mortality from exposure to
NOX, SO2 and paniculate matter
(PM25)
Avoided climate change /global
warming impacts
^
^
•/
•/
Benefit per ton of
air pollutant
removed (Chapter
8)
Social cost of
carbon (SCC)
(Chapter 8)
Reduced Water Withdrawal Benefits
Reduced groundwater
withdrawals
Reduced surface water
withdrawals
Increased availability of
groundwater resources
Reduced impingement and
entrainment mortality
^
•/
Avoided cost
(Chapter 9)
Qualitative
discussion
a. These values are implicit in the total WTP for water quality improvements.
April 19, 2013
2-13
-------
Benefit and Cost Analysis for Proposed ELGs
3: Human Health Benefits
Human Health Benefits
EPA expects that the proposed ELGs would yield a range of health benefits by reducing effluent discharges to
surface waters and, as a result, ambient pollutant concentrations in the receiving reaches. EPA's EA (U.S.
EPA, 2013a) provides details on the health effects caused by steam electric pollutants. Recreational anglers
and subsistence fishers (and their household members) who consume fish caught in the reaches receiving
steam electric discharges are expected to benefit from reduced pollutant concentrations in fish tissue.9 This
chapter presents EPA's analysis of human health benefits from reduced exposure to steam electric pollutants
via the fish consumption pathway.10 The analyzed health benefit categories include: 1) reduced cancer cases
from arsenic exposure, 2) reduced intelligence quotient (IQ) losses in children resulting from exposure to
lead, and 3) reduced IQ losses among children exposed to mercury in-utero.
The total quantified human health benefits included in this analysis represent only a subset of the potential
health benefits expected to result from the proposed ELGs due to lack of data on a dose-response
relationship11 between ingestion rates and potential other adverse health effects (such as kidney damage from
cadmium or selenium exposure, cardiovascular impacts from lead exposure, gastrointestinal problems from
zinc, thallium, or boron exposure, and others). Further, the analyses rely on several models that incorporate
various assumptions that contribute to uncertainty in the estimated benefits and are subject to the limitations
of the underlying water quality and fish tissue data developed in the EA (see U.S. EPA, 2013a). Beyond these
limitations, the methodologies used to assess the human health benefits involve other limitations and
uncertainties summarized in Section 3.6.
EPA's analysis of the monetary value of human health benefits is based on data from the EA (see U.S. EPA,
2013a for more details). The relevant data include COMIDs12 for receiving waters, ambient pollutant
concentrations in receiving reaches under the baseline and post-compliance scenarios,13 pollutant
concentrations in fish tissue, fish consumption rates among age cohorts for affected recreational anglers and
subsistence fishers, and the daily average dose and the Lifetime Average Daily Dose (LADD, with units
mg/kg BW/day) of pollutants for each age cohort for recreational anglers and subsistence fishers. Table 3-1
identifies the cohorts defined in the EA and used for this analysis.
9 Although reaches downstream of those receiving steam electric discharges may also be affected, this analysis focuses
only on the benefits accrued from reaches receiving discharges directly from steam electric plants.
10 The analysis of human health benefits focuses on the fish consumption pathway only, since EPA assumed that
drinking water is treated to reduce pollutant concentrations below MCLs. See Section 2.1.2: Drinking Water
Consumption for details.
11 A dose response relationship is an increase in incidences of an adverse health outcome per unit increase in exposure to
a toxin.
12 A COMID is a unique numeric identifier for a given waterbody, assigned by a joint effort of the United States
Geological Survey, EPA, and Horizon Systems, Inc.
13 The baseline and post-compliance conditions account for discharges from steam electric plants only. It does not
account for other pollutant sources to receiving waters (e.g., other industrial plant discharges, non-point sources, air
deposition), which may contribute to elevated ambient concentrations but would not be affected by the proposed ELG.
April 19, 2013
3-1
-------
Benefit and Cost Analysis for Proposed ELGs
3: Human Health Benefits
Table 3-1. Cohorts Used for the Human Health Benefits Analysis
Cohort
Age (years)
Recreational Anglers and
their Households
0-1
2
3-5
6-10
11-15
16-20
Adult (21 or higher)
Subsistence Fishers and their
Households
0-1
2
3-5
6-10
11-15
16-20
Adult (2 lor higher)
Source: U.S. EPA, 2013a.
3.1 Reduced Cancer Cases from Consumption of Fish Contaminated with Arsenic
Among steam electric pollutants analyzed in the EA, arsenic is the only confirmed carcinogen with a
published dose response function (see U.S. EPA, 2010c).14 EPA estimated the number of annual cancer cases
associated with consumption of fish contaminated with arsenic from steam electric discharges under the
baseline and each analyzed regulatory option. The reduction in the number of cancer cases from the baseline
to post-compliance represents human health benefits attributable to the proposed ELGs.
3.1.1 Methodology and Data
This section describes the methodology and data used to 1) identify the population of recreational and
subsistence fishers potentially exposed to steam electric pollutants (hereafter, affected population); 2) estimate
the difference in the number of avoided cancer cases due to the proposed ELGs under each of the regulatory
options; and 3) estimate the monetary value of avoiding cancer cases.
3.1.1.1 Affected Population
The affected population (i.e., individuals potentially exposed to steam electric pollutants via consumption of
contaminated fish tissue) includes recreational anglers and subsistence fishers who fish reaches receiving
steam electric discharges, and their household members. EPA estimated the number of people who are likely
to fish reaches receiving steam electric discharges based on typical travel distances to a fishing site, presence
of substitute fishing locations, data on the locations and status offish consumption advisories (FCAs) for
receiving reaches, and information on anglers' awareness and adherence to FCAs. To account for the family
members potentially exposed to steam electric pollutants, EPA multiplied the number of anglers estimated to
fish receiving reaches by the average household size in a given state. Appendix A provides details on the
assumptions and data used in estimating affected populations. Equation 3-1 shows the calculation of the
affected population, ExPop(i)(s), for waterbody / in state s.
Equation 3-1. ExPop(i)(s) = PoplOO(i) x %Fis/i(s) x L^J x 4(i) x
BufM(iy
The equation terms are defined as follows:
> Total population living within 100-mile distance of receiving reach - PoplOO(i): The number of
anglers who fish receiving reaches depends on how far anglers typically travel to their fishing
destinations. Viscusi et al. (2008) found that 78 percent of anglers live within 100 miles of their
14 Although other pollutants, such as cadmium, are also likely to be carcinogenic (see U.S. DHHS, 2008), EPA did not
identify dose-response functions to quantify the effects of changes in these other pollutants.
April 19, 2013
3-2
-------
Benefit and Cost Analysis for Proposed ELGs 3: Human Health Benefits
fishing destinations. Accordingly, EPA used a distance of 100 miles to estimate the number of
individuals potentially exposed to steam electric pollutants via fish consumption. As a first step in
estimating the affected population, EPA estimated the number of people residing within the 100-mile
buffer of receiving reaches by overlaying the receiving reach COMIDs and population data from the
2010 Census.
> Fraction of people who fish15 in the population - %Fish(s): To determine what percentage of the total
population participates in fishing, EPA used region-specific U.S. Fish and Wildlife Service (U.S.
FWS, 2006) estimates of the population 16 and older who fish in conjunction with U.S. Census
Bureau (2010b) data on the share of the total population that is over 16.16
> Adjustment for substitute sites - [M(i)/BuJM(i)j: Anglers may live within 100 miles of more than one
waterbody and may choose to fish in other nearby waterbodies rather than the receiving reach. To
account for the substitution effect of other sites, EPA estimated the fraction of the total stream miles
within a 100-mile buffer zone attributed to the receiving reach. EPA assumed that recreational fishing
efforts are uniformly distributed across all reach miles in the 100-mile buffer zone. Therefore, the
number of anglers fishing the receiving reach is estimated by multiplying the fraction of the total
reach miles attributed to the receiving reach by the angler population residing within the 100-mile
buffer zone. For example, if the buffer zone around a 0.5-mile receiving reach includes 500 total
miles of streams, EPA limited the estimate of potentially affected anglers to 0.1 percent of anglers
living within the 100-mile buffer zone. This analysis does not account for the quality of substitute
sites and, as a result, under- or overstates the fraction of anglers fishing the receiving reach.
> Adjustment for fish consumption advisories -A(i): EPA further adjusted the affected angler
population to reflect the presence of FCAs. Based on EPA's review of studies documenting anglers'
awareness of FCA and their behavioral responses to FCA,1? approximately 57.0 percent to
61.2 percent of anglers are aware of FCA, and 71.6 percent to 76.1 percent of those who are aware
ignore FCA. Conservatively assuming that 61.2 percent of anglers are aware of applicable FCA and
that 71.6 percent of aware anglers ignore them, the number of anglers exposed to steam electric
pollutants would be 17.4 percent lower for reaches with FCA.18 Therefore, for receiving reaches with
FCA (as identified in the EA), EPA reduced the affected populations by 17.4 percent. According to
U.S. FWS (2006) data, approximately 23.3 percent of anglers release all the fish they catch ("catch-
and-release" anglers). Anglers practicing "catch-and-release" would not be exposed to steam electric
pollutants via consumption of contaminated fish. EPA did not further adjust the affected population to
account for this practice since the data used to calculate the lifetime average daily dose (LADD) due
to fish consumption is based on actual consumption rates for all recreational anglers, including those
that practice "catch and release."19
> Number of exposed family members - PPH(s): To account for family members in anglers households
who may be exposed to steam electric pollutants via consumption of recreationally caught fish, EPA
used the U.S. Census Bureau (2010a) data on the average number of people per household by state.
15 Consistent with the assumed 95 percentile fish consumption rate for subsistence fishers, EPA assumed that 5 percent
of the fishing population consists of subsistence fishers, and that the remaining 95 percent of the fishing population
consists of recreational anglers.
16 See ExhibitA-l inAppendixA for details.
17 See ExhibitA-2 and ExhibitA-3 inAppendixA for details.
18 This is calculated as 61.2% aware of advisories times 28.4% (100%-71.6%) who choose not to fish or otherwise don't
eat fish caught in waterbodies affected by advisories.
19 See the EA (U.S. EPA, 2012a) for a more detailed discussion of the development of lifetime average daily dose
(LADD) estimates.
April 19, 2013 3^
-------
Benefit and Cost Analysis for Proposed ELGs 3: Human Health Benefits
To account for future population growth or decline, EPA also adjusted the affected population estimates based
on the projected population changes from EPA's Environmental and Benefits Mapping and Analysis Program
(BenMap) for the years 2014 (the promulgation year) through 2030. These projections are based on 2000
Census data as well as projections taken from the 2007 Woods and Poole projection estimates (Woods and
Poole Economics, 2007). Because BenMap (at the time this analysis was conducted) does not project
population estimates past 2030, EPA used stepwise autoregressive forecasting to estimate population changes
between 2030 and 2040. 20
Many of the 100-mile buffers around receiving reaches span more than one state. To estimate affected
populations within multi-state buffers, EPA used the share of the 100-mile buffer area in each state to
estimated weighted averages for the following parameters: percent of the population participating in fishing,
persons per household, and population growth.
Finally, EPA disaggregated the total affected population into each cohort based on 1) U.S. Census data
documenting the share of the total population in one-year age increments (U.S. Census Bureau, 2010b), and
2) assumptions about the share of anglers who practice subsistence fishing (versus recreational). EPA
assumed that 5 percent of the affected population would be considered subsistence fishers, consistent with the
assumed 95th percentile fish consumption rate for high-risk subsistence fishers.21
3. 1.1.2 Estimating Avoided Cancer Cases
In the EA (U.S. EPA, 2013a), EPA estimated the LADD for seven age cohorts for recreational and
subsistence anglers (14 cohorts total, Table 3-1) for the baseline and each analyzed regulatory option. EPA
used these data to calculate the total number of cancer cases for each cohort for each receiving reach under the
baseline and each of the regulatory options, based on Equation 3-2.
Equation 3-2. CC(Q(c) = ExPop(i)(c} * CSF * LADD(f)(c)
CC(/)(c) = the number of cancer cases for waterbody / in cohort c.
ExPop(/)(c) = the number of people affected for waterbody / in cohort c.
CSF = Cancer Slope Factor for skin cancer from arsenic [1.5 (mg/kg BW/day)"1].
LADD(/)(c) = Lifetime Average Daily Dose of arsenic for waterbody /' in cohort c (mg/kg
BW/day).
Summing the number of cancer cases across all cohorts and all receiving reaches yields the total number of
annual cancer cases under the baseline and each of the regulatory options.22 To estimate the number of
avoided cancer cases, EPA subtracted the estimated number of cancer cases for each analyzed regulatory
option from the estimated number of cancer cases under the baseline.
3.1.1.3 Monetary Value of Avoided Cancer Cases
EPA used the "value of a statistical life" (VSL) approach to estimate the monetary value of benefits
associated with avoided cancer cases. EPA's Guidelines for Preparing Economic Analyses (U.S. EPA, 2010a)
20 See Exhibit A-4 in Appendix A for population growth projections over the analysis period.
21 Recreational anglers consume less fish and would thus experience lower cancer risks than subsistence fishers.
22 EPA multiplied results for receiving reaches associated with surveyed steam electric plants by the sample weight to
estimate total cases for all 1,079 steam electric plants subject to the proposed ELG. Because benefits are incurred only
for waterbodies affected by steam electric plants with a sample weight of one, the use of sample weights does not
introduce additional uncertainty.
April 19, 2013
-------
Benefit and Cost Analysis for Proposed ELGs 3: Human Health Benefits
currently recommends a default VSL of $7.9 million (2008 dollars), based on a distribution fitted to 26
published VSL estimates. Updated to 2010 dollars,23 the VSL is $8.0 million.
The VSL is based on estimates of society's willingness-to-pay (WTP) to avoid the risk of premature mortality
from unforeseen instant mortality with no significant period of morbidity. The VSL approach may overstate
benefits of avoiding cancer cases because not all cancer cases are fatal. On the other hand, the use of VSL to
value an avoided cancer case may underestimate the benefits of reduced cancer risk, since it does not consider
the costs and other effects that usually precede premature mortality from cancer (e.g., pain, suffering, and
medical costs).
A reduction in pollutant loadings does not immediately result in cessation of adverse health effects. There is a
lag between the time when exposures are reduced and the time when a reduction in risk occurs. Additionally,
there may be a latency period between the initial exposure and the onset of the illness. The duration of the
cessation lag is unknown but may be as long as decades.24 For this analysis, EPA assumed that cancer cases
resulting from arsenic would not occur for ten years after exposure and discounted the value of avoided
cancer cases by an additional ten years.
The value of cancer risk reduction is a "normal good",25 and thus is expected to grow over time as real income
grows. EPA used historic state-specific median household income data from the U.S. Census Bureau's 2009
Community Population Survey (U.S. Census Bureau, 2010b) for the years 1984 to 2009, and applied a
stepwise autoregressive forecasting method to estimate future annual state level median household income
through 2040.26 For each year in the analysis, EPA adjusted the VSL to account for income growth
projections and the mid-range income elasticity assumptions27 from U.S. EPA (2010a).
3.1.2 Results
Table 3-2 shows the estimated changes in incidence of cancer cases from exposure to arsenic in fish tissue
under the proposed ELGs and the annualized benefits calculated using 3 percent and 7 percent discount rates.
The table lists the regulatory options in order of their total toxic-weighted pollutant removals.
EPA estimates that Option 3 would reduce the number of cancer cases between 2017 and 2040 from 0.9 under
the baseline to 0.4, for annual benefits of $0.09 million at a 3 percent discount rate ($0.05 million at a 7
percent discount rate). EPA expects Options 3a and 3b to have smaller human health effects and benefits than
Option 3. Option 4 would provide further reductions in the number of cancer cases (down to 0.2 over the
period of 2017 through 2040), for total annual benefits of $0.15 million at a 3 percent discount rate ($0.09
million at a 7 percent discount rate). The reductions in cancer cases and benefits of Option 4a are expected to
be between those of Options 3 and 4, i.e., 0.4 to 0.7 fewer cancer cases per year and benefits between $0.09
million and $0.15 million, at a 3 percent discount rate.
23 EPA adjusted the VSL using the Consumer Price Index (CPI); 2010 = 218.056; 2008 = 215.303.
24 U.S. EPA (2010c) notes that the cessation lag for skin cancer from arsenic is unknown, but that the cessation lag for
internal cancers from arsenic may be longer than for skin cancer, ranging from 15 to 50 years.
25 A "normal good" is one for which consumer demand increases as income increases.
26 EPA updated the median household income to 2010 dollars using the Consumer Price Index.
27 "Income elasticity" is the degree to which WTP increases as income increases. For this analysis, EPA assumed that for
every 1 percent increase in income, there is a 0.4 percent increase in VSL.
April 19, 2013 3^T
-------
Benefit and Cost Analysis for Proposed ELGs
3: Human Health Benefits
Table 3-2. Benefits from Reduced Cancer Cases
Scenario
Baseline
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Affected
Population
137,476
137,476
137,476
137,476
137,476
137,476
137,476
137,476
137,476
Total Cancer Cases,
2017 to 2040
0.9
0.9
a
0.9
a
0.5
b
0.2
0.2
Reduced Cancer
Cases, 2017 to 2040
NA
0.0
a
0.0
a
0.4
b
0.7
0.7
Annualized Benefits
(Millions; 2010$)
3% Discount
Rate
NA
$0.00
a
$0.00
a
$0.09
b
$0.15
$0.16
7% Discount
Rate
NA
$0.00
a
$0.00
a
$0.05
b
$0.09
$0.09
a. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Options 3a and 3b to be less
than those of Option 3, with values for Option 3a being lower than Option 3b. EPA does not know how the benefits of Options 3a
and 3b compare to those of Options 1 and 2.
b. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Option 4a to be between those
of Options 3 and 4.
Source: U.S. EPA Analysis, 2013
EPA conducted the analysis of health related benefits only for the reaches that receive steam electric
discharges directly. The proposed ELGs are expected to provide additional benefits by also reducing arsenic
concentration in reaches downstream from steam electric plant discharges. To evaluate the potential
significance of these benefits, EPA also conducted a sensitivity analysis using fish tissue concentrations
modeled for downstream reaches based on a simple pollutant dilution model. Appendix C presents the results
of this sensitivity analysis.
3.1.3 Revised Cancer Slope Factor
The Integrated Risk Information System (IRIS) reports a cancer slope factor (CSF) of 1.5 cases per mg/kg
BW/day, which is based on incidences of skin cancer. EPA applied the 1.5 cases per mg/kg BW/day CSF to
estimate the benefits shown in Table 3-2. EPA is currently revising its cancer assessment of arsenic to reflect
new data on internal cancers including bladder and lung cancers associated with arsenic exposure via oral
ingestion. The draft CSF is substantially higher - at 25.7 per mg/kg BW/day for women and 16.9 per mg/kg
BW/day for men (U.S. EPA, 2010c). Higher CSF values would lead to higher estimated reduction in the
number of cancer cases (lunch and bladder combined) and higher benefit estimates for the proposed ELG.
Exposure to lead can cause a variety of adverse health effects in adults and children (see U.S. EPA, 2013a).
Because of data limitations, EPA estimated only the benefits to pre-school (ages 0 to 7) children from reduced
lead exposure via consumption of contaminated fish tissue.
3.2.1 Methodology and Data
This section describes the methodology and data used to estimate benefits to pre-school children potentially
exposed to lead from consumption of fish caught in the reaches receiving steam electric discharges and that
suffer learning disabilities as a result of their exposure.
April 19, 2013
3-6
-------
Benefit and Cost Analysis for Proposed ELGs 3: Human Health Benefits
EPA estimated benefits from reduced exposure to lead to preschool children using blood lead concentration
(PbB) as a biomarker of lead exposure. EPA first modeled PbB under the baseline and post-compliance
scenarios, and then used a dose response relationship between PbB and IQ loss to estimate avoided IQ losses
in the affected population of children and reduced incidences of extremely low IQ scores (less than 70, or two
standard deviations below the mean). EPA calculated the monetary value of benefits to children based on the
impact of an additional IQ point on an individual's future earnings and the cost of compensatory education for
children with learning disabilities (including children with IQ less than 70 and PbB levels above 20 |o,g/dL).
3.2.1.1 Affected Populations
EPA used the methodology described in Section 3.1.1 to estimate the population of pre-school children who
live in recreational angler and subsistence fisher households and are potentially exposed to lead via
consumption of contaminated fish tissue. Since this benefit category applies to children up to the seventh
birthday only, EPA restricted the analysis to the relevant age cohorts of recreational anglers and subsistence
fishers household members (Table 3-1). EPA subdivided the age 6-10 cohort based on the age six percent of
the population (U.S. Census Bureau, 201 Ob), thereby restricting the analysis to the portion of the population
cohort under age 7.28
3.2.1.2 IQ Losses due to Lead Exposure
This analysis considers children who are born after implementation of the proposed ELGs and live in
recreational angler and subsistence fisher households. It relies on EPA's Integrated Exposure, Uptake, and
Biokinetics (IEUBK) Model for Lead in Children (U.S. EPA, 2009c), which uses lead concentrations in a
variety of media - including soil, dust, air, water, and diet - to estimate total exposure to lead for children in
seven one-year age cohorts from birth through the seventh birthday. Based on this total exposure, the model
generates a predicted geometric mean PbB for a population of children exposed to similar lead levels.
Appendix B provides for a more detailed description of the IEUBK model and describes EPA's application of
the model to estimating benefits to pre-school children from reduced exposure to lead contaminated fish.
For each receiving reach, EPA used the cohort-specific LADD provided in the EA (U.S. EPA, 2013a). Lead
bioavailability and uptake after consumption varies for different chemical forms. Many factors complicate the
estimation of bioavailability, including nutritional status and timing of meals relative to lead intake. For this
analysis, EPA used the default media-specific bioavailability factor provided in the IEUBK model, which is
50 percent for oral ingestion. EPA used the IEUBK model to generate the geometric mean PbB for each
cohort and each receiving reach under the baseline and post-compliance scenarios.
A linear dose-response relationship between PbB and IQ losses in the study by Schwartz (1994) suggests that
a decrease of 0.25 IQ points can be expected for every 1 ng/dL increase in PbB. More recent studies have
calculated a steeper slope factor for lead effects on cognitive abilities (i.e. the effect is greater) at lower PbB
levels. Therefore, using a linear dose-response function may underestimate lead impacts at lower PbB levels.
For example, in a pooled data analysis, Lanphear et al. (2005; as cited in ATSDR, 2007) found that the
greatest IQ losses per 1 ng/dL occur at the lowest ranges of PbB. When the authors grouped IQ losses data for
children with PbB below and above 7.5 ng/dL, they found that the IQ losses were 2.94 points per ng/dL for
children with PbB concentrations below 7.5 |o,g/dL and 0.16 points per |o,g/d for children with PbB
concentrations above7.5 |o,g/dL.
28 U.S. Census Bureau (2010b) indicates that 1.3 percent of the population is 6 years old.
April 19, 2013
-------
Benefit and Cost Analysis for Proposed ELGs 3: Human Health Benefits
The IEUBK model results show that the estimated mean PbB for children exposed to lead from fish
consumption in the baseline is approximately 2.7 (ig/dL.29 EPA therefore used the dose-response factor of
2.94 points per |o,g/dL from Lanphear et al. to estimate IQ losses. For each analyzed regulatory option, EPA
multiplied the estimated average decrease in PbB under the post-compliance scenario by 2.94 IQ points lost
(per ng/dL increase in PbB). This calculation provides the avoided IQ loss per child. Multiplying the result by
the number of affected pre-school children yields the total increase in the number of IQ points for the affected
population of children for a given regulatory option.
The IEUBK model estimates the mean of the PbB distribution in children, assuming a continuous exposure
pattern for children from birth through the seventh birthday. The Census Bureau (201 Ob) indicates that
children ages 0 to 7 are approximately evenly distributed by age. To get an annual estimate of the number of
children that would benefit from implementation of the proposed ELGs, EPA divided the estimated number of
affected pre-school children by 7. This division adjusts the equation to apply only to children age 0 to 1. The
estimated avoided IQ loss is thus an annual value (i.e., it would apply to the cohort of children born each year
after implementation).30 Equation 3-3 shows this calculation for the annual increase in total IQ points.
Equation 3-3. A/ = (ACM * 2. 94 *
A/<2 = the difference in total IQ points between the baseline and regulatory option scenarios
AGM= the change in the average PbB in affected population of children ((ig/dL)
2.94 = slope of the dose response function (IQ points lost per ug/dL increase in PbB)
ExCh = the number of affected children aged 0 to 7
The available economic literature provides little empirical data on society's overall WTP to avoid a decrease
in children's IQ. To determine the value of avoided IQ losses, EPA used estimates of the changes in a child's
future expected lifetime earnings per one IQ point reduction and the cost of additional education.
Salkever (1995) and Schwartz (1994) estimate that a one point IQ reduction reduces expected lifetime
earnings by 2.38 percent and 1.76 percent, respectively. Data from the U.S. Census Bureau for 2009 indicate
that lifetime earnings are approximately $138,030 when discounting future earnings at 7 percent, and
$614,729 when discounting future earnings at 3 percent. 31 The resulting estimated values of an IQ point are
summarized in Table 3-3.
Table 3-3. Value of an IQ Point3 (2010$)
Discount Rate
3 percent
7 percent
Assumed Reduction in Expected Lifetime Earnings
(percent per IQ point)
1.76 percent/IQ point
(Schwartz, 1994 )
$10,819
$2,429
2.38 percent/IQ point
(Salkever, 1995)
$14,631
$3,285
a. Values are not adjusted for the cost of education.
29 The value is based on IEUBK outputs.
30 Dividing by seven undercounts overall benefits. Children from ages 1 to 7 are not accounted for in the base year of the
analysis, although they are presumably affected by lead exposure.
31 EPA updated lifetime earnings to 2010 dollars using the Consumer Price Index (2010 = 218.056; 2009 = 214.537).
April 19, 2013 34
-------
Benefit and Cost Analysis for Proposed ELGs 3: Human Health Benefits
Decreased IQ also results in less education and, therefore, reduced education costs. Chambers (2004)
estimates that the annual cost of education for an average student is $9,724.32 Schwartz (1994) and Salkever
(1995) estimate that a one IQ point reduction results in 0.131 and 0.101 fewer education years, respectively;
this represents lifetime cost savings of $979 to $1,274. EPA subtracted these education costs from the value
of lifetime earnings per IQ point in Table 3-3.
The value of an IQ point reduction adjusted for the avoided cost of education ranges between $1,156
(following Schwarz (1994) and discounting future earnings at 7 percent) to $13,651 (following Salkever
(1995) and discounting future earnings at 3 percent). This effect represents only one component of society's
WTP to avoid IQ decreases, and thus underestimates the total value of benefits to children from reduced
exposure to lead.
3.2.1.3 Reduced Expenditures on Compensatory Education
Children whose PbB exceeds 20 ng/dL are more likely to have IQs less than 70, which means that they would
require compensatory education tailored to their specific needs. EPA's IEUBK model can generate
probabilities that a child would have a PbB in excess of a specific threshold.33
EPA estimated the number of children that would have PbB above 20 |o,g/dL for each receiving reach under
the baseline and each analyzed regulatory option. EPA assumed that 20 percent of children with PbB above
20 ug/dL would have IQs less than 70 and require compensatory education.34 Equation 3-4 shows the
calculation of the number of children requiring compensatory education for each receiving reach. Summing
across all receiving reaches provides the total number of children who would require special education.
Equation 3-4. CompEd(i) = ExCh(i) * Pr20(i) * 0. 20
CompEd(/) = the number of children with PbB over 20 |o,g/dL and IQ less than 70 (who
would need compensatory education) for reach /
ExCh(/) = the number of affected children for reach /
Pr20(/) = the probability that a child's PbB is above 20 |o,g/dL for reach /
0.20 = 20% (share of children with PbB over 20 |o,g/dL that would have IQ scores less than
70)
The U.S. Department of Education (Chambers, et al., 2003) estimated that average annual expenditures for a
student with mental retardation are approximately $8,484 higher than for an average student. Updating to
2010 dollars35 yields annual costs of $15,805 per child. EPA assumed that children with IQ less than 70
would incur these additional costs each year from age 7 to age 18. Discounting future costs using a 3 percent
discount rate yields a total compensatory education cost of approximately $157,327 per child with an IQ score
less than 70 ($125,537 per child if using a 7 percent discount rate).
32 Estimates of the cost of education were updated to 2010 dollars using the Consumer Price Index for education (2010 =
199.337; 2003 = 134.4)
33 See Appendix B for a detailed description of the IEUBK model and its application in this analysis.
34 This assumption follows the methodology used by EPA in Economic and Environmental Benefits Analysis Document
for the Final Effluent Limitations Guidelines and Standards for the Metal Products and Machinery Point Source
Category (U.S. EPA, 2003).
35 All values were updated to 2010 dollars using the Consumer Price Index for Education (2010 = 199.337; 1999 =
107.0).
April 19, 2013 3^9~
-------
Benefit and Cost Analysis for Proposed ELGs
3: Human Health Benefits
3.2.2 Results
EPA estimates that the annual benefits of avoided IQ losses range between $2.2 and $3.2 million at a
3 percent discount rate for Option 3, and $0.2 to $0.3 million at a 7 percent discount rate (Table 3-4). For
Option 4, these benefits range between $5.6 and $7.9 million at a 3 percent discount rate ($0.4 to $0.8 million
at a 7 percent discount rate).
EPA estimates that the preferred options for existing sources would result in annual cost savings from reduced
compensatory education requirements due to the decreased number of children with PbB above 20 ug/dL and
IQs less than 70 valued at $0.02 million and $0.07 million annually, respectively for Options 3 and 4, at a
3 percent discount rate ($0.01 million and $0.03 million annually at a 7 percent discount rate) (see Table 3-5).
As discussed in Section 3.1.2, EPA conducted the analysis of health related benefits only for the reaches that
receive steam electric discharges directly. The proposed ELGs are expected to provide additional benefits by
also reducing lead concentration in reaches downstream from steam electric plant discharges. To evaluate the
potential significance of these benefits, EPA also conducted a sensitivity analysis using fish tissue
concentrations modeled for downstream reaches based on a simple pollutant dilution model. Appendix C
presents the results of this sensitivity analysis.
Table 3-4. Estimated Benefits from Avoided IQ Losses for Children Ages 0 to 7
Scenario
Baseline
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Number of
Affected
Children 0 to 7
12,478
12,478
12,478
12,478
12,478
12,478
12,478
12,478
12,478
Total Avoided
IQ Losses, 2017
to 2040
NA
159.9
b
159.9
b
6,386.6
c
15,837.6
15,836.1
Annualized Value of Avoided IQ Point Losses a
(Millions; 2010$)
3% Discount Rate
Low Bound
High Bound
NA
$0.06
b
$0.06
b
$2.21
c
$5.55
$5.55
$0.08
b
$0.08
b
$3.17
c
$7.94
$7.94
7% Discount Rate
Low Bound
High Bound
NA
$0.00
b
$0.00
b
$0.16
c
$0.40
$0.40
$0.01
b
$0.01
b
$0.31
c
$0.80
$0.80
a. Low bound assumes that the loss of one IQ point results in the loss of 1.76% of lifetime earnings (following Schwartz, 1994);
high bound assumes that the loss of one IQ point results in the loss of 2.38% of lifetime earnings (following Salkever, 1995).
b. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Options 3a and 3b to be less
than those of Option 3, with values for Option 3a being lower than Option 3b. EPA does not know how the benefits of Options 3a
and 3b compare to those of Options 1 and 2.
c. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Option 4a to be between those
of Options 3 and 4.
Source: U.S. EPA Analysis, 2013
Table 3-5. Estimated Avoided Cost of Compensatory Education for Children with Blood Lead
Concentrations above 20 |ag/dl_ and IQ Less than 70
Scenario
Baseline
Option 1
Option 3 a
Number of
Affected
Children
Oto7
12,478
12,478
12,478
Number of Cases
of PbB >
20 ug/dL and IQ
< 70, in 2017 to
2040
15.1
15.1
a
Decrease in
Number of Cases
oflQ < 70, in
2017 to 2040
NA
0.0
a
Avoided Annual Cost
(Millions; 2010$)
3% Discount
Rate
NA
$0.00
a
7% Discount
Rate
NA
$0.00
a
April 19, 2013
3-10
-------
Benefit and Cost Analysis for Proposed ELGs
3: Human Health Benefits
Table 3-5. Estimated Avoided Cost of Compensatory Education for Children with Blood Lead
Concentrations above 20 |ag/dl_ and IQ Less than 70
Scenario
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Number of
Affected
Children
Oto7
12,478
12,478
12,478
12,478
12,478
12,478
Number of Cases
ofPbB>
20 ug/dL and IQ
< 70, in 2017 to
2040
15.1
a
11.1
b
2.6
2.6
Decrease in
Number of Cases
oflQ < 70, in
2017 to 2040
0.0
a
4.0
b
12.5
12.5
Avoided Annual Cost
(Millions; 2010$)
3% Discount
Rate
$0.00
a
$0.02
b
$0.07
$0.07
7% Discount
Rate
$0.00
a
$0.01
b
$0.03
$0.03
PbB = blood lead concentration.
a. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Options 3a and 3b to be less
than those of Option 3, with values for Option 3a being lower than Option 3b. EPA does not know how the benefits of Options 3a
and 3b compare to those of Options 1 and 2.
b. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Option 4a to be between those
of Options 3 and 4.
Source: U.S. EPA Analysis, 2013
3.3 Benefits to Infants from Reduced Exposure to Mercury
Mercury can have a variety of adverse health effects on adults and children (see U.S. EPA, 2013a). The
proposed ELGs are expected to reduce the discharge of mercury to surface waters by steam electric plants and
therefore provide a range of human health benefits. Due to data limitations, however, EPA estimated only the
benefits from reduced IQ losses among children exposed to mercury in-utero as a result of maternal
consumption of contaminated fish.
3.3.1 Methodology and Data
This section describes the methodology and data used to 1) identify the number of children in angler and
substance fisher households who are likely to have in-utero exposure to mercury; 2) estimate IQ losses in the
affected children under the baseline and each analyzed regulatory option; and 3) estimate the monetary value
of benefits from reducing IQ losses among affected children.
3.3.1.1 Affected Population
EPA identified the population of children exposed in-utero starting from the reach-specific affected
population described in Section 3.1.1 (i.e., based on the number of recreational and subsistence fishers and
adjusted to reflect fishing practices (advisories, catch and release), the availability of substitute locations,
household size, and age cohorts).
Because this analysis focuses only on infants born after implementation of the proposed ELGs, EPA further
limited the affected population by estimating the number of women between the ages of 15 and 50 potentially
exposed to contaminated fish caught in the waterbodies affected by mercury discharges from steam electric
plants, and multiplying the result by the state-specific average fertility rate.36 This yields the annual number of
births in the affected population. EPA acknowledges that fertility rates vary by age. However, the use of a
36 The average state-specific fertility rates range from 10.7 to 20.6 babies per year per 1,000 people. See Exhibit A-5 in
Appendix A for details.
April 19, 2013
3-11
-------
Benefit and Cost Analysis for Proposed ELGs 3: Human Health Benefits
single average fertility rate for all ages is not expected to bias results because the average fertility rate reflects
the underlying distribution of fertility rates by age.
Two of the population cohorts utilized in the EA (U.S. EPA, 2013a) spanned ages 11 to 15 (one for
recreational and one for subsistence, Table 3-1). EPA used the percentage of the population aged 11 to 15 that
is aged 15 (U.S. Census Bureau, 2010b) to subdivide the two cohorts and estimate the number of 15-year-olds
to be included as mothers to affected infants.37
3.3.1.2 IQ Losses due to in Utero Mercury Exposure
In this analysis, EPA used a linear dose-response relationship between maternal mercury hair content and
subsequent childhood IQ loss from Axelrad et al. (2007). Axelrad et al. (2007) developed a dose-response
function based on data from three epidemiological studies in the Faroe Islands, New Zealand, and Seychelle
Islands. According to their results, there is a 0.18 point IQ loss for each 1 part-per-million (ppm) increase in
maternal hair mercury.
To estimate maternal hair mercury concentrations based on the daily intake (see EA for details; U.S. EPA,
2013a), EPA used the median conversion factor derived by Swartout and Rice (2000), who estimated that a a
0.08 (ig/kg body weight increase in daily mercury dose is associated with a 1 ppm increase in hair
concentration. Equation 3-5 shows EPA's calculation of the total annual IQ decrement for a given receiving
reach.
Equation 3-5. '
-------
Benefit and Cost Analysis for Proposed ELGs
3: Human Health Benefits
values of an IQ point ranges from $1,156 (following Schwarz (1994) and discounting future earnings at
7 percent) to $13,651 (following Salkever (1995) and discounting future earnings at 3 percent). This effect
represents only one component of society's WTP to avoid IQ losses, and thus underestimates the total value
of benefits to children from reduced exposure to mercury.
3.3.2 Results
Table 3-6 shows the estimated value of avoided IQ losses due to in-utero mercury exposure. The estimated
annual benefits of avoided IQ losses from reduced maternal exposure to mercury under Option 3 range
between $4.1 million and $5.8 million at a 3 percent discount rate (between $0.3 million and $0.6 million at a
7 percent discount rate). These benefits accrue over the period of 2017 through 2040. Benefits for Option 4
are about 1.8 times higher than for Option 3, with annualized value of avoided IQ losses ranging between
$8.4 million and $12.1 million at a 3 percent discount rate ($0.6 million to $1.2 million at a 7 percent discount
rate).
As discussed above, EPA conducted the analysis of health related benefits only for the reaches that receive
steam electric discharges directly. The proposed ELGs may reasonably be expected to provide additional
benefits by also reducing mercury concentration in reaches downstream from steam electric plant discharges.
To evaluate the potential significance of these benefits, EPA also conducted a sensitivity analysis using fish
modeled tissue concentrations for downstream reaches based on a simple pollutant dilution model. Appendix
C presents the results of this sensitivity analysis.
Table 3-6. Estimated Benefits from Avoided IQ Losses Due to Reduced In-utero Mercury Exposure
Scenario
Baseline
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Number of
Births in
Affected
Population (per
year)
,932
,932
,932
,932
,932
,932
,932
,932
,932
Total Avoided
IQ Losses, 2017
to 2040
NA
8,895
b
9,096
b
11,547
c
23,970
24,004
Annualized Value of Avoided IQ Losses"
(Millions; 2010$)
3% Discount Rate
Low Bound
High
Bound
NA
$3.15
b
$3.22
b
$4.08
c
$8.42
$8.43
$4.51
b
$4.61
b
$5.83
c
$12.05
$12.06
7% Discount Rate
Low Bound
High
Bound
NA
$0.23
b
$0.24
b
$0.30
c
$0.61
$0.61
$0.46
b
$0.47
b
$0.59
c
$1.21
$1.22
a. Low bound estimate assumes that the loss of one IQ point results in the loss of 1.76 percent of lifetime earnings (following
Schwartz, 1994); high bound estimate assumes that the loss of one IQ point results in the loss of 2.38 percent of lifetime earnings
(following Salkever, 1995).
b. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Options 3a and 3b to be less
than those of Option 3, with values for Option 3a being lower than Option 3b. EPA does not know how the benefits of Options 3a
and 3b compare to those of Options 1 and 2.
c. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Option 4a to be between those
of Options 3 and 4.
Source: U.S. EPA Analysis, 2013
nefits to Subsistence Fishers
As noted in Section 3.1.1, EPA assumed for this analysis that 5 percent of the exposed population is
subsistence fishers, and that the remaining 95 percent is recreational anglers. This is based on the assumed
95th percentile fish consumption rate for subsistence fishers. These individuals consume more self-caught fish
April 19, 2013
3-13
-------
Benefit and Cost Analysis for Proposed ELGs 3: Human Health Benefits
than recreational anglers and would therefore experience higher health risks associated with the consumption
of contaminated fish tissue.
Table 3-7 shows the annual human health benefit estimates for two analyzed options - Options 3 and 4 -
disaggregated into those accruing to recreational anglers and to subsistence fishers. Although subsistence
fishers account for 5 percent of the total exposed population, they account for 18 percent to 50 percent of the
total benefits, depending on the regulatory option and benefit category.
April 19, 2013 3-14
-------
Benefit and Cost Analysis for Proposed ELG
3: Human Health Benefits
Table 3-7. Estimated Annualized Health Benefits to Recreational Anglers and Subsistence Fishers (Millions; 2010 $)
ELG
Regulatory
Option
Option 3
Option 4
Discount
Rate
3 percent
7 percent
3 percent
7 percent
Benefit Category
Avoided Cancer Cases from
Exposure to Arsenic
Avoided IQ Losses from
Exposure to Lead
Avoided Compensatory
Education from Exposure to Lead
Avoided IQ Losses from in-Utero
Exposure to Mercury
Avoided Cancer Cases from
Exposure to Arsenic
Avoided IQ Losses from
Exposure to Lead
Avoided Compensatory
Education from Exposure to Lead
Avoided IQ Losses from in-Utero
Exposure to Mercury
Avoided Cancer Cases from
Exposure to Arsenic
Avoided IQ Losses from
Exposure to Lead
Avoided Compensatory
Education from Exposure to Lead
Avoided IQ Losses from in-Utero
Exposure to Mercury
Avoided Cancer Cases from
Exposure to Arsenic
Avoided IQ Losses from
Exposure to Lead
Avoided Compensatory
Education from Exposure to Lead
Avoided IQ Losses from in-Utero
Exposure to Mercury
Recreational Anglers
Annual Benefits
Low
High
$0.08
$1.79
$2.56
$0.01
$3.28
$4.70
$0.04
$0.13
$0.25
$0.01
$0.24
$0.48
$0.13
$4.53
$6.48
$0.05
$6.78
$9.70
$0.07
$0.33
$0.65
$0.02
$0.49
$0.98
% of Total
81%
81%
50%
81%
81%
81%
50%
81%
50%
82%
70%
81%
50%
82%
71%
81%
Subsistence Fishers
Annual Benefits
Low
High
$0.02
$0.42
$0.60
$0.01
$0.79
$1.14
$0.01
$0.03
$0.06
$0.01
$0.06
$0.12
$0.13
$1.02
$1.46
$0.02
$1.64
$2.35
$0.07
$0.07
$0.15
$0.01
$0.12
$0.24
% of Total
19%
19%
50%
19%
19%
19%
50%
19%
50%
18%
30%
19%
50%
18%
29%
19%
Total Exposed
Population
Annual Benefits
Low
High
$0.09
$2.21
$3.17
$0.02
$4.08
$5.83
$0.05
$0.16
$0.31
$0.01
$0.30
$0.59
$0.25
$5.55
$7.94
$0.07
$8.42
$12.05
$0.14
$0.40
$0.80
$0.03
$0.61
$1.21
April 19, 2013
3-15
-------
Benefit and Cost Analysis for Proposed ELG
3: Human Health Benefits
3.5 Potential Additional Health Benefits
EPA expects that additional health benefits would arise from reduced exposure to pollutants in steam electric
plant discharges; however, monetary valuation of these other health benefits is not possible due to lack of data
on a dose-response relationship between pollutant ingestion rate and potential adverse health effects. To
provide an additional measure of the potential health benefits of the proposed ELGs, EPA also estimated the
expected reduction in the number of receiving reaches with pollutant concentrations in excess of human
health-based aquatic water quality criteria (AWQC). This analysis and its findings are not additive to the
preceding analyses of changes in cancer rates, lead-related effects, and mercury-related effects, but represent
another way of characterizing potential health benefits resulting from reduced exposure to steam electric
pollutants.
This analysis compares in-stream pollutant concentrations estimated for the baseline and each analyzed
regulatory option in the receiving reaches (see the EA; U.S. EPA, 2013a) to criteria established by EPA for
protection of human health. EPA compared in-water concentrations of arsenic, copper, nickel, selenium,
thallium, and zinc to EPA's national recommended water quality criteria protective of human health used by
states and tribes (U.S. EPA, 2012b). Pollutant concentrations in excess of these values indicate potential risks
to human health. For another four steam electric pollutants (cadmium, chromium, lead, and mercury) for
which there are no recommended criteria, EPA instead compared concentrations to MCLs (U.S. EPA, 2012a).
The analysis was performed on reaches that receive discharges from steam electric plants directly (i.e.,
excludes reaches located downstream of steam electric plants and that would also have changes in pollutant
concentrations under the proposed ELGs).
Table 3-8 shows the results of this analysis. EPA estimates that in-stream concentrations of steam electric
pollutants exceed human health criteria or MCLs for at least one pollutant in 146 receiving reaches
nationwide as the result of baseline steam electric pollutant discharges. EPA expects that Option 3 would
eliminate the occurrence of concentrations in excess of human health-based criteria for 27 of the receiving
reaches and reduce the number of exceedances in another 27 reaches. Options 3a and 3b would have effects
that will be smaller or equal to those of Option 3. Option 4 would reduce the number of exceedances in 24
reaches and eliminate them altogether in another 98 reaches. The Agency expects that Option 4a will reduce
the number of exceedances between Options 3 and 4, i.e., eliminate all exceedances in 27 to 98 reaches, and
reduce exceedances in 24 to 27 reaches.
Table 3-8. Receiving Reaches Exceeding Human Health Criteria for Steam Electric Pollutants
Scenario
Baseline
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Number of Receiving Reaches
with Steam Electric Pollutant
Concentrations Exceeding
Human Health Criteria for at
Least One Pollutant
146
146
a
146
a
119
b
48
50
Number of Receiving Reaches with Improved Water
Quality, Relative to Baseline
All Exceedances
Eliminated
NA
0
a
0
a
27
b
98
96
Number of Exceedances
Reduced
NA
9
a
12
a
27
b
24
21
April 19, 2013
3-16
-------
Benefit and Cost Analysis for Proposed ELG
3: Human Health Benefits
Table 3-8. Receiving Reaches Exceeding Human Health Criteria for Steam Electric Pollutants
Scenario
Number of Receiving Reaches
with Steam Electric Pollutant
Concentrations Exceeding
Human Health Criteria for at
Least One Pollutant
Number of Receiving Reaches with Improved Water
Quality, Relative to Baseline
All Exceedances
Eliminated
Number of Exceedances
Reduced
a. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Options 3a and 3b to be less
than those of Option 3, with values for Option 3a being lower than Option 3b. EPA does not know how the benefits of Options 3a
and 3b compare to those of Options 1 and 2.
b. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Option 4a to be between those
of Options 3 and 4.
Source: U.S. EPA Analysis, 2013
3.6 Limitations and Uncertainties
This analysis does not include all possible human health benefits associated with post-compliance reductions
in pollutant discharges due to lack of data on a dose-response relationship between ingestion rates and
potential adverse health effects. Therefore, the total quantified human health benefits included in this analysis
represent only a subset of the potential health benefits expected to result from the proposed ELGs.
Additionally, the methodologies and data used in the analysis of health benefits associated with reduced
incidences of adverse health outcomes due to consumption offish contaminated with steam electric pollutants
involve limitations and uncertainties. Table 3-9 summarizes the uncertainties and indicates the direction of the
potential bias.
Table 3-9. Uncertainties in the Analysis of Human Health Benefits
Uncertainty/Assumption
Background concentrations of
steam electric pollutants from
upstream sources, contaminated
sediments from previous
discharges, or natural sources, are
not considered in this analysis.3
The analysis of health benefits
does not include risks to
recreational anglers and
subsistence fishers downstream
from receiving reaches.3
The analysis does not consider the
suitability of alternate fishing sites.
The number of subsistence fishers
was assumed to equal 5 percent of
the total number of anglers fishing
the affected reaches.
Effect on Benefits
Estimate
Uncertain
Underestimate
Uncertain
Uncertain
Notes
Even if discharges of the pollutants are reduced or
eliminated as a result of the proposed ELGs, background
concentrations may persist for years. This may result in
an over- or underestimate of the number of reaches for
which the proposed ELGs would reduce the number of
health-based water concentration criteria exceedances.
By omitting downstream effects, this analysis potentially
understates baseline risks that could be reduced by the
proposed ELGs, and therefore underestimates the
benefits.
Estimating the number of anglers fishing receiving
reaches based on the ratio of reach length to the total
number of reach miles within the same 100-mile buffer
area recognizes the effects of the quantity of competing
fishing opportunities on the likelihood of fishing a given
reach, but does not account for the differential quality of
fishing sites. If the quality of substitute sites is distinctly
worse or better (e.g., some sites have better access or
designated fishing areas), the estimated affected
populations are likely to be overstated or understated.
The magnitude of subsistence fishing in the United States
or individual states is not known. Assuming 5 percent
may understate or overstate the number of potentially
affected subsistence fishers (and their households).
April 19, 2013
3-17
-------
Benefit and Cost Analysis for Proposed ELG
3: Human Health Benefits
Table 3-9. Uncertainties in the Analysis of Human Health Benefits
Uncertainty/Assumption
EPA used a CSF for arsenic of 1.5
cases per mg/kg BW/day.
There is a linear 0.18 point IQ loss
for each 1 ppm increase in
maternal hair mercury.
For the mercury- and lead-related
benefits analyses, EPA assumed
that IQ losses are an appropriate
endpoint for quantifying adverse
cognitive and neurological effects
resulting from childhood or in-
utero exposures to lead and
mercury (respectively).
The IEUBK model processes daily
intake from "alternative sources"
to 3 decimal places (ug/day).
EPA did not quantify the benefits
associated with reduced adult
exposure to lead and mercury.
Effect on Benefits
Estimate
Uncertain
Uncertain
Underestimate
Underestimate
Underestimate
Notes
This is the current IRIS value and was based on
incidences of skin cancer. However, EPA is currently
revising its cancer assessment of arsenic to reflect new
data on internal cancers. It is possible that the revised
combined (lung and bladder cancer) CSF would be higher
(draft value is 25.7 per mg/kg BW/day), suggesting that
the use of the current IRIS value may result in an
underestimate of benefits.
This dose-response function used in this analysis may
over- or underestimate IQ impacts arising from mercury
exposure if a linear function is not the best representation
of the relationship between maternal body burden and IQ
losses.
IQ may not be the most sensitive endpoint. Additionally,
there are deficits in cognitive abilities that are not
reflected in IQ scores, including acquisition and retention
of information presented verbally and many motor skills
(U.S. EPA, 2005). To the extent that these impacts create
disadvantages for children exposed to mercury at current
exposure levels or result in the absence of (or
independent from) measurable IQ losses, this analysis
may underestimate the benefits of the proposed ELG of
reduced lead and mercury exposure.
Since the intakes are very small, some variation is missed
by using the model (i.e., it does not capture very small
changes).
The scientific literature suggests that exposure to lead and
mercury may have significant adverse health effects for
adults; if measurable effects at occurring at current
exposure levels, excluding the benefits of reduced adult
exposure to these pollutants results in an underestimate of
benefits.
a. EPA is currently reviewing and revising
downstream reaches in a future revision to
its modeling methodology and plans to consider background concentrations and
this analysis.
April 19, 2013
3-18
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
4 Non-Market Benefits from Water Quality Improvements
As discussed in the environmental assessment (EA) document (U.S. EPA, 2013a), heavy metals, nutrients,
and other pollutants discharged by steam electric plants have a wide range of effects on water resources
located in the vicinity and downstream from the plants. These environmental changes affect environmental
goods and services valued by humans, including recreation; commercial fishing; public and private property
ownership; navigation; water supply and use; and existence services such as aquatic life, wildlife, and habitat
designated uses. Some environmental goods and services (e.g., commercially caught fish) are traded in
markets, thus their value can be directly observed. Other environmental goods and services (e.g., recreation
and support of aquatic life) that cannot be bought or sold directly and thus don't have observable market
values. These second types of environmental goods and services are classified as "nonmarket". The expected
changes in nonmarket values of the water resources affected by the proposed ELG (hereafter nonmarket
benefits)) are additive to the market benefits (e.g., avoided costs of producing various market goods and
services) and benefits from improved groundwater quality estimated in other chapters (Freeman, 2003).
EPA's approach to estimating the nonmarket benefits from water quality improvements resulting from the
proposed ELGs involves 1) characterizing baseline and post-compliance water quality using a water quality
index and 2) monetizing changes in the nonmarket value of affected water resources attributable to the
proposed ELGs using a meta-analysis of surface water valuation studies that provide data on the public's
willingness-to-pay (WTP) for water quality improvements.38 The analysis accounts for improvements in water
quality resulting from changes in nutrient and metals concentrations in reaches affected by discharges from
steam electric plants.
4.1 Water Quality
To link water quality changes from reduced metal and nutrient discharges to effects on human uses and
support for aquatic and terrestrial species habitat, EPA used a water quality index (WQI) which translates
water quality measurements, gathered for multiple parameters that are indicative of various aspects of water
quality, into a single numerical indicator.
The WQI provides the link between specific pollutant levels, as reflected in individual index parameters (e.g.,
dissolved oxygen (DO) concentrations), and the presence of aquatic species and suitability for particular uses.
The WQI value, which is measured on a scale from 0 to 100, reflects varying water quality, with 0 for poor
quality and 100 for excellent.
The WQI used in this analysis modifies the WQI used by EPA in the Environmental and Economic Benefits
Assessment for the Final Construction and Development Rule (also referred to as the C&D rule; U.S. EPA,
2009b), which builds on McClelland (1974) and on the methodology developed by Dunnette (1979) and
subsequently updated by Cude (2001) to better account for spatial and morphologic variability in the natural
characteristics of streams. A more detailed discussion of the history of the WQI framework is found in
Chapter 10 of the C&D report (U.S. EPA, 2009b). To account for reductions in loadings of heavy metals
The meta-analysis of surface water valuation studies for estimating benefits of water quality improvements resulting
from the regulation follows EPA's approach in the analysis of Construction and Development (C&D) regulations (U.S.
EPA, 2009b). The technical details involved in the estimation of meta-analyses are also presented in sources such as
Bateman and Jones (2003), Johnston et al. (2005, 2006), Shrestha et al. (2007), and Rosenberger and Phipps (2007). EPA
is presenting the meta-analysis approach in this analysis but there are other approaches that could be used to monetize
the benefits.
April 19, 2013 4T
-------
Benefit and Cost Analysis for Proposed ELGs 4: Benefits from Water Quality Improvements
resulting from the proposed ELGs, EPA modified the WQI for freshwater waterbodies39 from the C&D
analysis to include metals. This was done by incorporating elements of the WQI developed by the Canadian
Council of Ministers of the Environment (CCME) wherein the index values are calculated based on the scope,
frequency, and amplitude of exceedances of specified numeric thresholds (CCME, 2001). The WQI used in
this analysis, therefore, includes the six parameters of the WQI previously used for the Final Construction and
Development Rule - DO, biochemical oxygen demand (BOD), fecal coliform (FC), total nitrogen (TN), total
phosphorus (TP), and total suspended solids (TSS) - and one additional subindex for metals, for a total of
seven parameters.
4.1.1 WQI Calculation
Implementing the WQI methodology involves three key steps: 1) obtaining water quality levels for each of
seven parameters included in the WQI; 2) transforming parameter levels to subindex values expressed on a
common scale; and 3) aggregating the individual parameter subindices to obtain an overall WQI value that
reflects waterbody conditions across the seven parameters. These steps are repeated to calculate the WQI
value for the baseline (i.e., in the absence of the proposed ELGs), and for each analyzed regulatory option.
The first step in the implementation of the WQI involves obtaining water quality levels for each parameter,
and for each waterbody, under both baseline conditions and post-compliance conditions (see Sections 4.1.3
and 4.1.4). Some parameter levels are field measurements while others are modeled values.
The second step involves transforming the parameter measurements into subindex values that express water
quality conditions on a common scale of 0 to 100. EPA used the subindex transformation curves developed by
Dunnette (1979) and Cude (2001) for the Oregon WQI for BOD, DO, and FC. For TSS, TN, and TP
concentrations, EPA adapted the approach developed by Cude (2001) to account for the wide range of natural
or background nutrient and sediment concentrations that result from the variability in geologic and other
region-specific conditions, and to reflect the national context of the analysis. TSS, TN, and TP subindex
curves were developed for each Level III ecoregion (U.S. EPA, 2009b) using baseline TSS, TN, and TP
concentrations calculated in SPARROW at the E2RF1 reach level40'41'42 For each of the 85 Level III
ecoregions intersected by the E2RF1 reach network, EPA derived the transformation curves by assigning a
score of 100 to the 25th percentile of the reach-level TSS concentrations in the ecoregion (i.e., using the 25th
percentile as a proxy for "reference" concentrations), and a score of 70 to the median concentration. An
exponential equation was then fitted to the two concentration points following the approach used in Cude
(2001).
For this analysis, EPA also used a metals-specific subindex curve based on the number of Ambient Water
Quality Criteria (AWQC) exceedances for metals in each waterbody. National freshwater chronic AWQC
39 EPA analyzed changes to water quality resulting from the implementation of the proposed ELG on receiving
freshwater reaches. While steam electric plants also discharge to estuarine and coastal reaches, EPA did not estimate
benefits from reducing pollutant loadings to these water bodies due to the relatively small changes in concentrations
expected.
40 The SPARROW (SPAtially Referenced Regressions On Watershed attributes) model was developed by the United
States Geological Survey (USGS) for the regional interpretation of water-quality monitoring data. The model relates in-
stream water-quality measurements to spatially referenced characteristics of watersheds, including contaminant sources
and factors influencing terrestrial and aquatic transport. SPARROW empirically estimates the origin and fate of
contaminants in river networks and quantifies uncertainties in model predictions. More information on SPARROW can
be found at http://water.usgs.gOv/nawqa/sparrow/FAQs/faq.html#l
41 EPA's E2RF1 (Enhanced River File Version 2.0) is a digital stream networks used in SPARROW models. This dataset
extends over the continental United States and includes approximately 62,000 stream reaches.
42 The selected data exclude outlier TSS concentration values, defined as values that exceed the 95th percentile based on
the national population of all reaches.
April 19, 2013 4^
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
values are available for arsenic, cadmium, chromium, lead, mercury, nickel, selenium, and zinc. To develop
this subindex curve, EPA used an approach developed by the CCME (CCME 2001). The CCME water quality
index is based on three attributes of water quality that relate to water quality objectives: scope (number of
monitored parameters that exceed water quality standard or toxicological benchmark); frequency (number of
individual measurements that do not meet objectives, relative to the total number of measurements for the
time period of interest) and amplitude (i.e., amount by which measured values exceed the standards or
benchmarks). Following the CCME approach, EPA's metal subindex considers the number of parameters
with exceedances of the relevant water quality criterion. Unlike the CCME index, however, the metal
subindex does not differ depending on the frequency or amplitude. With regards to frequency, EPA modeled
long-term annual average concentrations in ambient water (see Section 4.1.2 and EA for details), and
therefore any exceedance of an AWQC may indicate that ambient concentrations exceed AWQCs most of the
time (assumed to be 100 percent of the time). EPA did not consider amplitude, because if the annual average
concentration exceeds the chronic AWQC then the water is impaired for that constituent and the level of
exceedance is of secondary concern. Using this approach, the subindex curve for metals assigns the lowest
subindex score of 0 to waters where exceedances are observed for all eight metals analyzed, and a maximum
score of 100 to waters where there are no exceedances.
Table 4-1 presents parameter-specific functions used for transforming water quality data into water quality
subindices for freshwater waterbodies for the six traditional pollutants. Table 4-2 presents the subindex values
for metals. The equation parameters for each of the 85 ecoregion-specific TSS, TN, and TP subindex curves
are provided in Appendix D.
Table 4-1. Freshwater Water Quality Subindices
Parameter
Concentrations
Concentration Unit
Subindex
Dissolved Oxygen (DO)
DO saturation 1,600
Lbs/100 mL
Lbs/100 mL
Lbs/100 mL
98
98 * exp(FC - 50) '
* -9.9178 E-4
10
Total Nitrogen (TN)a
TN
TN
TN
TN 100
Mg/L
Mg/L
Mg/L
10
a * exp(TN*b); where a and b are
ecoregion-specific values
100
Total Phosphorus (TP)
TP
TP
TP
TPS100
Mg/L
Mg/L
Mg/L
10
a * exp(TP*b); where a and b are
ecoregion-specific values
100
April 19, 2013
4-3
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
Table 4-1. Freshwater Water Quality Subindices
Parameter
Concentrations
Concentration Unit
Subindex
Total Suspended Solids (TSS)C
TSS
TSS
TSS
TSS100
Mg/L
Mg/L
Mg/L
10
a * exp(TSS*b); where a and b are
ecoregion-specific values
100
Biochemical Oxygen Demand, 5-day (BOD)
BOD
BOD
BOD<8
BOD>8
Mg/L
Mg/L
100 *exp(BOD* -0.1993)
10
a. TN10 and TN100 are ecoregion-specific TN concentration values that correspond to subindex scores of 10 and 100, respectively.
b. TP10 and TP100 are ecoregion-specific TP concentration values that correspond to subindex scores of 10 and 100, respectively.
c. TSS10 and TSS100 are ecoregion-specific TSS concentration values that correspond to subindex scores of 10 and 100, respectively.
Source: EPA analysis using methodology in Cude (2001) .
Table 4-2. Freshwater Water Quality Subindex for Heavy Metals
Parameter
Metals
Metals
Metals
Metals
Metals
Metals
Metals
Metals
Metals
Number of AWQC Exceedances
0
1
2
3
4
5
6
7
8
Subindex
100
87.5
75
62.5
50
37.5
25
12.5
0
The final step in implementing the WQI involves combining the individual parameter subindices into a single
WQI value that reflects the overall water quality across the parameters. EPA calculated the overall WQI for a
given reach using a geometric mean function and assigned all WQ parameters an equal weight of 0.143 (l/7th
of the overall score). Unweighted scores for individual metrics of a WQI have previously been used in Cude
(2001), CCME (2001), and Carruthers and Wazniak (2003).
Equation 4-1 presents EPA's calculation of the overall WQI score.
Equation 4-1.
WQIr =
Wi
n =
i=\
the multiplicative water quality index (from 0 to 100) for reach r
the water quality subindex measure for parameter /
the weight of the /-th parameter (0.143)
the number of parameters (i.e., seven)
Once an overall WQI value is calculated, it can be related to suitability for potential uses. Vaughan (1986)
developed a water quality ladder (WQL) that can be used to indicate whether water quality is suitable for
various human uses (i.e., boating, rough fishing, game fishing, swimming, and drinking without treatment).
Vaughan identified "minimally acceptable parameter concentration levels" for each of the five potential uses.
April 19, 2013
4-4
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
Vaughan used a scale of zero to 10 instead of the WQI scale of zero to 100 to classify water quality based on
its suitability for potential uses. Therefore, the WQI value corresponding to a given water quality use
classification equals the WQL value multiplied by 10. Table 4-3 presents water use classifications and the
corresponding WQL and WQI values.
Table 4-3. Water Quality Classifications
Water Quality Classification
. . . drinking without treatment
... swimming
. . . game fishing
... rough fishing
. . . boating
WQL Value
9.5
7.0
5.0
4.5
2.5
WQI Value
95
70
50
45
25
Source: Vaughan (1986)
4.1.2 Sources of Data on Ambient Water Quality
EPA used the following data sources to obtain ambient concentrations for the seven parameters included in
the WQI:
> Outputs from USGS's SPARROW models provided baseline and post-compliance concentrations of
total nitrogen and total phosphorus, and baseline concentrations of total suspended solids.43 See U.S.
EPA (2009b) for a discussion of how EPA used the national SPARROW models to estimate ambient
concentrations of TN, TP, and TSS.
> EPA estimated baseline and post-compliance metal concentrations using the water quality model
component of EPA's Risk-Screening Environmental Indicators (RSEI) model (U.S. EPA, 2012c).
EPA used estimates of metal loadings discharged from steam electric plants to each of 296 receiving
reaches under the baseline and five of the eight regulatory options (see EA for discussion of receiving
reaches; U.S. EPA, 2013a). In addition to these loadings, the RSEI model also incorporates 2010
Toxic Release Inventory (TRI) data on annual average discharges from other industrial sources.44
EPA input the loadings from steam electric plant and TRI discharges in the RSEI model to estimate
the long-term average concentrations in receiving and downstream reaches. The RSEI model uses a
simple dilution and first-order decay equation (metals are treated as conservative substances). In
the model, a plant is assumed to release its annual discharge at a constant rate throughout the year.
Annual average concentrations are then estimated and tracked downstream throughout the reach
network until one of three conditions occurs: 1) the release has traveled 300 km (186 miles)
downstream; 2) the release has traveled a distance equivalent to one week of travel time; or 3) the
concentration reaches 1 x 10"9 mg/L. The number of exceedances per waterbody (each receiving
reach) was calculated by comparing baseline and post-compliance concentrations with EPA's
freshwater chronic aquatic life criteria values for each metal45. If the concentration was greater than
43 EPA did not model changes in TSS concentrations as a result of this proposed ELG.
44 EPA removed from the 2010 TRI data the loadings associated with steam electric plants that report to TRI. For steam
electric plants, EPA used the loadings estimated in the EA under the baseline and the regulatory options. Loadings for
other TRI dischargers (non-steam electric plant and other industrial facilities) remained constant throughout all
scenarios.
45 RSEI utilizes the USGS's National Hydrology Dataset (NHD) which defines a reach as a continuous piece of surface
water with similar hydrologic characteristics. In the NHD each reach is assigned a reach code; a reach may be composed
of a single feature, like a lake or isolated stream, but reaches may also be composed of a number of contiguous features.
Each reach code occurs only once throughout the nation and once assigned, a reach code is permanently associated with
April 19, 2013
4-5
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
the criteria value for a given metal, EPA categorized the waterbody as having an AWQC exceedance
for that metal. EPA then summed the total number of AWQC exceedances (up to eight) for each
waterbody under the baseline and under each analyzed regulatory option.
> The USGS National Water Information System (NWIS) provided concentration data for three
parameters: 1) fecal coliform, 2) dissolved oxygen, and 3) biochemical oxygen demand. 46 EPA's
Storage and Retrieval (STORET) data warehouse provided additional data on fecal coliform counts
and biochemical oxygen demand when NWIS data was unavailable (U.S. EPA, 2008a).47
Note that the concentration data input into the WQI typically represent long-term average concentrations.
Table 4-4 summarizes the water quality modeling data used for estimating baseline and post-compliance
metal and nutrient concentrations for reaches directly receiving steam electric plant discharge and for
downstream reaches.
Table 4-4. Water Quality Modeling Data used in Calculating the Baseline and Policy Metal and
Nutrient Concentrations
Reach
Reach directly
receiving steam
electric plant
discharge (296
reaches) and
downstream
reaches
Input Data
Baseline and policy metal loadings
to reaches directly receiving steam
electric plant discharges (U.S.
EPA, 2013a).
Metal loadings from other TRI
dischargers in 2010.
Baseline and policy nutrient
loadings to reaches directly
receiving steam electric plant
discharges (U.S. EPA, 2013a)
Water Quality Model
Concentrations modeled in
RSEI
Concentrations modeled in
SPARROW
Model Output
In-steam metal
concentrations at the NHD
level
In-stream nutrient
concentrations at the
E2RF1 level
EPA used two different reach classification frameworks to assess in-stream water quality under the baseline
and each of the regulatory options: the National Hydrography Dataset (NHD) network and the USGS's
Enhanced River File 1 (E2RF1). Metal concentrations were estimated for reaches indexed to the NHD
network. In contrast, the SPARROW, NWIS, and STORET data are available for reaches indexed to the
E2RF1 network and to USGS's Hydrological Unit Code (HUC) watersheds. To conduct the analysis, EPA
analyzed the data at the level of spatial detail at which they exist, i.e., changes estimated at the level of E2RF1
reaches are attributed to these reaches; changes estimated at the more detailed NHD level are attributed only
to those (fewer) reach miles to which they apply. The WQI and benefits are ultimately calculated at the
resolution of E2RF1 reaches, but with adjustments made to data available at the NHD level to reflect
differences in spatial scale. Thus, to reconcile the two levels of resolution, EPA mapped all receiving reaches
from the NHD to the E2RF1 network using GIS and assigned the closest E2RF1 ID to each NHD catchment.
EPA then calculated the fraction of E2RF1 reach miles that correspond to the NHD reaches affected by steam
electric plant discharges either directly (receiving reach) or from upstream sources (downstream reach).
Figure 4-1 illustrates the differences in scale between the E2RF1 network and the NHD network.
its reach. If the reach is deleted, its reach code is retired. The NHD reaches in this analysis range from 0.003 miles to
9.11 miles in length.
46 USGS's NWIS dataset provides information on the occurrence, quantity, quality, distribution, and movement of
surface and underground waters based on data collected at approximately 1.5 million sites in all 50 States, the District of
Columbia, and U.S. territories. More information on NWIS can be found at http://waterdata.usgs.gov/nwis/
47 EPA's STORET (STOrage and RETrieval) Data Warehouse is a repository for water quality, biological, and physical
water monitoring data. More information can be found at http://www.epa. gov/storet/
April 19, 2013
4-6
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
Legend
— NHD Medium Resolution FlowlinesVs! (1 100000 scale)
^^~ Enhanced River File 1 Reacn Network
^ USGS HUC 8-Dlgit Watershed
0 5 10 20 30
40
• Km
Figure 4-1: Comparison between the NHD and E2RF1 Network in a Single Watershed.
A total of 121 E2RF1 IDs had multiple NHD catchments with changes in metal concentrations. For these
cases, EPA used a length weighted average to estimate E2RF1-level concentrations and the number of
exceedances attributed to the E2RF1 reach. Table 4-5 contains a summary of the methods used by EPA to
transfer metal concentrations from the NHD level to the E2RF1 level.
Table 4-5. Methods Used to Transfer Metal Concentration Reductions from the NHD to the E2RF1
Level
Scenario
Method
Number of E2RF1
Reaches
A single NHD reach is estimated to
have metal concentration
reductions along the E2RF1 reach
Water quality is calculated at E2RF1 level and the
change is attributed to only the portion of the
E2RF1 reach comprised of the NHD.
196
Multiple NHD reaches are
estimated to have metal
concentration reductions along the
E2RF1 reach
The number of AWQC exceedances is calculated at
the NHD level and then a reach length weighted
average is used to calculate the number of AWQC
exceedances at the E2RFllevel using NHD lengths
along the E2RF1 reach.
121
Of the 296 NHD reaches directly receiving steam electric plant discharges, 61 are projected to experience
changes in metal concentrations sufficient to eliminate one or more AWQC exceedances under the regulatory
options EPA analyzed. Additionally, a total of 639 downstream NHD reaches are estimated to have fewer
AWQC exceedances under the regulatory options. The combination of these NHD reaches and the E2RF1
reaches with changes in nutrient concentrations modeled in SPARROW amount to a total of 3,030 unique
April 19, 2013
4-7
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
benefiting E2RF1 reaches comprising atotal of 23,129 reach miles, the E2RF1 network in the contiguous
United States is comprised of approximately 60,878 reaches (636,863 reach miles) .
Baseline concentrations for all WQI parameters were available for atotal of 2,965 of the 3,030 reaches. EPA
used a successive average approach to address the data gaps in WQI parameters not modeled above (i.e., DO,
BOD, fecal coliform) in the remaining freshwater reaches. The approach involves assigning the average of
ambient concentrations for a WQI parameter within a hydrologic unit to reaches within the same hydrologic
unit with missing data, and progressively expanding the geographical scope of the hydrologic unit (HUC8,
HUC6, HUC4, and HUC2) to fill in all missing data.48 This approach assumes that reaches located in the
same watershed generally share similar characteristics. Using this estimation approach, EPA was able to
compile baseline water quality data for all freshwater reaches. Table 4-6 summarizes the data sources used to
estimate baseline and post-compliance values by water quality parameters.
Table 4-6. Water Quality Data used in Calculating the Baseline and Policy WQI
Parameter
TN
TP
TSS
DO
BOD
Fecal Coliform
Metals
Baseline value
From SPARROW output (baseline run)
From SPARROW output (baseline run)
From SPARROW output (baseline run)
Baseline value at the E2RF1 level3
Baseline value at the E2RF1 level3
Baseline value at the E2RF1 level3
Baseline exceedances at NHD level
Post-compliance value
From SPARROW output (regulatory option run)
From SPARROW output (regulatory option run)
No change. Regulatory option value equal
baseline value
No change. Regulatory option value equal
baseline value
No change. Regulatory option value equal
baseline value
No change. Regulatory option value equal
baseline value
Regulatory option exceedances at NHD level
when a single NHD reach is estimated to have
reductions in AWQC exceedances along a given
E2RF1. When multiple NHD reaches along the
same E2RF1 reach have reductions in AWQC
exceedance, then regulatory option exceedances
are estimated at theE2RFl level using a weighted
average (see Table 4-5)
a. Values based on STORET and NWIS data, complemented with data available for progressively larger geographical units
(HUCS, HUC6, HUC4, and HUC2), as needed to fill in all missing data.
4.1.3 Baseline WQI
Based on the estimated WQI value under the baseline scenario, EPA categorized each E2RF1 reach using five
WQI ranges (WQI < 25, 25
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
WQI values greater than 70 indicate that waters are swimmable (the recreational use with the highest required
WQI).49
Table 4-7. Percentage of Affected Reach Miles by WQI Classification: Baseline Scenario
Water
Quality
Classification
Unusable
Suitable for
Boating
Suitable for
Rough Fishing
Suitable for
Game Fishing
Suitable for
Swimming
Baseline
WQ
WQK25
25
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
Table 4-8. Water Quality Improvements from Proposed ELGs in All Benefiting Reaches
Change In WQI
Total Number of Improved
Reaches
Total Improved Reach
Miles
Percentage of Benefiting
Reach Miles
(23,129 Miles)
Percentage of E2RFlState
River Miles (547,352 Miles)
Benefiting
Option 1
0 < AWQI < 1
1
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
jness to Pay for Water
To estimate nonmarket benefits of water quality improvements resulting from the proposed ELG options,
EPA used results of the meta-analysis of stated preference studies described in the Environmental and
Economic Benefits Assessment for the Final Construction and Development Rule (U.S. EPA, 2009a). EPA's
benefit transfer approach follows standard methods described by Johnston et al. (2005), Shrestha et al.
(2007), and Rosenberger and Phipps (2007). EPA's meta-analysis results imply a simple benefit function of
the following general form:
Equation 4-2. \\\(WTPY,A,S) = Intercept + ^(coefficient^ x (independent variable value{)
ln(WTPYAS) =
The predicted natural log of annual household WTP for a given year
(Y), analytic category (A) and a given state (S)). Each analytic group
corresponds to a combination of the baseline water quality category (WQI
baseline) and the expected change in water quality under the regulatory
option for a given year(AWQI)
coefficient,
A vector of variable coefficients from the meta-regression with i
elements
independent variable values; = A vector of independent variable values with i elements
Here, ln(WTPY,A,s) is the dependent variable in the meta-analysis—the natural log of annual state-level
household WTP for water quality improvements in a given year (Y). Water quality improvements are
measured as a change in WQI from the baseline to post-compliance scenario (AWQI). Household WTP vary
by year, because household income (an independent variable) changes over time. The metadata include
independent variables characterizing specific details of the resource(s) valued, such as the baseline resource
conditions; the extent of resource improvements (i.e., AWQI) and whether they occur in estuarine or
freshwater; the geographic region and scale of resource improvements (e.g., regional improvements versus
improvements in one or several waterbodies); characteristics of surveyed populations (e.g., users, nonusers);
and other specific details of each study. Table 4-9 provides the estimated regression equation intercept (5.71),
variable coefficients (coefficient?), and the corresponding independent variable names. This meta-regression
allows the Agency to forecast WTP based on assigned values for model variables that are chosen to represent
a resource change in the proposed ELGs' policy context.
Table 4-9. Independent Variable Assignments for Surface Water Quality Meta-Analysis
Variable
Type
Study and
Methodology
Variables
Variable
intercept
year_index
discrete
volunt
mail
Coefficient
5.7109
-0.08043
-0.1248
-1.3233
-0.2013
Assigned
Value
9.68
1
0
0
Explanation
Set to 9.68 to reflect the mean year that the studies
in the dataset were conducted.
Set to one to reflect survey efforts that employed
discrete choice elicitation methods, which are
preferred over other approaches, such as open-
ended and payment card methods.
Set to zero because hypothetical voluntary
payment mechanisms are not incentive compatible
(Mitchell and Carson 1989).
Set to zero because in-person interviews are
preferred (if feasible).
April 19, 2013
4-11
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
Table 4-9. Independent Variable Assignments for Surface Water Quality Meta-Analysis
Variable
Type
Surveyed
Population
Waterbody
Type Variables
Geographic
Region and
Scale
Variables
Resource
Improvement
Variables
Variable
lump_sum
WQI
nonparam
non reviewed
median WTP
outlier_bids
income
nonusers
single river
single_lake
multiple_river
regional_fresh
salt_pond
num riv
pond
mr
mp
allmult
Coefficient
0.5569
-0.3275
-0.6698
-0.2718
-0.5358
-0.8837
0.0000027
-0.4036
-0.4279
-0.06316
-1.4752
0.1588
0.9849
0.1173
-0.8846
1.6337
-0.3728
Assigned
Value
0
1
0
0
0
1
Varies
0
0
0
0
1
0
0
0
0
1
Explanation
Set to zero because the policy option would be
paid for over a period of years.
Set to one because of the methodological use of
the WQI in the meta-analysis.
Set to zero because most studies used in the meta
analysis used regression analysis to calculate
willingness-to-pay values.
Set to zero because studies published in peer-
reviewed journals are preferred.
Set to zero because only average or mean WTP
values in combination with the number of affected
households will mathematically yield total benefits
if the distribution of WTP is not perfectly
symmetrical.
Set to one because survey data that exclude such
responses are preferable; outlier bids are often
excluded from the analysis of stated preference
data because these bids (often identified as greater
than a certain percentage of a respondent's
income) may indicate that a respondent did not
consider his or her budget constraints and or
supplementary goods.
Median household income values assigned
separately for each state and varies by year based
on the estimated income growth in future years.
Set to zero in order to estimate the total value for
aquatic habitat improvements, including both use
and nonuse values; for nonuser population, the
total value of water resource improvements
includes nonuse values only (Freeman 2003).
The expectation is that multiple freshwater bodies
within a state would be affected by the proposed
ELGs. Therefore regional Afresh is set to one.
Indicates the number of rivers or salt ponds
affected by the proposed ELGs, and is set to zero
because this analysis assumes that the effluent
guidelines would affect the entire
watershed/region; this variable assignment is
constant across study regions.
Regional variables are omitted from the predictive
portion of the analysis (i.e., set to zero) because
the regression results suggest that these variables
may be picking spurious or other unexplained
effects (e.g., author's effect).
Set to one because it is assumed that multiple
species would benefit from water quality
improvements
April 19, 2013
4-12
-------
Benefit and Cost Analysis for Proposed ELGs
4: Benefits from Water Quality Improvements
Table 4-9. Independent Variable Assignments for Surface Water Quality Meta-Analysis
Variable
Type
Variable
nonspec
lnquality_ch
fish_use
fishplus
Inbase
Coefficient
-0.4042
0.4065
-0.3317
0.4432
0.02610
Assigned
Value
0
Varies
1
0
Varies
Explanation
Set to zero because it is assumed that multiple
species would benefit from water quality
improvements.
Set to the natural log of the average change in
water quality for each analytic scenario.
Set to one because the proposed ELGs are
expected to benefit a variety of aquatic species and
therefore enhance recreational fishing
opportunities.
Set to zero because the proposed ELGs are not
expected to result in a fish population change of 50
percent or greater.
Set to the natural log of the average baseline water
quality for each analytic scenario.
EPA assigned a value to each model variable corresponding with theory, characteristics of the water
resources, and sites affected by the proposed ELGs and the policy context. This follows general guidance
from Bergstrom and Taylor (2006) that meta-analysis benefit transfer should incorporate theoretical
expectations and structures, at least in a weak form. In this instance, three of the methodology variables,
discrete, WQI_study, and outlier_bids are all included with an assigned value of one. Yearjndex is given the
value of 9.68, which corresponds to the mean year that the studies were conducted, 2002. Nonparam is set to
zero because most studies included in metadata used parametric methods to estimate WTP values. Other study
and methodology variables (volunt, mail, lump_sum, non_reviewed, median_WTP) are assigned a zero value.
The median household income (income) varies by state and year. To estimate state level household income
EPA used historic state-specific median household income data from the U.S. Census Bureau's 2009
Community Population Survey (U.S. Census Bureau, 2010b). The data contained median household for the
years 1984 to 2009. The Consumer Price Index (CPI) was used to inflate all values to 2010 dollars (U.S. BLS,
2010) and a stepwise autoregressive forecasting method was used to estimate future annual state level median
household income. The variable nonusers was set to zero because water quality improvements are expected to
enhance both use and non-use values of the affected resources and thus benefit both users and nonusers. For
example, recreational users may benefit from enhanced recreational opportunities and non-users may benefit
from the knowledge that water quality in their state has improved.
The proposed ELGs are expected to affect water quality at a regional level because steam electric plants are
located in 39 continental states and water quality improvements are expected in 42 continental states. The
Agency set a dummy variable denoting multiple regions (mr) to zero because the expected magnitude of
water quality improvement is relatively modest and, as a result, EPA's analysis focuses primarily on in-state
or local resource improvements. The Mountain Plain regional dummy variable (mp) is set to zero because the
magnitude of the regional effect suggests that spurious or otherwise unexplained effects (e.g., the effect of
specific researchers who appear more than once in the data) may drive their overall magnitude.
To account for the regional scale of the water quality effect resulting from the proposed ELGs, the variable
regionalAfresh is assigned a value of one. Other variables relating to waterbody type (i.e., single Jake,
single_river, salt_pond, multiple_river, num_riv_pond) are set to zero.
As discussed in Chapter 1, benefits from water quality changes are estimated for all years between 2017 and
2040. For reaches directly receiving steam electric plant discharges, benefits are expected to begin accruing
April 19, 2013
4-13
-------
Benefit and Cost Analysis for Proposed ELGs 4: Benefits from Water Quality Improvements
3 years after renewal of a plant's NPDES permit, after promulgation of the rule in 2014. This corresponds to
the technology implementation window of 2017 through 2021. For example if a plant's implements control
technologies to meet the revised effluent limits in 2020, the receiving reach would begin accumulating
benefits in 2020 and have no benefits in 2017, 2018, and 2019. Downstream reaches, however, are affected
not only by discharges from this plant, but also by any discharge from upstream. Consequently, for this
analysis, EPA assumed that all benefits begin accruing in 2019, which is the midpoint of the compliance
period.51
The majority of the studies in the meta-analysis valued improvements in multiple waterbodies or across a
geographic region, and the studies that values a single resource generally valued such large waterbodies, for
example the entirety the Minnesota River (355 miles) or the Chesapeake Bay (4,479 square miles), that if
these waterbodies were translated to the E2RF1 or NHD network they would be comprised of multiple reach
codes. Therefore, the valuation function from the meta-analysis cannot appropriately be applied to value
changes in a single small waterbody. To estimate benefits, EPA grouped similar reaches located in the same
state together for valuation purposes. These groupings correspond to 16 analytic categories reflecting
combinations of four levels of baseline water quality conditions (25
-------
Benefit and Cost Analysis for Proposed ELGs 4: Benefits from Water Quality Improvements
Equation 4-3. WTPYAS = exp(\n_WTPY AS +<72e /2)xPercentRiverMiles
exp(-) = the exponential operator
ln_WTPYjA,s = the predicted natural log of household WTP for a
given year (Y), analytic category (A), and state (S)
oe2 = the model residual variance (0.1876)
PercentRiverMiles = the percent of state E2RF1 Miles within each analytic
category
EPA used a procedure described by Krinsky and Robb (1986) to estimate the 5th, average, and 95th percentiles
of total WTP for each state, based on the results of the total WTP regression model.53 These bounds
characterize the uncertainty associated with the benefit results. This analysis provides confidence limits for
WTP estimates related to the covariance of meta-analysis parameter estimates. It does not, however, assess
the sensitivity of results to changes in meta-regression model assumptions or specifications (cf. Johnston et al.
2005, 2006) or assumptions implied in benefit aggregation (cf. Loomis 1996; Loomis et al. 2000; Bateman et
al. 2006). As noted above, EPA has made assumptions and specifications based on best benefit transfer
practices (e.g., Johnston et al. (2005), Shrestha et al. (2007), and Rosenberger and Phipps (2007)).
To calculate average household WTP for a given year, household WTP values were summed for all reaches
for both users and nonusers. Equation 4-4 provides the discount formula used to calculate annualized benefits.
Equation 4-4. AWTPS =
AWTPS= Annualized benefits in 2010$ for a given state (S)
WTPYA,s = WTP value for water quality improvements for a given year (Y),
analytic category (A), and state (S)
Y = Year when benefits are realized
i = Discount rate (3 or 7 percent)
n = Duration of the analysis (24 years)
As shown in Table 4-10 and Table 4-11, average annual household WTP estimates for the proposed ELGs
range from $0.00 for EPA Regions 1, 2, 9, and 10 under Option 1, to $2.43 in EPA Region 4 under Option 4
(with a 3 percent discount rate), for the five regulatory options EPA analyzed. Under Option 3, the midpoint
estimate of annualized household WTP values ranges from $0.01 per household in Region 10 to $2.19 per
household in Region 4 with a 3 percent discount rate ($0.01 to $1.93 with a 7 percent discount rate). Option 4
has the highest the midpoint estimates of $2.43 per household in Region 4, with a 3 percent discount rate.
53 The procedure involves sampling the variance-covariance matrix of the estimated coefficients (matrix containing the
variance of a random variable across the diagonal and the covariance, how random variables vary or change together, off
the diagonal). WTP values are then calculated for each drawing from the variance-covariance matrix and an empirical
distribution of WTP values is constructed.
April 19, 2013 "
-------
Benefit and Cost Analysis for Proposed ELG Regulation
4: Benefits from Water Quality Improvements
Table 4-10. Average Household WTP for Water Quality Improvements (2010$; 3% Discount Rate)
EPA Region
1
2
3
4
5
6
7
8
9
10
Option 1
5%
$0.00
$0.00
$0.00
$0.09
$0.01
$0.07
$0.00
$0.06
$0.00
$0.00
Mean
$0.00
$0.00
$0.03
$0.39
$0.09
$0.36
$0.08
$0.18
$0.00
$0.00
95%
$0.00
$0.00
$0.08
$1.03
$0.31
$1.05
$0.30
$0.48
$0.00
$0.00
Option 2
5%
$0.01
$0.00
$0.09
$0.31
$0.15
$0.17
$0.03
$0.06
$0.00
$0.00
Mean
$0.10
$0.02
$0.48
$1.61
$1.00
$0.79
$0.26
$0.25
$0.05
$0.01
95%
$0.36
$0.07
$1.34
$4.49
$3.03
$2.13
$0.84
$0.68
$0.19
$0.04
Option 3
5%
$0.01
$0.00
$0.10
$0.44
$0.30
$0.18
$0.03
$0.07
$0.00
$0.00
Mean
$0.10
$0.02
$0.54
$2.19
$1.46
$0.89
$0.29
$0.35
$0.07
$0.01
95%
$0.36
$0.07
$1.52
$6.00
$4.04
$2.43
$0.96
$1.04
$0.26
$0.04
Option 4
5%
$0.04
$0.02
$0.55
$0.54
$0.58
$0.32
$0.13
$0.18
$0.03
$0.00
Mean
$0.31
$0.19
$1.96
$2.43
$2.34
$1.47
$0.77
$0.79
$0.21
$0.01
95%
$0.96
$0.67
$4.76
$6.43
$5.99
$3.98
$2.31
$2.11
$0.60
$0.04
Option 5
5%
$0.04
$0.02
$0.54
$0.52
$0.57
$0.31
$0.13
$0.18
$0.03
$0.00
Mean
$0.31
$0.19
$1.94
$2.39
$2.32
$1.45
$0.76
$0.79
$0.21
$0.01
95%
$0.95
$0.66
$4.74
$6.35
$5.95
$3.94
$2.30
$2.11
$0.60
$0.04
Source: U.S. EPA Analysis, 2013
Table 4-11. Average Household WTP for Water Quality Improvements (2010$; 7% Discount Rate)
EPA Region
1
2
3
4
5
6
7
8
9
10
Option 1
5%
$0.00
$0.00
$0.00
$0.08
$0.01
$0.06
$0.00
$0.05
$0.00
$0.00
Mean
$0.00
$0.00
$0.02
$0.34
$0.08
$0.32
$0.07
$0.16
$0.00
$0.00
95%
$0.00
$0.00
$0.07
$0.90
$0.27
$0.92
$0.27
$0.42
$0.00
$0.00
Option 2
5%
$0.01
$0.00
$0.08
$0.27
$0.13
$0.15
$0.03
$0.06
$0.00
$0.00
Mean
$0.09
$0.02
$0.42
$1.41
$0.88
$0.69
$0.23
$0.22
$0.05
$0.01
95%
$0.32
$0.06
$1.18
$3.95
$2.66
$1.87
$0.73
$0.59
$0.16
$0.03
Option 3
5%
$0.01
$0.00
$0.09
$0.39
$0.26
$0.16
$0.03
$0.06
$0.00
$0.00
Mean
$0.09
$0.02
$0.47
$1.93
$1.29
$0.78
$0.26
$0.31
$0.06
$0.01
95%
$0.32
$0.06
$1.33
$5.27
$3.55
$2.13
$0.84
$0.91
$0.23
$0.03
Option 4
5%
$0.03
$0.01
$0.48
$0.47
$0.51
$0.28
$0.12
$0.16
$0.03
$0.00
Mean
$0.27
$0.17
$1.72
$2.14
$2.05
$1.29
$0.68
$0.69
$0.18
$0.01
95%
$0.84
$0.58
$4.18
$5.65
$5.26
$3.49
$2.03
$1.85
$0.53
$0.04
Option 5
5%
$0.03
$0.01
$0.48
$0.46
$0.50
$0.27
$0.12
$0.16
$0.03
$0.00
Mean
$0.27
$0.17
$1.71
$2.10
$2.04
$1.27
$0.67
$0.69
$0.18
$0.01
95%
$0.84
$0.58
$4.16
$5.58
$5.23
$3.46
$2.02
$1.85
$0.53
$0.04
Source: U.S. EPA Analysis, 2013
April 19, 2013
4-16
-------
Benefit and Cost Analysis for Proposed ELG Regulation 4: Benefits from Water Quality Improvements
4.3 Total WTP for Water Quality Improvements
To estimate total WTP (TWTP) for water quality improvements in a given state, EPA multiplied the per-
household WTP values for the estimated water quality improvement by the number of households within each
state in a given year. The total WTP equation for each reach is provided below:
Equation 4-5. TWTPY s = WTPYAS(WQIbaselme,MVQI
Y,S
TWTPY,s= the total state-level welfare change from improved water quality for all 16
analytic categories (A) for a given year (Y)
WTP Y,A,s = the estimated state-level per-household WTP for a given water quality
improvement (i.e.,analytic category A) in year Y; analytic category A
corresponds to a combination of the baseline water quality category(WQI
baseline) and the expected change in water quality under the regulatory
option for a given year(AWQI)
HH Y,S = the number of households residing in state (S) in a given year (Y)
EPA generated annual state level population estimates through the period of analysis using the method
described in Chapter 3. For the purposes of this analysis, EPA assumed that any changes in average
household size overtime would be negligible and average household size was kept constant for all years.
EPA then calculated annualized total WTP values for each state with both a 3 and 7 percent discount rate
using Equation 4-4. Annualized values were then summed across all states to calculate total annualized
benefits of the proposed ELGs for each analyzed regulatory option, by EPA region and across all states. Table
4-12 and Table 4-13 present the results for the 3 percent and 7 percent discount rates, respectively.
As detailed by Loomis (1996), Loomis et al. (2000) and Bateman et al. (2006), among others, there are
numerous uncertainties and assumptions in aggregating WTP across spatial jurisdictions. While these
uncertainties are well known, the literature does not agree on appropriate, standardized guidance for benefit
aggregations, and applied benefit-cost analysis almost universally requires simplifying assumptions in order
to generate defensible welfare aggregations. In analyzing benefits of the proposed ELGs, EPA assumed that
households would gain no benefits from water quality improvements in aquatic resources located outside of
their state of residence. Consequently, the population considered in the benefits analysis of the proposed
ELGs does not represent all the households that are likely to hold values for water resources in a given state.
Residents of other states may hold values for water resources outside of their home state, in particular if such
resources have personal, regional, or national significance. Even if per household WTP for out-of-state
residents are small they can be very large in the aggregate if these values are held by a substantial fraction of
the population.
April 19, 2013 4-17
-------
Benefit and Cost Analysis for Proposed ELG Regulation
4: Benefits from Water Quality Improvements
Table 4-12. Estimated Annualized Benefits for Water Quality Improvements (3% Discount Rate, Millions 2010$)
EPA
Region
1
2
3
4
5
6
7
8
9
10
Total
Option 1
5%
$0.00
$0.00
$0.04
$0.49
$0.15
$1.23
$0.02
$0.04
$0.00
$0.00
$1.97
Mean
$0.00
$0.00
$0.25
$2.36
$1.20
$3.87
$0.45
$0.14
$0.00
$0.00
$8.28
95%
$0.00
$0.00
$0.73
$6.43
$3.92
$9.21
$1.68
$0.38
$0.00
$0.00
$22.35
Option 2
5%
$0.03
$0.04
$0.86
$2.03
$1.52
$2.45
$0.13
$0.05
$0.00
$0.01
$7.13
Mean
$0.55
$0.41
$4.36
$11.21
$10.70
$8.56
$1.33
$0.27
$0.46
$0.09
$37.96
95%
$1.97
$1.33
$11.89
$31.79
$32.60
$20.84
$4.45
$0.84
$1.61
$0.31
$107.65
Option 3
5%
$0.03
$0.04
$0.93
$3.25
$3.17
$2.52
$0.15
$0.09
$0.01
$0.01
$10.20
Mean
$0.55
$0.41
$4.91
$16.20
$16.00
$8.95
$1.48
$0.69
$0.63
$0.09
$49.90
95%
$1.97
$1.33
$13.59
$44.21
$44.68
$21.98
$4.90
$2.29
$2.31
$0.31
$137.59
Option 4
5%
$0.00
$0.22
$4.54
$3.94
$6.07
$3.42
$0.67
$0.47
$0.29
$0.01
$19.63
Mean
$1.65
$2.51
$16.22
$17.80
$24.91
$12.10
$3.46
$2.26
$1.82
$0.11
$82.83
95%
$5.05
$8.51
$39.46
$47.05
$64.28
$29.73
$10.00
$6.11
$5.23
$0.35
$215.79
Option 5
5%
$0.00
$0.22
$4.49
$3.82
$6.00
$3.31
$0.67
$0.47
$0.29
$0.01
$19.27
Mean
$1.64
$2.49
$16.09
$17.48
$24.72
$11.83
$3.44
$2.26
$1.82
$0.11
$81.89
95%
$5.04
$8.48
$39.20
$46.51
$63.95
$29.29
$9.96
$6.11
$5.23
$0.35
$214.12
Source: U.S. EPA Analysis, 2013
Table 4-13. Estimated Annualized Benefits for Water Quality Im
EPA
Region
1
2
3
4
5
6
7
8
9
10
Total
Option 1
5%
$0.00
$0.00
$0.03
$0.41
$0.12
$1.02
$0.01
$0.03
$0.00
$0.00
$1.64
Mean
$0.00
$0.00
$0.20
$1.97
$1.01
$123
$O38
$67i2
$o766
$o766
$6.91
95%
$0.00
$0.00
$0.61
$5.36
$3.29
$7.67
$1.41
$0.32
$0.00
$0.00
$18.65
Option 2
5%
$0.03
$0.03
$0.73
$1.69
$1.27
$2.05
$0.11
$0.04
$0.00
$0.01
$5.96
Mean
$0.46
$0.35
$3.66
$9.35
$8.98
$7U4
$TTi
$6723
$038
$6708
$31.73
95%
$1.65
$1.12
$9.98
$26.50
$27.34
$1737"
$3/72"
$0/70"
$134"
$6726"
$89.98
jrovements (7% Discount Rate, Millions 2010$)a
Option 3
5%
$0.03
$0.03
$0.78
$2.71
$2.66
$2.10
$0.12
$0.07
$0.01
$0.01
$8.53
Mean
$0.46
$0.35
$4.11
$13.50
$13.43
$7.46
$1.23
$0.57
$0.52
$0.08
$41.71
95%
$1.65
$1.12
$11.39
$36.85
$37.47
$1832
$4769
$L90
$L91
$6726
$114.97
Option 4
5%
$0.00
$0.18
$3.80
$3.29
$5.10
$2785
$6756
$039
$6724
$6761
$16.42
Mean
$1.38
$2.10
$13.57
$14.84
$20.89
$10.09
$2.89
$1.87
$1.50
$0.09
$69.23
95%
$4.23
$7.13
$33.00
$39.22
$53.87
$24.78
$8.36
$5.06
$4.33
$0.29
$180.28
Option 5
5%
$0.00
$0.18
$3.76
$3.19
$5.04
$2776
$6756
$039
$6724
$6761
$16.12
Mean
$1.38
$2.09
$13.46
$14.58
$20.73
$9787
$2788
$1787
$1750
$6769
$68.45
95%
$4.23
$7.11
$32.78
$38.77
$53.59
$24741
$832
$5765
$433
$6729
178.89
Source: U.S. EPA Analysis, 2013
April 19, 2013
4-18
-------
Benefit and Cost Analysis for Proposed ELG Regulation
4: Benefits from Water Quality Improvements
EPA estimated that between 3,945 and 22,447 reach miles would improve under the proposed ELGs,
depending on the regulatory option. The total annualized benefits of water quality improvements resulting
from reduced metal and nutrient pollution in these reaches range from $8.3 million to $82.8 million (with a 3
percent discount rate) and from $6.9 million to $69.2 million with a 7 percent discount rate.
""'Op'tiorf'3' is expected to improve atotal of 15,682 reach miles. The midpoint estimate of annualized benefits of
improving these waterbodies is $49.9 million per year with a 3 percent discount rate ($41.7 million with a
7 percent discount rate). EPA regions 4 and 5 are expected to accrue the most benefits with $16.2 and $16.0
million annually (with a 3 percent discount rate). Options 3a and 3b are expected to result in the improvement
of fewer miles, and provide lower estimated benefits that those of Option 3. Option 4a is expected to have
benefits between those of Options 3 and 4, i.e., improve between 15,682 and 22,447 reach miles and provide
annualized benefits estimated at between $49.9 and $82.8 million (midpoint estimate; 3 percent discount rate).
4.4 Limitations and Uncertainties
Table 4-14 summarizes the uncertainties in the analysis of benefits associated with improved surface water
quality and indicates the direction of the potential bias.
Table 4-14. Uncertainties in the Analysis of Nonmarket Water Quality Benefits
Issue
Effect on Benefits
Estimate
Notes
Limitations inherent to the meta-analysis model and benefit transfer
Assumption that
households will only
value improvements of
waterbodies located in
their state.
Underestimate
Residents of other states may hold values for water resources
outside of their home state, in particular if such resources have
personal, regional, or national significance. Even if per household
WTP for out-of-state residents are small they can be very large in
the aggregate if these values are held by a substantial fraction of
.!.h? Eoj>ulation
Potential hypothetical
bias underlying
contingent valuation
method (CVM) results
Uncertain
Following standard benefit transfer approaches, including meta-
analytic transfers, this analysis proceeds under the assumption that
each source study provides a valid, unbiased estimate of the
welfare measure under consideration (cf Moeltner et al. 2007;
Rosenberger and Phipps 2007). To minimize potential
hypothetical bias underlying stated preference studies included in
meta-data, EPA set independent variable values to reflect best
. transfer practices.
Use of different water
quality measures in
underlying meta-data
Underestimate
The estimation of WTP may be sensitive to differences in the
environmental water quality measures across studies in the meta
data. Studies that did not use the WQI were mapped to the WQI so
a comparison could be made across studies. The dummy variable
(WQI) captures the effect of a study using (WQI=1) or not using
(WQI=0) the WQI. EPA observed that studies that did not use the
WQI had lower WTP values. This indicates a potential systematic
bias in the mapping of studies that did not use the WQI. In
analyzing the benefits of the proposed ELGs, EPA set the WQI to
one to reduce uncertainty in WTP estimates associated with
studies that did not include WQI as a native survey instrument.
April 19, 2013
4-19
-------
Benefit and Cost Analysis for Proposed ELG Regulation
4: Benefits from Water Quality Improvements
Table 4-14. Uncertainties in the Analysis of Nonmarket Water Quality Benefits
Issue
Transfer error
Effect on Benefits
Estimate
Uncertain
Notes
Transfer error may occur when benefit estimates from a study site
are adopted to forecast the benefits of a policy site. Rosenberger
and Stanley (2006) define transfer error as the difference between
the transferred and actual, generally unknown, value. While meta-
analysis is fairly accurate when estimating benefit function,
transfer error may be a problem in cases where the sample size is
small. While meta-analyses have been shown to outperform other
function-based transfer methods in many cases, this result is not
universal (Shrestha et al. 2007). This notwithstanding, results
reviewed by Rosenberger and Phipps (2007) are "very promising"
for the performance of meta-analytic benefit transfers relative to
alternative transfer methods.
Use of the WQI to link water quality changes discharges to effects on human uses and support for aquatic and
terrestrial species
Changes in WQI
reflect only reductions
in metal and nutrient
concentrations
In-stream metal
concentrations are
based only on loadings
from steam electric
plants and other TRI
dischargers
Use of nonlinear
subindex curves
Underestimate
Uncertain
Underestimate
The estimated changes in WQI reflect only water quality
improvements resulting directly from reductions in metal and
nutrient concentrations. They do not include improvements in
water quality indicators associated with other pollutant loadings
(e.g., BOD, dissolved oxygen), nor do they consider
improvements in other water quality variables such as TSS.
Omitting some water quality parameters from the analysis is likely
to result in underestimation of the expected water quality changes.
In-stream concentrations for heavy metals were estimated based
on loadings from steam electric plant and other TRI dischargers
only and, as a result, do not account for background
concentrations of these pollutants from other sources, such as
contaminated sediments, non-point sources, point sources that are
not required to report to TRI, air deposition, etc. Not including
other contributors to background metal concentrations in the
analysis is likely to result in understatement of baseline
concentrations of these pollutants. The overall impact of this
limitation on the estimated WTP for water quality improvement is
uncertain but is expected to be small since the WTP function used
in this analysis is most sensitive to the change in water quality.
The methodology used to translate in-stream sediment and nutrient
concentrations into subindex scores employs nonlinear
transformation curves. Water quality changes that fall outside of
the sensitive part of the transformation curve (i.e., above^elow
the upper/lower bounds, respectively) yield no benefit in the
analysis.
April 19, 2013
4-20
-------
Benefit and Cost Analysis for Proposed ELGs
5: Threatened and Endangered Species Benefits
5 Impacts and Benefits to Threatened and Endangered Species
Threatened and endangered (T&E) species are species vulnerable to future extinction or at risk of extinction
in the near future, respectively. These designations reflect low or rapidly declining population levels, loss of
essential habitat, or life history stages that are particularly vulnerable to environmental alteration. In many
cases, T&E species are given special protection due to inherent vulnerabilities to habitat modification,
disturbance, or other human impacts. This chapter examines the environmental impacts of steam electric plant
discharges on T&E species and the benefits associated with improvements resulting from the proposed ELGs.
As described in the EA (U.S. EPA, 2013a), the chemical constituents of steam electric waste streams can pose
serious threats to ecological health due to the bioaccumulative nature of many constituents, high
concentrations, and high loadings. Pollutants such as selenium, arsenic and mercury have been associated
with fish kills, disruption of growth and reproductive cycles and behavioral and psychological alterations in
aquatic organisms (U.S. EPA, 2009a; Appendix F). Additionally, high nutrient loads can lead to the
eutrophication of waterbodies. Eutrophication can lead to increases in the occurrence and intensity of water
column phytoplankton, including harmful algal blooms (e.g., nuisance and/or toxic species), which have been
found to cause fatal poisoning in other animals, fish, and birds (Williams et al., 2001). Eutrophication may
also result in the loss of critical submerged rooted aquatic plants (or macrophytes), and reduced dissolved
oxygen (DO), levels, leading to anoxic or hypoxic waters.
For species vulnerable to future extinction, even minor changes to growth and reproductive rates and small
levels of mortality may represent a substantial portion of annual population growth. Consequently, steam
electric plant discharges may either lengthen recovery time, or hasten the demise of these species. For this
reason, the proposed ELGs may have a significant impact on T&E species populations.
From an economic perspective, T&E species affected by steam electric plant discharges may have both use
and nonuse values. However, given the protected nature of T&E species and the fact that the majority of T&E
species do not have direct uses, the majority of the economic value for T&E species comes from nonuse
values. Species-specific estimates of nonuse values held for the protection of T&E species can be most
accurately derived by primary research using stated preference techniques. However, the cost, administrative
burden, and time required to develop primary research estimates to value effects of the proposed regulation on
T&E species are beyond the schedule and resources available to EPA for this rulemaking. As an alternative,
EPA considered a benefit transfer approach that relies on information from existing studies (U.S. EPA,
201 Ob).
In this chapter, EPA explores the current status of major freshwater taxa, identifies the extent to which the
proposed ELGs can be expected to benefit species protected by the Endangered Species Act, and applies
economic valuation studies to these T&E species to estimate WTP for these benefits.
5.2 Baseline Status of Freshwater Fish Species
Reviews of aquatic species' conservation status over the past three decades have documented the effect of
cumulative stressors on freshwater aquatic ecosystems, resulting in a significant decline in the biodiversity
and condition of indigenous communities (Deacon et al., 1979; Williams et al., 1989; Williams et al., 1993;
Taylor et al., 1996; Taylor et al., 2007; Jelks et al., 2008). Overall, aquatic species are disproportionately
imperiled relative to terrestrial species. For example, while 39 percent of freshwater and diadromous fish
April 19, 2013
5-1
-------
Benefit and Cost Analysis for Proposed ELGs 5: Threatened and Endangered Species Benefits
species (Jelks et al., 2008) are classified as T&E, a similar status review found that only 7 percent of North
American bird and mammal species are currently imperiled (Wilcove and Master, 2005).
Approximately 39 percent of described fish species in North America are imperiled, with 700 fish taxa
classified as vulnerable (230), threatened (190), or endangered (280) in addition to 61 taxa presumed extinct
or functionally extirpated from nature (Jelks et al., 2008). These data show that the number of T&E species
have increased by 98 percent and 179 percent when compared to similar reviews conducted by the American
Fisheries Society in 1989 (Williams, Johnson et al. 1989) and 1979 (Deacon et al., 1979), respectively.
Despite recent conservation efforts, including the listing of several species under the Endangered Species Act
(ESA), only 6 percent of the fish taxa assessed in 2008 had improved in status since the 1989 inventory (Jelks
et al., 2008).
Several families offish have strikingly high proportions of T&E species. Approximately 46 percent and
44 percent of species within families Cyprinidae (carps and true minnows) and Percidae (darters and perches)
are imperiled, respectively. Some families with few, wide-ranging species have even higher rates of
imperilment, including the Acipenseridae (sturgeons; 88 percent) and Polyodontidae (paddlefish;
100 percent). Families with species important to sport and commercial fisheries ranged from a low of
22 percent for Centrarchidae (sunfishes) to a high of 61 percent for Salmonidae (salmon) (Jelks et al., 2008).
To assess the potential effects of the proposed ELGs on T&E species, EPA constructed databases to
determine which species are found in waters expected to improve due to a reduction in pollutant discharge
from steam electric plants. Notably, these databases exclude all species considered threatened or endangered
by scientific organizations [e.g., the American Fisheries Society (Williams et al., 1993; Taylor et al., 2007;
Jelks et al., 2008)] but not protected by the ESA. These databases allowed EPA to estimate the potential for
adverse impacts of steam electric plant discharges on T&E species, as well as benefits associated with the
proposed ELGs.
5.3.1 Identifying T&E Species Potentially Affected by the Proposed ELGs
To estimate the effects of the proposed ELGs on T&E species, all affected species must first be identified.
EPA identified all species currently listed or in consideration for listing under the ESA (as of August 4, 2012)
using the US Fish and Wildlife Service Environmental Conservation Online System (U.S. FWS, 2010a).
Whenever possible, the geographical distribution of T&E species was obtained in geographic information
system (GIS) format as polygon (shape) files, line files (for inhabitants of small creeks and rivers) and as a
subset of geodatabase files. Data sources include the US Fish and Wildlife Service (U.S. FWS, 2010b),
NCAA's Office of Response and Restoration (NOAA, 2010), NatureServe (NatureServe, 2009), and NOAA
National Marine Fisheries Service (NMFS, 2010a; NMFS, 2010b; NMFS, 2010c). For several freshwater
species, geographic ranges were available only as 6-digit hydrologic unit codes (HUC) (NatureServe, 2009;
U.S. FWS, 2010b). For these species, EPA created GIS data layers using a GIS HUC database obtained from
the USGS (Steeves and Nebert, 1994).
To determine the probability that individual T&E species could benefit from the proposed ELGs, EPA
compiled data on locations of steam electric plants and receiving waterbodies. The Agency used plant and
outfall coordinates it had obtained through its 2010 Questionnaire for the Steam Electric Power Generating
Effluent Guidelines (the industry survey) and georeferenced these coordinates to waterbodies (see EA for
details; U.S. EPA, 2013a). The result of this analysis consists of the National Hydrography Dataset (NHD)
Plus (COMIDs) identifiers of waterbodies that receive discharges from steam electric plants and indicators of
water quality under the baseline and each analyzed regulatory option. EPA queried these data to identify
April 19, 2013 5-2
-------
Benefit and Cost Analysis for Proposed ELGs
5: Threatened and Endangered Species Benefits
"affected areas" as those habitats where 1) receiving waters do not meet water quality metrics recognized to
cause harm in organisms under baseline conditions; and 2) receiving waters exceed water quality metrics
under the most stringent regulatory option EPA analyzed (Option 5). EPA used these data in ArcGIS to
determine the T&E species with habitat extents overlapping the affected areas.
EPA constructed two screening databases using the spatial data:
> Database 1 - all T&E species whose habitat is within 5 miles of waterbodies affected by steam
electric plant discharges. The 5-mile buffer was chosen to account for any minor errors in outfall
location and habitat maps, and as a very rough approximation of a distance where benefits from the
rule are most likely to affect species life history parameters (e.g., survival, reproduction).
> Database 2 - all T&E species whose habitat overlaps those waterbodies affected by the effluent
discharges from steam electric plants.
After identifying T&E species potentially affected by the proposed ELGs, EPA classified the species on the
basis of their vulnerability to changes in water quality. Species were classified as follows:
> High vulnerability - species living in aquatic habitats for several life history stages and/or species that
obtain a majority of their food from aquatic sources.
> Moderate vulnerability - species living in aquatic habitats for one life history stage and/or species
that obtain some of their food from aquatic sources.
> Low vulnerability - species whose habitats overlap bodies of water, but whose life history traits and
food sources are terrestrial.
Life history data used to classify species were obtained from a wide variety of sources (Froese and Pauly,
2009; NatureServe, 2009; AFSC, 2010; ASMFC, 2010; NEFSC, 2010; PIFSC, 2010a; PIFSC, 2010b; SEFSC,
2010; SWFSC, 2010; U.S. FWS, 2010c).
The results of the spatial analysis and vulnerability classification process for Database 1 and Database 2 (as
described above) are presented in Table 5-1 and Table 5-2, respectively. More species were identified in
Database 2 than in Database 1, because the impacted waterbodies are geographically larger than the area
encompassed by the five-mile radius. Appendix F lists all T&E species potentially affected by the proposed
ELGs.
Table 5-1. T&E Species With Affected Habitat Within 5 Miles of Steam Electric Power Plants
Species Group
Amphibians
Arachnids
Birds
Clams
Crustaceans
Fishes
Insects
Mammals
Reptiles
Snails
Total
Species Vulnerability
Low
0
4
6
0
0
0
9
15
2
9
45
Moderate
3
0
5
0
1
0
1
2
0
0
12
High
1
0
1
29
1
14
2
0
4
16
68
Species Count
4
4
12
29
2
14
12
17
6
25
125
Source: U.S. EPA Analysis, 2013
April 19, 2013
5-3
-------
Benefit and Cost Analysis for Proposed ELGs
5: Threatened and Endangered Species Benefits
Table 5-2. T&E Species with Habitat Occurring within Waterbodies Affected by Steam Electric Power
Plants
Species Group
Amphibians
Arachnids
Birds
Clams
Crustaceans
Fishes
Insects
Mammals
Reptiles
Snails
Total
Species Vulnerability
Low
1
10
11
0
0
0
14
26
2
10
74
Moderate
5
0
6
0
3
0
3
6
0
0
23
High
3
0
1
65
3
31
3
1
4
16
127
Species Count
9
10
18
65
6
31
20
33
6
26
224
Source: U.S. EPA Analysis, 2013
For the purposes of estimating benefits, EPA excluded all species with low and moderate vulnerability
potentials based upon life history traits. For all species with high potential vulnerability, EPA conducted
further analyses to identify those species likely to be affected by the proposed ELGs, rather than all species
whose life histories make them vulnerable. High vulnerability species meeting the following criteria were
removed from further consideration:
> Species presumed to be extinct, including those not collected for a minimum of 30 years (e.g.,
Notunts trautmani).
> Endemic species living in waterbodies (e.g., isolated headwaters, natural springs, etc.) unlikely to be
affected by steam electric plant discharges (e.g., Gambusia georgei).
> Species protected by the ESA whose recovery plans i) do not include pollution or water quality issues
as factors preventing recovery, and ii) identify habitat destruction (due to damming, stream
channelization, water impoundments, wetland drainage, etc.) as a primary factor preventing recovery
(e.g., Erimystax cahni}.
> Listings due to non-native species introductions and/or hybridization with native or non-native
congeners (e.g., Oncorhynchus clarki somias)
> Listings where water quality issues are identified as the primary issue preventing recovery, but where
a specific industry or entity not within the scope of the proposed ELGs is identified as the culprit.
(e.g., Erimystax. cahni due to siltation from coal mining activity).
> Species about which very little is known, including geographic distribution.
After eliminating the T&E species meeting these criteria, EPA identified a total of 15 species whose recovery
may be enhanced by the proposed ELGs.
5.3.2 Assessing Benefits of T&E Species Improvements from the Proposed ELGs
The proposed ELGs have the potential to positively affect the recovery trajectory for 15 T&E species. For
each of these species, EPA estimated the magnitude of potential benefits by identifying inhabited waterbodies
likely to meet ambient water quality criteria (AWQC) for aquatic life as a consequence of the proposed ELGs
and comparing these areas to the overall area of habitat occupied by T&E species.
First, for each T&E species affected by steam electric plant discharges, EPA examined water quality in each
of the waterbodies inhabited by each T&E species under baseline conditions, and under conditions projected
to exist following implementation of the proposed ELGs. For each analyzed regulatory option, EPA identified
April 19, 2013
5-4
-------
Benefit and Cost Analysis for Proposed ELGs
5: Threatened and Endangered Species Benefits
waterbodies that 1) do not meet AWQC for wildlife under baseline conditions, but 2) have no wildlife AWQC
exceedances following implementation of the proposed ELGs. For seven species, there were no waterbodies
that met these conditions, leaving seven T&E fish species and one dragonfly species in 18 states and the
District of Columbia that may experience increases in population growth rates as a result of the proposed
ELGs (Table 5-3).
Table 5-3. T&E Species Whose Recovery May Benefit from the Proposed ELGs
Species
Acipenser brevirostrum
Acipenser oxyrinchus destoi
Etheostoma chermocki
Notropis cahabae
Percina aurolineata
Percina rex
Percina tanasi
Somatochlora hineana
Common Name
Shortnose sturgeon
Gulf sturgeon
Vermilion darter
Cahaba shiner
Goldline darter
Roanoke logperch
Snail darter
Mine's Emerald Dragonfly
State(s)
DE, DC, FL, MD, ME, NH,
PA, SC, and VT
LA and MS
GA
GA
GA
NC and VA
AL, SC, and TN
IL and IN
Source: U.S. EPA Analysis, 2013
EPA did not identify data sufficient to explicitly model population growth rates as a function of water quality
for any of these species. Therefore, to estimate proportionate population increases as a result of the proposed
ELGs, EPA identified the fraction of inhabited waterbodies that meet wildlife AWQC as a consequence of the
proposed ELGs. This fraction was used to estimate relative population changes in estimating the willingness-
to-pay (WTP) for T&E species recovery.
Estimating WTP for T&E Species Population Increases
5.4.1 Economic Valuation Methods
For several reasons, it is difficult to estimate the benefits of improving T&E species habitats resulting from
the proposed ELGs. First, data required to estimate the response of T&E populations to improved habitats are
rarely available. Second, the contribution of T&E species to ecosystem stability, ecosystem function, and life
history remains relatively unknown. Third, much of the wildlife economic literature focuses on commercial
and recreational benefits that are not relevant for many protected species (i.e., use values).There is a paucity
of economic data focused on the benefits of preserving habitat for T&E species because nonuse values
comprise the principal source of benefit estimates for most T&E species.
Analysis of nonuse benefits for T&E species affected by pollutant discharges from steam electric plants from
the proposed ELGs involves the following steps: 1) quantifying the impacts of pollutant discharges from
steam electric plants on T&E species and estimating the change in these impacts as a consequence of reducing
steam electric discharges; and 2) estimating an economic value of improving T&E habitats and populations as
a consequence of the proposed ELGs.
Benefit transfer involves extrapolating existing estimates of nonmarket values to the policy sites that
potentially differ from the original analytical situation in terms of geographic locations or affected species.
Ideally, the resource in question (i.e., T&E species), policy variables (e.g., change in species status, recovery
interval, population size, etc.), and the geographic location and benefitting population (i.e., defined human
population) are identical. Such a match rarely occurs. Despite discrepancies in these variables, however, a
benefit transfer approach can provide useful insights into the social benefits gained by reducing impacts on
T&E species.
April 19, 2013
5-5
-------
Benefit and Cost Analysis for Proposed ELGs
5: Threatened and Endangered Species Benefits
5.4.2 Estimating WTP for Improved Protection of T&E Species
To estimate the potential economic values of increased T&E species populations affected by the proposed
ELGs, EPA used a benefit transfer approach based on a meta-analysis of 31 stated preference studies eliciting
WTP for changes in T&E populations (Richardson and Loomis 2009). This meta-analysis is based on studies
conducted in the United States that valued threatened, rare, or endangered fish, bird, reptile, or mammal
species. Because the underlying meta-data does not contain insect valuation studies, EPA was unable to
monetize any benefits for potential population increases of Hine's Emerald Dragonfly as a result of the
proposed ELGs. Equation 5-1 contains the estimated WTP equation from the Richardson and Loomis (2009)
paper that EPA used to monetize potential population increases resulting from the proposed ELGs.
Equation 5-1.
In WTP (2006$) = -153.231 + 0.870 In CHANCESIZE + 1.256VISITOR + 1.020 FISH + 0.772MARINE +
0.826 BIRD - 0.603 In RESPONSERATE + 2.767 CONJOINT + 1.024 CHARISMATIC - 0.903 MAIL +
0.078 STUDYYEAR.
Table 5-4 lists the assigned variable values and definitions used in estimating per household WTP for
improved protection of T&E species resulting from the proposed ELGs.
Table 5-4. Independent Variable Assignments for the T&E Meta-Regression
Variable
Intercept
In ChangeSize
Visitor
Fish
Marine
Bird
Charismatic
Conjoint
In ResponseRate
Mail
StudyYear
Description
Intercept
Natural log of percentage change in the
population of the species of interest
Dummy variable indicating if survey
respondents are visitors rather than full-
time residents
Dummy variable indicating population
increases for fish species
Dummy variable indicating population
increases for marine mammals
Dummy variable indicating population
increases for bird
Dummy variable indicating a
charismatic species
Dummy variable indicating conjoint
method surveys
Natural log of the survey response rate
Indicates mail surveys
Year of study
Value
-153.231
Varies
0
1
0
0
Varies
0
3.912
0.851
1992
Explanation
-
Log of percentage change in fish
population
Primary beneficiaries are expected to be
full-time state residents
Only freshwater T&E fish species are
expected to be affected
Sturgeon species are considered
charismatic; minnow species are not
Default value from Richardson and
Loomis (2009) as only one underlying
meta-study used conjoint analysis; the rest
were CV studies
Mean value from Richardson and Loomis
(2009) following the Johnston et al. (2006)
approach where values for methodological
attributes are set at mean values from the
metadata
EPA does not currently have either species-specific estimates of the population effects of the proposed ELGs
or population models to estimate future population changes for the affected T&E species due to improved
aquatic habitat conditions. In the absence of such estimates, EPA used best professional judgment to assign a
range of potential improvements in the T&E populations based on the expected reductions in AWQC
exceedances under the post-compliance scenario. To estimate total population increases as a result of each
analyzed regulatory option, EPA assumed minimal increases in population size of 0.5, 1, or 1.5 percent. EPA
then weighted these population growth estimates within states by the proportion of reaches used by T&E
species expected to meet wildlife-based AWQC under each option. The natural log of these weighted
April 19, 2013
5-6
-------
Benefit and Cost Analysis for Proposed ELGs 5: Threatened and Endangered Species Benefits
population growth estimates under each scenario was used to assign a value to the ChangeSize parameter
estimate. EPA used the approach described in Johnston et al. (2006) and assigned mean study values from
Richardson and Loomis (2009) for the methodological variables (In ResponseRate, Mail, and Study Year).
EPA believes that its low (0.5 percent), medium (1.0 percent) and high (1.5 percent) estimates of population
growth for T&E species occurring because of the proposed ELGs are reasonable and plausible. This is
because few individuals must be saved to attain these growth estimates. For example, a T&E species with a
state-level population of 10,000 individuals (which is reasonable for threatened species, and likely an over-
estimate for endangered species and endemic species) and a population growth rate of 0 in the baseline would
achieve low, medium and high population increases if the proposed ELGs result in 2.1, 4.4 and 6.5 fewer
premature mortalities per year between 2017 and 2040, respectively. Additionally, the number of avoided
premature mortalities would scale with habitat utilization, which decreases the number of avoided premature
mortalities necessary to achieve these results.
Two states (South Carolina and Georgia) have multiple T&E species with potential population growth
resulting from the proposed ELGs. For these two states, EPA estimated benefits only for the species expected
to see the greatest proportional population increase. EPA did not sum WTP estimates for multiple species at
the state level, because the underlying meta-data are based on single species valuations and these estimates do
not account for species substitution or complementary effects. Therefore, these WTP estimates may be
affected by the availability of related species (Hoehn and Loomis, 1993). Further, WTP varies by species
characteristics such as size, charisma, and endemism (Loomis and Ekstrand, 1997; Morse-Jones et. al, 2012;
Metrick and Weitzman, 1996). The approach used by EPA for this analysis may understate WTP for
protection of multiple species as T&E valuation studies have shown that WTP is greater for the preservation
of multiple species as opposed to a single species (Stanley, 2005; Loomis and Ekstrand, 1997; Nunes and van
den Bergh, 2001; White et al. 1997; White et al. 2001). EPA was unable to account for benefits to multiple
species as the value of multiple species is not equal to the aggregated value of a single species. For example,
Stanley (2005) estimated an average annual household WTP value of $25 for the preservation of the riverside
fairy shrimp in Orange County CA, and an average annual household WTP value of $52 to preserve all 32
locally endangered species. Moreover, wildlife valuation studies have shown that there are diminishing
marginal benefits for wildlife preservation (Morse-Jones et. al, 2012; Rollins and Lyke 1998).
Because population growth was assessed at the state level, EPA was unable to attribute benefits to a specific
steam electric plant and therefore to account for the timing of benefits based on the assumed control
technology implementation year. EPA assumed that benefits begin accruing in 2019 for all states. This year is
the midpoint of the period of 2017 through 2021 when plants are assumed to implement control technologies
to comply with the revised effluent limits and standards.
For each state, EPA estimated household WTP for improved protection of T&E species resulting from the
proposed ELGs using Equation 5-1 and the independent variable assignments presented in Table 5-4. EPA
estimated total annual benefits for the years between 2019 and 2040 by multiplying household WTP by the
number of households in each state for a given year. EPA then calculated the value of benefits for each year
and the annualized total WTP values for each state using 3 percent and 7 percent discount rates.
5.5 Results
Table 5-5 and Table 5-6 present the annualized total benefits calculated using a 3 percent and 7 percent
discount rate, respectively. Annualized benefits for Options 1 through 5 range from $3.9 million to $47.3
million with a 3 percent discount rate ($3.2 million to $39.5 million with a 7 percent discount rate), depending
on the regulatory option and estimated population increase. Under Option 3, eight states have potential
increases in T&E species populations, with annualized benefits of $10.0 million with a 3 percent discount rate
April 19, 2013 5-7
-------
Benefit and Cost Analysis for Proposed ELGs 5: Threatened and Endangered Species Benefits
($8.4 million with a 7 percent discount rate). Benefits of Option 4 are higher and more broadly distributed,
with 17 states having potential increases in T&E species populations and annualized benefits at $33.3 million
(medium population increase estimate) being more than three times larger than those of Option 3.
As for the other benefit categories discussed in this report, EPA expects the benefits of Options 3a and 3b to
be below those of Option 3, whereas Option 4a will have benefits between those of Options 3 and 4.
April 19, 2013 5-8
-------
Benefit and Cost Analysis for Proposed ELGs
5: Threatened and Endangered Species Benefits
Table 5-5. Estimated Annualized Benefits to T&E Species from WQ Improvements (3% Discount Rate, Millions 2010$)a
State
AL
DC
DE
GA
FL
LA
MA
MD
ME
MS
NC
NH
PA
SC
TN
VA
VT
Total
Option 1
Low
$0.47
-
-
$0.80
-
$1.46
-
-
-
$0.46
$0.36
-
-
-
-
$0.29
-
$3.85
Medium
$0.87
-
-
$1.46
-
$2.66
-
-
-
$0.85
$0.66
-
-
-
-
$0.54
-
$7.03
High
$1.23
-
-
$2.07
-
$3.79
-
-
-
$1.21
$0.93
-
-
-
-
$0.77
-
$10.00
Option 2
Low
$0.47
-
-
$0.80
-
$1.46
-
-
-
$0.46
$0.36
-
-
-
-
$0.29
-
$3.85
Medium
$0.87
-
-
$1.46
-
$2.66
-
-
-
$0.85
$0.66
-
-
-
-
$0.54
-
$7.03
High
$1.23
-
-
$2.07
-
$3.79
-
-
-
$1.21
$0.93
-
-
-
-
$0.77
-
$10.00
Option 3
Low
$0.47
-
-
$0.80
-
$1.46
-
-
-
$0.46
$0.36
-
-
$1.03
$0.35
$0.54
-
$5.47
Medium
$0.87
-
-
$1.46
-
$2.66
-
-
-
$0.85
$0.66
-
-
$1.88
$0.65
$0.99
-
$10.00
High
$1.23
-
-
$2.07
-
$3.79
-
-
-
$1.21
$0.93
-
-
$2.68
$0.92
$1.40
-
$14.23
Option 4
Low
$0.67
$0.07
$0.15
$0.80
$3.89
$2.66
$0.86
$0.60
$0.80
$1.21
$1.46
$0.47
$1.36
$1.03
$0.35
$1.60
$0.22
$18.20
Medium
$1.23
$0.12
$0.28
$1.46
$7.11
$4.86
$1.57
$1.10
$1.46
$2.21
$2.66
$0.86
$2.49
$1.88
$0.65
$2.93
$0.40
$33.26
High
$1.75
$0.17
$0.40
$2.07
$10.11
$6.92
$2.23
$1.56
$2.07
$3.14
$3.79
$1.22
$3.54
$2.68
$0.92
$4.17
$0.56
$47.32
Option 5
Low
$0.67
$0.07
$0.15
$0.80
$3.89
$2.66
$0.86
$0.60
$0.80
$1.21
$1.46
$0.47
$1.36
$1.03
$0.35
$1.60
$0.22
$18.20
Medium
$1.23
$0.12
$0.28
$1.46
$7.11
$4.86
$1.57
$1.10
$1.46
$2.21
$2.66
$0.86
$2.49
$1.88
$0.65
$2.93
$0.40
$33.26
High
$1.75
$0.17
$0.40
$2.07
$10.11
$6.92
$2.23
$1.56
$2.07
$3.14
$3.79
$1.22
$3.54
$2.68
$0.92
$4.17
$0.56
$47.32
' Low, Medium, and High increases in population are based on a 0.5
the T&E species in the state. EPA weighted these population growth
ELGs.
Source: U.S. EPA Analysis, 2013
1, and 1.5 percent increase in T&E species population growth resulting from WQ improvements in all reaches used by
estimates by the percent of all reaches used by T&E species that are expected to improve as a result of the proposed
April 19, 2013
5-9
-------
Benefit and Cost Analysis for Proposed ELGs
5: Threatened and Endangered Species Benefits
Table 5-6. Estimated Annualized Benefits to T&E Species from WQ Improvements (7% Discount Rate, Millions 2010$)a
State
AL
DC
DE
GA
FL
LA
MA
MD
ME
MS
NC
NH
PA
SC
TN
VA
VT
Total
Option 1
Low
$0.40
-
-
$0.66
-
$1.22
-
-
-
$0.39
$0.30
-
-
-
-
$0.25
-
$3.22
Medium
$0.73
-
-
$1.22
-
$2.23
-
-
-
$0.71
$0.55
-
-
-
-
$0.45
-
$5.88
High
$1.03
-
-
$1.73
-
$3.17
-
-
-
$1.01
$0.78
-
-
-
-
$0.64
-
$8.37
Option 2
Low
$0.40
-
-
$0.66
-
$1.22
-
-
-
$0.39
$0.30
-
-
-
-
$0.25
-
$3.22
Medium
$0.73
-
-
$1.22
-
$2.23
-
-
-
$0.71
$0.55
-
-
-
-
$0.45
-
$5.88
High
$1.03
-
-
$1.73
-
$3.17
-
-
-
$1.01
$0.78
-
-
-
-
$0.64
-
$8.37
Option 3
Low
$0.40
-
-
$0.66
-
$1.22
-
-
-
$0.39
$0.30
-
-
$0.86
$0.30
$0.45
-
$4.57
Medium
$0.73
-
-
$1.22
-
$2.23
-
-
-
$0.71
$0.55
-
-
$1.57
$0.54
$0.82
-
$8.36
High
$1.03
-
-
$1.73
-
$3.17
-
-
-
$1.01
$0.78
-
-
$2.23
$0.77
$1.17
-
$11.90
Option 4
Low
$0.57
$0.06
$0.13
$0.66
$3.23
$2.23
$0.72
$0.50
$0.67
$1.01
$1.22
$0.39
$1.15
$0.86
$0.30
$1.34
$0.18
$15.20
Medium
$1.03
$0.10
$0.23
$1.22
$5.90
$4.07
$1.32
$0.92
$1.22
$1.85
$2.22
$0.71
$2.09
$1.57
$0.54
$2.44
$0.33
$27.79
High
$1.47
$0.15
$0.33
$1.73
$8.40
$5.80
$1.88
$1.31
$1.74
$2.64
$3.16
$1.01
$2.98
$2.23
$0.77
$3.48
$0.47
$39.54
Option 5
Low
$0.57
$0.06
$0.13
$0.66
$3.23
$2.23
$0.72
$0.50
$0.67
$1.01
$1.22
$0.39
$1.15
$0.86
$0.30
$1.34
$0.18
$15.20
Medium
$1.03
$0.10
$0.23
$1.22
$5.90
$4.07
$1.32
$0.92
$1.22
$1.85
$2.22
$0.71
$2.09
$1.57
$0.54
$2.44
$0.33
$27.79
High
$1.47
$0.15
$0.33
$1.73
$8.40
$5.80
$1.88
$1.31
$1.74
$2.64
$3.16
$1.01
$2.98
$2.23
$0.77
$3.48
$0.47
$39.54
' Low, Medium, and High increases in population are based on a 0.5,1, and 1.5 percent increase in T&E species population growth resulting from WQ improvements in all reaches used by
the T&E species in the state. These population growth estimates were then weighted by the percent of all reaches used by T&E species that are expected to improve as a result of the
proposed ELGs.
Source: U.S. EPA Analysis, 2013
April 19, 2013
5-10
-------
Benefit and Cost Analysis for Proposed ELGs
5: Threatened and Endangered Species Benefits
5.6 Limitations and Uncertainties
Table 5-7 summarizes the caveats, omissions, biases, and uncertainties known to affect EPA's estimates of
the benefits to T&E species and indicates the direction of the potential bias.
—Table 5-7. Uncertainties in the Analysis of T&E Species Benefits
Issue
Change in T&E populations
due to the effect of steam
electric ELGs is uncertain
Only those T&E species
listed as threatened or
endangered on the
Endangered Species Act are
included in the analysis
A 5 -mile buffer was used to
screen T&E species. This is a
very rough approximation of
a distance where benefits
may be most apparent
Benefit estimates do not
include monetized values for
potential population increases
in Mine's Emerald Dragonfly
(Somatochlora hineana)
Benefit estimates are likely to
include only a subset of
species that may be affected
Benefit transfer introduces
uncertainties
Ecological roles filled by
T&E species
Overlap between WTP
estimates for T&E species
and the WTP estimates for
improvements in water
quality
Effect on Benefits
Estimate
Uncertain
Underestimate
Uncertain
Underestimate
Underestimate
Uncertain
Underestimate
Overestimate
Notes
Data necessary to quantitatively estimate population changes
are unavailable. Therefore, EPA used best professional
judgment to assess reasonable changes in T&E populations.
Actual effects of the proposed ELGs may be larger or smaller
than projected changes in the population of T&E species
assumed in this analysis.
The databases used to estimate benefits to T&E species
exclude all species considered threatened or endangered by
scientific organizations but not protected by the ES A. The
magnitude of the underestimate is likely to be significant, since
the proportion of imperiled fish and mussel species is high
(e.g., Jelks et al 2008, Taylor et al 2007)
Effects are likely to be species and waterbody specific. For
some species, 5 miles may be an underestimate of the distance
at which acute effects of pollution are felt. For other species, 5
miles may be an overestimate.
It is likely that population increases in Somatochlora hineana
have value to the public. In addition to bequest, altruistic, and
existence values, dragonflies may have aesthetic or cultural
values. Dragonflies also provide beneficial ecological services.
They are voracious insectivores that prey on mosquitoes, flies,
and other small insects. The estimated annual benefits of pest
control attributable to insects are $4.5 billion in the United
States (Losey and Vaughan, 2006).
EPA was conservative when applying benefit transfer of values
for species potentially affected. Water quality issues may be
important to species recovery even if not listed explicitly in
species recovery plans.
Value may over- or understate true WTP values (See Section 4
for more details).
WTP values are unlikely to include changes to food-webs and
ecosystem stability as a consequence of the restoration (or loss)
of T&E species.
There may be some overlap between WTP estimates for T&E
species and the WTP estimates for improvements in water
quality because WTP values for improvements in water quality
may inherently include benefits to T&E species. However,
none of the studies in EPA's meta-analysis of WTP for water
quality improvements specifically mentioned or otherwise
prompted respondents to include benefits to T&E species
populations (see Chapter 4); therefore, any overlap is likely to
be minimal.
April 19, 2013
5-11
-------
Benefit and Cost Analysis for Proposed ELGs
5: Threatened and Endangered Species Benefits
Table 5-7. Uncertainties in the Analysis of T&E Species Benefits
Issue
WTP estimates represent
population increases for only
one species
WTP estimates do not take
into account possible
substitution for effects for
similar species.
Effect on Benefits
Estimate
Underestimate
Overestimate
Notes
WTP for multiple species may be greater than WTP to preserve
a single species (Stanley, 2005, Nunes and van den Bergh,
2001, White et al. 1997, White et al. 2001; Loomis and
Ekstrand, 1997). However, only a small number of benefiting
states have potential population increases in multiple species.
Also, EPA's analysis values species that are likely to
experience the largest population increase, Therefore, while
estimating WTP for a single species may underestimate
benefits, this underestimate is likely to be small.
WTP estimates may be affected by the availability of related
species (Hoehn and Loomis, 1993), however Kahneman and
Knetsch (1992) argue that substitution effects may not apply to
the values associated with endangered species, because their
uniqueness is the essence of their existence value.
April 19, 2013
5-12
-------
Benefit and Cost Analysis for Proposed ELGs
6: Benefits from Improved Groundwater Quality
6 Benefits from Groundwater Quality Improvements
Steam electric plants manage coal combustion residuals (CCR) such as flue gas desulfurization (FGD) solids,
fly ash, and bottom ash through either wet or dry handling. For plants that use wet handling, the waste is
typically sluiced to one or more surface impoundment (e.g., settling pond), where the solids settle out of the
water. In addition to solids, these impoundments typically contain water with high concentrations of steam
electric pollutants, including dissolved metals. These pollutants may leach into groundwater aquifers,
degrading water quality and potentially creating health hazards to households drawing drinking water from
affected aquifers. The health hazards may arise from water consumption or dermal contact with pollutants in
the water, and include a wide range of human health effects (e.g., cancer, kidney damage, nervous system
damage and others). The EA (U.S. EPA, 2012a) describes the health effects caused by steam electric
pollutants in more detail.
EPA expects that some of the regulatory options would eliminate the future leaching of steam electric
pollutants from surface impoundments to groundwater aquifers by prompting changes in the way steam
electric plants handle CCR. The proposed ELGs are expected to reduce contaminations in both public and
private drinking water wells. However, because public drinking water systems must treat water to reduce
pollutant concentrations below maximum contaminant levels (MCLs) the proposed ELGs are not expected to
generate significant benefits to the population served by public drinking water sources. Therefore, EPA's
analysis of benefits from improved groundwater quality focused only on households relying on private wells
that draw water from the aquifers located in the vicinity of steam electric plants. To estimate the monetary
value of benefits from reduced groundwater contamination EPA used results of Poe et al.'s (2001) meta-
analysis.54
Methodology and Data
This section describes the methodology and data used to estimate 1) the baseline groundwater quality in the
aquifers affected by leachate from steam electric impoundments; 2) changes in groundwater pollutant
concentrations resulting from the proposed ELGs; 3) the number of households potentially exposed to
impoundment leachate; and 4) the households' willingness-to-pay (WTP) for improving the quality of
groundwater used for drinking water supply.
6.1.1 Baseline Water Quality
EPA modeled baseline groundwater concentrations of steam electric pollutants that would be reduced by the
proposed ELGs (including lead, mercury, thallium, cadmium, selenium, arsenic, and chromium) within a one-
mile radius from steam electric plant impoundments. Within the one-mile radius, EPA used 0.2 mile bands to
account for differences in groundwater concentrations of pollutants for drinking water wells located in
varying distances from an impoundment.
6.1.2 Water Quality Improvements
EPA used drinking water MCLs55 as a benchmark to assess groundwater contamination in the aquifers
affected by steam electric plant impoundment leachate. An MCL for drinking water is the highest level of a
54 The meta-analysis is defined as a "statistical analysis of a large collection of results for individual studies for the
purpose of integrating findings" (Poe et al, 2001; p. 138).
55 See U.S. EPA (2012a) for details.
April 19, 2013
6-1
-------
Benefit and Cost Analysis for Proposed ELGs 6: Benefits from Improved Groundwater Quality
contaminant that is allowed in drinking water. The MCL is based on the MCL Goal (MCLG), which is the
level of a contaminant in drinking water below which there is no known or expected health risk. EPA sets the
MCL as close to the MCLG as possible, considering the best available treatment technologies and costs. EPA
compared the modeled baseline concentrations of steam electric pollutants in groundwater to MCL for
drinking water to identify those aquifers exceeding MCLs in the baseline.
Closing the impoundment (as may occur if the impoundment is no longer used to manage CCR waste under
the proposed ELG regulatory options) would remove the source of future contamination for groundwater
aquifers affected by leachate from steam electric impoundments and thus improve future groundwater quality.
EPA was not able to model changes in groundwater concentrations that would result from implementation of
the proposed ELGs due to data and practical limitations. To assess the potential magnitude of benefits from
the proposed ELGs, EPA used a simplified approach in which EPA assumed that water quality in affected
aquifers that exceed MCLs under the baseline would improve by 70 percent given that the source of
contamination is removed.56 EPA assumed that groundwater quality in the affected aquifers that do not exceed
MCLs under the baseline would improve by 30 percent, as a result of impoundment closures.
Some impoundments are expected to see a reduction, but not the complete elimination, of the CCR waste they
handle under the regulatory options (TDD; U.S. EPA, 2013b). For these impoundments, EPA assumed that
the benefits would be proportional to the reduction in the amount of CCR waste managed by the
impoundment (for example, benefits are proportional to the share of ash that is fly ash for regulatory options
that address fly ash only).
6.1.3 Affected Households
As noted above, EPA expects that, by reducing pollutant leaching from impoundments, the proposed ELGs
would benefit households relying on private drinking water wells that draw water from groundwater aquifers
in the vicinity of impoundments.
EPA utilized a synthetic population database developed by RTI International to estimate the number of adults
and children living in 0.2 mile radial distance bands surrounding impoundments, up to 1 mile away.57 The
1990 Census provides the most recent data on groundwater users and indicates that, nationally, approximately
15 percent of households utilize private groundwater wells. For this analysis, EPA applied the site-specific
ratio of households using groundwater reflected in the 1990 Census data (the most recent Census dataset
including private well data) at the block group level to the RTI synthetic population data (which are based on
the 2000 Census) to determine the share of the impoundment-specific population using private groundwater
wells.
To estimate the number of households in each distance band, EPA divided the number of persons residing in
each distance band by the state-specific average number of persons per household (based on U.S. Census
Bureau, 2010a).58
56 Although the decommissioning and closure of the impoundment would remove the source of contamination for these
aquifers, EPA did not assume that they would be improved by 100 percent (i.e., complete removal of contamination)
because there may be some other nearby sources of contamination, and it may take a very long time for legacy
contaminations from the impoundments to be completely eliminated.
57 The synthetic population database provides higher resolution population estimates than typical census block data. For
details on the database, see Wheaton et al. (2009).
58 EPA does not have data to determine the number of households for 4 to 18 plants with impoundments affected by the
proposed ELG, depending on the regulatory option. For these plants, EPA applied the average per-plant benefits of
plants with sufficient data.
April 19, 2013 &
-------
Benefit and Cost Analysis for Proposed ELGs 6: Benefits from Improved Groundwater Quality
6.1.4 Monetary Values of Groundwater Quality Improvements
EPA used the results of Poe et al.'s (2001) meta-analysis of groundwater valuation studies to estimate WTP
for groundwater quality improvements resulting from the proposed ELGs. Poe et al. (2001) compiled meta-
data from 13 stated preference studies that elicit household WTP values for improvements in groundwater
quality or quantity. In these studies, the estimated household's WTP range from $80 and $1,788 (updated to
2010 dollars using the Consumer Price Index (CPI)). Poe et al. (2001) grouped the studies into three
categories:
> Studies that place values on specific changes in water quality,
> Studies that evaluate water quality changes presented in more general terms, and
> Studies that frame the valuation problem as affecting water quantity (as opposed to quality).
Poe et al. (2001) used meta-data to model WTP for groundwater improvements as a function of "core
economic variables" (change in supply, changes in contamination probability, use versus nonuse values,
income, and prices of substitutes), and "methodological variables" (whether cancer is mentioned as a concern,
whether locality is emphasized, percent of population on public water supply versus private wells, elicitation
method, and analytical methods).59 The authors report three separate regression functions: 1) core economic
variables only, 2) "full complete," which includes all core economic and methodological variables, and 3)
"short complete," which retains all core economic variables, but removes the least significant methodological
variables. To estimate per household WTP for groundwater quality improvements that would result from the
proposed ELGs, EPA applied the "full complete" regression function shown in Equation 6-1.
Equation 6-1.
WTP (1997$) = -491.1078 + -440.3059(D(use)) + -174.140l(D(Aprob)) + 1085.1610(Aprob) +
235.4335(D(Asupp(y)) + 289.1667(Asupp(y) + -83.0452(D($sub)) + 8.4215(/(t/iou)) + 186.8805(D(canc)) +
-121.1955(D((ocaO) + 210.4727(puWic%) + 73.8525(D(0£a/tDC)) + -93.5884(D(pcard)) +
185.2880(D(PC)) + -181.0027(D(DCCam)) + 227.4003(D(DCmiO)
This function allows EPA to forecast WTP for improvements in groundwater quality based on assigned
values to model variables that are chosen to represent a resource change (probability of contamination, kprob)
and other methodological variables. EPA assigned a value to each model variable following general guidance
for meta-analysis benefit transfer from Bergstrom and Taylor (2006). Table 6-1 describes the explanatory
variables, estimated coefficients, and assigned variable values used in this analysis. EPA assigned the value to
the change in probability of groundwater contamination variable (kprob) using the following assumptions:
> EPA assumed a 70 percent change in the probability of contamination for aquifers that exceed MCLs
in the baseline, and
> EPA assumed a 30 percent change for aquifers that do not exceed MCLs in the baseline.
EPA used state-specific median household income from the U.S. Census Bureau (2010a) to assign the income
value (Income) for the households associated with each impoundment.
EPA set the use variable [D(use)} to zero to capture both use and nonuse values, and set the change in supply
[D(ksupply)} and quantity of change in supply (^supply) variables to zero to isolate the effect of a change in
groundwater quality (rather than quantity). EPA also set the price of substitutes [D($ sub)] value to zero since
the substitute drinking water sources (such as bottled water) and their prices are not considered in this
59 The authors used a weighted least square procedure with standard errors derived using the Huber-White consistent
covariance estimator. This approach treats each study as the equivalent of a sample cluster with the potential for
heteroskedasticity (i.e. differences in variance across clusters).
April 19, 2013
-------
Benefit and Cost Analysis for Proposed ELGs
6: Benefits from Improved Groundwater Quality
analysis. EPA set the percent on public water supply (Public%) variable to zero since the benefits are
expected to accrue only to private well users. EPA assigned zero values to the methodology variables
(D(OEaftDC), D(pcard), D(DC), D(DC-cam), and D(DC-util)). This assignment results in the application of
the default methodology assumption of an open-ended contingent valuation question (with no dichotomous
choice).The majority of studies in the meta-data used an open-ended survey format.
Table 6-1. Estimated Meta-Regression for Household WTP for Groundwater Quality Improvements
Variable Type
Core Economic
Variables
Methodological
Variables
Variable
Constant
D(use)
D(Aprob)
Aprob
D(Asupply)
Asupply
D($ sub)
I(thou)
D(canc)
D(local)
Public%
D(OEaftDC)
D(pcard)
D(DC)
D(DC-cam)
D(DC-util)
Coefficient
-491.1078
-440.3059
-174.1401
1085.1610
235.4335
289.1667
-83.0452
8.4125
186.8805
-121.1955
210.4727
73.8525
-93.5884
185.2280
-181.0027
227.4003
Assigned
Value
0
1
varies (0.3
or 0.7)
0
0
0
varies
1
0
0
0
0
0
0
0
Description
Focus on filtration devices or containment strategies;
excluding non-use values; set to zero to reflect
residents' use and nonuse values of groundwater
Whether a change in the probability of contamination
was specified; set to 1 to reflect that site-specific
changes were specified
Site specific change in probability of contamination (0.7
if modeled concentration exceeds a human health
criteria; 0.3 if not)
Whether a change in supply was specified; set to zero
since a change in supply is not expected
Change in groundwater supply; set to zero since a
change in supply is not expected
Whether the price of substitutes was specified; set to
zero since substitute prices are not considered in this
analysis
Site specific median household income (in thousands)
Whether cancer was mentioned as a concern; set to 1
because steam electric discharges include arsenic which
is a carcinogen
Whether a specific locality was emphasized; set to 0
because benefits extend to other areas surrounding
impoundments
Percentage of respondents on public water supplies; set
to zero since benefits analysis is restricted to private
well users
Open ended value elicitation following dichotomous
choice; set to zero to reflect default survey design values
Payment card; set to zero to reflect default survey design
values
Dichotomous choice; set to zero to reflect default survey
design values
Cameron approach; set to zero to reflect default survey
design values
Utility theoretical approach; set to zero to reflect default
survey design values
Source: Poe et al. (2001); "foil complete" model
EPA used the meta-regression to estimate per household WTP for groundwater quality improvements for
households residing in 0.2-mile radial bands within 1 mile of impoundments affected by the proposed ELGs.
EPA then calculated the total benefits associated with reduced contamination of groundwater by multiplying
the band-specific household WTP by the number of households residing in the 0.2 mile bands surrounding
April 19, 2013
6-4
-------
Benefit and Cost Analysis for Proposed ELGs
6: Benefits from Improved Groundwater Quality
each impoundment and summing across all bands of impoundments expected to close under a given
regulatory option.
6.2
EPA estimated that, on average, 43 households use private wells within 1-mile of affected aquifers. As
described above, EPA assumed that the proposed ELGs would reduce the probability of groundwater
contamination in aquifers associated with 147 to 623 impoundments by 30 percent or 70 percent60. The
estimated household's WTP for groundwater improvements varied by aquifer depending on income and the
expected change in probability of groundwater contamination (i.e., 70 percent or 30 percent). The average
WTP for reducing the probability of groundwater contamination across all affected aquifers is approximately
$442 per household, with a range of $111 (for aquifers with low income populations and smaller
improvements in quality) to $976 (for aquifers with high income populations and larger improvements in
quality). As shown in Table 6-2, the estimated annual benefits from reducing the probability of groundwater
contamination are $1.6 million and $6.5 million, respectively for Options 3 and 4, using a 3 percent discount
rate ($1.4 million and $5.5 million using a 7 percent discount rate). EPA expects the annualized benefits of
Options 3a and 3b to be less than those of Option 3, (i.e., less than $0.7 million, at 3 percent discount rate)
and the annualized benefits of Option 4a to be between those of Options 3 and 4 (i.e., between $1.6 and $6.5
million, at 3 percent discount rate) .
Table 6-2. Estimated Annualized Benefits of Reduced Probability of Groundwater Contamination from
Impoundment Leaching (Millions; 2010$)
Regulatory Option
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Number of
Impoundments"
147
b
147
b
239
c
623
623
3% Discount Rate
$0.65
b
$0.65
b
$1.64
c
$6.49
$6.49
7% Discount Rate
$0.56
b
$0.56
b
$1.40
c
$5.52
$5.52
a. Impoundments managing reduced CCR waste quantities as a result of the proposed ELG option (out of a total of 1,070 impoundments at steam
electric plants).
b. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Options 3a and 3b to be less than
those of Option 3, with values for Option 3a being lower than Option 3b. EPA does not know how the benefits of Options 3a and 3b
compare to those of Options 1 and 2.
c. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Option 4a to be between those of
Options 3 and 4.
Source: U.S. EPA Analysis, 2013
6.3 Limitations and Uncertainties
Table 6-3 summarizes the limitations and uncertainties in the analysis of benefits associated with the reduced
probability of groundwater contamination from impoundment leachate. While the degree of uncertainty in the
estimated WTP for improved groundwater quality is greater than desirable, this approach allowed EPA to
assess whether benefits are likely to occur and assess the potential magnitude of benefits to the affected
groundwater users. EPA believes that benefits to groundwater are likely to occur and therefore this approach
provides a useful insight into the benefits of the proposed ELGs.
60 Approximately 40 percent of 0.2-mile radial bands around affected impoundments may be improved by 70 percent,
while 60 percent may be improved by 30 percent.
April 19, 2013
6-5
-------
Benefit and Cost Analysis for Proposed ELGs
6: Benefits from Improved Groundwater Quality
Table 6-3. Uncertainties in the Analysis of Groundwater Quality Benefits
Uncertainty/Assumption
Background concentrations of steam
electric pollutants are not
considered in this analysis, nor are
the design of impoundments or the
lag between reductions in leachate
loads and water quality
improvements.
EPA used current MCL
exceedances as a benchmark to
determine whether the probability of
contamination would be reduced by
70 percent (for aquifers exceeding
the health criteria in the baseline) or
by 30 percent (for aquifers not
exceeding the health criteria in the
baseline) as a result of the proposed
ELGs.
EPA assumed that the benefits
accruing to aquifers surrounding
impoundments affected by the
proposed ELGs would be
proportional to the reduction in the
quantity of CCR waste managed by
the impoundment.
EPA relied on 1990 Census data on
the site-specific share of the
population using private wells.
To estimate benefits associated with
decreased groundwater
contamination, EPA relied on a
meta-analysis of groundwater
valuation studies (Poe et al., 2001).
Effect on Benefits
Estimate
Overestimate
Uncertain
Uncertain
Overestimate
Uncertain
Notes
The analysis does not account for the presence of a liner
or other design characteristics that reduce leaching.
Even if leaching of steam electric pollutants from an
unlined impoundment is reduced or eliminated as a
result of the proposed ELGs, background concentrations
may persist for years. The benefits of reduced
groundwater leaching may be overestimated in the short
term.
Due to data limitations on the aquifers surrounding
affected impoundments, it is not possible to model
changes in pollutant concentrations resulting from
implementation of the proposed ELGs. The overall
effect of this assumption on benefit estimates is
uncertain.
The relationship between the quantity of CCR waste
managed by the impoundments, pollutant loads in the
leachate, and groundwater quality in surrounding
aquifers may not be linear.
The fraction of households using private wells has
declined since 1990; as such this assumption may result
in an overestimate of the number affected households.3
EPA used the best data available in the absence of more
recent location-specific data.
Poe et al., (2001) reported wide variations in WTP for
groundwater quality improvements, as well as variations
in the methods used to elicit and present those values.
Therefore, actual household WTP for groundwater
quality improvements may be higher or lower.
a. For example, data from the 2009 American Housing Survey indicates that approximately one out of every 12 homes, or 8.3 percent of
households, rely on private wells, while the 1990 Census data indicated that the share of households using private wells was closer to 15 percent.
Data from the American Housing Survey is not detailed enough for use in this analysis.
April 19, 2013
6-6
-------
Benefit and Cost Analysis for Proposed ELGs
7: Benefits from Avoided Impoundment Failures
Benefits from Avoided Impoundment Failures
The operational changes prompted by the proposed ELGs are expected to cause some plant owners to reduce
their reliance on impoundments to handle coal combustion residuals. These changes would affect the future
probability and/or magnitude of impoundment failures and the resulting accidental, and sometimes
catastrophic releases, of coal combustion residuals (CCR). Benefits from the reduced risk of impoundment
failures include avoided cleanup costs, environmental damage, and transaction costs.
EPA's analysis of the monetary value of avoided impoundment failures is based on the identification of
impoundments that would be affected by each of the regulatory options. EPA estimated benefits from avoided
impoundment failures based on the probability of a release for each impoundment in a given year, the
capacity of the impoundment, and the cost (including cleanup costs, natural resource damages (NRD) and
transaction costs) per gallon of CCR slurry spilled. Benefits are calculated as the difference between expected
failure costs for a regulatory option and expected failure costs under baseline conditions, over the period of
2014 through 2040.
7.1 Methods and Data
This section describes the methodology and data used to determine the baseline and post-compliance
probability of impoundment failures, assign costs to failure events, and estimate the total present and
annualized values of benefits from avoided impoundment failures resulting from the proposed ELGs.
7.1.1 Failure Probability
EPA determined future probability of impoundment failures based on historical trends. EPA used data from a
survey of impoundments conducted in support of regulations being developed by EPA's Office of Resource
Conservation and Recovery (ORCR; see U.S. EPA, 2010d) governing the disposal of CCR in impoundments.
The survey sought information from plant owners and operators about impoundment releases that occurred
during the period of 2000 through 2009 for a total of 726 impoundments. Responses to the survey provided
historical data on 38 releases that occurred during the 10-year period; additionally, some plant owners
included information about 4 release incidents that occurred during the period of 1995 through 1999 at 4
impoundments.61 EPA used this data to estimate an overall average annual failure rate, as shown in Equation
7-1.
Equation 7-1.
Annual Failure Rate =
Number of Releases
Number of Impoundments x Number of Years
42
(4x15)+ (722 xlO)
= 0.58%
EPA applied this baseline average failure rate to estimate an expected number of releases for the universe of
1,070 impoundments at steam electric plants over the years 2019 through 2040. This expected number of
releases is then used to calculate the expected failure costs under the baseline and each analyzed regulatory
61 Thus, the survey responses provided 10 years of data for 722 impoundments and 15 years of data for 4 impoundments,
for a total of 7,280 observations (4 impoundments times 15 years of data, plus 722 impoundments times 10 years of
data).
April 19, 2013
7-1
-------
Benefit and Cost Analysis for Proposed ELGs 7: Benefits from Avoided Impoundment Failures
option. Note that this analysis does not account for the effect of best management practices (BMPs) -
including integrity inspections and preventive maintenance - that would help reduce the probability of
impoundment failures.62
Under all scenarios (baseline and each analyzed regulatory option), the risk of failure was assumed to be zero
for years 2014 through 2018 due to integrity site assessments conducted by EPA in 2009 through 2012, which
are expected to prevent all failures for the first five years after the recommended "action plan" to improve
impoundment structures is completed (2014 through 2018).
As described above, the baseline average failure rate is assumed to be uniform across all impoundments and
years, irrespective of impoundment characteristics. In practice, the probability of a failure may depend on
impoundment characteristics, including the amount of CCR managed, and may therefore change as a result of
the proposed ELGs. To test the sensitivity of the benefit estimates to alternative assumptions about
impoundment failure probability, EPA also used the survey data to develop a statistical model of the
probability of impoundment failure as a function of the impoundment capacity and age. This approach
resulted in impoundment-specific failure rates for each impoundment in each year of the analysis. Appendix G
presents the results of this sensitivity analysis. The sensitivity analysis highlights the potential magnitude of
benefits from avoided impoundment failures (sensitivity benefits are 2.5 to 2.7 times higher than estimated
below); however, because of the limited data available to develop and validate the statistical model, EPA did
not use the model in the primary benefit estimates presented in this chapter but instead relied on historical
failure rates.
7.1.2 Capacity Factor
The impoundment survey conducted by ORCR (see U.S. EPA, 2010d) and described above provide data on
the spill volume and impoundment capacity for 15 of the 42 documented releases. These data show spill
volumes ranging from less than 0.01 percent to 37.9 percent of impoundment capacity, with an average of
6.45 percent. EPA used this average ratio (spill volume is 6.45 percent of impoundment capacity) to estimate
the average volume spilled in expected releases from each of the 1,070 Steam Electric plant impoundments,
i.e., in the event of a failure, EPA assumed that a volume of CCR equal to 6.45 percent of the impoundment
capacity would be released.
7.1.3 Failure Costs
The following sections discuss three categories of costs associated with impoundment failures: cleanup,
natural resources damages, and transaction costs. All dollar values are presented in year 2010 dollars.63
7.7.3.7 Cleanup Costs
EPA estimated per-gallon cleanup costs based on three historical impoundment failures. The average unit cost
associated with these historical incidents is $0.62 per gallon spilled.
> The Massey Coal slurry spill involved the collapse of a 2,000 acre-foot (651.7 million gallon) surface
impoundment on top of an idled underground mine, causing 250 million gallons of slurry to spill into
two nearby streams. The cleanup costs of this spill were $65 million (Pincock, et al., 2001; Geotimes,
2003), or approximately $0.26 per gallon spilled.
2 The preamble accompanying this proposed rule describes the BMPs.
3 As needed, costs were updated to 2010 dollars using the Construction Cost Index from the Engineering News Report.
April 19, 2013
-------
Benefit and Cost Analysis for Proposed ELGs 7: Benefits from Avoided Impoundment Failures
> The Martins Creek spill involved the discharge of 100 million gallons of slurry over the course of 3
days, resulting from the failure of a wooden stop log. Cleanup costs were approximately $44 million
(Nixon, 2006; Barker, 2009; Dewberry and Davis, 2009), or approximately $0.44 per gallon spilled.
> The TVA Kingston spill involved the release of approximately 1.1 billion gallons of slurry. The
cleanup costs were approximately $1.27 billion (TVA, 2010; 2011), or approximately $1.16 per
gallon spilled.
To estimate the expected cleanup costs of a potential future impoundment release as a function of each steam
electric impoundment's capacity, EPA multiplied the impoundment capacity by 6.45 percent and applied the
average per gallon cost ($0.62/gallon), as shown in Equation 7-2.
Equation 7-2. CCLEANUPI = ICAPACITYI x 0.0645 x $0.62
Where:
CCLEANUPI = the cleanup cost associated with a release from impoundment /
ICAPACITYI = the capacity of impoundment /'.
7.7.3.2 Natural Resource Damages
Israel (2006) provides a detailed state-by-state summary of NRD programs, including some prominent NRD
cases (arising from oil spills, chemical spills, and other incidents) in each state.64 The median dollar value of
NRD settlement from these cases was approximately $4 million in 2010 dollars.65 Releases resulting from
impoundment failures may affect resources similar to those that were affected by the NRD cases identified in
Israel (2006) and therefore have similar NRD costs.
To calculate expected NRD costs for future impoundment releases, EPA assumed that the NRD costs vary
depending on the magnitude of the release, as indicated by cleanup costs, and represent a certain percentage
of cleanup costs. To derive a per-gallon estimate of the expected NRD value for future releases, EPA divided
the median NRD settlement of $4 million by the median cleanup cost for cases discussed in the previous
section ($65 million). This calculation yields an NRD estimate of 6 percent of cleanup costs, or $0.04 per
gallon spilled (6 percent of $0.62/gallon).66 This estimate is meant to reflect an approximate value for NRD
resulting from impoundment failures as a function of the volume of CCR slurry released; actual damages
would be highly location- and failure-specific.
EPA calculated the NRD costs associated with an expected impoundment failure as a function of
impoundment capacity using Equation 7-3.
64 In a separate study, Lemley and Skopura (2012) attempt to estimate the total damages to affected fish and wildlife
from coal combustion waste at twenty-one surface impoundment sites. Estimation of the economic values is explained in
the study's support document (http://pubs.acs.org/doi/suppl/10.1021/es301467q/suppl_file/es301467q_si_001.pdf). The
results of the study, while perhaps useful for providing a general sense of the scale and types of potential damages, are
not appropriate for use in this benefits analysis. The estimates are based on broad and uniform assumptions, and use
approaches and values that do not meet EPA guidelines for economic analyses to support rulemaking.
65 Natural resource damages do not include cleanup costs (or legal and transaction costs, if reported) but include only the
resource restoration and compensation values. For example, in one case, Israel (2006) reported that "In total, the State's
claim was $764 million, $342 million of which was restoration cost damages, $410 million of which was compensable
value damages, and $12 million of which was assessment and legal costs." For this case, EPA used the sum of
$342 million and $410 million (excluded legal costs) as NRD.
66 Note that this is a rough assumption that does not account for differences in the resources affected in each case.
April 19, 2013 T^T
-------
Benefit and Cost Analysis for Proposed ELGs 7: Benefits from Avoided Impoundment Failures
Equation 7-3. CNRD, = ICAPACITYI x 0.0645 x $0.04
Where:
CNRDI = the cost of natural resource damages associated with a release from impoundment /'
ICAPACITYI = the capacity of impoundment /'.
7.1.3.3 Transaction Costs
For this analysis, transaction costs include the costs associated with negotiating NRD, determining
responsibility among potentially responsible parties, and litigating details regarding settlements and
remediation.6? EPA estimated transaction costs based on data showing transaction costs as a share of total
cleanup costs at Superfund sites and the share of spending that represents total transaction costs. Table 7-1
shows the data sources. On average, transaction costs account for 27 percent of total spending.
Table 7-1: Studies Summarizing Transaction Costs as a Share of Superfund
Spending (for potentially responsible parties)
Acton (1995)
Acton and Dixon (1992)
Dixon,etal. (1993)
Steinhardt et al. (1994)
Average
Multiplier (transaction costs as a share of cleanup cost)3
27%
17%
32%
33%
27%
37%
a. Multiplier is calculated as Average/(l - Average)
These data indicate that transaction costs represent 27 percent of total costs, or an additional 37 percent of
cleanup costs. Therefore, the estimated transaction costs per gallon of slurry spilled is $0.23 (37 percent of
$0.62/gallon). Using these assumptions, EPA calculated the transaction costs associated with an expected
impoundment failure as a function of impoundment capacity using Equation 7-4.
Equation 7-4. CTRANSACTIOM = ICAPACITYI x 0.0645 x $0.23
Where:
CTRANSACTIOM = the transaction costs associated with a release from impoundment /
ICAPACITY = the capacity of impoundment /'.
7.7.3.4 Total Costs
Table 7-2 summarizes unit costs for cleanup, NRD, and transaction. Total impoundment failure costs are
$0.06 per gallon of impoundment capacity. In applying unit costs to any impoundment, EPA set a maximum
cost of $1.3 billion for any single incident based on the total estimated cleanup cost of the TVA Kingston
spill.
67 These activities involve services, whether performed by the complying entity or other parties, that EPA expects would
be required in the absence of this regulation in the event of an impoundment failure. Accordingly, it is appropriate to
account for the avoided resource cost of these services as social benefits in the benefit-cost analysis for the proposed
rule. Note that the transaction costs do not include fines, cleanup costs, damages, or other costs that constitute transfers
or are already accounted for in the other categories analyzed separately.
April 19, 2013 74
-------
Benefit and Cost Analysis for Proposed ELGs
7: Benefits from Avoided Impoundment Failures
Table 7-2: Unit Costs (2010$)
Cost Component
Unit Cost
Cleanup ($/gallon spilled)
NRD ($/gallon spilled)
Transaction costs ($/gallon spilled)
Total costs ($/gallon spilled)
Total costs based on impoundment capacity ($/gallon capacity)3
$0.62
$0.04
$0.23
$0.89
$0.06
a. Total release costs times the average volume spilled as a percentage of impoundment capacity (6.45
percent)
Source: U.S. EPA Analysis, 2013
The proposed ELGs would provide benefits by reducing the impact of impoundment failures for
impoundments that are expected to see reduced utilization as a result of the proposed ELGs, but would
continue to operate in the absence of the regulation.68 For each of the 1,070 impoundments included in the
analysis, EPA calculated the difference between the annualized costs from future expected failures under the
baseline and each analyzed regulatory option; the costs were estimated by multiplying the estimated cost per
failure by the failure probability for a given impoundment and year (0.58 percent), discounting for future
years, aggregating across the analysis time horizon (2017 to 2040), and annualizing over a 24-year period
using rates of 3 percent and 7 percent.69 The benefits of each analyzed regulatory option therefore are the
costs of expected releases under the baseline minus the costs of expected releases under the regulatory
scenario.
7.2 Results
Table 7-3 summarizes total benefits of avoided impoundment failures, calculated as the sum of the avoided
failure costs across all impoundments expected to be affected by the proposed ELGs under each analyzed
regulatory option.
EPA estimates that Option 3 would generate annual benefits valued at $114.8 million, using a 3 percent
discount rate. The results also provide relevant information for understanding the potential benefits of the
other preferred options. Options 3a and 3b are expected to provide smaller benefits than Option 3. Option 4a
is expected to provide benefits between those estimated for Options 3 and 4, i.e., between $114.8 million and
$295.1 million, using a 3 percent discount rate.
§ "Table 7-3. Estimated Annualized Benefits of Avoided Surface Impoundment Failures (Millions;
010$)a
Regulatory Option
Option 1
Option 3 a
Option 2
Option 3b
3% Discount Rate
$62.1
b
$62.1
b
7% Discount Rate
$52.2
b
$52.2
b
As described in Section 7.1.1, EPA assumed a uniform probability of failure of 0.58 percent per impoundment per
year. In this analysis, the costs of impoundment failures vary in proportion to the expected volume of CCR waste that
may be released, which is assumed to change as a result of the proposed ELG. The sensitivity analysis described in
Appendix G instead assumes that the probability of failure depends on impoundment age and capacity; for this alternate
analysis, EPA assumed that changes in the volume of CCR waste managed using impoundments due to the proposed
ELG reduce impoundment capacity by an equivalent fraction and therefore reduce the probability of a failure, in addition
to changing the volume of CCR waste that may be released in any given incident and the cost of that incident.
69 Note that while the analysis period starts in 2014, plants are assumed to close or reduce use of impoundments
according to the same schedule as control technology implementation (2017-2021). Additionally, as noted in Section
7.1.1, EPA assumed zero probability of failure in years 2014 through 2019 as a result of integrity site assessments.
April 19, 2013
7-5
-------
Benefit and Cost Analysis for Proposed ELGs
7: Benefits from Avoided Impoundment Failures
Table 7-3. Estimated Annualized Benefits of Avoided Surface Impoundment Failures (Millions;
2010$)a
Regulatory Option
Option 3
Option 4a
Option 4
Option 5
3% Discount Rate
$114.8
c
$295.1
$295.1
7% Discount Rate
$95.9
c
$245.9
$245.9
a. Baseline value of total failure costs minus option value of total failure costs.
b. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Options 3a and 3b to be less
than those of Option 3, with values for Option 3a being lower than Option 3b. EPA does not know how the benefits of Options 3a
and 3b compare to those of Options 1 and 2.
c. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Option 4a to be between those
of Options 3 and 4.
Source: U.S. EPA Analysis, 2013
As described in Section 7.1.1, these estimates do not account for the effect of BMPs that would further reduce
the probability of a failure and lead to higher benefits for the proposed ELGs. EPA does not have sufficient
information to accurately quantify and monetize the benefits of implementing BMPs. Preventing all future
impoundment failures would provide annual benefits estimated at up to $378 million (using a 3 percent
discount rate), relative to the baseline probability of failure and costs.
EPA does not anticipate BMPs to fully realize such benefits, however. Inspections and other BMPs aim to
prevent future failures by identifying conditions that have contributed to past impoundment failures and
releases (e.g., slope instability, structural defects, seepage, overtopping, and inadequate management
practices; see NRC, 2002); they may not be as effective at preventing impoundment failures caused by
unusual weather events or earthquakes. Further, data from the Steam Electric Industry Survey and 2010 CCR
Impoundment Survey suggest that steam electric plants already implement inspections and monitoring
programs of varying scope and frequency (U.S. EPA, 2010e; U.S. EPA, 2012d), but field assessments
conducted subsequent to the 2010 CCR Impoundment Survey indicate that the existing inspection programs at
some plants failed to identify embankment erosion, seepage and other conditions (U.S. EPA; 2012e).
imitations and Uncertainties
Table 7-4 summarizes the limitations and uncertainties in the analysis of benefits associated with reduced
impoundment failures arising from the proposed ELGs. The methodologies used in this analysis involve
several simplifications and sources of uncertainties, as described below. Whether these simplifications and
uncertainties, taken together, are likely to result in an understatement or overstatement of the estimated
benefits is not known.
Table 7-4. Uncertainties in Analysis of Avoided Risk of Impoundment Failure Benefits
I
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
The analysis assumes that, in the
absence of the proposed ELGs, all
impoundments would continue to
operate in the baseline during the entire
Esriod of analysis.
Overestimate
Plant owners may close existing impoundments or
make other changes to their operations that would
reduce the baseline probability of failure. Not
accounting for these baseline conditions may
overstate the benefits of the proposed ELGs.
April 19, 2013
7-6
-------
Benefit and Cost Analysis for Proposed ELGs
7: Benefits from Avoided Impoundment Failures
I. Uncertainties in Analysis of Avoided Risk of Impoundment Failure Benefits
Uncertainty/Assumption
EPA estimated expected future
impoundment failures based on a
uniform probability of failure
(0.58 percent per impoundment per
year). In practice, the probability of
failure may depend on impoundment
characteristics and management
practices.
The analysis uses a uniform cost of
$0.89 per gallon spilled and assumes
that, on average, each release involves a
volume equal to 6.45 percent of the
impoundment capacity. The analysis
caps total costs at $1.3 billion per
incident. In practice, failure costs depend
on incident-specific characteristics. Cost
estimates for future failures are therefore
highly uncertain.
The analysis does not include benefits
from inspection, monitoring and other
BMPs
Effect on Benefits
Estimate
Uncertain
Uncertain
Underestimate
Notes
EPA used historical data to evaluate benefits
assuming the best-fit relationship between the
probability of failure and impoundment age and
capacity. This sensitivity analysis is described in
Appendix G. The results suggest that using a uniform
failure rate may understate benefits of the proposed
ELGs. Conversely, the historical failure rate may
overstate projected failures under baseline conditions
by not reflecting the effects of any recent changes in
impoundment management practices (e.g., revised
inspection and monitoring programs).
There is significant uncertainty involved in
estimating the costs of unknown future release
incidents, and these estimates are based on a small
and highly variable sample of historic releases. The
costs of future releases may be substantially higher
or lower than the cost estimates applied in this
analysis, depending on site-specific factors,
including the ecosystems, infrastructure, and other
resources damaged by the release.
BMPs are expected to help further reduce the
probability of impoundment failures by identifying
conditions associated with past releases. EPA does
not have data to accurately estimate the additional
benefits of implementing these BMPs.
April 19, 2013
7-7
-------
Benefit and Cost Analysis for Proposed ELGs
8: Air-Related Benefits
8 Air-Related Benefits
The proposed ELGs are expected to affect air pollution through three main mechanisms: 1) additional
auxiliary electricity use by steam electric plants to operate wastewater treatment, ash handling, and other
systems needed to comply with the new effluent limits and standards; 2) additional transportation-related
emissions due to the increased trucking of coal combustion residue (CCR) waste to on-site or off-site
landfills; and 3) the change in the profile of electricity generation due to relatively higher cost to generate
electricity at plants incurring compliance costs for the proposed ELGs. The different profile of generation can
result in lower or higher air pollutant emissions due to differences in emission factors. Thus, small reductions
in coal-based electricity generation as a result of the proposed ELGs are compensated by increases in
generation using other fuels or energy sources - biomass, landfill gas, natural gas, nuclear power, oil, and
wind power. The changes in air emissions reflect the differences in emissions factors for these other fuels or
sources of energy, as compared to coal.
In this analysis, EPA estimated the human health and other benefits resulting from net changes in emissions
of three pollutants: Nitrogen oxides (NOX), sulfur dioxide (SO2), and carbon dioxide (CO2).
NOX and SOX (which include SO2 emissions quantified in this analysis) are known precursors to fine particles
(PM2 5) air pollution, a criteria air pollutant that has been associated with a variety of adverse health effects -
most notably, premature mortality.70 CO2 is an important greenhouse gas that is linked to climate changes
effects, including: an increase in temperature; sea level rise; changes in weather patterns toward an intensified
water cycle with stronger floods and droughts; and stress on ecosystems, especially in the Arctic, mountain
and tropical areas, resulting in the shift of species habitat range. The expected economic losses from climate
change include reduced agricultural yields, human health risks, property damages from increased flood
frequencies, the loss of ecosystem services, etc. Increased CO2 levels also affect biological systems
independent of climate change. For example, oceans become markedly more acidic, endangering coral reefs
and potentially harming fisheries and other marine life.
ata and Methodology
8.1.1 Changes in A ir Emissions
As discussed in the RIA (Chapter 5: Electricity Market Analyses), EPA used the Integrated Planning Model
(IPM) to estimate the electricity market-level effects of two of the eight regulatory options (Options 3 and 4;
see Chapter 5 in RIA (U.S. EPA, 2013c)). IPM outputs include NOX, SO2, and CO2 emissions to air from
electricity generating units (EGU). Comparing these emissions to those projected for the base case provides
an assessment of the changes in air emissions resulting from changes in the profile of electricity generation
under the proposed ELGs. EPA used two run years, 2020 and 2030, to represent the periods of 2017-2024,
and 2025-2035, respectively. For this analysis, EPA assumed that changes in emissions estimated for the
period of 2025-2035 continue through 2040.
EPA developed separate estimates of air emissions associated with increases in electricity generation to power
wastewater treatment systems by multiplying plant-specific additional electricity consumption estimated as
part of the engineering analysis by plant- or NERC-specific emission factors obtained from IPM for each
70 Sulfur oxides (SOx) include sulfur monoxide (SO), sulfur dioxide (SO2), sulfur trioxide (SO3) and other sulfur oxides.
In this analysis, EPA analyzed changes in emissions of SO2 only.
April 19, 2013 s
-------
Benefit and Cost Analysis for Proposed ELGs
8: Air-Related Benefits
analysis year. EPA estimated air emissions associated with increased trucking by multiplying the number of
miles by average emission factors. Details of these two analyses are provided in the TDD.
Table 8-1 through Table 8-3 summarize the estimated changes in emissions for the three mechanisms, the
three pollutants, and the two regulatory options covered in this particular analysis. As shown in the tables,
EPA estimates that changes in auxiliary service (Table 8-1) and transportation (Table 8-2) would result in an
increase in emissions (positive values), while changes in the profile of electricity generation (Table 8-3)
would reduce CO2 and SO2 emissions (negative values), but increase NOx emissions. Table 8-4 presents the
net emissions changes across the three mechanisms.
The largest effect on projected air emissions is due to the change in the emissions profile of electricity
generation at the market level. As presented in the RIA (Section 10.6: Executive Order 13211: Actions
Concerning Regulations That Significantly Affect Energy Supply, Distribution, or Use), IPM projects small
reductions in the use of coal for electricity generation as a result of the proposed ELGs (0.1 percent for
Option 3; 0.3 percent for Option 4), which is compensated by increases in generation using other fuels or
energy sources - biomass, landfill gas, natural gas, nuclear power, oil, and wind power. The changes in air
emissions reflect the differences in emissions factors for these other fuels, as compared to coal.
Table 8-1. Estimated Changes in Electricity Consumption and Air Pollutant Emissions due to Increase
in Auxiliary Service at Steam Electric Plants, Relative to Baseline
Year
Electricity
Consumption (MWh)
CO2 (Metric
Tonnes/Year)
NOx (Tons/Year)
SO2 (Tons/Year)
Options
2014-2016
2017
2018
2019
2020
2021-2024
2025-2040
0
77,536
161,756
206,958
276,720
303,332
303,332
0.0
71,038.4
149,500.9
191,097.8
254,786.9
276,960.1
276,680.4
0.0
89.9
133.5
162.9
215.4
239.6
197.1
0.0
56.4
146.2
185.7
237.9
257.7
274.0
Option 4
2014-2016
2017
2018
2019
2020
2021-2024
2025-2040
0
139,221
300,317
421,580
572,273
673,780
673,780
0.0
102,517.2
250,761.7
362,607.3
501,679.5
592,847.9
619,790.1
0.0
88.4
190.7
296.3
420.2
512.8
527.3
0.0
98.2
275.4
412.5
564.0
675.9
709.0
Source: U.S. EPA Analysis, 2013; see TDD for details.
Table 8-2. Estimated Changes in Annual Air Pollutant Emissions due to Increased Trucking at Steam
Electric Plants, Relative to Baseline
Year
CO2 (Metric Tonnes/Year)
NOx (Tons/Year)
SO2 (Tons/Year)
Options
2014-2016
2017
2018
2019
2020
2021-2040
0.0
4,699.4
9,818.2
13,595.7
18,683.8
22,548.7
0.0
2.1
4.3
5.9
8.2
9.9
0.0
0.0
0.1
0.1
0.2
0.2
April 19, 2013
8-2
-------
Benefit and Cost Analysis for Proposed ELGs
8: Air-Related Benefits
Table 8-2. Estimated Changes in Annual Air Pollutant Emissions due to Increased Trucking at Steam
Electric Plants, Relative to Baseline
Year
CO2 (Metric Tonnes/Year)
NOx (Tons/Year)
SO2 (Tons/Year)
Option 4
2014-2016
2017
2018
2019
2020
2017-2040
0.0
8,484.0
20,826.9
28,277.0
37,520.9
44,167.9
0.0
3.7
9.1
12.4
16.4
19.3
0.0
0.1
0.2
0.3
0.4
0.5
Source: U.S. EPA Analysis, 20131; see TDD for details
Table 8-3. Estimated Changes in Annual Air Pollutant Emissions due to Changes in Electricity
Generation Profile, Relative to Baseline
Year
CO2 (Metric Tonnes/Year)
NOx (Tons/Year)
SO2 (Tons/Year)
Option 3
2014-2016
2017-2024
2025-2040
0.0
-1,354,732.8
-1,461,681.0
0.0
1
1
,714.6
,119.1
0.0
-2,869.8
-3,096.
6
Option 4
2014-2016
2017-2024
2025-2040
0.0
-2,779,337.5
-4,412,837.1
0.0
1
1
,216.8
,052.5
0.
0
-706.8
-4,247.
1
Source: U.S. EPA Analysis, 2013; see Chapter 5 in PJAfor details.
Table 8-4. Estimated Net Changes in Air Pollutant Emissions due to
Steam Electric Plants, Increased Trucking at Steam Electric Plants,
Generation Profile, Relative to Baseline
Year
CO2 (Metric Tonnes/Year)
Increase in Auxiliary Service at
and Changes in Electricity
NOx (Tons/Year)
SO2 (Tons/Year)
Option 3
2014-2016
2017
2018
2019
2020
2021-2024
2025-2040
0.0
- ,278,995.0
- ,195,413.8
- ,150,039.3
- ,081,262.1
- ,055,224.0
- ,162,451.8
0.0
1,806.5
1,852.4
1,883.4
1,938.2
1,964.0
1,326.1
0.0
-2,813.4
-2,723.6
-2,684.0
-2,631.8
-2,611.9
-2,822.4
Option 4
2014-2016
2017
2018
2019
2020
2021-2024
2025-2040
0.0
-2,668,336.3
-2,507,748.9
-2,388,453.3
-2,240,137.1
-2,142,321.8
-3,748,879.1
0.0
1,308.9
1,416.6
1,525.5
1,653.4
1,748.9
1,599.0
0.0
-608.4
-431.1
-294.0
-142.4
-30.4
-3,537.7
Source: U.S. EPA Analysis, 2013.
8.1.2 NOxandSO2
Detailed human health benefits analyses for air regulations typically involve the use of a sophisticated air
quality model, such as the Community Multiscale Air Quality (CMAQ) Model, and the Environmental
April 19, 2013
8-3
-------
Benefit and Cost Analysis for Proposed ELGs 8: Air-Related Benefits
Benefits Mapping and Analysis Program (BenMAP), EPA's state-of-the-art air pollution benefits analysis
modeling tool. The air quality model estimates the changes in concentrations of criteria air pollutants in each
cell of a grid resulting from changes in emissions to air (e.g., of NOX and SO2) under various policy scenarios.
These criteria air pollutant changes are then input to BenMAP, which estimates the resulting changes in
incidence in the population of the adverse health effects associated with the pollutants and the corresonding
monetized benefits (see Abt Associates (2010) for additional description of BenMAP). This detailed approach
for human health benefits analysis of air regulations tends to be time- and resource-intensive.
Recognizing that a less resource- and time-intensive approach is sometimes desirable, EPA developed
estimates of national monetized benefits per ton of emissions avoided for use in estimating benefits without
the need to conduct detailed air quality and human health benefits modeling. Because the benefits per ton of
emissions depend on both the type of emissions (e.g., NOX vs. SO2) and the geographic distribution (relative
to population centers) of the emitting sources, benefits per ton estimates must be specific to a combination of
emission type and source category. EPA developed benefits per ton estimates for specific combinations of
emission source categories and PM2 5 precursors. EPA used this approach, for example, in its assessment of
the benefits of PM and SO2 reductions for the Industrial Boiler and Process Heaters NESHAP rule (U.S. EPA
2004a) and its analysis of the Mobile Source Area Toxics Rule (U.S. EPA 2004b) (See also Fann et al., 2012;
Fann et al., 2009; Levy et al., 2009; and Muller et al., 2009).
EPA's calculation of the benefits per ton values involved three principal steps, as described by Fann et al.
(2009) and the Technical Support Document for the calculation of benefit per-ton estimates (U.S. EPA
2008b):
1. An air quality model was used to estimate the changes in ambient PM2 5 concentrations resulting from
specified precursor emissions reductions under various scenarios; under each scenario the total tons
(of the precursor emissions) reduced was calculated.
2. BenMAP was used to estimate the changes in incidence of the associated health effects and the
monetized benefits of those incidence reductions under each scenario; and
3. National benefits per ton estimates were calculated by dividing the national monetized benefits by the
total tons of emissions reduced under each scenario.
In the current analysis, benefits per ton estimates are needed for four combinations of emission type and
source category involving NOX or SOX:
> NOX from EGUs (to be applied to changes in market-level NOX emissions projected by IPM, and
changes in emissions from auxiliary service);
> SOX from EGUs (to be applied to changes in market-level SOX emissions projected by IPM, and
changes in emissions from auxiliary service);
> NOX from mobile sources (to be applied to changes in NOX emissions associated with transporting
CCR waste to landfills); and
> SOX from mobile sources (to be applied to changes in SOX emissions associated with transporting
CCR waste to landfills).
As described by Fann et al. (2009), "ambient PM2 5 is a complex mixture of primary and secondarily formed
particles, resulting from interactions in the atmosphere and physical transport of emissions of particulate
matter precursors, including available SO2, NOX, and NH3, meteorology (particularly temperature), and
baseline levels and composition of PM25" (Fann et al. 2009, p. 170). NOX and SOX differ in their propensity for
April 19, 2013 8-4
-------
Benefit and Cost Analysis for Proposed ELGs 8: Air-Related Benefits
becoming PM25. The benefits per ton estimates assume, however, that all fine participates have the same
potency for causing premature mortality (U.S. EPA 2011).71
Fann et al. (2012) reported benefits per ton estimates for a variety of emission type/source category
combinations, including all of those listed above, that are relevant to the current analysis, for the years 2005
and 2016. Although they are not reported in Fann et al. (2012), EPA obtained benefits per ton estimates for
each of these categories for the years 2020, 2025, and 2030 directly from one of the study co-authors.72 For
these additional years, the benefits per ton were calculated assuming the same change in ambient air quality as
the author's forecast for 2016, but accounting for the 2020, 2025 and 2030 projected baseline mortality and
population to estimate the change in mortality risk
For this analysis, benefits per ton estimates are needed for each year from 2017 through 2040. Because the
benefits per ton estimates for the years 2016, 2020, 2025, and 2030 are almost linear as a function of year, EPA
interpolated benefits per ton values for the intermediate years (e.g., between 2020 and 2025) and projected
values for the years from 2031 through 2040 by linear regression, using (year, benefits per ton) data points for
the years 2016, 2020, 2025, and 2030. Note, however, that the approximate linearity of the (year, benefits per
ton) data points may be an artifact of the inability to project meteorological changes and thus changes in air
quality for all years after 2016, noted above. Thus, although it was necessary to generate benefits per ton
estimates for each year from 2017 through 2040 to be consistent with the rest of the analysis, additional
uncertainty was generated by using benefits per ton estimates for the future years that did not account for
meteorological and air quality changes.
Assuming that the geographic distribution of controlled emitting sources in a source category (e.g., EGUs)
and of emissions reductions in the current analysis are similar to the geographic distribution of emitting
sources and emissions reduction in the analysis in Fann et al. (2012), EPA can derive a rough estimate of
benefits from changes in air emissions by applying these benefits per ton estimates to the changes (in tons) of
emissions resulting from compliance with the proposed ELGs. For example, the benefits from reduced
emissions of NOX from EGUs under Option 3 can be estimated by multiplying emissions avoided under the
regulatory option by the appropriate benefits per ton value.
As noted above, NOX and SOX are known precursors to PM2 5 Several adverse health effects have been
associated with PM2 5 including premature mortality, non-fatal heart attacks, hospital admissions, emergency
department visits, upper and lower respiratory symptoms, acute bronchitis, aggravated asthma, lost work days
and acute respiratory symptoms. All of these health effects were included in the estimation of benefits that
went into the calculation of benefits per ton in Fann et al. (2012).
A very large percentage of the total monetized benefits of reducing PM2 5 concentrations are attributable to
avoided premature mortality. Fann et al. (2012) used data from Krewski et al. (2009), a study of mortality and
long-term exposure to PM2 5, to estimate the change in incidence of premature mortality associated with a
given change in PM25 concentrations. This study is one of several credible peer-reviewed long-term exposure
studies that EPA has used in benefits analyses of PM25. Appendix H provides more details on the
concentration response function used for this analysis.
When using long-term exposure studies, EPA has traditionally assumed that premature mortality avoided as a
result of a reduction in PM2 5 concentrations in a given year do not all occur in that year. Instead, EPA
assumes that the avoided PM2 5-related premature mortalities are distributed over a 20-year period, with most
occurring in the earlier years. EPA values avoided premature mortality and then discounts that value back to
71 Benefits per ton estimates are available for other pollutants, such as direct PM2 5 emissions, but they were not included
in this analysis because emissions factors were not available. The chemistry of PM formation is complex and nonlinear.
72 Provided in personal communication with Charles Fulcher, EPA Office of Air Quality Planning and Standards
(OAQPS), on October 19, 2012.
April 19, 2013 84
-------
Benefit and Cost Analysis for Proposed ELGs
8: Air-Related Benefits
the year of the analysis. Thus the numerator (the benefits) of the benefits per ton estimate for a given year is
the value of morbidity avoided in that year plus the present discounted value of the stream of avoided
premature mortalities over a twenty year period, discounted back to that year. For example, a benefits per ton
estimate from Fann et al. (2012) for 2016 is the value of avoided morbidity in 2016 plus the present
discounted value of the stream of avoided premature mortalities from 2016 to 2036, discounted back to 2016.
EPA obtained two sets of benefits per ton estimates for this analysis for the years 2005, 2016, 2020, 2025, and
2030: one set using a 3 percent discount rate and the other using a 7 percent discount rate.73 All benefits per
ton estimates for years 2016 through 2040 were further discounted back to the year 2014 (using a 3 percent or
7 percent discount rate, as appropriate). Because avoided premature mortalities are assumed to occur over a
twenty-year period, and real income is likely to increase over time, these benefits per ton estimates reflect
EPA's estimated increases in WTP for mortal risk reductions with respect to increases in real income. The
income growth adjustment factors used are those in BenMAP (Abt Associates Inc., 2010). Table 8-5
summarizes the benefits per ton estimates EPA used for the different emission type and source category
combinations involving NOX and SO2 in the analysis of the proposed ELGs.
Table 8-5. National Benefits per Ton Estimates for NOx and SO2 Emissions (2010$/ton) from the
Benefits per Ton Analysis Reported by Fann et al. (2012)a'b'°
Year
ECU
NOx
SO2
Mobile Source (Onroad)
NOx
SO2
3% Discount Rate
2005
2016
2020
2025
2030
$3,700
$5,200
$5,400
$5,900
$6,200
$27,000
$35,000
$36,000
$40,000
$42,000
$4,500
$7,400
$7,800
$8,500
$9,200
$20,000
$20,000
$21,000
$24,000
$26,000
7% Discount Rate
2005
2016
2020
2025
2030
$3,300
$4,700
$4,900
$5,300
$5,600
$24,000
$31,000
$33,000
$36,000
$38,000
$4,000
$6,700
$7,100
$7,700
$8,300
$18,000
$18,000
$19,000
$21,000
$24,000
a Provided for this analysis by Charles Fulcher, EPA/OAQPS on October 19, 2012.
b Mortality benefits based on Krewski et al. (2009).
'Estimation of benefits per ton for 2016, 2020, 2025, and 2030 were based on year 2016 emissions modeling.
Source: U.S. EPA Analysis, 2013 based on Fann etal. (2012)
8.1.3 CO2
Estimates of benefits per ton for CO2 were derived using estimates of the social cost of carbon (SCC)
obtained from the Interagency Working Group on Social Cost of Carbon (IWGSCC, 2010).74
The SCC is an estimate of the monetized damages associated with an incremental increase in carbon
emissions in a given year. It is intended to include (but is not limited to) changes in net agricultural
productivity, human health, property damages from increased flood risk, and the value of ecosystem services
due to climate change.
The interagency group selected four SCC values for use in regulatory analyses: $4.86, $22.12, $37.28, and
$67.07 per metric ton of CO2 emissions in the year 2010, in 2010 dollars. The first three values are based on
the average SCC from three integrated assessment models, at discount rates of 5, 3, and 2.5 percent,
73 Provided in personal communication with Charles Fulcher, EPA/OAQPS, on October 19, 2012
74 The interagency process included EPA, the Department of Transportation, and other executive branch entities.
April 19, 2013
8-6
-------
Benefit and Cost Analysis for Proposed ELGs
8: Air-Related Benefits
respectively. SCC values are provided for several discount rates because the literature shows that the SCC is
quite sensitive to assumptions about the discount rate, and because no consensus exists on the appropriate rate
to use in an intergenerational context. The fourth value is the 95th percentile of the distribution of SCC from
all three models using a 3 percent discount rate. It is included to represent higher-than-expected impacts from
temperature change further out in the tails of the SCC distribution. Low probability, high impact events are
incorporated into all of the SCC values through explicit consideration of their effects in two of the three
integrated assessment models as well as the use of a probability density function for equilibrium climate
sensitivity in all three models. Treating climate sensitivity probabilistically allows the estimation of SCC at
higher temperature outcomes, which lead to higher projected damages. The SCC increases over time because
future emissions are expected to produce larger incremental damages as physical and economic systems
become more stressed in response to greater climatic change. The interagency group estimated the growth rate
of the SCC directly using the three integrated assessment models rather than assuming a constant annual
growth rate. This helps to ensure that the estimates are internally consistent with other modeling assumptions.
The IWGSCC report provides SCC estimates at 5-year intervals, starting with the year 2010 and ending with
the year 2050. EPA used linear interpolation to estimate SCC values for the intermediate years. Table 8-6
presents the SCC estimates used in this analysis. As with benefits per ton estimates for NOX and SO2, all
estimates for years 2015 through 2040 were discounted back to the year 2014.
Table 8-6. SCC Values (201 OS/metric tonne CO2)
Year
2017
2018
2019
2020
2025
2030
2035
2040
5% Discount Rate,
Average
$6.30
$6.51
$6.72
$7.03
$8.47
$10.02
$11.57
$13.13
3% Discount Rate,
Average
$25.11
$25.63
$26.15
$26.66
$30.59
$33.90
$37.21
$40.51
2.5 Discount Rate,
Average
$41.03
$41.75
$42.37
$43.10
$47.44
$51.67
$56.01
$60.36
3% Discount Rate,
95th Percentile
$78.54
$80.09
$81.75
$83.40
$93.43
$103.35
$113.37
$123.29
Source: IWGSCC, 2010 (values updated to 2010 dollars).
EPA estimates the dollar value of the CO2-related benefits for each analysis year between 2017 and 2040 by
applying the global SCC estimates, shown in Table 8-6, to the estimated reductions in CO2 emissions under
the proposed ELGs.
8.1.4 Estimating Total Air-Related Benefits
EPA calculated the monetized air-related benefits of the proposed ELGs, under options 3 and 4, in any given
year (discounted back to the year 2014) by (1) multiplying the tons of emissions avoided for a given
emissions type/source category combination in that year by the benefits per ton for that emissions type/source
category combination for that year, and then (2) summing the benefits across all emissions type/source
category combinations. The total benefit for year y, then, is calculated using Equation 8-1.
Equation 8-1.
Where:
^ (Tons avoided) yj x BPTy
2014
7=1
j = 1,2, and 3 denote NOX, SO2, and CO2, respectively, from market-level EGUs;
j = 4, 5, and 6 denote NOX, SO2, and CO2, respectively, associated with auxiliary service;
April 19, 2013
8-7
-------
Benefit and Cost Analysis for Proposed ELGs
8: Air-Related Benefits
j = 7, 8, and 9 denote NOX, SO2, and CO2, respectively, associated with transportation; and
is the present discounted value, discounted to the year 20 14, of the benefits per ton for the
jth emissions type/source category combination.
The total present discounted value of benefits, discounted to the year 2014, PDV2oi4, is calculated using
Equation 8-2.
Equation 8-2.
2040 9
PDF
2014
avoided) y] x BPT
™14 .
y=2014j=l
Table 8-7 shows the estimated benefits from reductions in emissions of NOX, SO2, and CO2 in each of several
selected years for the two regulatory options EPA analyzed.
Table 8-7. Estimated Benefits from Reduced Air Emissions for Selected Years (millions; 2010$)
Year
2017a
2020
2025
2030
Option 3
3% Discount Rate
$121.9
$113.7
$140.6
$149.7
7% Discount Rateb
$86.8
$84.9
$104.4
$111.5
Option 4
3% Discount Rate
$82.9
$57.1
$246.7
$265.7
7% Discount Rateb
$29.5
$12.3
$150.6
163.0
a The benefits per ton values used for year 2017 benefit calculation is assumed to be the same as the 2016 benefits per ton values.
b EPA used SCC values based on a 5 percent discount rate to calculate total benefit values presented for the 7 percent discount rate.
Table 8-8 shows the annualized benefits from reductions in emissions of NOX, SO2, and CO2 for the two
regulatory options EPA analyzed. EPA annualized benefit estimates to enable consistent reporting across
benefit categories (e.g., benefits from improvement in water quality). The total air-related benefits include
benefits from CO2 emissions reductions calculated using average SCC values discounted at 3 percent or
5 percent, depending on the discount rate used for other categories of benefits (3 percent or 7 percent).
Appendix I presents estimates of CO2-related benefits calculated using alternate SCC values (average at
2.5 percent, average at 5 percent, and the 95th percentile at 3 percent).
The annualized benefits of Options 3 and 4 are $127.6 million and $170.5 million, respectively, using a
discount rate of 3 percent ($82.3 million and $74.6 million, respectively, using a discount rate of 7 percent).
These estimates provide relevant information for understanding the potential benefits of the preferred options,
including Options 3a, 3b, and 4a. As discussed in Chapter 5 of the RIA, EPA expects the impacts of Option
4a on the profile of electricity generation - and therefore air emissions - to be between those of Options 3 and
4. Accordingly, EPA expects the benefits of Option 4ato be between those of Options 3 and 4. Further, the
Agency expects Options 3a and 3b to result in smaller changes than those projected under Option 3, resulting
in smaller benefits than those estimated for Option 3.
April 19, 2013
8-8
-------
Benefit and Cost Analysis for Proposed ELGs
8: Air-Related Benefits
Table 8-8. Estimated Annualized Benefits from Reduced Air Emissions
(Millions; 2010$)
Regulatory
Option
Option 3
Option 4
Pollutant
NOX
S02
CO2a
TOTAL
NOX
S02
CO2a
TOTAL
3% Discount Rate
-$8.5
$102.5
$33.6
$127.6
-$8.8
$84.7
$94.6
$170.5
7% Discount Rate"
-$7.0
$80.5
$8.8
$82.3
-$6.9
$56.8
$24.7
$74.6
8
a. EPA used SCC values based on a 5 percent discount rate, and discounted at 5 percent, for total
benefit values presented for the 7 percent discount rate.
Source: U.S. EPA Analysis, 2013
.3 Limitations and Uncertainties
This analysis includes only those benefits associated with avoided mortality due to PM2 5 reduction.
Therefore, the quantified human health benefits included in this analysis represent only a subset of the total
potential air-related benefits expected to result from the proposed ELGs. There are also the standard sources
of uncertainty found in any air pollution benefits analysis - uncertainties surrounding the estimated emissions
changes, the estimated changes in air pollutant concentrations resulting from changes in emissions, the
estimated concentration-response relationships between the air pollutant and various health effects in the
exposed population, and the estimated value of each health effect avoided. There is additional uncertainty in
the SCC estimates, which reflect the projection of future harm from climate change, and the benefits per ton
estimates. More details about the limitations and uncertainties associated with the air-benefit analysis are
discussed in Table 8-9.
Table 8-9. Uncertainties in Analysis of Air-related Benefits
Issue
Effect on Benefits
Estimate
Notes
Analysis included only premature
mortality.
Underestimate
This analysis does not include all of the human health
benefits associated with air pollution reductions because
the benefits per ton estimates that were used are based
only on mortality, and not morbidity endpoints. Thus the
quantified human health benefits included in this
analysis represent only a subset of the total potential
health benefits expected to result from the proposed
ELGs.
There is uncertainty in projecting
the future harm from climate
change.
Uncertain
When attempting to assess the incremental economic
impacts of carbon dioxide emissions, the analyst faces a
number of serious challenges. A report from the
National Academies of Science (NRC, 2009) points out
that any assessment will suffer from uncertainty,
speculation, and lack of information about (1) future
emissions of greenhouse gases, (2) the effects of past
and future emissions on the climate system, (3) the
impact of changes in climate on the physical and
biological environment, and (4) the translation of these
environmental impacts into economic damages. As a
result, any effort to quantify and monetize the harms
associated with climate change will raise serious
questions of science, economics, and ethics and should
be viewed as provisional.
April 19, 2013
8-9
-------
Benefit and Cost Analysis for Proposed ELGs
8: Air-Related Benefits
Table 8-9. Uncertainties in Analysis of Air-related Benefits
Issue
Effect on Benefits
Estimate
Notes
The Interagency Working Group on Social Cost of
Carbon (IWGSCC, 2010) noted a number of limitations
to the SCC analysis, including the incomplete way in
which the integrated assessment models capture
catastrophic and noncatastrophic impacts, their
incomplete treatment of adaptation and technological
change, uncertainty in the extrapolation of damages to
high temperatures, and assumptions regarding risk
aversion. The limited amount of research linking climate
impacts to economic damages makes the interagency
modeling exercise even more difficult. This said, the
SCC estimates were developed using a defensible set of
input assumptions that are grounded in the existing
literature. As noted in the SCC TSD, the U.S.
government intends to revise these estimates over time,
taking into account new research findings that were not
available in 2010.
There is uncertainty associated with
the effects of compliance costs on
the forecast change in emissions
from the electricity sector.
Uncertain
Compliance costs (capital, fixed or variable) will
influence marginal generation decisions of plants
affected by the proposed ELGs. In order to model the
electricity market effects of the proposed ELGs, EPA
made certain modeling assumptions that may influence
the pattern of generation across the electricity sector,
and therefore emissions. For example, EPA converted
engineering capital costs to annual fixed operation and
maintenance (O&M) costs in order to model the cost of
complying with the proposed ELGs in IPM.
See RIA Chapter 5, and Section 10.6: Executive Order
13211: Actions Concerning Regulations That
Significantly Affect Energy Supply, Distribution, or Use
for additional discussion of how modeling assumptions
may influence the forecast air pollution changes from
the IPM modeling.
Differences between modeled and actual quantities of
electricity generated and emission factors of dispatched
generating units would affect the changes in air
pollutants emissions and therefore the benefits resulting
from these changes. EPA does not have information to
Quantify the magnitude of this uncertainty.
EPA used a reduced form approach
(benefits per ton) to value air-
related benefits of emissions
changes.
Uncertain
As Fann et al. (2012) note, "... implicit in the benefit
per ton assessment is that the key attributes of the
modeling — e.g. population distribution, source
parameters, etc. — are not so different from the policy
scenario as to affect the estimated benefits appreciably.
Reduced form approaches assume a linear relationship
between changes in emissions and benefits, an
assumption that may not be valid for large changes in
emissions" (Fann et al., 2012, p. 142).
April 19, 2013
8-10
-------
Benefit and Cost Analysis for Proposed ELGs
8: Air-Related Benefits
Table 8-9. Uncertainties in Analysis of Air-related Benefits
Issue
EPA used year-specific benefits per
ton estimates to derive values for
each year within the analysis period.
Effect on Benefits
Estimate
Uncertain
Notes
Use of year-specific benefits per ton estimates from
which to generate annual estimates introduces another
layer of uncertainty into the analysis. In particular,
because actual air quality modeling was carried out only
for 2005 and 2016, the approximate linearity seen in the
benefits per ton estimates for 2016, 2020, 2025, and
2030 may be an artifact of assuming that air quality
remains constant at 2016 levels. The benefits per ton
estimates for intermediate years also do not take into
account the likely non-linearity involved. If each year-
specific benefits per ton is uncertain, then an annual
estimate incorporating benefits per ton-based estimates
may be more uncertain. As a result, the annual estimates
can be considered only rough estimates.
April 19, 2013
8-11
-------
Benefit and Cost Analysis for Proposed ELGs
9: Benefits from Reduced Water Withdrawals
Benefits from Reduced Water Withdrawals
Steam electric plants use vast quantities of water for ash transport and for operating wet flue gas
desulfurization (FGD) scrubbers.
By eliminating or reducing water used in sluicing operations or prompting the recycling of water in FGD
wastewater treatment systems, the proposed ELGs are expected to reduce water withdrawal from both surface
waterbodies and aquifers. The reduction in water use depends on the regulatory option.75 EPA estimates that
power plants would reduce the use of water by 54 billion gallons per year (147 million gallons per day) under
Option 3 and by 158 billion gallons per year (432 million gallons per day) under Option 4 (see Chapter 77 of
TDD for details).
Reduced surface water intake would reduce impingement and entrainment mortality. Due to data limitations,
EPA did not quantify and monetize these benefits as part of this analysis. Accordingly, the rest of this chapter
focuses solely on the benefits associated with reduced groundwater withdrawals.
Reduced water intake from groundwater sources by steam electric plants would result in increased availability
of groundwater for local municipalities that rely on groundwater aquifers for drinking water supplies. These
municipalities are expected to avoid the cost of supplementing drinking water supplies through alternative
means, such as bulk drinking water purchases. The following sections describe EPA's estimate of reduced
groundwater withdrawal benefits.
9.1
Reduced Groundwater Withdrawals
EPA estimated the benefits of reduced groundwater withdrawals based on avoided costs of purchasing
drinking water during periods of shortages in groundwater supply.
9.1.1 Methods
EPA's analysis of the proposed ELG options (U.S. EPA, 2013b) indicate that up to two plants (one in Florida
and one in Nebraska) would reduce the volume of groundwater withdrawn as a result of the proposed ELGs.
Because the states are potentially or currently water-stressed (Tetra Tech, 2011), the proposed ELGs are likely
to generate benefits from improved groundwater recharge. To estimate the value of improved groundwater
supply, EPA relied on state-specific prices of bulk drinking water supplies, since municipalities may need to
purchase supplementary supplies in response to groundwater shortages arising from excessive withdrawals.
EPA recognizes that the assumption that a reduction in groundwater withdrawals in the water-stressed states
may result in reduced groundwater shortages is somewhat speculative, but used this assumption to provide
screening-level estimates of the potential benefits.
To estimate the monetary value of reduced groundwater withdrawal, EPA relied on current state-specific
water prices ($730.94 per acre/foot for Florida and $1,169.53 per acre/foot for Nebraska). For each affected
plant and regulatory option, EPA multiplied the reduction in groundwater withdrawal (in gallons per year) by
the estimated price of drinking water per gallon. EPA used a conversion factor of 325,851 to convert acre foot
to gallons.
75 The policy options for fly and bottom ash would eliminate or reduce water use associated with current wet sluicing
operating systems at steam electric plants. Reductions in intake flow would occur at plants which convert to dry handling
or recycle FGD wastewater under regulatory options as part of their treatment system.
April 19, 2013 JM~
-------
Benefit and Cost Analysis for Proposed ELGs
9: Benefits from Reduced Water Withdrawals
9.1.2 Results
Table 9-1 shows estimated annual benefits from reduced groundwater withdrawals. The annual benefits from
the preferred options for existing sources are $0.03 million and $0.13 million, respectively for Options 3 and
4, using a 3 percent discount rate ($0.01 million and $0.08 million using a 7 percent discount rate). As for
other categories of benefits, EPA expects that Options 3a and 3b would have lower benefits than Option 3,
and Option 4a would have benefits between those of Options 3 and 4.
Table 9-1. Estimated Annualized Benefits from Reduced Groundwater Withdrawals (Millions; 2010$)
Regulatory Option
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Reduction in
Groundwater Intakes
(million gallons per year;
full implementation)
0
a
0
a
16.0
b
47.5
47.5
3% Discount Rate
$0.00
a
$0.00
a
$0.03
b
$0.13
$0.13
7% Discount Rate
$0.00
a
$0.00
a
$0.01
b
$0.08
$0.08
a. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Options 3a and 3b to be less
than those of Option 3, with values for Option 3a being lower than Option 3b. EPA does not know how the benefits of Options 3a
and 3b compare to those of Options 1 and 2.
b. Value was not estimated for this regulatory option. EPA expects the effects - and the benefits - of Option 4a to be between those
of Options 3 and 4
Source: U.S. EPA Analysis, 2013
9.1.3 Limitations and Uncertainties
Table 9-2 summarizes the limitations and uncertainties in the analysis of benefits associated with reduced
groundwater withdrawals.
Table 9-2. Uncertainties in Analysis of Reduced Groundwater Withdrawals
Uncertainty/Assumption
EPA assumed that municipalities
would need to replace lost
groundwater supplies with bulk
drinking water purchases.
EPA assumed a direct relationship
between groundwater withdrawals
in water-stressed states and
groundwater shortages, i.e., that
reducing demand for limited
groundwater supplies would result
in avoided costs for purchased
water.
Affected aquifer characteristics
Effect on Benefits
Estimate
Uncertain
See below.
Overestimate
Uncertain
Notes
Municipalities may not need to replace groundwater
withdrawn by steam electric plants (in which case the
benefits of the ELG may be overstated), or they may
choose to replace the groundwater through other means,
such as desalinization (in the case of Florida, in which
case the benefits of the ELG may be understated).
EPA assumed that demand for additional water supply
exists in the affected areas (Florida and Nebraska) due to
potential draughts. However, the extent of this demand is
uncertain.
If the affected aquifers are used for private wells only, the
estimated benefits of improved groundwater recharge
could be under- or overstated, depending on households
WTP for protecting groundwater quantity.
April 19, 2013
9-2
-------
Benefit and Cost Analysis for Proposed ELGs
10: Total Benefits
10 Summary of Total Benefits
'
0.1
Table 10-1 summarizes the total annual monetized benefits using 3 percent and 7 percent discount rates.
Table 10-2 and Table 10-3 compile, for each of the five analyzed regulatory options, the time profiles of total
(non-discounted) monetized benefits. The tables also report the calculated present and annualized values of
benefits at 3 percent and 7 percent discount rates, respectively.
The estimated total monetized benefits of the five regulatory options EPA analyzed explicitly range from
$82.0 million to $605.5 million per year using a 3 percent discount rate, depending on the regulatory option
($65.9 million to $424.8 million per year using a 7 percent discount rate). Option 3 has total benefits of
approximately $311.7 million, using a 3 percent discount rate ($230.4 million using a 7 percent discount rate),
when including air-related benefits.
EPA did not analyze the benefits of three preferred options for existing sources (Option 3a, Option 3b and
Option 4a), but generally expects the annual total benefits of Options 3a and 3b to be less than those of
Option 3 and the benefits of Option 4a to be between those of Option 3 and 4. Section 10.3 provides estimated
total annualized benefits of Options 3a, 3b, and 4a as inferred from EPA's analyses of Options 1 through 5.
The monetized benefits of the proposed ELGs do not account for all benefits because they omit various
sources of benefits to society from reduced steam electric pollutant discharges, such as non-cancer health
benefits (other than IQ benefits from reduced childhood exposure to lead and in-utero exposure to mercury)
and reduced cost of drinking water treatment. See Chapter 2 for a discussion of categories of benefits EPA
did not monetize. In addition, EPA's analysis of human health benefits associated with water quality
improvements includes only partial benefits for directly receiving reaches. Finally, EPA was able to estimate
air-related benefits for Options 3 and 4 only (see Chapter 8). Benefits for options 1, 2 and 5 are therefore
understated; in particular, EPA expects that the benefits for Option 5 would be higher than those for Option 4
if air-related benefits were included.
Chapters 3 through 9 provide more detail on limitation and uncertainty inherent in the analysis of each benefit
category.
April 19, 2013
10-1
-------
Benefit and Cost Analysis for Proposed ELGs
10: Total Benefits
Table 10-1. Summary of Total Annualized Benefits (Millions; 2010$)
Benefit Category
Option 1 | Option 3a | Option 2 | Option 3b | Option 3 | Option 4a | Option 4 | Option 5
3% Discount Rate
Human Health Benefits
Avoided cancer cases from exposure to arsenic
Reduced IQ losses in children from exposure to lead3
Avoided Cost of Compensatory Education for
Children with Blood Lead Concentrations above
20 |ag/dL and IQ Less than 70a
Reduced IQ losses associated with in-utero exposure
to mercury3
Improved Ecological Conditions and Recreational Uses
Use and nonuse values for water quality
improvements'3
Nonuse values of T&E species0
Groundwater Quality Benefits
Market and Productivity Benefits (Avoided
Impoundment Failures)
Air-related Benefits
Reduced Water Withdrawals
Total (excluding air-related benefits)"1
Range
Total (including air-related benefits)"1
Range
$0.0
$0.1
$0.0
$3.8
NE
NE
NE
NE
$0.0
$0.1
$0.0
$3.9
NE
NE
NE
NE
$0.1
$2.7
$0.0
$5.0
NE
NE
NE
NE
$0.2
$6.8
$0.1
$10.2
$0.2
$6.8
$0.1
$10.2
$8.3
$7.0
$0.7
$62.1
NE
$0.0
$82.0
($71.1 to
$99.0)
e
NE
NE
NE
NE
NE
NE
f
f
$38.0
$7.0
$0.7
$62.1
NE
$0.0
$111.7
($76.4 to
$184.4)
e
NE
NE
NE
NE
NE
NE
f
f
$49.9
$10.0
$1.6
$114.8
$127.6
$0.0
$184.1
($136.9 to
$275.8)
$311.7
($264.5 to
$403.3)
NE
NE
NE
NE
NE
NE
f
f
$82.8
$33.3
$6.5
$295.1
$170.5
$0.1
$435.0
($347.3 to
$578.6)
$605.5
($5 17.8 to
$749.1)
$81.9
$33.3
$6.5
$295.1
NE
$0.1
$434.1
($346.9 to
$576.9)
e
April 19, 2013
10-2
-------
Benefit and Cost Analysis for Proposed ELGs
10: Total Benefits
Table 10-1. Summary of Total Annualized Benefits (Millions; 2010$)
Benefit Category
Option 1 | Option 3a | Option 2 | Option 3b | Option 3 | Option 4a | Option 4 | Option 5
7% Discount Rate
Human Health Benefits
Avoided cancer cases from exposure to arsenic
Reduced IQ losses in children from exposure to lead3
Avoided Cost of Compensatory Education for
Children with Blood Lead Concentrations above
20 |ag/dL and IQ Less than 70a
Reduced IQ losses associated with in-utero exposure
to mercury3
Improved Ecological Conditions and Recreational Uses
Use and nonuse values for water quality
improvements'3
Nonuse values of T&E species0
Groundwater quality benefits
Market and Productivity Benefits (Avoided
Impoundment Failures)
Air-related Benefits
Reduced Water Withdrawals
Total (excluding air-related benefits)"1
Range
Total (including air-related benefits)"1
Range
$0.0
$0.0
$0.0
$0.3
NE
NE
NE
NE
$0.0
$0.0
$0.0
$0.4
NE
NE
NE
NE
$0.1
$0.3
$0.0
$0.4
NE
NE
NE
NE
$0.1
$0.6
$0.0
$0.9
$0.1
$0.6
$0.0
$0.9
$6.9
$5.9
$0.6
$52.2
NE
$0.0
$65.9
($57.3 to
$79.7)
e
NE
NE
NE
NE
NE
NE
f
f
$31.7
$5.9
$0.6
$52.2
NE
$0.0
$90.7
($6 1.6 to
$151.0)
e
NE
NE
NE
NE
NE
NE
f
f
$41.7
$8.4
$1.4
$95.9
$82.3
$0.0
$148.1
($109.5 to
$223.7)
$230.4
($193.2 to
$307.4)
NE
NE
NE
NE
NE
NE
f
f
$69.2
$27.8
$5.5
$245.9
$74.6
$0.1
$350.2
($278.7 to
$467.9)
$424.8
($358.9 to
$548.1)
$68.5
$27.8
$5.5
$245.9
NE
$0.1
$349.4
($284.0 to
$472.1)
e
NE = not estimated
a. Values provided are the mean for each option; for full range of benefits estimated, see Chapter 3.
b. Values provided are the mean for each option; for full range of benefits estimated, see Chapter 4. Benefits for Option 5 are slightly less than those for Option 4 because Option 5
improves slightly fewer miles than Option 4 (22,441 vs. 22,447 miles). Under Option 4, plants with both leachate and FGD wastestreams are assumed to implement chemical
precipitation and biological treatment for the combined streams whereas, under Option 5, these same plants are assumed to treat the two streams separately: FGD wastewater by
evaporation and leachate using chemical precipitation (which removes less pollutant load than biological treatment).
c. Values provided are the mean for each option; for full range of benefits estimated, see Chapter 5.
d. Values for individual benefit categories may not sum to the total due to independent rounding.
e. The total monetized benefits for options 1,2, and 5 do not include air-related benefits. This category of benefits was analyzed for Options 3 and 4 only (see Chapter 8). Section 10.3
describes EPA's calculation of inferred air-related benefits for Options 1, 2, and 5 based on results for Options 3 and 4.
f. EPA did not estimate benefits for Options 3a, 3b and 4a, but expects the benefits of Options 3a and 3b to be less than those of Option 3 and the benefits of Option 4a to be between
those of Options 3 and 4. Section 10.3 describes EPA's calculation of inferred total benefits for Options 3a, 3b, and 4a based on results for Options 1 through 5.
Source: U.S. EPA Analysis, 2013
April 19, 2013
10-3
-------
Benefit and Cost Analysis for Proposed ELGs
10: Total Benefits
10.2 Time Profile of Benefits
Table 10-2 and Table 10-3 compile, for each of the regulatory options, the time profiles of total (non-
discounted) monetized benefits. The tables also report the calculated present and annualized values of benefits
at 3 percent and 7 percent discount rates, respectively.
Table 10-2: Time Profile of Benefits at 3 Percent (Millions; 2010$) (Including Air-Related Benefits for
Year
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Annualized Benefits, 3%
Option la
$0.0
$0.0
$0.0
$1.1
$3.0
$82.9
$88.2
$97.9
$98.8
$100.2
$100.8
$101.1
$101.4
$101.6
$101.8
$102.1
$102.3
$102.5
$102.7
$102.9
$103.1
$103.3
$103.5
$103.8
$104.0
$104.2
$104.4
$82.0
i 1
Option 3a
b
Option 2a
$0.0
$0.0
$0.0
$1.2
$3.1
$116.5
$122.2
$132.1
$133.4
$135.2
$136.0
$136.7
$137.3
$137.9
$138.4
$139.0
$139.6
$140.1
$140.6
$141.1
$141.6
$142.1
$142.6
$143.2
$143.7
$144.2
$144.8
$111.7
Option 3b
b
Option 3
$0.0
$0.0
$0.0
$123.1
$122.9
$285.3
$291.9
$337.3
$340.7
$345.4
$348.9
$369.2
$371.3
$374.4
$377.3
$380.3
$383.0
$386.1
$389.1
$391.9
$394.9
$397.8
$400.8
$403.6
$406.5
$409.5
$412.4
$311.7
Option 4a
c
Option 4
$0.0
$0.0
$0.0
$85.1
$84.2
$434.1
$467.4
$579.6
$584.1
$591.3
$596.1
$790.4
$795.9
$801.9
$807.6
$813.3
$819.0
$824.9
$830.8
$836.3
$842.2
$847.8
$853.7
$859.3
$864.9
$870.8
$876.4
$605.5
Option 5a
$0.0
$0.0
$0.0
$3.2
$11.5
$367.4
$409.5
$527.4
$530.7
$536.4
$539.9
$542.6
$545.8
$547.4
$549.0
$550.6
$552.2
$553.5
$554.9
$556.3
$557.8
$559.2
$560.6
$562.1
$563.6
$565.0
$566.5
$434.1
a. Estimates for Options 1,2 and 5 do not include air-related benefits. This category of benefits was only estimated for Options 3 and
4 (see Chapter 8)
b. EPA did not estimate the benefits of this regulatory option. EPA expects the benefits of Options 3a and 3b to be less than those of
Option 3, with values for Option 3a being lower than those for Option 3b. See Section 10.3 for EPA's calculation of inferred total
benefits for Options 3 a and 3b.
c. EPA did not estimate the benefits of this regulatory option. EPA expects the benefits of Option 4a to be between those of Options 3
and 4. See Section 10.3 for EPA's calculation of inferred total benefits for Option 4a.
Source: U.S. EPA Analysis, 2013
April 19, 2013
10-0
-------
Benefit and Cost Analysis for Proposed ELGs
10: Total Benefits
Table 10-3: Time Profile of Benefits at 7 Percent (Millions; 2010$) (Including Air-Related Benefits for
Options 3 and 4)
Year
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Annualized Benefits, 7%
Option la
$0.0
$0.0
$0.0
$0.3
$0.8
$81.8
$87.1
$96.8
$97.0
$97.1
$97.3
$97.5
$97.7
$97.9
$98.1
$98.3
$98.5
$98.7
$98.8
$99.0
$99.2
$99.4
$99.5
$99.7
$99.9
$100.1
$100.3
$65.9
Option 3a
b
Option 2a
$0.0
$0.0
$0.0
$0.3
$0.8
$115.5
$121.0
$131.0
$131.5
$132.0
$132.5
$133.0
$133.5
$134.1
$134.6
$135.2
$135.7
$136.2
$136.6
$137.1
$137.6
$138.1
$138.6
$139.1
$139.6
$140.1
$140.6
$90.7
Option 3b
b
Option 3
$0.0
$0.0
$0.0
$122.2
$119.7
$283.7
$290.0
$335.1
$337.7
$340.4
$343.0
$362.5
$363.9
$366.9
$369.7
$372.6
$375.3
$378.4
$381.2
$384.0
$386.9
$389.7
$392.6
$395.4
$398.3
$401.2
$404.0
$230.4
Option 4a
c
Option 4
$0.0
$0.0
$0.0
$82.9
$77.1
$430.5
$463.3
$574.8
$577.3
$580.1
$582.7
$775.7
$779.4
$785.3
$790.9
$796.4
$802.0
$807.7
$813.4
$818.8
$824.5
$829.9
$835.7
$841.2
$846.6
$852.4
$857.9
$424.8
Option 5C
$0.0
$0.0
$0.0
$1.1
$4.4
$363.8
$405.4
$522.5
$523.9
$525.2
$526.5
$527.9
$529.3
$530.7
$532.2
$533.6
$535.1
$536.3
$537.5
$538.8
$540.0
$541.3
$542.6
$544.0
$545.3
$546.6
$547.9
$349.4
a. Estimates for Options 1,2 and 5 do not include air-related benefits. This category of benefits was only estimated for Options 3 and
4 (see Chapter 8)
b. EPA did not estimate the benefits of this regulatory option. EPA expects the benefits of Options 3a and 3b to be less than those of
Option 3, with values for Option 3a being lower than those for Option 3b. See Section 10.3 for EPA's calculation of inferred total
benefits for Options 3 a and 3b.
c. EPA did not estimate the benefits of this regulatory option. EPA expects the benefits of Option 4a to be between those of Options 3
and 4. See Section 10.3 for EPA's calculation of inferred total benefits for Option 4a.
Source: U.S. EPA Analysis, 2013
April 19, 2013
10-2
-------
Benefit and Cost Analysis for Proposed ELGs 11: Total Social Costs
10.3 Inferred Benefits for Regulatory Options not Analyzed Explicitly
As noted throughout this report, EPA calculated benefits for some of the options considered for this proposal.
Benefits for these options, however, provide information relevant to understanding the potential magnitude of
benefits under all proposed options, including Options 3a, 3b, and 4a. The facilities affected by Option 3a are
a subset of Option 3 facilities; Option 3 benefit estimates therefore provide an upper bound estimate of
benefits anticipated under Options 3a and 3b. In a similar way, EPA expects Option 4 to provide an upper
bound estimate of benefits anticipated under Option 4a.
As an illustrative analysis, EPA inferred the potential benefits associated with Options 3a and 3b by
subtracting the benefits for Option 2 (scaled up to include a rough estimate of air emissions benefits) from the
benefits for Option 3, because Option 3 includes a combination of the wastestreams and control technologies
in Options 3a and 2. EPA inferred the potential benefits associated with Option 3b based on the pollutant
loading reductions (pounds) projected for Option 3b relative to pollutant loading reductions projected for
Option 2 (plus the fly ash dry handling benefits of Option 3a) because Option 3b includes both fly ash
requirements and the Option 2 FGD wastewater treatment requirements for a subset of facilities. This is
equivalent to interpolating benefits linearly between Options 3a and 3 based on relative pollutant removals.
Similarly, EPA inferred the potential benefits associated with Option 4a based on the bottom ash pollutant
loading reductions projected for this option, relative to bottom ash pollutant loading reductions projected for
Option 4, plus the benefits of Option 3, because Option 4a includes all of the requirements of option 3 plus
the bottom ash requirements of Option 4 for a subset of facilities. This is equivalent to interpolating benefits
linearly between Options 3 and 4 based on relative pollutant removals.
EPA used total benefits, including air-related benefits, to infer benefits of Options 3a, 3b, and 4a. The first
step in the analysis involved adjusting the total benefits of Option 2 shown in Table 10-1 to reflect anticipated
air-related benefits. EPA made this adjustment using the results for Options 3 and 4, which show the total
benefits including air-related benefits being, on average, 1.54 times the total benefits excluding air-related
benefits using a 3 percent discount rate. Figure 10-1 displays the annual benefits of the regulatory options in
relation to their total pollutant loading reductions.
April 19, 2013 10-1
-------
Benefit and Cost Analysis for Proposed ELGs
11: Total Social Costs
$800.0 -
$700.0
01 $600.0 -
S $500.0 -
CO
re $400.0 -
w
s
o
-------
Benefit and Cost Analysis for Proposed ELGs
11: Total Social Costs
Table 10-4 summarizes total annualized benefits estimated, adjusted or inferred using the calculations
described above for the eight options discussed in today's proposal. Note that there is significant uncertainty
in values inferred because the methodology used does not account for differences in the pollutants, receiving
waterbodies, and exposed populations between the options.
Table 10-4. Total Monetized Benefits for the Proposed ELGs (Millions; 2010$)
Regulatory Option
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Method
Adjusted Estimate3
Inference15
Adjusted Estimate3
inference15
Estimate
inference15'0
Estimate
Adjusted Estimate3
Total Monetized Benefits
at 3% Discount Rate
$126.5
$139.4
$172.3
$205.5
$311.7
$482.5
$605.5
$669.6
Total Monetized Benefits
at 7% Discount Rate
$91.2
$104.8
$125.6
$153.0
$230.4
$343.4
$424.8
$483.7
a. Estimated total benefits for Options 1, 2, and 5 do not include air-related benefits (see Section 10.1). To infer benefits for Options
3a, 3b, and 4a, EPA calculated adjustments to the benefits of Options 1, 2, and 5 to include inferred air-related benefits. The
calculation adjusts the total benefits presented in Table 10-1 based on the average ratio of [total benefits including air-related
benefits] / [total benefits excluding air-related benefits] for Options 3 and 4 calculated using 3% and 7% discount rates.
b. EPA did not estimate benefits for Options 3a, 3b and 4a. EPA inferred benefits for Options 3a, 3b, and 4a for illustrative purposes
using elements of the more rigorous analysis done to estimate benefits for Options 3 and 4.
c. At the time this analysis was conducted, EPA inferred benefits for Option 4a based on pollutant removals for this option of
2,630 million pounds per year. As presented in the TDD (U.S. EPA, 2013b), however, the pollutant removals for this option are
2,620 million pounds per year, which would result in total annual benefits of $480.8 million (at 3 percent discount and $342.3
million (at 7 percent discount).
Source: U.S. EPA Analysis, 2013
April 19, 2013
10-3
-------
Benefit and Cost Analysis for Proposed ELGs
11: Total Social Costs
11 Summary of Total Social Costs
This chapter develops EPA's estimates of the costs to society resulting from the proposed ELGs. As analyzed
in this chapter, the social costs of regulatory actions are the opportunity costs to society of employing
resources to prevent the environmental damage otherwise occurring from discharges of wastewater containing
metals, nutrients, and other pollutants.
'
1.1 Overview of Social Costs Analysis Framework
RIA Chapter 3: Compliance Costs presents EPA's development of costs to the 1,079 steam electric plants
subject to the proposed ELGs (U.S. EPA, 2013c). These costs are used as the basis of the social cost analysis.
However, the compliance costs used to estimate total social costs differ in their consideration of taxes from
those reported in RIA Chapter 3, which were calculated for the purpose of estimating the private costs and the
economic impacts of the proposed ELGs. In the analysis of costs to society, compliance costs are considered
without accounting for any tax effects. The costs to society are the full value of the resources used, whether
they are paid for by the regulated plants or by all taxpayers in the form of lost tax revenues.77
As described in Chapter 1, EPA assumed that steam electric plants, in the aggregate, would implement
control technologies during a 5-year period from 2017 to 2021. For this analysis, EPA developed a year-
explicit schedule of compliance outlays over the period of 2017 through 2040.78 After creating a cost-
incurrence schedule for each cost component, EPA summed the costs expected to be incurred in each year for
each plant, then aggregated these costs to estimate the total costs for each year in the analysis period.
After compliance costs were assigned to the year of occurrence, the Agency adjusted these costs for real
change between their stated year and the year(s) of their incurrence as follows:
> All technology costs, except planning, were adjusted to their incurrence year(s) using the
Construction Cost Index (CCI) from McGraw Hill Construction and the Gross Domestic Product
(GDP) deflator index published by the U.S. Bureau of Economic Analysis (BEA);
> Planning costs were adjusted to their incurrence year(s) using the Employment Cost Index (ECI)
Bureau of Labor Statistics (BLS) and GDP deflator.
Note that the CCI and ECI adjustment factors were developed only through the year 2017; after these years,
EPA assumed that the real change in prices is zero - that is, costs are expected to change in line with general
inflation. EPA judges this to be a reasonable assumption, given the uncertainty of long-term future price
projections.
After developing the year-explicit schedule of total social costs and adjusting them for predicted real change
to the year of their incurrence, EPA calculated the present value of these cost outlays as of the promulgation
year by discounting the cost in each year back to 2014, using both 3 percent and 7 percent discount rates.
These discount rate values reflect guidance from the Office of Management and Budget (OMB) regulatory
analysis guidance document, Circular A-4 (OMB, 2003; updated 2009). EPA calculated the constant annual
equivalent value (annualized value), again using the two values of the discount rate, 3 percent and 7 percent,
77 For the impact analyses, compliance costs are measured as they affect the financial performance of the regulated plants
and firms. The economic impact analyses therefore consider the tax deductibility of compliance expenditures, as
appropriate depending on the tax status of the complying entity.
78 The end of the analysis period, 2040, was determined based on the life of the longest-lived compliance technology
implemented at any steam electric plant (20 years), and the last year of technology implementation (2021).
April 19, 2013 TU4
-------
Benefit and Cost Analysis for Proposed ELGs
11: Total Social Costs
over a 24-year social cost analysis period. EPA assumed no re-installation of compliance technology during
the period covered by the social cost analysis.
To assess the economic costs of the proposed ELGs to society, EPA relied first on the estimated costs to
steam electric plants for the labor, equipment, material, and other economic resources needed to comply with
the proposed ELGs. In this analysis, EPA assumed that the market prices for labor, equipment, material, and
other compliance resources represent the opportunity costs to society for use of those resources in regulatory
compliance. Finally, EPA assumed in its social cost analysis that the regulation does not affect the aggregate
quantity of electricity that would be sold to consumers and, thus, that the regulation's social cost would
include no loss in consumer and producer surplus from lost electricity sales by the electricity industry in
aggregate. Given the small impact of the regulation on electricity production cost for the total industry, EPA
believes that this assumption is reasonable for the social cost analysis (for more details on the impacts of the
proposed ELGs on electricity production cost, see RIA Chapter 5: Electricity Market Analyses).The social
cost analysis considers costs on an as-incurred, year-by-year basis - that is, this analysis associates each cost
component to the year(s) in which they are assumed to occur relative to the assumed promulgation and
technology implementation years.79
Finally, as discussed in RIA Chapter 10 (Section 10.7: Paperwork Reduction Act of 1995), the proposed ELGs
are not expected to result in additional administrative costs for plants to implement, and State and federal
NPDES permitting authorities to administer, the proposed ELGs. As a result, the social cost analysis focuses
on the resource cost of compliance as the only direct cost incurred by society as a result of the proposed
ELGs.
1.2 Key Findings for Regulatory Options
Table 11-1 presents annualized social costs for each of the eight regulatory options, in order of increasing
toxic-weighted pollutant removals. At a 3 percent discount rate, estimated annualized social costs range
between $185 million under Option 3a and $2,329 million under Option 5.
The preferred options for existing sources have annualized costs of $185 million, $281 million, $572 million
and $954 million, respectively for Options 3a, 3b, 3 and 4a, at a 3 percent discount rate. At a 7 percent
discount rate, annualized costs range between $165 million under Option 3a and $2,209 million under Option
5, with the preferred options having costs of $165 million, $257 million, $545 million and $915 million,
respectively for Options 3a, 3b, 3 and 4a.
Table 11-1: Summary of Annualized Social Costs (Millions; $2010)
Regulatory Option
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
3% Discount Rate
$268.3
$185.2
$386.8
$281.4
$572.0
$954.1
$1,381.2
$2,328.7
7% Discount Rate
$259.2
$164.6
$380.8
$257.2
$545.3
$914.7
$1,323.2
$2,209.4
Source: U.S. EPA Analysis, 2013.
Table 11-2 provides additional detail on the social cost calculations. The table compiles, for each of the eight
regulatory options, the time profiles of compliance costs incurred. The table also reports the calculated
79 The specific assumptions of when each cost component is incurred can be found in Chapter 3: Compliance Costs of
the RIA.
April 19, 2013
11-5
-------
Benefit and Cost Analysis for Proposed ELGs
11: Total Social Costs
annualized values of costs at 3 percent and 7 percent discount rates. The maximum compliance outlays are
incurred over the years 2017 through 2021, i.e., during the estimated window when steam electric plants are
expected to implement compliance technologies.
Table 11-2: Time Profile of Costs to Society (Millions; $2010)
Year
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Annualized Costs, 3%
Annualized Costs, 7%
Option 1
$0.0
$0.0
$0.0
$523.6
$615.2
$437.3
$533.5
$314.5
$214.0
$215.7
$219.7
$217.9
$220.8
$224.8
$223.3
$225.9
$226.2
$225.3
$217.1
$216.7
$218.7
$221.5
$223.4
$224.6
$224.3
$223.8
$223.3
$268.3
$259.2
Option 3a
$0.0
$0.0
$0.0
$105.9
$169.3
$203.8
$300.0
$274.0
$198.7
$195.9
$196.4
$197.8
$196.4
$203.7
$204.3
$203.9
$203.9
$203.8
$199.1
$197.3
$198.3
$199.8
$199.6
$203.2
$203.3
$203.1
$203.0
$185.2
$164.6
Option 2
$0.0
$0.0
$0.0
$841.4
$1,003.5
$688.3
$799.3
$503.3
$286.0
$287.7
$291.7
$289.9
$292.8
$296.8
$295.2
$297.9
$298.2
$297.3
$289.1
$288.7
$290.7
$293.4
$295.4
$296.6
$296.3
$295.8
$295.2
$386.8
$380.8
Option 3b
$0.0
$0.0
$0.0
$168.1
$431.4
$339.5
$650.1
$370.1
$276.4
$271.8
$273.5
$272.5
$274.4
$281.5
$282.0
$281.7
$282.0
$281.8
$277.6
$273.6
$275.3
$275.6
$277.4
$281.1
$281.3
$280.8
$280.6
$281.4
$257.2
Option 3
$0.0
$0.0
$0.0
$947.3
$1,172.8
$892.1
$1,099.3
$777.3
$484.7
$483.6
$488.1
$487.7
$489.2
$500.5
$499.5
$501.8
$502.1
$501.1
$488.2
$486.0
$489.0
$493.3
$495.0
$499.8
$499.6
$498.8
$498.2
$572.0
$545.3
Option 4a
$0.0
$0.0
$0.0
$1,402.0
$1,852.1
$1,613.0
$2,007.6
$1,757.5
$778.7
$777.0
$782.1
$783.9
$781.6
$800.6
$801.8
$803.0
$807.3
$804.1
$782.3
$782.0
$785.0
$792.4
$793.4
$798.2
$798.0
$795.8
$793.7
$954.1
$914.7
Option 4
$0.0
$0.0
$0.0
$1,939.0
$2,628.7
$2,432.2
$2,798.9
$2,716.0
$1,123.7
$1,126.0
$1,132.2
$1,132.4
$1,133.0
$1,163.6
$1,160.0
$1,164.1
$1,169.6
$1,160.9
$1,133.5
$1,135.5
$1,132.1
$1,148.5
$1,150.9
$1,153.8
$1,157.5
$1,149.7
$1,140.1
$1,381.2
$1,323.2
Option 5
$0.0
$0.0
$0.0
$3,179.5
$4,411.5
$3,769.9
$4,418.4
$4,012.1
$2,002.9
$2,005.9
$2,011.9
$2,012.4
$2,013.1
$2,043.4
$2,039.2
$2,044.0
$2,049.3
$2,040.8
$2,013.5
$2,015.2
$2,011.3
$2,028.4
$2,030.5
$2,033.5
$2,037.2
$2,029.1
$2,019.2
$2,328.7
$2,209.4
Source: U.S. EPA Analysis, 2013.
April 19, 2013
11-6
-------
Benefit and Cost Analysis for Proposed ELGs
12: Benefits and Social Costs
12 Benefits and Social Costs
This chapter compares total monetized benefits and social costs for the eight regulatory options considered for
the proposed ELGs. Benefits and costs are compared on two bases: (1) for each of the options analyzed and
(2) incrementally across options. The comparison of benefits and costs also satisfies the requirements of
Executive Order 12866: Regulatory Planning and Review and Executive Order 13563: Improving Regulation
and Regulatory Review (see Chapter 9: Other Administrative Requirements of the RIA; U.S. EPA, 2013c).
12.1 Comparison of Benefits and Social Costs by Option
Chapter 10 and Chapter 11 present estimates of the benefits and social costs, respectively, for the regulatory
options evaluated in developing the proposed ELGs.
Table 12-1 presents EPA's estimates of benefits and social costs of the regulatory options for existing steam
electric plants, at 3 percent and 7 percent discount rates, and annualized over 24 years. These values are all in
2010 dollars and are based on the discounting of costs and benefits to 2014, the rule promulgation year. As
described in Section 10.3, EPA did not estimate benefits of Options 3a, 3b and 4a; the values included in
Table 12-1 represent benefits inferred based on EPA's analyses of Options 1 through 5.
Table 12-1. Total Annualized Benefits and Social Costs by Regulatory Option and Discount Rate
(Millions; 2010$)
Regulatory Option
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Total Monetized Benefits"
3%
$82.0
$111.7
$311.7
$605.5
$434.1
7%
$65.9
$90.7
$230.4
$424.8
$349.4
Total Monetized Benefits,
Including Adjusted or Inferred
Values
3%
$126.5 b
$139.4 c
$172.3 *
$205.5 c
$311.7
$482.5 c'd
$605.5
$669.6 L
7%
$91.2 b
$104.8 c
$125.6 *
$153.0 c
$230.4
$343.4 c'd
$424.8
$483.7 l
Total Social Costs
3%
$268.3
$185.2
$386.8
$281.4
$572.0
$954.1
$1,381.2
$2,328.8
7%
$259.2
$164.5
$380.8
$257.2
$545.3
$914.7
$1,323.2
$2,209.4
a. The benefit values are the estimated "mean" values. Additional "low" and "high" value estimates are presented in Chapters 3 through
9.
b. EPA did not analyze air-related benefits for Options 1, 2, and 5. This category of benefits was only estimated for Options 3 and 4
(see Chapter 8). To infer benefits of Options 3a, 3b, and 4a, EPA adjusted the total benefits estimated for Options 1, 2 and 5 by
multiplying the totals without air-related benefits by the average ratio of [total with air-related benefits]/[total without air-related
benefits] for Options 3 and 4. See Section 10.3 for details.
c. EPA did not estimate the benefits of this regulatory option. EPA inferred benefits for Options 3a, 3b, and 4a for illustrative purposes
using elements of the more rigorous analysis done to estimate benefits for Options 2, 3 and 4. See Section 10.3 for details.
d. At the time this analysis was conducted, EPA inferred benefits for Option 4a based on pollutant removals for this option of 2,630
million pounds per year. As presented in the TDD (U.S. EPA, 2013b), however, the pollutant removals for this option are 2,620 million
pounds per year, which would result in total annual benefits of $480.8 million (at 3 percent discount and $342.3 million (at 7 percent
discount).
Source: U.S. EPA Analysis, 2013.
April 19, 2013
12-2
-------
Benefit and Cost Analysis for Proposed ELGs 12: Benefits and Social Costs
'
2.2 Analysis of Incremental Benefits and Social Costs
In addition to comparing benefits and costs for each regulatory option, as presented in the preceding section,
EPA also analyzed the benefits and costs of the options on an incremental basis. The comparison in the
preceding section addresses the simple quantitative relationship between estimated benefits and costs for each
option by itself: for a given option, which is greater - costs or benefits - and by how much in relative terms?
In contrast, incremental analysis looks at the differential relationship of benefits and costs across options and
poses a different question: as increasingly more costly options are considered, by what amount do benefits,
costs, and net benefits (i.e., benefits minus costs) change from option to option? Incremental net benefit
analysis provides insight into the net gain to society from imposing increasingly more costly requirements and
can help regulatory decision-makers choose among a set of regulatory proposals that otherwise have a similar
quantitative relationship between benefits and costs based on a one-option-at-a-time comparison.
EPA conducted the incremental net benefit analysis by calculating, for the eight regulatory options, the
change in net benefits, from option to option, in moving from the least stringent option to successively more
stringent options. As described in Chapter 1, the regulatory options differ in the technology basis used to
determine effluent limits and standards for different wastestreams. Thus, the difference in benefits and costs
across the options derives from the characteristics of the wastestreams controlled by an option, the relative
effectiveness of the control technology in reducing pollutant loads, and the distribution and characteristics of
steam electric plants that would implement the technologies and of the receiving waterbodies.
As noted previously, however, the total monetized benefits for Options 1, 2, and 5 do not include air-related
benefits; this benefit category is included in results for Options 3 and 4 only. Further, EPA did not estimate
the benefits of Options 3a, 3b and 4a, but instead inferred these benefits based on the results for Options 1
through 5. Therefore, to allow for consistent calculation of incremental benefits as one moves from one option
to the next, EPA adjusted the total benefits estimated for Options 1, 2, and 5 to account for potential air-
related benefits of these options. EPA used the inferred benefits for Options 3a, 3b and 4a discussed in
Section 10.3.
As reported in Table 12-2, at a 3 percent discount rate, EPA estimates that annual social costs exceed mean
annual monetized benefits by $142 million for Option 1 to $1.6 billion for Option 5. At a 7 percent discount
rate, annual social costs exceed mean annual monetized benefits by $168 million to $1.7 billion for Options 1
and 5.
At a 3 percent discount rate, the incremental change in mean net benefits in moving from Option 1 to
Option 3a is $96 million (the positive value indicates that the increase in benefits is larger than the increase in
costs). Moving from Option 3ato Option 2, the incremental change is an additional -$169 million (the
negative value indicates that the increase in costs is larger than the increase in benefits), while the increment
is $139 million when moving from Option 2 to Option 3b, -$185 million when moving from Option 3b to
Option 3, and -$211 million when moving from Option 3 to Option 4a.
April 19, 2013 12-3
-------
Benefit and Cost Analysis for Proposed ELGs
12: Benefits and Social Costs
Table 12-2. Incremental Net Benefit Analysis (Millions; $2010)
Regulatory
Option"
Option 1
Option 3 a
Option 2
Option 3b
Option 3
Option 4a
Option 4
Option 5
Total Annual Monetized
Benefits, Including
Adjusted or Inferred
Values
3%
$126.5 d
$139.4 e
$172.3 d
$205.5 e
$311.7
$482.5 e'f
$605.5
$669.6 d
7%
$91.2 d
$104.8 e
$125.6 d
$153.0 e
$230.4
$343.1 e'f
$424.8
$483.7 d
Total Social Costs
3%
$268.3
$185.2
$386.8
$281.4
$572.1
$954.1
$1,381.2
$2,328.8
7%
$259.2
$164.6
$380.8
$257.2
$545.3
$914.7
$1,323.2
$2,209.4
Net Annual Monetized
Benefits'1
3%
-$141.9
-$45.8
-$214.5
-$75.9
-$260.4
-$471.6
-$775.7
-$1,659.1
7%
-$167.9
-$59.8
-$255.2
-$104.2
-$314.9
-$571.6
-$898.5
-$1,725.7
Incremental Net Annual
Monetized Benefits0
3%
-$141.9
$96.1
-$168.7
$138.6
-$184.5
-$211.3
-$304.1
-$883.4
7%
-$167.9
$108.1
-$195.4
$151.0
-$210.7
-$256.7
-$326.9
-$827.2
a. Options are presented in order of increasing benefits.
b. Net benefits are calculated by subtracting total annualized costs from total annual monetized benefits.
c. Incremental net benefits are equal to the difference between net benefits of an option and net benefits of the previous, less stringent
option.
d. EPA did not analyze air-related benefits for Options 1, 2, and 5. This category of benefits was only estimated for Options 3 and 4
(see Chapter 8). To infer benefits of Options 3a, 3b, and 4a, EPA adjusted the total benefits estimated for Options 1,2 and 5 by
multiplying the totals without air-related benefits by the average ratio of [total with air-related benefits]/[total without air-related
benefits] for Options 3 and 4. See Section 10.3 for details.
e. EPA did not estimate the benefits of this regulatory option. EPA inferred benefits for Options 3a, 3b, and 4a for illustrative purposes
using elements of the more rigorous analysis done to estimate benefits for Options 2, 3 and 4. See Section 10.3 for details.
f At the time this analysis was conducted, EPA inferred benefits for Option 4a based on pollutant removals for this option of 2,630
million pounds per year. As presented in the TDD (U.S. EPA, 2013b), however, the pollutant removals for this option are 2,620 million
pounds per year, which would result in total annual benefits of $480.8 million (at 3 percent discount and $342.3 million (at 7 percent
discount).
Source: U.S. EPA Analysis, 2013.
April 19, 2013
12-4
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
13 References
Abt Associates Inc., 2010. BenMAP: Environmental Benefits Mapping and Analysis Program, User's
Manual, Section 4.1.10. Bethesda, MD: Abt Associates Inc. Available:
http://www.epa.gov/air/benmap/docs.html [accessed 14 October 2010].
Acton, J.P. 1995. CBO Testimony on the Comprehensive Environmental Response, Compensation, and
Liability Act before the Subcommittee on Superfund, Waste Control, and Risk Assessment
Committee on Environment and Public Works, United States Senate. Congressional Budget Office.
Acton J.P., L.S. Dixon. 1992. Superfund and Transaction Costs: The Experiences of Insurers and Very Large
Industrial Firms. RAND Institute for Civil Justice.
Agency for Toxic Substances and Disease Registry (ATSDR), United States Department of Health and
Human Services. 2007. Toxicological Profile for Lead.
Aiken, R.A. 1985. Public Benefits of Environmental Protection in Colorado. Masters thesis, Colorado State
University.
Alaska Fisheries Science Center (AFSC) of NOAA National Marine Fisheries Service (2010). Alaska
Fisheries Science Center Publications Database.
Anderson, G.D. and S.F. Edwards. 1986. Protecting Rhode Island's Coastal Salt Ponds: "An Economic
Assessment of Downzoning." Coastal Zone Management 14( Yi): 67-91.
Arrow, K., R. Solow, E. Learner, P. Portney, R. Rander, and H. Schuman. 1993. Report of the NOAA Panel
on Contingent Valuation, Federal Register 58 (Jan.): 4602-4614.
Atlantic States Marine Fisheries Commission (ASMFC). (2010). "Managed Species." from
http ://www.asmfc .org/managedspecies .htm.
Axelrad, D.A., D.C. Bellinger, L.M. Ryan, and T.J. Woodruff. 2007. Dose-Response Relationship of Prenatal
Mercury Exposure and IQ: An Integrative Analysis of Epidemiological Data. Environmental Health
Perspectives 115 (4): 609-615.
Azevedo, C., J.A. Herriges, and C.L. Kling. 2001. Valuing Preservation and Improvements of Water Quality
in Clear Lake. Center for Agricultural and Rural Development (CARD), Iowa State University, Staff
Report 01-SR 94.
Barker, S. 2009. Leaks, Water Worries Shared with PA., but TVA Breach Bigger: Sizing Up Fly Ash Ponds'
Spills. Koxnews.com, January 4.
Bateman, I.J. and A.P. Jones. 2003. "Contrasting Conventional with Multi-Level Modeling Approaches to
Meta-Analysis: Expectation Consistency in U.K. Woodland Recreation Values." Land Economics
79(2): 235-258.
Bateman, I.J., B.H. Day, S. Georgiou and I. Lake. 2006. "The Aggregation of Environmental Benefit Values:
Welfare Measures, Distance Decay and Total WTP." Ecological Economics 60(2): 450-460.
Bergstrom, J.C. and L.O. Taylor. 2006. "Using Meta-Analysis for Benefits Transfer: Theory and Practice."
Ecological Economics 60(2):351-360.
Bergstrom, J.C. and P. De Civita. 1999. "Status of Benefits Transfer in the United States and Canada: A
Review." Canadian Journal of Agricultural Economics 47(1): 79-87.
April 19, 2013 13-1
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
Bockstael, N.E., K.E. McConnell and I.E. Strand. 1989. "Measuring the Benefits of Improvements in Water
Quality: The Chesapeake Bay." Marine Resource Economics 6: 1-18.
Bockstael, N.E., McConnell, K.E., and Strand, I.E. 1988. Benefits from improvements in Chesapeake Bay
Water Quality. Department of Agricultural & Resource Economics, University of Maryland, 1988.
Boyle, K.J. and J.C. Bergstrom. 1992. "Benefit Transfer Studies: Myths, Pragmatism and Idealism." Water
Resources Research 28(3):657-663.
Boyle, K.J., P.J. Poor, and L.O. Taylor. 1999. Estimating the Demand for Protecting Freshwater Lakes from
Eutrophication. American Journal of Agricultural Economics 81: 1118-1122.
Breffle, William S., Morey, E., Rowe, R., Waldman, D. and Wytinck, S. 1999. Recreational Fishing Damages
from Fish Consumption Advisories in the Waters of Green Bay. Prepared by Stratus Consulting Inc.
for U.S. Fish and Wildlife Service, U.S. Department of Interior and U.S. Department of Justice, 1999.
Burger, J. 2004. Fish Consumption Advisories: Knowledge, Compliance and Why People Fish in an Urban
Estuary. Journal of Risk Research, 7(5): 463-479.
Cameron, T.A. and D.D. Huppert. 1989. "OLS versus ML Estimation of Non-market Resource Values with
Payment Card Interval Data." Journal of Environmental Economics and Management 17, 230-246.
Cameron, Trudy Ann. 1988. Using the Basic "Auto-validation" Model to Assess the Effect of Environmental
Quality on Texas Recreational Fishing Demand: Welfare Estimates. EPA National Center for
Environmental Economics, 1988.
Canadian Council of Ministers of the Environment (CCME). 2001. Canadian water quality guidelines for the
protection of aquatic life: CCME Water Quality Index 1.0, Technical Report. In: Canadian
environmental quality guidelines, 1999, Canadian Council of Ministers of the Environment,
Winnipeg.
Carruthers, T. and Wazniak, C. 2003. Development of a Water Quality Index for the Maryland Coastal Bays.
Maryland's Coastal Bays: Ecosystem Health Assessment, Chapter 4.4. 2003.
Carson, R.T. and R.C. Mitchell. 1993. "The value of clean water: the public's willingness to pay for beatable,
fishable, and swimmable quality water." Water Resources Research 29(7): 2445-2454.
Carson, R.T., T. Groves, and M.J. Machina. 2000. Incentive and Informational Properties of Preference
Questions. Draft Paper. University of California, La Jolla.
Carson, R.T., W.M. Hanemann, R.J. Kopp, J.A. Krosnick, RC. Mitchell, S. Presser, P.A. Ruud, and V.K.
Smith. 1994. Prospective Interim Lost Use Value due to DDT and PCB Contamination in the
Southern California Bight. Volume 2. Report to the National Oceanic and Atmospheric
Administration, Produced by Natural Resources Damage Assessment Inc., LA Jolla, CA.
Center for Disease Control (CDC), Department of Health and Human Services. 2009. Fourth National Report
on Human Exposure to Environmental Chemicals.
Chambers, J.F., J. Shkolnik, and M. Perez. 2003. Total Expenditures for Students with Disabilities, 1999-
2000: Spending Variation by Disability. Special Education Expenditure Project, United States
Department of Education.
Chambers, J. G., T. B. Parrish, J. J. Harr. 2004. What Are We Spending on Special Education Services in the
United States, 1999-2000? Special Education Expenditure Project, United States Department of
Education.
April 19, 2013 13-2
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
Chiang, A. 1998. A seafood consumption survey of the Laotian community of West Contra Costa County,
CA. Asian Pacific Environmental Network, Oakland.
Clark, E., J.A. Haverkamp, and W.Chapman. 1985. "Eroding Soils: The Off-Farm Impacts." Washington,
DC: The Conservation Foundation.
Clonts, H.A., and J.W. Malone. 1990. Preservation Attitudes and Consumer Surplus in Free Flowing Rivers.
In: Social Science and Natural Resource Recreation Management, Joanne Vining, editor. Westview
Press, Boulder, CO. pp. 301-317.
Croke, Kevin, R.G. Fabian, and G. Brenniman. 1986-87. Estimating the Value of Improved Water Quality in
an Urban River System. Journal of Environmental Systems 16, No. 1. pp. 13-24.
Cude, C. G. 2001. Oregon Water Quality Index: "A Tool for Evaluating Water Quality Management
Effectiveness." Journal of the American Water Resources Association 37(1): 125-137.
Dalhuisen, J., M., R. J. G. M. Florax, H. L. F. de Groot, P. Nijkamp. 2003. Price and Income Elasticities of
Residential Water Demand: A Meta-analysis. Land Economics 79(2): 292-308.
De Zoysa, A.D.N. 1995. A Benefit Evaluation of Programs to Enhance Groundwater Quality, Surface Water
Quality and Wetland Habitat in Northwest Ohio. Dissertation, Ohio State University.
Deacon, J. E., J. D. Kobetich, et al. 1979. "Fishes of North America endangered, threatened, or of special
concern: 1979." Fisheries 4(2): 1-69.
Desvousges, W.H., F.R. Johnson, and H.S. Banzhaf. 1998. Environmental Policy Analysis with Limited
Information: Principles and Applications of the Transfer Method. Northampton, MA: Edward Elgar
Publishers.
Desvousges,W.H., V.K. Smith and A. Fisher. 1987. "Option Price Estimates for Water Quality
Improvements: A Contingent Valuation Study for the Monongahela River." Journal of Environmental
Economics and Management. 14: 248-267.
Dewberry & Davis LLC. 2009. Coal Combustion Waste Impoundment Dam Assessment Report: Martins
Creek Steam Electric Station PPL Generation Bangor, Pennsylvania. Project # 0-381: Assessment of
Dam Safety Coal Combustion Surface Impoundments for the REAC Program.
Dixon L.S., D.S. Drezner, J.K. Hammitt. 1993. Private-Sector Cleanup Expenditures and Transaction Costs at
18 Superfund Sites. RAND Corporation.
Dunnette, D. A. 1979. A Geographically Variable Water Quality Index Used in Oregon. As cited in Cude
2001.
Fann N, Fulcher C, and Hubbell B. 2009. The influence of location, source, and emission type in estimates of
the human health benefits of reducing a ton of air pollution. Air Qual Atmos Health 2:169-76.
Fann N, Baker K.R. and Fulcher C. 2012. Characterizing the PM2.5-related health benefits of emission
reductions for 17 industrial, area and mobile emission sectors across the U.S. Environment
International 49: 141-151.
Freeman III, A.M. 2003. The Measurement of Environmental and Resource Values: Theory and Methods.
Washington, DC: Resources for the Future.
Friedmann, A.S., M.C. Watzin, T. Brinck-Johnsen, J.C. Leiter. 1996. Low levels of dietary methylmercury
inhibit growth and gonadal development in juvenile walleye (Stizostedion vitreum). Aquatic
Toxicology 35: 265-278.
Froese, R. and D. Pauly. 2009. "Fishbase (version 07/2009)." http://www.fishbase.org.
April 19, 2013 "
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
Geotimes. 2003. Settlement Reached on Coal Slurry Spill.
http://www.geotimes.org/oct03AVebExtral01703.html
Glass, G.V. 1976. "Primary, Secondary, and Meta-Analysis of Research." Educational Researcher 5(10): 3 8.
Hayes, K.M., T.J. Tyrell and G. Anderson. 1992. "Estimating the Benefits of Water Quality Improvements in
the Upper Narragansett Bay." Marine Resource Economics 7: 75-85.
Herriges, J.A. and J.F. Shogren. 1996. "Starting point bias in dichotomous choice valuation with follow-up
questioning." Journal of Environmental Economics and Management 30(1):112-131.
Kite, D. 2002. Willingness to Pay for Water Quality Improvements: The Case of Precision Application
Technology. Department of Agricultural Economics and Rural Sociology, Auburn University,
Auburn, AL; 8 August 2002.
Hoehn, J.P., Loomis, J.B., 1993. Substitution effects in the valuation of multiple environmental programs.
Journal of Environmental Economics and Management 25 (1), 56-75.
Huang, J.C., T.C. Haab and J.C. Whitehead. 1997. "Willingness to Pay for Quality Improvements: Should
Revealed and Stated Preference Data Be Combined?" Journal of Environmental Economics and
Management 34(3): 240-255.
Hushak, L. and M. Bielen. 1999. Valuing the Ottawa River: The Economic Values and Impacts of
Recreational Boating. Ohio Sea Grant College Program Technical Bulletin Series Publication OSHU-
TB-042 2000, 2002.
Interagency Working Group on Social Cost of Carbon (IWGSCC), United States Government. Technical
Support Document: Social Cost of Carbon for Regulatory Impact Analysis under Executive Order
12866. February 2010.
Israel, B.D. 2006. State-by-State Guide to NRD Programs in All 50 States. Arnold & Porter LLP.
Jakus, P., M. Downing, M.S. Bevelhimer, J.M. Fly. 1997. Do Sportfish Consumption Advisories Affect
Reservoir Anglers' Site Choice? Agricultural and Resource Economics Review. Vol. 26, 2, Oct.
1997.
Jakus, P., M. McGuinness, and A. Krupnick. 2002. The Benefits and Costs of Fish Consumption Advisories
for Mercury. Resources for the Future Discussion Paper 02-55.
Jelks, H. L., S. J. Walsh, et al. 2008. "Conservation Status of Imperiled North American Freshwater and
Diadromous Fishes." Fisheries 33(8): 372-407.
Johnston, R.J., E.Y. Besedin and R.F. Wardwell. 2003. "Modeling Relationships Between Use and Nonuse
Values for Surface Water Quality: A Meta-Analysis." Water Resources Research 39(12): 1363-1372.
Johnston, R.J., E.Y. Besedin, R. lovanna, C.J. Miller, R.F. Wardwell and M.H. Ranson. 2005. "Systematic
Variation in Willingness to Pay for Aquatic Resource Improvements and Implications for Benefit
Transfer: A Meta-Analysis." Canadian Journal of Agricultural Economics 53(2-3): 221-248.
Johnston, R.J., J.J. Opaluch, T.A. Origanums, and M.J. Mazzotta. 2001. "Estimating Amenity Benefits of
Coastal Farmland." Growth and Change 32(summer): 305-325.
Johnston, R.J., M.H. Ranson, E.Y. Besedin and E.G. Helm. 2006. "What Determines Willingness To Pay Per
Fish? A Meta-Analysis of Recreational Fishing Values." Marine Resource Economics 21: 1-32.
Kahneman, D. and Knetsch, J. 1992. "Valuing Public Goods: The Purchase of Moral Satisfaction" Journal of
Environmental Economics and Management 22: 57-70
April 19, 2013 13-4
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
Kaoru, Y. 1993. "Differentiating Use and Nonuse Values for Coastal Pond Water Quality Improvements."
Environmental and Resource Economics 3: 487-494.
Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi Y, et al. Extended follow-up and spatial analysis of
the American Cancer Society study linking particulate air pollution and mortality. Research report.
Health Effects Institute; 2009. p. 5-114. (discussion 115-36).
Krinsky, I. and A.L. Robb. 1986. "Approximating the statistical properties of elasticities." Review of
Economics and Statistics 68(4):715-719.
Lanphear, B.P., R. Hornung, J. Khoury, et al. 2005. Low-level Environmental Lead Exposure and children's
Intellectual Function: An International Pooled Analysis. Environmental Health Perspectives 113(7):
894-899. As cited in ATSDR (2007).
Lant, C.L., and G.A. Tobin. 1989. "The Economic Value of Riparian Corridors in Cornbelt Floodplains: A
Research Framework." Professional Geographer (41): 337-349.
Lant, C.L., and R. S. Roberts. 1990. "Greenbelts in the Cornbelt: Riparian Wetlands, Intrinsic Values, and
Market Failure." Environment and Planning 22: 1375-1388.
Leggett, C.G. and N.E. Bockstael. 2000. Evidence of the Effects of Water Quality on Residential Land Prices.
Journal of Environmental Economics and Management 39: 121-144.
Lemly, A.D. and J.P. Skopura. 2012. Wildlife and the Coal Waste Policy Debate: Proposed Rules for Coal
Waste Disposal Ignore Lessons from 45 Years of Wildlife Poisoning. Environmental Science and
Technology.
Levy JI, Baxter LK, Schwartz J. Uncertainty and variability in health-related damages from coal-fired power
plants in the United States. Risk Anal (29): 1000-14.
Lichtkoppler, F.R. and T.W. Elaine. 1999. "Environmental Awareness and Attitudes of Ashtabula County
Voters Concerning the Ashtabula River Area of Concern: 1996-1997." Journal of Great Lakes
Resources 25: 500-514.
Lindsey, G. 1994. "Market Models, Protest Bids, and Outliers in Contingent Valuation." Journal of Water
Resources Planning and Management 12: 121-129.
Lipton, D. 2004. "The Value of Improved Water Quality to Chesapeake Bay Boaters." Marine Resource
Economics 19: 265-270
Loomis, J., P. Kent, L. Strange, K. Fausch and A. Covich. 2000. "Measuring the total economic value of
restoring ecosystem services in an impaired river basin: results from a contingent valuation survey."
Ecological Economics 33: 103-117.
Loomis, J.B. and Ekstrand, E. 1997. "Economic Benefits of Critical Habitat for the Mexican Spotted Owl: A
Scope Test Using A Multiple-Bound Contingent Valuation Survey" Journal of Agricultural and
Resource Economics 22(2): 356-366.
Loomis, J.B. 1996. "How Large is the Extent of the Market for Public Goods: Evidence from a Nation Wide
Contingent Valuation Survey." Applied Economics 28(7): 779-782.
Losey John E and Mace Vaughan. 2006. The Economic Value of Ecological Services Provided by Insects.
BioScience 56(4): 311-323.
Lyke, A.J. 1993. Discrete Choice Models to Value Changes in Environmental Quality: A Great Lakes Case
Study. Dissertation submitted to the Graduate School of the University of Wisconsin, Madison.
April 19, 2013 13-5
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
Matthews, L.G., F.R. Romans, and K.W. Easter. 1999. Reducing Phosphorous Pollution in the Minnesota
River: How Much is it Worth? Staff Paper. Department of Applied Economics, University of
Minnesota.
McClelland, N. I. 1974. Water Quality Index Application in the Kansas River Basin. U.S. Environmental
Protection Agency Region 7, Kansas City, Missouri.
Mergler, D., H.A. Anderson, L.H.M. Chan, K.R. Mahaffey, M. Murray, M. Sakamoto, and A.H. Stern. 2007.
Methylmercury Exposure and Health Effects in Humans: A Worldwide Concern. AMBIO: A Journal
of the Human Environment 36: 3-11.
Metrick, A and Weitzman, M. 1996. "Patterns of Behavior in Endangered Species Preservation" Land
Economics 72(1): 1-16.
Minnesota Department of Natural Resources (MDNR). 2010. Managing Minnesota's Shallow Lakes for
Waterfowl and Wildlife. Shallow Lakes Program, Minnesota Division of Fish and Wildlife.
Mitchell, R.C. and R.T. Carson. 1989. Using Surveys to Value Public Goods: The Contingent Valuation
Method. Resources for the Future, Washington, DC.
Moeltner, K., K.J. Boyle, and R.W. Paterson. 2007. "Meta-Analysis and Benefit Transfer for Resource
Valuation—Addressing Classical Challenges with Bayesian Modeling." Journal of Environmental
Economics and Management 53(2): 250-269.
Morse-Jones, S., Bateman, I. J., Kontoleon, A, Ferrini, S, Burgess, N. D., and Turner, R. K. 2012 "Stated
preferences for tropical wildlife conservation amongst distant beneficiaries: Charisma, endemism,
scope and substitution effects" Ecological Economics 78: 9 - 18.
Muller N.Z. and R. Mendelsohn. 2009. Efficient pollution regulation: getting the prices right. American
Economic Review 99: 1714-1739.
National Marine Fisheries Service (NMFS) of the National Oceanic and Atmospheric Administration
(NOAA). 2010a. Fish Stock Sustainability Index: 2009 Quarter 4 Update Through December 31,
2009. Silver Spring, Maryland, U.S. Department of Commerce, NOAA, National Marine Fisheries
Service: 3.
National Marine Fisheries Service (NMFS) of the National Oceanic and Atmospheric Administration
(NOAA). 201 Ob. NOAA Fisheries Geographical Information Systems, Fisheries Data: Critical
Habitat.
National Marine Fisheries Service (NMFS) of the National Oceanic and Atmospheric Administration
(NOAA). 2010c. "NOAA National Marine Fisheries Service Southwest Regional Office, GIS Data."
from http://swr.nmfs.noaa.gov/salmon/layers/finalgis.htm.
National Oceanic and Atmospheric Administration (NOAA). 2010. Environmental Sensitivity Index (ESI)
Maps, Office of Response and Restoration.
National Research Council (NRC). 2002. Coal Waste Impoundments: Risks, Responses, and Alternatives.
Committee on Coal Waste Impoundments, Committee on Earth Resources, Board on Earth Sciences
and Resources, 244 pages, http://www.nap.edu/catalog/10212.html
National Research Council (NRC). 2009. Informing Decisions in a Changing Climate: Panel on Strategies and
Methods for Climate-Related Decision Support. Committee on the Human Dimensions of Global
Change Division of Behavioral and Social Sciences and Education, 198 pages.
April 19, 2013 13-6
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
NatureServe. 2009. "NatureServe Explorer: An Online Encyclopedia of Life." from
http://www.natureserve.org/explorer/.
Navrud, S. and R. Ready. 2007. "Review of Methods for Value Transfer." In: Navrud, S. and R. Ready (eds.),
Environmental Value Transfer: Issues and Methods. Dordrecht, The Netherlands: Springer, pp.1-10.
Niimi, A. J. and G. P. Kissoon. 1994. Evaluation of the critical body burden concept based on inorganic and
organic mercury toxicity to Rainbow Trout (Oncorhynchus mykiss). Archives of Environmental
Contamination and Toxicology 26: 169 - 178.
Nixon, J. 2006. EPA Oks PPL's Cleanup of Delaware River Fly-Ash Spill: Assessment Concludes Additional
Federal Action is Not Required. The Morning Call, http://articles.mcall.com/2006-04-
06/news/3676115_l_cleanup-effort-superfund-law-river-bottom
Northeast Fisheries Science Center (NEFSC) of NOAA National Marine Fisheries Service. 2010. "Northeast
Fisheries Science Center Publications." from http://www.nefsc.noaa.gov/publications/.
Nunes, P.A.D., van der Bergh, J.C.J.M., 2001. Economic valuation of biodiversity: sense or nonsense?
Ecological Economics. 39, 203-222.
Olsen, D., Richards, Jack, and Scott, R. D. 1991. Existence and Sport Values for Doubling the Size of
Columbia River Basin Salmon and Steelhead Runs. Rivers. Vol. 2, No. 1. pp. 44-56.
Opaluch, J.J., T. Origanums, M.J. Mazzotta, J. Diamantides, and R. Johnston. 1998. Resource and
Recreational Economic Values for the Peconic Estuary. Report prepared for Peconic Estuary
Program, Suffolk County Department of Health Services, Riverhead, NY, by Economic Analysis,
Inc., Peace Dale, Rhode Island.
Pacific Islands Fisheries Science Center (PIFSC) of the NOAA National Marine Fisheries Service. 2010a.
"Fishery Biology and Stock Assessment Division." from http://www.pifsc.noaa.gov/fbsad/index.php.
Pacific Islands Fisheries Science Center (PIFSC) of the NOAA National Marine Fisheries Service. 2010b.
Pacific Islands Fisheries Science Center Staff Publications Database.
Phaneuf, D. J., C. L. Kling, and J. A. Herriges. 1998. Valuing Water quality Improvements Using Revealed
Preference Methods When Corner Solutions are Present. American Journal of Agricultural
Economics 80: 1025-1031.
Pincock, Allen & Holt. 2001. Financial Implications in Tailings Dam Management. Pincock Perspectives: 24
(November).
Pittsburgh Water and Sewer Authority/University of Pittsburgh School of Engineering. 2012. Bromides, TDS,
and Radionuclides in the Allegheny River: A Possible Link with Marcellus Shale Operations.
Poe, G.L., K.J. Boyle, and J.C. Bergstrom. 2001. A Preliminary Meta-Analysis of Contingent Values for
Ground Water Quality Revisited. The Economic Value of Water Quality. J.C. Bergstrom, K.J. Boyle,
and G.L. Poe, eds; pp 137-162.
Poor PJ, Boyle KJ, Taylor LO, Bouchard R. 2001. Water clarity in hedonic property value models. Land
Economics 77:482A93.
Richardson, L. and J. Loomis. 2009. "The total economic value of threatened, endangered and rare species:
An updated meta-analysis." Ecological Economics 68(5): 1535-1548.
Roberts, L.A., and J.A. Leitch. 1997. Economic Valuation of Some Wetland Outputs of Mud Lake.
Agricultural Economics Report No. 381, Department of Agricultural Economics, North Dakota
Agricultural Experiment Station, North Dakota State University.
April 19, 2013 1X7"
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
Rollins, K. and Lyke, A. 1998. "The Case for Diminishing Marginal Existence Values" Journal of
Environmental Economics and Management 36: 324-344.
Rosenberger, R.S. and J.B. Loomis. 2000a. "Using Meta-Analysis for Benefit Transfer: In-Sample
Convergent Validity Tests of an Outdoor Recreation Database." Water Resources Research 36(4):
1097-1107.
Rosenberger, R.S. and J.B. Loomis. 2000b. "Panel Stratification in Meta-Analysis of Economic Studies: An
Investigation of Its Effects in the Recreation Valuation Literature." Journal of Agricultural and
Applied Economics 32(3): 459-470.
Rosenberger, R.S. and R.J. Johnston. 2007. Selection Effects in Meta-Valuation Function Transfers. Meta-
Analysis in Economics Conference, September 27-30, 2007, Sandbjerg Manor, S0nderborg,
Denmark.
Rosenberger, R.S. and T.D. Stanley. 2006. "Measurement, Generalization, and Publication: Sources of Error
in Benefit Transfers and Their Management." Ecological Economics 60(2): 372-378.
Rosenberger, R.S. and T.T. Phipps. 2007. "Correspondence and Convergence in Benefit Transfer Accuracy:
Meta-Analytic Review of the Literature." In: Navrud, S. and R. Ready (eds.), Environmental Value
Transfer: Issues and Methods. Dordrecht, The Netherlands: Springer, pp.23-43.
Rowe, R.D., W.D. Schulze, B. Hurd and D. Orr. 1985. Economic Assessment of Damage Related to the Eagle
Mine Facility. Energy and Resource Consultants, Inc., Boulder, CO.
Ruhl, L., A. Vengosh, G.S. Dwyer, H. Hsu-Kim, G. Schwartz, A. Romanski, S.D. Smith. 2012. The Impact of
Coal Combustion Residue Effluent on Water Resources: A North Carolina Example. Environmental
Science and Technology 2012 Oct 15.
Salkever, D.S. 1995. Updated Estimates of Earnings Benefits from Reduced Exposure of Children to
Environmental Lead. Environmental Research 70: 1-6.
Sanders, L. B., R.G. Walsh, and J.B. Loomis. 1990. "Toward Empirical Estimation of the Total Value of
Protecting Rivers." Water Resources Research. 26(7): 1345-1357.
Santos, J.M.L. 1998. The Economic Valuation of Landscape Change. Cheltenham, UK: Edward Elgar.
Schulze, W.D., RD. Rowe, W.S. Breffle, R.R Boyce, and G.H. McClelland. 1995. Contingent Valuation of
Natural Resource Damages Due to Injuries to the Upper Clark Fork River Basin. State of Montana,
Natural Resource Damage Litigation Program. Prepared by: RCG/Hagler Bailly, Boulder, CO.
Schoch, N., M. Glennon, D. Evers, M. Duron, A. Jackson, C. Driscoll, X. Yu, H. Simonin. 2011. Long-term
monitoring and assessment of mercury based on integrated sampling efforts using the common loon,
prey fish, water, and sediment. New York State Energy Research and Development Authority Report
No. 12-06, 116pp.
Schwartz, J. 1994. Low-Level Lead Exposure and Children's IQ: A Meta-analysis and Search for a
Threshold. Environmental Research 65: 42-55.
Shrestha, R. K., J.R.R. Alavalapati. 2004. "Valuing Environmental Benefits of Silvopasture Practice: A Case
Study of the Lake Okeechobee Watershed in Florida." Ecological Economics 49: 349-359.
Shrestha, R., R. Rosenberger and J. Loomis. 2007. "Benefit Transfer Using Meta-Analysis in Recreation
Economic Valuation, in Benefit Transfer Accuracy: Meta-Analytic Review of the Literature." In:
Navrud, S. and R. Ready (eds.). Environmental Value Transfer: Issues and Methods. Dordrecht, The
Netherlands: Springer, pp. 161-177.
April 19, 2013 13-8
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
Singer, J.D. 1998. "Using SAS PROC MIXED to fit multilevel models, hierarchical models, and individual
growth models." Journal of Educational and Behavioral Statistics 23(4): 323-355.
Smith, V.K., and L. Osborne. 1996. "Do Contingent Valuation Estimates Pass the Scope Test? A Meta
Analysis." Journal of Environmental Economics and Management 31(3): 287-301.
Smith, V.K., and S.K. Pattanayak. 2002. "Is Meta-Analysis a Noah's Ark for Non-Market Valuation?"
Environmental and Resource Economics 22: 271-296.
Smith, V.K., and Y. Kaoru. 1990. "Signals or Noise? Explaining the Variation in Recreation Benefit
Estimates." American Journal of Agricultural Economics 42(2): 419-433.
Smith, V.K., G. Van Houtven, and S.K. Pattanayak. 2002. "Benefit Transfer via Preference Calibration:
'Prudential Algebra' for Policy." Land Economics 78(1): 132-152.
Southeast Fisheries Science Center (SEFSC) of NOAA National Marine Fisheries Service. 2010. SouthEast
Data, Assessment and Review (SEDAR).
Southwest Fisheries Science Center (SWFSC) of NOAA National Marine Fisheries Service. 2010.
"Southwest Fisheries Science Center Fisheries Biology Publications Database." from
http://swfsc.noaa.gov//publications/fedbin/qrypublications.asp?ParentMenuId=54.
Spash, C.L. and A. Vatn. 2006. "Transferring Environmental Value Estimates: Issues and Alternatives."
Ecological Economics 60(2): 379-388.
Stanley DL. 2005. Local perception of public goods: recent assessments of willingness-to-pay for endangered
species. Contemp. Econ. Policy 23:165-79
Steeves, P. and D. Nebert. 1994. l:250,000-scale Hydrologic Units of the United States. U.S. Geological
Survey, Reston, VA.
Steinhardt, B., J.F. Donaghy, F.A. Featherston. 1994. Report to Congressional Requesters: Superfund Legal
Expenses for Cleanup-Related Activities of Major U.S. Corporations. United States General
Accounted Office (U.S. GAO).
Stumborg, B.E., K.A. Baerenklau and R.C. Bishop. 2001. "Nonpoint Source Pollution and Present Values: A
Contingent Valuation of Lake Mendota." Review of Agricultural Economics. 23(1): 120-132.
Sutherland, R. J., and R.G. Walsh. 1985. "Effect of Distance on the Preservation Value of Water Quality."
Land Economics 61(3): 282-291.
Swartout, J. and G. Rice. 2000. Uncertainty Analysis of the Estimated Ingestion Rates Used to Derive the
Methylmercury Reference Dose. Drug and Chemical Toxicology 23 (1): 293-306.
Taylor, C. A., G. A. Schuster, et al. 2007. "A Reassessment of the Conservation Status of Crayfishes of the
United States and Canada after 10+ Years of Increased Awareness." Fisheries 32(8): 372-389.
Taylor, C. A., M. L. Warren, et al. 1996. "Conservation Status of Crayfishes of the United States and
Canada." Fisheries 21(4): 25-38.
Tennessee Valley Authority (TVA). 2011. Form 10-Q: Quarterly Report Pursuant to Sections 13 or 15(d).
Filing Date: 2011-08-11, Period of Report: 2011-06-30.
Tennessee Valley Authority (TVA). 2010. Kingston Ash Recovery Project Non-Time-Critical Removal
Action Embayment/Dredge Cell Action Memorandum. Document No. EPA-AO-024.
Tetra Tech. 2011. Evaluating Sustainability of Projected Water Demands on Future Climate Change
Scenarios. Prepared for Natural Resources Defense Council. Prepared by Tetra Tech.
April 19, 2013 ~
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
Tilden, J., et al. 1997. Health Advisories for Consumers of Great Lakes Sport Fish: Is the Message Being
Received?. Environmental Health Perspectives, Vol 105, 12.
United States Census Bureau. 1990. Census of Housing: Detailed Housing Characteristics. "Source of Water".
From http://www.census.gov/prod/cenl990/ch2/ch-2.html
United States Bureau of Labor Statistics (U.S BLS). 2010. "Consumer Price Index." Retrieved October 3,
2010, from http://www.bls.gov/cpi/data.htm.
United States Census Bureau. 2010a. State & County QuickFacts. http: //quickfacts .census. gov/qfd/index.html
United States Census Bureau. 2010b. QT-P2 Single Years of Age and Sex: 2010, Summary File.
http://factfinder2.census.gov/face s/tableservices/jsf/pages/productview.xhtml?pid=DEC_10_SFl_QT
P2&prodType=table
United States Department of Health and Human Services (U.S. DHHS). 1995. Final report: health study to
assess the human health effects of mercury exposure to fish consumed from the Everglades. Florida
Department of Health and Rehabilitative Services.
United States Department of Health and Human Services (U.S. DHHS). 2008. Draft Toxicological Profile for
Cadmium.
United States Environmental Protection Agency (U.S. EPA). 2003. Economic and Environmental Benefits
Analysis Document for the Final Effluent Limitations Guidelines and Standards for the Metal
Products and Machinery Point Source Category. EPA-821-B-03-002.
United States. Environmental Protection Agency (U.S. EPA). 2004a. Final regulatory impact analysis:
industrial boilers rule. Prepared by Office of Air and Radiation.
United States. Environmental Protection Agency (U.S. EPA). 2004b. Final regulatory impact analysis: clean
air nonroad diesel rule. Prepared by Office of Air and Radiation.
United States Environmental Protection Agency (U.S. EPA). 2004c. Regional Analysis for the Final Section
316(b) Phase II Existing Facilities Rule. U.S. EPA, Office of Water, Washington, DC, February
2004. EPA-821-R-04-006
United States Environmental Protection Agency (U.S. EPA). 2005. Regulatory Impact Analysis of the Clean
Air Mercury Rule. EPA-452/R-05-003.
United States Environmental Protection Agency (U.S. EPA). 2008a. Welcome to STORET, EPA's largest
computerized environmental data system. Available at http://www.epa.gov/storet/.
United States Environmental Protection Agency (U.S. EPA). 2008b. Technical Support Document:
Calculating Benefit Per-Ton Estimates. Final Ozone NAAQS regulatory impact analysis. Office of
Air Quality Planning and Standards, Office of Air and Radiation.
United States Environmental Protection Agency (U.S. EPA). 2009a. Steam Electric Power Generating Point
Source Category: Final Detailed Study. EPA 821-R-09-008.
United States Environmental Protection Agency (U.S. EPA). 2009b. Environmental Impact and Benefits
Assessment for Final Effluent Guidelines and Standards for the Construction and Development
Category. November, 2009. http://www.epa.gov/guide/construction/files/environment.pdf.
United States Environmental Protection Agency (U.S. EPA). 2009c. Software and User's Manuals.
http://www.epa.gov/superfund/lead/products.htm
April 19, 2013 13-10
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
United States Environmental Protection Agency (U.S. EPA). 2010a. 75 Federal Register 4173- 4226. 2010.
40 CFR Part 131. Water Quality Standards for the State of Florida's Lakes and Flowing Waters;
Proposed Rule. Environmental Protection Agency; January 26, 2010.
United States Environmental Protection Agency (U.S. EPA). 2010b. Guidelines for Preparing Economic
Analyses. EPA 240-R-10-001.
United States Environmental Protection Agency (U.S. EPA). 2010c. IRIS Toxicological Review of Inorganic
Arsenic (Cancer) (External Draft Review).
http://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=219111
United States Environmental Protection Agency (U.S. EPA). 2010d. Regulatory Impact Analysis for EPA's
Proposed RCRA Regulation of Coal Combustion Residues (CCR) Generated by the Electric Utility
Industry.
United States Environmental Protection Agency (U.S. EPA). 2010e. Questionnaire for the Steam Electric
Power Generating Effluent Guidelines, OMB Control 2040-0281, May 20, 2010.
United States Environmental Protection Agency (U.S. EPA). 2011. "Ecological Toxicity Information"
Available at: http://www.epa.gov/reg5sfun/ecology/toxprofiles.htm
United States Environmental Protection Agency (U.S. EPA). 2012a. National Primary Drinking Water
Regulations, http://water.epa.gov/drink/contaminants/index.cfm
United States Environmental Protection Agency (U.S. EPA). 2012b. National Recommended Water Quality
Criteria, http://water.epa.gov/scitech/swguidance/standards/criteria/current/index.cfmtfhhtable
United States Environmental Protection Agency (U.S. EPA). 2012c. Risk-Screening Environmental Indicators
(RSEI). http://www.epa.gov/oppt/rsei/
United States Environmental Protection Agency (U.S. EPA). 2012d. Information Request Responses from
Electric Utilities. Database of Survey Response (April 12, 2012).
http://www.epa.gov/osw/nonhaz/industrial/special/fossil/surveys/index.htm
United States Environmental Protection Agency (U.S. EPA). 2012e. Information Request Responses from
Electric Utilities. Summary Table for Impoundment Reports (October 10, 2012).
http://www.epa.gov/osw/nonhaz/industrial/special/fossil/surveys2/index.htm
United States Environmental Protection Agency (U.S. EPA). 2013a. Environmental Assessment for the
Proposed Effluent Limitations Guidelines and Standards for the Steam Electric Power Generating
Point Source Category (40 CFR 423). EPA-821-R-13-003
United States Environmental Protection Agency (U.S. EPA). 2013b. Technical Development Document for
the Proposed Effluent Limitations Guidelines and Standards for the Steam Electric Power Generating
Point Source Category. EPA-821-R-13-002
United States Environmental Protection Agency (U.S. EPA). 2013c. Regulatory Impact Analysis for Proposed
Effluent Limitations Guidelines and Standards for the Steam Electric Power Generating Point Source
Category. EPA-821-R-13-005. DCN SE03170
United States Fish and Wildlife Service (U.S. FWS). 2009. Federal and State Endangered and Threatened
Species Expenditures: Fiscal Year 2008. US Fish and Wildlife Service Endangered Species Program:
252.
United States Fish and Wildlife Service (U.S. FWS). 2006. National Survey of Fishing, Hunting, and
Wildlife-Associated Recreation. FHW/06-NAT.
April 19, 2013 13-11
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
United States Fish and Wildlife Service (U.S. FWS). 2010a. "Environmental Conservation Online System."
United States Fish and Wildlife Service (U.S. FWS). 2010b. "Critical Habitat Portal."
United States Fish and Wildlife Service (U.S. FWS). 2010c. "North Florida Ecological Services Office:
Loggerhead Sea Turtle (Caretta caretta)." from
http://www.fws.gov/northflorida/seaturtles/turtle%20factsheets/loggerhead-sea-turtle.htm.
United States Geological Survey (USGS). 2007. National Hydrography Dataset (NHD). Available at
http://nhd.usgs.gov/data.html
United States Geological Survey (USGS). 2008. National Water Information System - Web Interface. USGS
Surface Water Data for the Nation Available at http://waterdata.usgs.gov/nwis/sw. Accessed June 5,
2008.
United States Office of Management and Budget (OMB). 2003. Circular A-4: Regulatory Analysis.
Utah DNR, 2005. Utah Department of Natural Resources. "Utah Waterfowl Advisories"
http://waterfowladvisories.utah.gov/advisories.htm
Vandenberg, T.P., G.L. Poe and J.R. Powell. 2001. "Assessing the accuracy of benefits transfers: Evidence
from a multi-site contingent valuation study of ground water quality." In: The Economic Value of
Water Quality. Bergstrom, J.C., K.J. Boyle and G.L. Poe (eds.). Edward Elgar Publishers,
Northampton, MA.
Vaughan, W.J. 1986. "The RFF Water Quality Ladder." In: Mitchell, RC. and RT. Carson. The Use of
Contingent Valuation Data for Benefit/Cost Analysis in Water Pollution Control, Final Report.
Washington: Resources for the Future. Appendix B.
Viscusi, K.W, J. Huber and J. Bell. 2008. "The Economic Value of Water Quality." Environmental and
Resource Economics 41: 169 - 187.
Welle, P.G. 1986. Potential Economic Impacts of Acid Deposition: A Contingent Valuation Study of
Minnesota. Dissertation, University of Wisconsin-Madison.
Wey, K.A. 1990. Social Welfare Analysis of Congestion and Water Quality of Great Salt Pond, Block Island,
Rhode Island. Dissertation, University of Rhode Island.
Wheaton, W.D., J.C. Cajka, B.M. Chasteen, O.K. Wagener, P.C. Cooley, L. Ganapathi, D.J. Roberts, and J.L.
Allpress. 2009. Synthesized Population Databases: A U.S. Geospatial Database for Agent-Based
Models. Methods Report. RTI, International Press.
White, P.C.L., Bennett, A.C., Hayes, E.J.V., 2001. The use of willingness-to-pay approaches in mammal
conservation. Mammal Review 31 (2), 151-167.
White, P.C.L., Gregory, K.W., Lindley, P.J. & Richards, G. (1997) Economic values of threatened mammals
in Britain: a case study of the otter Lutra lutra and the water vole Arvicola terrestris. Biological
Conservation, 82, 345-354.
Whitehead, J.C. and G.C. Blomquist. 1991a. "A Link Between Behavior, Information, and Existence Value."
Leisure Sciences 13: 97 -109.
Whitehead, J.C. and G.C. Blomquist. 1991b. "Measuring contingent values for wetlands: Effects of
information about related environmental goods." Water Resources Research 27(10): 2523-2531.
Whitehead, J.C., and P.A. Groothuis. 1992. "Economic Benefits of Improved Water Quality: a case study of
North Carolina's Tar-Pamlico River." Rivers 3: 170-178.
April 19, 2013 13-12
-------
Benefit and Cost Analysis for Proposed ELGs 13: References
Whitehead, J.C., G.C. Blomquist, T.J. Hoban and W.B. Clifford. 1995. "Assessing the Validity and
Reliability of Contingent Values: A Comparison of On-Site Users, Off-Site Users, and Non-users."
Journal of Environmental Economics and Management 29: 238-251.
Whitehead, J.C., T. Hoban and W. Clifford. 2002. Landowners' Willingness to Pay for Water Quality
Improvements: Jointly Estimating Contingent Valuation and Behavior with Limited Information.
White paper developed in part by U.S. EPA, NCDENR and the College of Agriculture and Life
Sciences at NSCU.
Whittington, D., G. Cassidy, D. Amaral, E. McClelland, H. Wang and C. Poulos. 1994. The Economic Value
of Improving the Environmental Quality of Galveston Bay. Department of Environmental Sciences
and Engineering, University of North Carolina at Chapel Hill. GBNEP-38, 6/94.
Wilcove, D. S. and L. L. Master. 2005. "How many endangered species are there in the United States?"
Frontiers in Ecology and the Environment 3(8): 414-420.
Williams, C.D., Burns, J.W., Chapman, A.D., Flewelling, L., Pawlowicz, M. and W. Carmichael. 2001.
Assessment of cyanotoxins in Florida's lakes, reservoirs and rivers. White paper produced for Florida
Harmful Algal Bloom Task Force.
Williams, J. D., M. L. Warren, et al. 1993. "Conservation Status of Freshwater Mussels of the United States
and Canada." Fisheries 18(9): 6-22.
Williams, J. E., J. E. Johnson, et al. 1989. "Fishes of North America Endangered, Threatened, or of Special
Concern: 1989." Fisheries 14(6): 2-20.
Williams, R., et al. 2000. An examination offish consumption by Indiana recreational anglers: An onsite
survey. Technical Report 99-D-HDFW-2. Dept. of Forestry and Natural Resources, Purdue
University.
Woods & Poole Economics Inc. 2011. Complete Demographic Database. Washington, DC.
Woodward, R.T. and Y.S. Wui. 2001. "The economic value of wetland services: ameta-analysis." Ecological
Economics 37(2): 257-270.
April 19, 2013 13-13
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix A: Affected Population
Appendix A: Detail on Estimating Affected Population
Exhibit A-1 shows the state- and region-specific population data and percent of anglers in the state or regional
population. These data are used to identify the affected recreational angler population and their household
members who are potentially exposed to steam electric pollutants via consumption of contaminated fish
caught in receiving reaches (as described in Section 3.1.1).
Exhibit A-1. Population Data and Recreational Fishing Rates by State
East North Central Average
Illinois
Indiana
Michigan
Ohio
Wisconsin
East South Central Average
Alabama
Kentucky
Mississippi
Tennessee
Middle Atlantic Average
New Jersey
New York
Pennsylvania
Mountain Average
Arizona
Colorado
Idaho
Montana
Nevada
New Mexico
Utah
Wyoming
New Englan d A verage
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
Pacific Average
Alaska
California
Hawaii
Oregon
Washington
South Atlantic Average
Delaware
% of Population Who
Fish3
11.51%
11.51%
11.51%
11.51%
11.51%
11.51%
14.02%
14.02%
14.02%
14.02%
14.02%
6.39%
6.39%
6.39%
6.39%
10.52%
10.52%
10.52%
10.52%
10.52%
10.52%
10.52%
10.52%
10.52%
8.76%
8.76%
8.76%
8.76%
8.76%
8.76%
8.76%
6.66%
6.66%
6.66%
6.66%
6.66%
6.66%
10.99%
10.99%
Persons per Household1"
2.57
2.62
2.49
2.53
2.47
2.42
2.57
2.48
2.47
2.6
2.49
2.59
2.68
2.64
2.46
2.66
2.76
2.53
2.64
2.49
2.66
2.61
3.14
2.45
2.48
2.55
2.36
2.54
2.54
2.52
2.39
2.72
2.82
2.91
2.84
2.49
2.52
2.50
2.58
Population1"
12,830,632
6,483,802
9,883,640
11,536,504
5,686,986
4,779,736
4,339,367
2,967,297
6,346,105
8,791,894
19,378,102
12,702,379
6,329,017
5,029,196
1,567,582
989,415
2,700,551
2,059,179
2,763,885
563,626
3,518,288
1,328,361
6,547,629
1,316,470
1,052,567
625,741
710,231
36,756,666
1,360,301
3,831,074
6,724,540
897,934
April 19, 2013
A-1
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix A: Affected Population
Exhibit A-1. Population Data and Recreational Fishing Rates by State
District of Columbia
Florida
Georgia
Maryland
North Carolina
South Carolina
Virginia
West Virginia
West North Central Average
Iowa
Kansas
Minnesota
Missouri
Nebraska
North Dakota
South Dakota
West South Central Average
Arkansas
Louisiana
Oklahoma
Texas
Average
% of Population Who
Fish3
10.99%
10.99%
10.99%
10.99%
10.99%
10.99%
10.99%
10.99%
16.78%
16.78%
16.78%
16.78%
16.78%
16.78%
16.78%
16.78%
12.29%
12.29%
12.29%
12.29%
12.29%
11.10%
Persons per Household1"
2.21
2.52
2.7
2.63
2.47
2.52
2.54
2.37
2.41
2.36
2.46
2.45
2.47
2.45
2.24
2.43
2.60
2.48
2.61
2.49
2.81
2.55
Population1"
601,723
18,801,310
9,687,653
5,773,552
9,535,483
4,625,364
8,001,024
1,852,994
3,046,355
2,853,118
5,303,925
5,988,927
1,826,341
672,591
814,180
2,855,390
4,533,372
3,751,351
25,145,561
6,101,167
a. Source: U.S. FWS (2006); based on data at Census Region.
b. Source: U.S. Census Bureau (2010a).
EPA conducted a literature review of anglers' awareness of and responses to fish consumption advisories
(FCA), as summarized in Exhibit A-2 and Exhibit A-3.
Exhibit A-2. Angler Awareness of Fish Consumption Advisories
Source
U.S. DHHS, 1995
May and Burger, 1996
Connelly et al., 1996 (as cited in Jakus, et
al., 1997)
Tildenetal., 1997
Chiang, 1998
Burger et al., 1999 (as cited in Burger,
2004)
Williams etal., 2000
Jakus et al., 2002
Burger, 2004
Average
Survey Area
Florida Everglades
New York/New Jersey Harbor
Lake Ontario
Great Lakes
San Francisco Bay
New York/New Jersey Harbor
Indiana
Maryland Chesapeake Bay
New York/New Jersey Harbor
-
Percentage of Anglers Aware of Fish
Consumption Advisories
71.0%b
__,,
>95.0%a
__,,
__!,
60.0%b
__,,
48.0%c
__!,
57.0% to 61.2%d
a. Based on mail-in surveys of anglers with fishing licenses.
b. Based on with interviews with anglers on-site.
c. Based on a compilation of figures from the literature.
d. Lower estimate represents average without inclusion of estimate from Connelly et al. (1996).
April 19, 2013
A-2
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix A: Affected Population
Exhibit A-3. Percentage of Anglers who "Ignore" Fish Consumption Advisories
Source
Diana et al., 1993 (as cited in Jakus, et al,
1997)
Velicer et al., 1994 (as cited in Jakus, et
al., 1997)
U.S. DHHS, 1995
May and Burger, 1996 (as cited in Jakus,
etal., 1997)
MacDonald et al., 1997 (as cited in Jakus,
etal., 1997)
Chiang, 1998
Jakus et al., 2002
Average
Survey Area
Lake Ontario (New York)
Lake Ontario (New York)
Florida Everglades
New York/New Jersey Harbor
Maine
San Francisco Bay
Maryland Chesapeake Bay
-
Percentage of Anglers who "Ignore"
Fish Consumption Advisories"
70%
Lowb
74%
70%-88%
> 75%
40%
74%
71.6% to 76.1%c
a. "Ignore", in this case, means consuming fish caught from contaminated waters with knowledge that a waterbody has FCA.
b. Sample was composed of angler group 'opinion leaders' whose actions may not be representative of the general angling population.
c. Higher estimate represents average without inclusion of the estimate from Chiang (1998), which is an outlier.
EPA calculated site-specific population growth projections (relative to 2010) based on data from the 2000
Census and projections from Woods and Poole Economics (2007). Exhibit A- 4 shows the average population
growth for years 2010 through 2040.
Exhibit A- 4. Average Population Growth, Relative to 201 Oa
Year
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
Population Growth
.00
.01
.02
.02
.03
.04
.05
.06
.06
.07
.08
.09
.10
.11
.11
.12
Year
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Population Growth
.13
.14
.15
.16
.17
.18
.19
.19
.20
.21
.22
.23
.23
.24
.25
a. Estimates are based on data from the 2000 Census and projections from the 2007 Woods & Poole projection estimates (Woods
and Pool Economics, 2007).
Exhibit A-5 shows the state- and region-specific data on fertility and the number of births per year. These data
were used to identify the number of affected infants susceptible to in-utero exposure to mercury due to
maternal consumption of contaminated fish (as described in Section 3.3.1).
April 19, 2013
A-3
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix A: Affected Population
Exhibit A-5. Fertility Data and the Number of Birth per Year
East North
Central
Illinois
Indiana
Michigan
Ohio
Wisconsin
East South
Central
Alabama
Kentucky
Mississippi
Tennessee
Middle Atlantic
New Jersey
New York
Pennsylvania
Mountain
Arizona
Colorado
Idaho
Montana
Nevada
New Mexico
Utah
Wyoming
New England
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
Pacific
Alaska
California
Hawaii
Oregon
Washington
South Atlantic
Delaware
District of
Columbia
Florida
Population"
12,830,632
6,483,802
9,883,640
11,536,504
5,686,986
4,779,736
4,339,367
2,967,297
6,346,105
8,791,894
19,378,102
12,702,379
6,329,017
5,029,196
1,567,582
989,415
2,700,551
2,059,179
2,763,885
563,626
3,518,288
1,328,361
6,547,629
1,316,470
1,052,567
625,741
710,231
36,756,666
1,360,301
3,831,074
6,724,540
897,934
601,723
18,801,310
Number of
Women 15
to50b
2,309,405
3,268,056
1,573,390
2,484,965
2,835,019
1,385,594
1,122,715
1,155,351
1,057,585
734,070
1,543,855
3,397,994
2,168,090
5,010,234
3,015,658
667,432
1,548,377
1,253,964
366,226
228,795
632,088
484,932
696,422
128,655
603,410
869,841
317,790
1,676,571
330,453
271,493
154,314
2,485,786
173,966
9,399,440
305,513
917,183
1,632,826
1,613,949
218,161
174,240
4,301,117
Percent of
Population
who are
Women 15
to50c
25%
25%
24%
25%
25%
24%
24%
24%
24%
25%
24%
25%
25%
26%
24%
24%
24%
25%
23%
23%
23%
24%
25%
23%
25%
25%
24%
26%
25%
26%
25%
24%
24%
26%
22%
24%
24%
25%
24%
29%
23%
Number of
Women (15
to 50) who
had a Birth
in the Last
12 Months'1
128,918
185,373
92,540
131,068
157,318
78,292
68,758
67,289
61,009
46,721
100,013
183,325
131,141
256,497
162,336
42,416
92,263
73,670
25,718
11,761
39,633
30,999
56,952
8,329
29,668
42,687
14,234
83,335
15,368
14,722
7,660
150,044
12,233
562,297
19,404
55,854
100,430
89,988
11,059
6,915
236,097
Fertility
Rated
5.6%
5.7%
5.9%
5.3%
5.5%
5.7%
6.1%
5.8%
5.8%
6.4%
6.5%
5.5%
6.0%
5.1%
5.4%
6.4%
6.0%
5.9%
7.0%
5.1%
6.3%
6.4%
8.2%
6.5%
4.9%
4.9%
4.5%
5.0%
4.7%
5.4%
5.0%
6.3%
7.0%
6.0%
6.4%
6.1%
6.2%
5.4%
5.1%
4.0%
5.5%
Babies per
year per
1000 People6
13.9
14.4
14.3
13.3
13.6
13.8
14.9
14.1
14.1
15.7
15.8
13.6
14.9
13.2
12.8
15.3
14.6
14.6
16.4
11.9
14.7
m
2O6
14.8
12.2
12.1
10.7
12.7
11.7
14.0
12.2
15.3
17.2
15.3
14.3
14.6
14.9
13.3
12.3
11.5
12.6
April 19, 2013
A-4
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix A: Affected Population
Exhibit A-5. Fertility Data and the Number of Birth per Year
Georgia
Maryland
North Carolina
South Carolina
Virginia
West Virginia
West North
Central
Iowa
Kansas
Minnesota
Missouri
Nebraska
North Dakota
South Dakota
West South
Central
Arkansas
Louisiana
Oklahoma
Texas
Average
Population"
9,687,653
5,773,552
9,535,483
4,625,364
8,001,024
1,852,994
3,046,355
2,853,118
5,303,925
5,988,927
1,826,341
672,591
814,180
2,855,390
4,533,372
3,751,351
25,145,561
6,101,167
Number of
Women 15
to50b
2,518,541
1,464,034
2,324,058
1,108,257
1,987,524
429,609
705,711
715,128
682,702
1,303,385
1,465,150
429,867
154,315
189,429
2,218,672
690,561
1,124,309
885,806
6,174,011
1,504,688
Percent of
Population
who are
Women 15
to50c
26%
25%
24%
24%
25%
23%
24%
23%
24%
25%
24%
24%
23%
23%
24%
24%
25%
24%
25%
24%
Number of
Women (15
to 50) who
had a Birth
in the Last
12 Months'1
144,020
80,639
131,969
68,314
107,196
23,683
43,565
46,953
43,792
84,639
81,734
25,825
9,820
12,190
146,147
39,910
71,036
55,386
418,254
87,974
Fertility
Rated
5.7%
5.5%
5.7%
6.2%
5.4%
5.5%
6.3%
6.6%
6.4%
6.5%
5.6%
6.0%
6.4%
6.4%
6.3%
5.8%
6.3%
6.3%
6.8%
5.9%
Babies per
year per
1000 People6
14.9
14.0
13.8
14.8
13.4
12.8
14.9
15.4
15.3
16.0
13.6
14.1
14.6
15.0
15.3
14.0
15.7
14.8
16.6
14.3
a. Source: U.S. Census Bureau (2010a).
b. Source: U.S. Census Bureau (2010c).
c. Number of Women of childbearing age (15 to 50) divided by Population.
d. Number of Women (15 to 50) who had a Birth in the Last 12 Months divided by the Number of Women 15 to 50.
e. Number of Women (15 to 50) who had a Birth in the Last 12 Months divided by Population times 1000.
April 19, 2013
A-5
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix B: IEUBK Model
Appendix B: IEUBK Model Description and Application
The Integrated Exposure Uptake Biokinetic (IEUBK) model uses exposure, uptake, and biokinetic response
information to estimate the blood lead concentration (PbB) level distribution for a population of children
receiving similar exposures (U.S. EPA, 2009c). The estimated distribution may be used to predict the
probability of elevated PbB levels in children exposed to environmental lead. The model has four
components. Each component reflects a different aspect of the overall biological process:
> The multimedia nature of exposure to lead via soil, house dust, drinking water, air, and food (i.e.,
the exposure component},
> The differential bioavailability of various sources of lead (i.e., the uptake component),
> The pharmacokinetics of internal distribution of lead to bone, blood, and other tissues (i.e., the
biokinetic component}, and
> Inter-individual variability of PbB levels (i.e., the variability component}.
The model uses estimated or measured lead concentration in fish tissues and other media to estimate a
continuous exposure pattern for children from birth through the seventh birthday and predicts the geometric
mean PbB for a population of children receiving similar exposures. The model also generates the percentage
of children with PbB levels in excess of a user-specified level of concern (in the case of this analysis, 20
ug/dL).
The IEUBK model has default values for exposure and uptake parameters, which are shown in Exhibit B-l.
However, users can also change these default values according to site- or population-specific factors. For this
analysis, EPA entered the receiving reach-specific LADD for each cohort (a product of the EA; U.S. EPA,
2012a) as an "alternate source," and assumed an uptake rate of 50 percent (consistent with the default dietary
bioavailability rate). Additionally, EPA changed the default "cutoff PbB level from 10 ug/dL to 20 ug/dL to
estimate the percent of children who would have PbB above that level. For all other variables, EPA used the
default IEUBK values.
Exhibit B-1. Default Values for IEUBK Model Parameters
Parameter
Indoor air lead concentration (% of outdoor)
Default
Value
30
Units
%
Air (by year)
Air concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
Time outdoors
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-7 years (36-84 mo)
0.10
0.10
0.10
0.10
0.10
0.10
0.10
1
2
3
4
jig/ms
jig/ms
jig/ms
jig/ms
jig/ms
jig/ms
jig/ms
h/day
h/day
h/day
h/day
April 19, 2013
B-1
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix B: IEUBK Model
Exhibit B-1. Default Values for IEUBK Model Parameters
Parameter
Ventilation rate
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
Lung absorption
Default
Value
2
3
5
5
5
7
7
32
Units
ms/day
ms/day
ms/day
ms/day
ms/day
ms/day
ms/day
%
Diet (by year)
Dietary lead intake
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
2.26
1.96
2.13
2.04
1.95
2.05
2.22
ug Pb /day
ug Pb /day
ug Pb /day
ug Pb /day
ug Pb /day
ug Pb /day
ug Pb /day
Alternate Diet Sources (by food class)
Concentration
Home-grown fruits
Home-grown vegetables
Fish from fishing
Game animals from hunting
Percent of food class
Home-grown fruits
Home-grown vegetables
Fish from fishing
Game animals from hunting
0
0
0
0
0
0
0
0
ugPb/g
ugPb/g
ugPb/g
ugPb/g
%
%
%
%
Drinking Water
Lead concentration in drinking water
Ingestion rate
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
4
0.20
0.50
0.52
0.53
0.55
0.58
0.59
ug/L
L/day
L/day
L/day
L/day
L/day
L/day
L/day
Alternate Drinking Water Sources
Concentration
First-draw water
Flushed water
Fountain water
Percentage of total intake
First-draw water
Flushed water
Fountain water
4
1
10
50
35
15
ug/L
ug/L
ug/L
%
%
%
April 19, 2013
B-2
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix B: IEUBK Model
Exhibit B-1. Default Values for IEUBK Model Parameters
Parameter
Default TT .
Units
Value
Soil/Dust (constant over time)
Concentration
Soil
Dust
Soil/dust ingestion weighting factor (percent soil)
200 ug/g
200 ug/g
45 %
Soil/Dust Ingestion (by year)
Soil/Dust ingestion
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0.085 g/day
0.135 g/day
0.135 g/day
0.135 g/day
0.100 g/day
0.090 g/day
0.085 g/day
Soil/Dust Multiple Source Analysis (constant)
Fraction of indoor dust lead attributable to soil (MSD)
Ratio of dust lead concentration to outdoor air lead concentration
Soil/Dust Multiple Source Analysis with Alternative Household
Concentration
Household dust
Secondary occupational dust
School dust
Daycare center dust
Second home
Interior lead-based paint
Percentage
Household dust
Secondary occupational dust
School dust
Daycare center dust
Second home
Interior lead-based paint
0.70 unitless
100 ug Pb/g dust per ug
Pb/ms air
Dust Lead Sources (constant)
150 ug/L
1,200 ug/L
200 ug/L
200 ug/L
200 ug/L
1,200 ug/L
100 %
0 %
0 %
0 %
0 %
0 %
Bioavailability for All Gut Absorption Pathways
Total lead absorption (at low intake)
Diet
Drinking water
Soil
Dust
Alternate source
Fraction of total net absorption at low intake rate that is attributable to non
saturable (passive) processes
50 %
50 %
30 %
30 %
0 %
0.2 unitless
Alternate Sources (by year)
Total lead intake
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
0 ug/day
0 ug/day
0 ug/day
0 ug/day
0 ug/day
April 19, 2013
B-3
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix B: IEUBK Model
Exhibit B-1. Default Values for IEUBK Model Parameters
Parameter
Default
Value
Units
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
ug/day
ug/day
Maternal-to-Newborn Lead Exposure
Mothers blood lead concentration at childbirth
1.0
ug/L
Plotting and Risk Estimation
Geometric standard deviation (GSD) for blood lead
1.6
Unitless
Blood lead level of concern, or cutoff
10
ug/L
Computational Options
Iteration time step for numerical integration
Source: U.S. EPA (2007).
April 19, 2013
B-4
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix C: Sensitivity Analysis - Human Health Benefits
Appendix C: Human Health Benefits Including Downstream Reaches
As discussed in Chapter 3, EPA analyzed the human health benefits of reducing steam electric pollutant
discharges using only data for the reaches that receive steam electric discharges directly. However, the
proposed ELGs are expected to provide additional human health benefits by also reducing pollutant
concentrations in reaches downstream from steam electric plant discharges. To evaluate the potential
significance of these benefits, EPA conducted a sensitivity analysis using fish tissue concentrations modeled
for downstream reaches, based on ambient metal concentrations estimated using a simple dilution model.
To expand the analysis to include downstream reaches, EPA first estimated baseline and post-compliance
metals concentrations using the water quality model component of EPA's Risk-Screening Environmental
Indicators (RSEI) model, which incorporates 2010 Toxic Release Inventory (TRI) data on annual average
discharges.80 EPA input the loadings from steam electric plant and TRI discharges in the RSEI model to
estimate the long-term average metals concentrations in receiving and downstream reaches. The RSEI model
uses a simple dilution and first-order decay equation wherein metals are modeled as conservative substances.
For each reach, EPA then used the modeled ambient water concentrations to derive fish tissue concentrations
and average daily intake rates for each of the 14 age cohorts.
Due to data and resources limitations, EPA restricted the downstream analysis to reaches that have baseline
concentrations above 0.001 ug/L. EPA thus identified a set of 9,626 receiving and downstream reaches
potentially affected by steam electric plant discharges, as compared with 296 reaches for the analysis
discussed in Chapter 3.
A series of tables provide the results of the sensitivity analysis for the five regulatory options EPA analyzed
(Options 1, 2, 3, 4, and 5): Exhibit C-l shows the benefits from reduced cancer cases, Exhibit C-2 shows the
benefits from avoided IQ losses from lead exposure, Exhibit C-3 shows the benefits from a reduced need for
compensatory education from lead exposure, and Exhibit C-4 shows the benefits from avoided IQ losses from
in-utero mercury exposure.
As shown by the results below, monetized human health benefits for the sensitivity analysis are significantly
greater than those estimated for directly receiving reaches only, suggesting that changes in water quality
downstream from steam electric plant discharges may provide non-trivial benefits to other affected
populations. For example, total annualized human health benefits for Option 3 across the four benefit
categories as presented in Exhibits C-l through C-4 are $139 million (3 percent discount rate), as compared to
$7.7 million when considering only directly receiving reaches in Chapter 3. EPA will continue to seek ways
to refine its estimates of downstream effects to characterize associated human health benefits.
Exhibit C-1. Benefits from Reduced Cancer Cases, Including Downstream Reaches
Scenario
Baseline
Option 1
Option 2
Affected
Population
1,738,165
1,738,165
1,738,165
Total Cancer
Cases, 2017 to
2040
24.6
24.5
24.5
Reduced Cancer
Cases, 2017 to
2040
NA
0.2
0.1
Annualized Benefits
(Millions; 2010$)
3% Discount
Rate
NA
$0.04
$0.03
7% Discount
Rate
NA
$0.02
$0.02
80 EPA removed from the 2010 TRI data the loadings associated with steam electric plants that report to TRI. For steam
electric plants, EPA used the loadings estimated in the EA under the baseline and the regulatory options. Loadings for
other TRI dischargers (non-steam electric plant and other industrial facilities) remained constant throughout all
scenarios.
April 19, 2013
C-1
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix C: Sensitivity Analysis - Human Health Benefits
Option 3
Option 4
Option 5
1,738,165
1,738,165
1,738,165
10.6
2.2
2.1
14.1
22.4
22.5
$3.22
$5.12
$5.15
$1.83
$2.92
$2.93
Exhibit C-2. Benefits from Avoided IQ Losses for Children Ages 0 to 7, Including Downstream
Reaches
Scenario
Baseline
Option 1
Option 2
Option 3
Option 4
Option 5
Number of
Affected
Children 0 to 7
158,173
158,173
158,173
158,173
158,173
158,173
Total Avoided IQ
Losses, 2017 to
2040
0
1,014
1,014
24,805
85,462
85,455
Annualized Value of Avoided IQ Point Losses a
(Millions; 2010$)
3% Discount Rate
Low Bound
High Bound
NA
$0.35
$0.35
$8.65
$29.79
$29.79
$0.51
$0.51
$12.37
$42.61
$42.61
7% Discount Rate
Low Bound
High Bound
NA
$0.03
$0.03
$0.62
$2.12
$2.12
$0.05
$0.05
$1.23
$4.24
$4.23
a. Low bound assumes that the loss of one IQ point results in the loss of 1.76 percent of lifetime earnings (following Schwartz,
1994); high bound assumes that the loss of one IQ point results in the loss of 2.38 percent of lifetime earnings (following Salkever,
1995).
Source: U.S. EPA Analysis, 2013
libit C-3. Avoided Cost of Compensatory Education for Children with Blood Lead Concentrations
above 20 |ag/dL and IQ Less than 70, Including Downstream Reaches
Scenario
Baseline
Option 1
Option 2
Option 3
Option 4
Option 5
Number of
Affected Children
Oto7
158,173
158,173
158,173
158,173
158,173
158,173
Number of Cases
ofPbB>20
ug/dL and IQ <
70, 2017 to 2040
20.0
20.0
20.0
13.6
3.7
3.7
Decrease in
Number of Cases
of
IQ < 70, 2017 to
2040
NA
0.0
0.0
6.4
16.3
16.3
Avoided Annual Cost
(Millions; 2010$)
3% Discount
Rate
NA
$0.00
$0.00
$0.04
$0.09
$0.09
7% Discount
Rate
NA
$0.00
$0.00
$0.02
$0.04
$0.04
PbB = blood lead concentration.
Source: U.S. EPA Analysis, 2013
Exhibit C-4. Benefits from Avoided IQ Losses Due to Reduced In-utero Mercury Exposure, Including
Downstream Reaches
Scenario
Baseline
Option 1
Option 2
Option 3
Option 4
Option 5
Number of Births
in Affected
Population (per
year)
25,087
25,087
25,087
25,087
25,087
25,087
Total Avoided IQ
Losses, 2017 to
2040
NA
151,415
154,843
302,903
565,449
566,493
Annualized Value of Avoided IQ Losses"
(Millions; 2010$)
3% Discount Rate
Low Bound
High Bound
NA
$52.79
$53.98
$105.60
$197.13
$197.49
$75.49
$77.20
$151.02
$281.92
$282.44
7% Discount Rate
Low Bound
High Bound
NA
$3.76
$3.85
$7.52
$14.04
$14.07
$7.50
$7.67
$15.01
$28.02
$28.07
a. Low bound estimate assumes that the loss of one IQ point results in the loss of 1.76% of lifetime earnings (following Schwartz,
1994); high bound estimate assumes that the loss of one IQ point results in the loss of 2.38% of lifetime earnings (following
April 19, 2013
C-2
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix C: Sensitivity Analysis - Human Health Benefits
Exhibit C-4. Benefits from Avoided IQ Losses Due to Reduced In-utero Mercury Exposure, Including
Downstream Reaches
Scenario
Number of Births
in Affected
Population (per
year)
Total Avoided IQ
Losses, 2017 to
2040
Annualized Value of Avoided IQ Losses"
(Millions; 2010$)
3% Discount Rate
Low Bound
High Bound
7% Discount Rate
Low Bound
High Bound
Salkever, 1995).
Source: U.S. EPA Analysis, 2013
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
C-3
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix D: Subindex Curves
Appendix D: TSS, TN, and TP Ecoregion-Specific Subindex Curves
This appendix provides the ecoregion-specific parameters used in estimating the TSS, TN, or TP water
quality subindex, as follows:
If [WQ Parameter] < WQ Parameter 100
If WQ Parameter 100 < [WQ Parameter] < WQ Parameter 10
If [WQ Parameter] > WQ Parameter 10
Subindex =100
Subindex = a exp(b [WQ Parameter])
Subindex =10
Where [WQ Parameter] is the measured concentration of either TSS, TN, or TP and WQ Parameter 10,
WQ Parameter 100, a, and b are specified in Exhibit D-l for TSS, Exhibit D-2 for TN, and Exhibit D-3 for
TP.
Exhibit D-1: TSS Subindex Curve Parameters, by Ecoregion
ID
10.1.2
10.1.3
10.1.4
10.1.5
10.1.6
10.1.7
10.1.8
10.2.1
10.2.2
10.2.4
11.1.1
11.1.2
11.1.3
12.1.1
13.1.1
15.4.1
5.2.1
5.2.2
5.3.1
5.3.3
6.2.10
6.2.11
6".2""l2
6".2""l3
6".2""l4
6.2.15
6.2.3
6.2.4
6.2.5
6.2.7
6.2.8
6.2.9
7.1.7
7.1.8
Ecoregion Name
Columbia Plateau
Northern Basin and Range
Wyoming Basin
Central Basin and Range
Colorado Plateaus
Arizona/New Mexico Plateau
Snake River Plain
Mojave Basin and Range
Sonoran Desert
Chihuahuan Desert
California Coastal Sage, Chaparral, and Oak Woodlands
Central California Valley
Southern and Baja California Pine-Oak Mountains
Madrean Archipelago
Arizona/New Mexico Mountains
Southern Florida Coastal Plain
Northern Lakes and Forests
Northern Minnesota Wetlands
Northern Appalachian and Atlantic Maritime Highlands
North Central Appalachians
Middle Rockies
Klamath Mountains
Sierra Nevada
Wasatch and Uinta Mountains
Southern Rockies
Idaho Batholith
Columbia Mountains/Northern Rockies
Canadian Rockies
North Cascades
Cascades
Eastern Cascades Slopes and Foothills
Blue Mountains
Strait of Georgia/Puget Lowland
Coast Range
a
126.56
112.42
123.36
121.22
144.44
126.76
146.39
119.34
112.39
214.39
127.97
171.86
115.12
261.35
120.98
116.95
157.76
154.99
174.99
245.15
144.64
238.9
18536
124.28
15142
184.23
180.7
396.62
240.95
192.94
178.82
148.35
181.06
174.78
b
-0.0038
-0.0007
-0.001
-0.0018
-0.001
-0.0004
-0.0027
-0.0015
-0.0002
-0.0005
-0.0012
-0.0044
-0.0007
-0.0005
-0.0004
-0.0405
-0.0233
-0.0186
-0.0261
-0.0176
-0.0038
-0.0068
:5"oYi6
:o;"ooi4
logos'!
-0.0142
-0.0168
-0.0308
-0.0193
-0.0181
-0.0145
-0.0037
-0.0224
-0.0114
TSSioo
63
160
220
109
363
668
142
121
567
1,419
205
122
197
2,053
477
4
20
24
21
51
98
129
53
160~
140~
43
35
45
46
36
40
107
27
49
TSS10
668
3,457
2,513
1,386
2,670
6,349
994
1,653
12,097
6,130
2,124
646
3,491
6,527
6,233
61
118
147
110
182
703
467
252
l$OQ
881
205
172
119
165
164
199
729
129
251
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
D-1
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix D: Subindex Curves
Exhibit D-1: TSS Subindex Curve Parameters, by Ecoregion
ID
7.1.9
8.1.1
8.1.2
8.1.3
8"'i:::4
s'Ts
8.1.6
8.1.7
8.1.8
8.1.10
8.2.1
8.2.2
8.2.3
8.2.4
8.3.1
8.3.2
8.3.3
8.3.4
8.3.5
8.3.6
8.3.7
8.3.8
8.4.1
8.4.2
8.4.3
8.4.4
8.4.5
8.4.6
8.4.7
8.4.8
8.4.9
8.5.1
8.5.2
8.5.3
8.5.4
9.2.1
9.2.2
9.2.3
9.2.4
9.3.1
9.3.3
9.3.4
9.4.1
9.4.2
9.4.3
9.4.4
9.4.5
9.4.6
9.4.7
Ecoregion Name
Willamette Valley
Eastern Great Lakes and Hudson Lowlands
Lake Erie Lowland
Northern Appalachian Plateau and Uplands
North Central Hardwood Forests
Driftless Area
S. Michigan/N. Indiana Drift Plains
Northeastern Coastal Zone
Maine/New Brunswick Plains and Hills
Erie Drift Plain
Southeastern Wisconsin Till Plains
Huron/Erie Lake Plains
Central Corn Belt Plains
Eastern Corn Belt Plains
Northern Piedmont
Interior River Valleys and Hills
Interior Plateau
Piedmont
Southeastern Plains
Mississippi Valley Loess Plains
South Central Plains
East Central Texas Plains
Ridge and Valley
Central Appalachians
Western Allegheny Plateau
Blue Ridge
Ozark Highlands
Boston Mountains
Arkansas Valley
Ouachita Mountains
Southwestern Appalachians
Middle Atlantic Coastal Plain
Mississippi Alluvial Plain
Southern Coastal Plain
Atlantic Coastal Pine Barrens
Aspen Parkland/Northern Glaciated Plains
Lake Manitoba and Lake Agassiz Plain
Western Corn Belt Plains
Central Irregular Plains
Northwestern Glaciated Plains
Northwestern Great Plains
Nebraska Sand Hills
High Plains
Central Great Plains
Southwestern Tablelands
Flint Hills
Cross Timbers
Edwards Plateau
Texas Blackland Prairies
a
210.3
144.62
112.79
322.68
148.68
117.97
191.44
158.48
156.02
133.08
121.34
145.17
187.95
235.18
175.82
149.68
220.47
224.11
205.3
492.49
184.36
162.32
186.83
166.76
183.67
216.16
175.16
329.77
283.25
212.77
207.09
182.17
131.35
138.62
283.76
136.43
174.13
135.01
201.19
133.98
130.6
289.85
125.61
156.84
137.77
270.93
134.97
173.77
134.23
b
-0.0114
-0.0104
-0.0049
-0.0113
-0.0108
-0.0012
-0.0143
-0.0164
-0.025
-0.0037
-0.0042
-0.0058
-0.0033
-0.003
-0.0042
-0.0013
-0.0037
-0.0048
-0.0085
-0.0048
-0.0045
-0.0013
-0.0063
-0.0062
-0.0032
-0.0087
-0.0018
-0.0062
-0.004
-0.0048
-0.0071
-0.0178
-0.0029
-0.0144
-0.0463
-0.0005
-0.0042
-0.0009
-0.001
-0.0006
-0.0004
-0.0066
-0.0005
-0.0005
-0.0003
-0.0009
-0.0006
-0.001
-0.0005
TSSjoo
65
36
25
103
37
141
46
28
18
78
46
65
191
282
135
303
217
169
85
333
136
362
99
82
190
89
317
193
261
157
103
34
93
23
23
640
131
347
673
483
636
162
507
925
1,280
1,084
523
544
624
TSS10
267
257
494
307
250
2,057
206
168
110
700
594
461
889
1,053
683
2,081
836
648
356
812
648
2,144
465
454
910
353
1,591
564
836
637
427
163
888
183
72
5,226
680
2,892
3,002
4,325
6,424
510
5,061
5,505
8,743
3,666
4,337
2,855
5,194
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
D-2
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix D: Subindex Curves
Exhibit D-1: TSS Subindex Curve Parameters, by Ecoregion
ID
9.5.1
9.6.1
Ecoregion Name
Western Gulf Coastal Plain
Southern Texas Plains/Interior Plains and Hills with
Xerophytic Shrub and Oak Forest
a
124.47
166.67
b
-0.0025
-0.0003
TSSjoo
88
1,602
TSS10
1,009
9,378
Exhibit D-2: TN Subindex Curve Parameters, by Ecoregion
ID
10.1.2
10.1.3
10.1.4
10.1.5
10.1.6
10.1.7
10.1.8
10.2.1
10.2.2
10.2.4
11.1.1
11.1.2
11.1.3
12.1.1
13.1.1
15.4.1
5.2.1
5.2.2
5.3.1
5.3.3
6".2"T6
6".2'Tl
6".2""l2
6.2.13
6.2.14
6.2.15
6.2.3
6.2.4
6.2.5
6.2.7
6.2.8
6.2.9
7. .7
7. .8
7. .9
8. .1
8. .2
8. .3
8. .4
8. .5
8. .6
8. .7
Ecoregion Name
Columbia Plateau
Northern Basin and Range
Wyoming Basin
Central Basin and Range
Colorado Plateaus
Arizona/New Mexico Plateau
Snake River Plain
Mojave Basin and Range
Sonoran Desert
Chihuahuan Desert
California Coastal Sage, Chaparral, and Oak Woodlands
Central California Valley
Southern and Baja California Pine-Oak Mountains
Madrean Archipelago
Arizona/New Mexico Mountains
Southern Florida Coastal Plain
Northern Lakes and Forests
Northern Minnesota Wetlands
Northern Appalachian and Atlantic Maritime Highlands
North Central Appalachians
Middle Rockies
Klamath Mountains
Sierra Nevada
Wasatch and Uinta Mountains
Southern Rockies
Idaho Batholith
Columbia Mountains/Northern Rockies
Canadian Rockies
North Cascades
Cascades
Eastern Cascades Slopes and Foothills
Blue Mountains
Strait of Georgia/Puget Lowland
Coast Range
Willamette Valley
Eastern Great Lakes and Hudson Lowlands
Lake Erie Lowland
Northern Appalachian Plateau and Uplands
North Central Hardwood Forests
Driftless Area
S. Michigan/N. Indiana Drift Plains
Northeastern Coastal Zone
a
116.58
126.97
124.89
116.66
146.41
116.33
129.93
136.69
117.99
104.2
123.22
126.07
122.76
130.61
141.64
1000000
141.98
142.55
142.6
180.92
136.51
140.34
14102
129.75
131.07
149.42
136.14
151.95
155.86
143.07
123.99
125.19
121.09
136.15
135.01
158.18
156.27
431.78
163.4
126.18
130.25
125.75
b
-0.663
-0.626
-0.445
-0.335
-0.588
-0.286
-0.594
-0.593
-0.495
-0.45
-0.889
-0.548
-0.564
-0.325
-0.541
-29.36
-0.985
-0.781
-0.854
-0.897
-0.991
-1.805
-1.424
-0.452
-0.66
-1.775
-1.599
-2.098
-1.231
-1.473
-1.07
-0.786
-0.723
-1.021
-0.809
-0.563
-0.38
-0.435
-0.599
-0.272
-0.149
-0.159
TN100
0.23
0.38
0.50
0.46
0.65
0.53
0.44
0.53
0.33
0.09
0.23
0.42
0.36
0.82
0.64
0.33
0.36
0.45
0.42
0.66
0.31
0.19
6.25
0.58
0.41
0.23
0.19
0.20
0.36
0.24
0.20
0.29
0.26
0.30
0.37
0.81
1.18
3.36
0.82
0.85
1.78
1.44
TN10
3.70
4.06
5.67
7.33
4.56
8.58
4.32
4.41
4.99
5.21
2.82
4.62
4.45
7.91
4.90
0.39
2.69
3.40
3.11
3.23
2.64
1.46
L87
5.67
3.90
1.52
1.63
1.30
2.23
1.81
2.35
3.22
3.45
2.56
3.22
4.90
7.23
8.66
4.66
9.32
17.23
15.92
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
D-3
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix D: Subindex Curves
Exhibit D-2: TN Subindex Curve Parameters, by Ecoregion
ID
8.1.8
8.1.10
8.2.1
8.2.2
8.273
8.2.4
8.3.1
8.3.2
sTis
JTi4
sTis
8.3.6
8.3.7
8.3.8
8.4.1
8.4.2
8.4.3
8.4.4
8.4.5
8.4.6
8.4.7
8.4.8
8.4.9
8.5.1
8.5.2
8.5.3
8.5.4
9.2.1
9.2.2
9.2.3
9.2.4
9.3.1
9.3.3
9.3.4
9.4.1
9.4.2
9.4.3
9.4.4
9.4.5
9.4.6
9.4.7
9.5.1
9.6.1
Ecoregion Name
Maine/New Brunswick Plains and Hills
Erie Drift Plain
Southeastern Wisconsin Till Plains
Huron/Erie Lake Plains
Central Corn Belt Plains
Eastern Corn Belt Plains
Northern Piedmont
Interior River Valleys and Hills
Interior Plateau
Piedmont
Southeastern Plains
Mississippi Valley Loess Plains
South Central Plains
East Central Texas Plains
Ridge and Valley
Central Appalachians
Western Allegheny Plateau
Blue Ridge
Ozark Highlands
Boston Mountains
Arkansas Valley
Ouachita Mountains
Southwestern Appalachians
Middle Atlantic Coastal Plain
Mississippi Alluvial Plain
Southern Coastal Plain
Atlantic Coastal Pine Barrens
Aspen Parkland/Northern Glaciated Plains
Lake Manitoba and Lake Agassiz Plain
Western Corn Belt Plains
Central Irregular Plains
Northwestern Glaciated Plains
Northwestern Great Plains
Nebraska Sand Hills
High Plains
Central Great Plains
Southwestern Tablelands
Flint Hills
Cross Timbers
Edwards Plateau
Texas Blackland Prairies
Western Gulf Coastal Plain
Southern Texas Plains/Interior Plains and Hills with
Xerophytic Shrub and Oak Forest
a
139.55
148.99
134.85
119.06
135.57
149.12
146.34
120.48
IIjl^SjT
IIllML
138.73
123.15
149.84
136
158.11
161.22
125.25
158.16
145.69
168.59
135.4
162.34
143.42
123.43
119.57
118.73
110.04
141.62
119.49
129.28
142.81
120.91
125.65
113.81
121.41
129.36
136.03
142.74
130.87
141.98
133.84
106.22
102.35
b
-0.553
-1.256
-0.16
-0.091
-0.087
-0.122
-0.314
-0.131
II:0-446]
II:<163T
-0.727
-0.379
-0.706
-0.344
-0.659
-0.907
-0.44
-0.777
-0.513
-1.108
-0.47
-0.942
-0.645
-0.444
-0.31
-0.701
-0.482
-0.086
-0.082
-0.074
-0.184
-0.386
-0.404
-0.324
-0.161
-0.178
-0.413
-0.343
-0.278
-0.588
-0.243
-0.301
-0.374
TN100
0.60
0.32
1.87
1.91
3.50
3.28
1.21
1.43
I!ZM£
IIIj^E
0.45
0.55
0.57
0.89
0.70
0.53
0.51
0.59
0.73
0.47
0.64
0.51
0.56
0.47
0.58
0.24
0.20
4.06
2.18
3.48
1.93
0.49
0.56
0.40
1.21
1.44
0.74
1.04
0.97
0.60
1.20
0.20
0.06
TN10
4.77
2.15
16.26
27.22
29.96
22.15
8.55
19.00
IIIji°2~
Illili]
3.62
6.62
3.83
7.59
4.19
3.07
5.74
3.55
5.22
2.55
5.54
2.96
4.13
5.66
8.00
3.53
4.98
30.82
30.25
34.59
14.45
6.46
6.26
7.51
15.51
14.38
6.32
7.75
9.25
4.51
10.68
7.85
6.22
Exhibit D-3: TP Subindex
ID
10.1.2
Curve Parameters,
by
Ecoregion
Ecoregion Name
Columbia Plateau
a
147.39
b
-2.211
TPioo
0.18
TP10
1.22
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
D-4
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix D: Subindex Curves
Exhibit D-3: TP Subindex Curve Parameters, by Ecoregion
ID
10.1.3
10.1.4
10.1.5
10.1.6
10:t:7
10.1.8
10.2.1
10.2.2
T6""2."4
iiTI
TTT.2'
11.1.3
12.1.1
13.1.1
15.4.1
5.2.1
5.2.2
5.3.1
5.3.3
6.2.10
6.2.11
6.2.12
6.2.13
6.2.14
6.2.15
6.2.3
6.2.4
6.2.5
6.2.7
6.2.8
6.2.9
7. .7
7. .8
7. .9
8. .1
8. .2
8. .3
8. .4
8. .5
8. .6
8. .7
8. .8
8.1.10
8.2.1
8.2.2
8.2.3
8.2.4
8.3.1
8.3.2
Ecoregion Name
Northern Basin and Range
Wyoming Basin
Central Basin and Range
Colorado Plateaus
Arizona/New Mexico Plateau
Snake River Plain
Mojave Basin and Range
Sonoran Desert
Cninuahuan Desert
California Coastal Sage, Chaparral, and Oak Woodlands
Central California Vailey
Southern and Baja California Pine-Oak Mountains
Madrean Archipelago
Arizona/New Mexico Mountains
Southern Florida Coastal Plain
Northern Lakes and Forests
Northern Minnesota Wetlands
Northern Appalachian and Atlantic Maritime Highlands
North Central Appalachians
Middle Rockies
Klamath Mountains
Sierra Nevada
Wasatch and Uinta Mountains
Southern Rockies
Idaho Batholith
Columbia Mountains/Northern Rockies
Canadian Rockies
North Cascades
Cascades
Eastern Cascades Slopes and Foothills
Blue Mountains
Strait of Georgia/Puget Lowland
Coast Range
Willamette Valley
Eastern Great Lakes and Hudson Lowlands
Lake Erie Lowland
Northern Appalachian Plateau and Uplands
North Central Hardwood Forests
Driftless Area
S. Michigan/N. Indiana Drift Plains
Northeastern Coastal Zone
Maine/New Brunswick Plains and Hills
Erie Drift Plain
Southeastern Wisconsin Till Plains
Huron/Erie Lake Plains
Central Corn Belt Plains
Eastern Corn Belt Plains
Northern Piedmont
Interior River Valleys and Hills
a
165.9
143.83
167.24
123.74
168.68
140.75
139.89
122.92
132^89
125"05
12632
212.01
140.62
555.88
157.9
152.78
171.4
260.92
157.84
188.95
205.2
142.56
141.72
185.94
168.85
197.1
289.57
227.85
154.67
141.59
165.33
185.34
159.54
148.02
230.09
3440.2
317.21
132.65
141.49
184.34
174
174.73
151.79
141.21
247.17
223.41
196
160.97
156.71
b
-2.78
-1.57
-2.541
-0.784
-3.39
-1.106
-0.978
-1.578
3/737
-L918
-2.T38
-0.941
-1.331
-306
-26.64
-16.37
-21.87
-21.53
-6.439
-15.04
-19.13
-2.752
-5.463
-21.89
-17.88
-27.87
-47.06
-26.77
-10.55
-3.31
-13.83
-14.77
-9.053
-7.95
-9.614
-8.887
-13.87
-4.905
-2.261
-5.59
-9.944
-28.94
-3.59
-1.577
-2.666
-3.555
-3.734
-2.567
-3.616
TPioo
0.18
0.23
0.20
0.27
0.15
0.31
0.34
0.13
0'.08
0/L2~
o'jr
0.80
0.26
0.01
0.02
0.03
0.02
0.04
0.07
0.04
0.04
0.13
0.06
0.03
0.03
0.02
0.02
0.03
0.04
0.11
0.04
0.04
0.05
0.05
0.09
0.40
0.08
0.06
0.15
0.11
0.06
0.02
0.12
0.22
0.34
0.23
0.18
0.19
0.12
TP10
1.01
1.70
1.11
3.21
0.83
2.39
2.70
1.59
6769
132
L19
3.25
1.99
0.01
0.10
0.17
0.13
0.15
0.43
0.20
0.16
0.97
0.49
0.13
0.16
0.11
0.07
0.12
0.26
0.80
0.20
0.20
0.31
0.34
0.33
0.66
0.25
0.53
1.17
0.52
0.29
0.10
0.76
1.68
1.20
0.87
0.80
1.08
0.76
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
D-5
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix D: Subindex Curves
Exhibit D-3: TP Subindex Curve Parameters, by Ecoregion
ID
8.3.3
8.3.4
8.3.5
8.3.6
8.3.7
8.3.8
8.4.1
8.4.2
8A3
8A4
8A5
8.4.6
8.4.7
8.4.8
8.4.9
8.5.1
8.5.2
8.5.3
8.5.4
9.2.1
9.2.2
9.2.3
9.2.4
9.3.1
9.3.3
9.3.4
9.4.1
9.4.2
9.4.3
9.4.4
9.4.5
9.4.6
9.4.7
9.5.1
9.6.1
Ecoregion Name
Interior Plateau
Piedmont
Southeastern Plains
Mississippi Valley Loess Plains
South Central Plains
East Central Texas Plains
Ridge and Valley
Central Appalachians
Western Allegheny Plateau
Blue Ridge
Ozark Highlands
Boston Mountains
Arkansas Valley
Ouachita Mountains
Southwestern Appalachians
Middle Atlantic Coastal Plain
Mississippi Alluvial Plain
Southern Coastal Plain
Atlantic Coastal Pine Barrens
Aspen Parkland/Northern Glaciated Plains
Lake Manitoba and Lake Agassiz Plain
Western Corn Belt Plains
Central Irregular Plains
Northwestern Glaciated Plains
Northwestern Great Plains
Nebraska Sand Hills
High Plains
Central Great Plains
Southwestern Tablelands
Flint Hills
Cross Timbers
Edwards Plateau
Texas Blackland Prairies
Western Gulf Coastal Plain
Southern Texas Plains/Interior Plains and Hills with
Xerophytic Shrub and Oak Forest
a
197.72
223.39
177.2
168
166.39
178.13
225.67
187.73
I74~12
152/7
204788
287.21
158.46
169.72
153.95
141.25
144.74
126.84
156.13
132.19
197.19
200.96
134.12
143.32
185
153.07
188.57
139.57
218.92
131.7
159.98
149.63
127.17
104.21
147.39
b
-5.623
-9.266
-5.69
-4.659
-1.677
-6.407
-16.59
-8.367
-ioTs
-2.889
-7.364
-5.786
-6.821
-7.296
-6.816
-3.807
-7.676
-8.388
-0.69
- .087
- .683
- .994
- .646
- .267
-3.788
-0.946
-1.178
-0.972
-2.351
-0.78
-1.384
-1.064
-1.863
-0.513
-2.211
TPioo
0.12
0.09
0.10
0.11
0.30
0.09
0.05
0.08
O05~
o'TJ'"
o'lo""
0.18
0.07
0.07
0.06
0.09
0.05
0.03
0.65
0.26
0.40
0.35
0.18
0.28
0.16
0.45
0.54
0.34
0.33
0.35
0.34
0.38
0.13
0.08
0.18
TP10
0.53
0.34
0.51
0.61
1.68
0.45
0.19
0.35
6727
6794
O41
0.58
0.41
0.39
0.40
0.70
0.35
0.30
3.98
2.38
1.77
1.50
1.58
2.10
0.77
2.88
2.49
2.71
1.31
3.31
2.00
2.54
1.36
4.57
1.22
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
D-6
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix E: Meta-Analysis
Appendix E: Meta-Analysis Results
EPA used function-based benefit transfer to estimate benefits of surface water quality improvements due to
reductions in in steam electric pollutant discharges due to the proposed ELGs. The benefit function was
derived using meta-analysis, following the general approach of Johnston et al. (2005), Shrestha et al. (2007),
and others, following conceptual methods outlined by Bergstrom and Taylor (2006). The recent literature has
given increasing emphasis to the potential use of meta-analysis to conduct and inform function-based benefit
transfer (Johnston et al. 2005; Bergstrom and Taylor 2006; Rosenberger and Stanley 2006; Shrestha et al.
2007). For the present analysis, the meta-regression model was based on a model specification and data
developed originally for the 316(b) Phase II Cooling Water Intake rule, and revised based on more recent
studies to support benefits assessment of the Construction and Development ELGs (U.S. EPA, 2009b).
Chapter A12, "Methods for Estimating Non-use Benefits," in the Regional Analysis document for the final
Phase II rule provides details on the original meta-analysis (U.S. EPA, 2004c); revisions are detailed below.
These revisions included adding new studies to the metadata, re-estimating the willingness-to-pay (WTP)
function to better account for ecological services potentially affected by heavy metal contamination and
nutrients, and testing additional functional forms and statistical approaches.
As stated by Rosenberger and Johnston (2007, p. 1-2):
One of the primary advantages of meta-analysis as a benefit transfer tool relates to its capacity to allow more appropriate
adjustments of welfare measures based on patterns observed in the literature. Within a benefit transfer context, transfer error
is often inversely related to the correspondence between a study site and a policy site among various dimensions
(Rosenberger and Phipps 2007). The probability of finding a good fit between a single (or multiple) study site and a policy
site, however, is usually low (Boyle and Bergstrom 1992; Spash and Vatn 2006). If, on the other hand, empirical studies
contribute to a body of WTP estimates (i.e., metadata), and if empirical value estimates are systematically related to
variations in resource, study, and site characteristics, then meta-regression analysis may provide a viable tool for estimating
a more universal transfer function with distinct advantages over unit value or other function-based transfer methods
(Johnston et al. 2003; Rosenberger and Loomis 2000a; Rosenberger and Stanley 2006). More specifically, Rosenberger and
Phipps (2007) posit a meta-valuation function as the envelope of a set of empirically-defined valuation functions reported in
the literature.
In the present case, EPA identified 45 valuation studies that use stated preference techniques to elicit benefit
values for water quality improvements. To examine the relative influence of study, economic, and resource
characteristics on WTP for improving surface water quality, the Agency conducted a regression-based meta-
analysis of 115 estimates of WTP for water quality improvements, provided by the 45 original studies.
Analytic methods and model specifications follow established methods in the published literature (e.g.,
Johnston et al. 2005; Bergstrom and Taylor 2006). The estimated econometric model is used as the basis of a
function-based benefit transfer, to calculate WTP for improving water quality in waterbodies affected by
discharge from steam electric plants.
The following discussion summarizes the results of EPA's meta-analysis of surface water valuation studies
and the use of the resulting benefit function for transfer.
E.1 Literature Review of Water Resource Valuation Studies
As outlined in the introduction, EPA conducted a meta-analysis of water resource valuation studies to
examine the relative influence of study, economic, and resource characteristics on total WTP for water quality
improvements. The Agency analyzed 45 studies, published between 1981 and 2008, that applied generally
accepted, standard valuation methods to determine total (including use and nonuse) values associated with
aquatic habitat improvements. These 45 studies all used stated preference techniques to assess WTP, but
varied in other respects, including the survey administration methods used, the specific environmental change
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
E-1
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
valued, and the geographic region affected by the environmental changes. Studies using stated preference
approaches are preferred to studies using revealed preference approaches, because they elicit total household
WTP (including use and nonuse values). Revealed preference studies allow to estimate use values only. Data
from the 45 studies result in a total of 115 observations for the meta-analysis because 30 studies provide more
than one usable estimate of total WTP for aquatic habitat improvements.
When constructing metadata for subsequent meta-analysis, analysts must determine the optimal scope of the
metadata (Rosenberger and Johnston 2007), interpreted as the exact definition of the dependent variable in the
meta-regression model, which, in turn, defines the set of source studies that can be considered for inclusion in
the metadata. The primary tradeoff is often between maintaining close similarity among dependent variables
versus including additional information (i.e., observations) in the metadata. Similarity in dependent variable
definition and study attributes within the metadata can be important for two reasons. First, theory may dictate
that certain types of estimated values are not strictly comparable (e.g., Hicksian compensating variation from
a stated preference model versus Marshallian consumer surplus from a travel cost model). Second, model fit
may be improved by narrowing the metadata, for example to include only valuation studies that use a
particular valuation approach (e.g., stated preference methods, as in Johnston et al. (2005); or travel cost
model estimates, as in Smith and Kaoru (1990)). Such study selection issues may be framed in terms of a
requirement that at a minimum, studies included in metadata satisfy both commodity consistency and welfare
consistency (Bergstrom and Taylor 2006). The former implies that "the commodity (Q) being valued should
be approximately the same within and across studies" (Bergstrom and Taylor 2006, p. 353). The latter implies
that "measures of WTP within and across studies ... should represent the same ... welfare change measure, or
ex-post calibrations [are] made to account for theoretical differences between welfare change measures"
(Bergstrom and Taylor 2006, p. 355).
The requirement of welfare consistency implies that—outside of preference calibration approaches that
explicitly account for theoretical differences between welfare constructs (e.g., Smith et al. 2002)—meta-
analyses should not combine data representing theoretically distinct welfare measures. For example, the ad
hoc combination of stated and revealed data for meta-analysis would generally violate the condition of
welfare consistency. Although one may introduce independent variables in regression models (typically
dummy variable intercept shifters) to account for such theoretical differences, Smith and Pattanayak (2002)
argue that such methods are unlikely to represent appropriate adjustments for fundamental differences in
theoretical welfare measures.
E.1.1 Identifying Water Resource Valuation Studies
EPA identified surface water valuation studies used in the total WTP meta-analysis by conducting an in-depth
search of the economic literature. EPA used a variety of sources and search methods to identify relevant
studies:
> Review of EPA's research and bibliographies dealing with non-market benefits associated with
water quality changes
> Systematic review of recent issues of key resource economics journals (e.g., Land Economics,
Marine Resource Economics, Journal of Environmental Economics and Management)
> Searches of online reference and abstract databases (e.g., Environmental Valuation Resource
Inventory (EVRI), Benefits Use Valuation Database (BUVD), AgEcon Search)
> Visits to home pages of authors known to have published stated preference studies and/or water
quality valuation research
> Searches of Web sites of agricultural and resource economics departments at several colleges
and universities
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-2
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
> Searches of Web sites of organizations and agencies known to publish environmental and
resource economics valuation research (e.g., Resources for the Future (RFF), National Center for
Environmental Economics (NCEE)).
From this review, EPA identified approximately 300 surface water valuation studies that were potentially
relevant for this analysis and compiled a bibliographic database to organize the literature review process.
Sixty-seven of these studies met the criteria identified for inclusion in the meta-analysis.81 These criteria were
designed to ensure both commodity and welfare consistency as noted above, and include:
> Specific amenity valued: Selected studies were limited to those in which the environmental
quality change being valued affects ecological services provided by surface waterbodies,
including aquatic life support, recreational activities (such as fishing, boating, and swimming),
and nonuse value
> Values estimated: Selected studies were limited to those that used stated preference techniques
to elicit household WTP.
> Study location: Selected studies were limited to those that surveyed U.S. and Canadian
populations to value resources
> Research methods: Selected studies were limited to those that applied research methods
supported by journal literature.
The Agency compiled extensive information from the 67 selected studies. Of these studies, 45 were utilized in
the model estimation. Reasons for the difference between the total number of studies in the final metadata
(67) and studies represented in the final model (45) include unavailability of information for certain key
regressors for all studies.
The tradeoff between the number of regressors or independent variables that may be included in a meta-
regression analysis (K) and the number of studies that are included in the metadata (N) is a fundamental
tradeoff in most meta-analyses in the valuation literature (Moeltner et al. 2007). That is, if a study considered
for inclusion in the metadata does not provide information for a certain regressor that analysts might wish to
include in the meta-regression, analysts must generally choose between omitting the regressor from the meta-
regression or omitting the study from the metadata. As a result, researchers wishing to increase the number of
explanatory variables in meta-regression models (increasing K) often do so at the cost of reductions in the
number of studies or observations in the metadata (reducing N). Conversely, increases in the number of
observations in meta-regression models are sometimes only possible if one reduces the number of
independent variables in the model. Hence, there is a tradeoff between the quantity of information in the
metadata (i.e., the number of observations or studies) and the possible risk of omitted variables bias due to the
omission of influential regressors. See Moeltner et al. (2007) for additional discussion of this "N versus K"
tradeoff in meta-analysis.
The data set used in the meta-analysis includes the following information:
> Full study citation
> Study location
> Sample data and description (e.g., size, response rate, income)
> Resource characteristics (e.g., affected waterbody type, recreational uses, baseline quality)
81 The remaining studies were either earlier unpublished versions or slightly modified versions of the included studies,
focused on water resources outside of the United States and Canada, used secondary research methods, or valued
environmental quality changes that were not directly linked to changes in water quality (e.g., change in recreational catch
rates).
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-3
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
> Description of environmental quality change, including geographic scale, affected species, and
affected recreational uses (e.g., water quality change from fishable to beatable)
> Quantitative measure of environmental quality change measured in terms of improvements in
Water Quality Index (WQI)82 and/or percentage reduction in pollutant concentration
> Study WTP values updated to 2007 dollars
> WTP estimation characteristics (i.e., parametric versus non-parametric, inclusion of protest bids
and outlier bids, WTP description).
£.7.2 Description of Studies Selected for Total WTP Meta-Analysis
The 45 studies that EPA used in the total WTP meta-analysis were conducted between 1981 and 2008, and
applied standard, generally accepted stated preference valuation methods to assess WTP. Studies were
excluded if they did not conform to general tenets of economic theory, or if they applied methods not
generally accepted in the literature.
All selected studies focus on environmental quality changes that affect surface water resources in the United
States. Beyond this general similarity, the studies vary in several respects. Differences include the specific
environmental change valued, scale of environmental improvement, geographic region affected by
environmental changes, types of values estimated, survey administration methods, demographics of the
survey sample, and statistical methods employed. The 45 studies include 25 journal articles, 6 reports, 5 Ph.D.
dissertations, 7 academic or staff papers, 1 book, and 1 master's thesis.
The 45 studies selected for the meta-analysis provided 115 observations in the final data set because multiple
estimates of WTP were available from 30 studies. The availability of multiple observations from single
studies is common in meta-analyses of this type (e.g., Bateman and Jones 2003; Johnston et al. 2005). Some
of the characteristics that allowed multiple observations to be derived from a single study include the extent of
the amenity change, the respondent population type, elicitation method(s), waterbody type, number of
waterbodies affected, recreational activities affected by the quality change, and species affected by the quality
change. These variations are often due to experimental design driven by the key research questions or
hypotheses. Exhibit E-l lists key study and resource characteristics and indicates the number of observations
derived from each study.
Surveys in 26 studies were administered by mail; 9 studies collected information through personal interviews
in the home, onsite, or in a centralized location, 9 surveys were conducted by telephone, and 1 conducted by
computer administration. Study sample sizes range from 96 to 4,033 responses.
82Additional details on the WQI and the use of the WQI in survey instruments are provided by McClelland (1974),
Vaughan (1986), and Mitchell and Carson (1989, p. 342). This index is linked to specific pollutant levels, which in turn
are linked to presence of aquatic species and suitability for particular recreational uses. The WQI allows the use of
objective water quality parameters (e.g., dissolved oxygen concentrations) to characterize ecosystem services or uses
provided by a given waterbody.
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-4
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix E: Meta-Analysis
Exhibit E-1: Selected Summary Information for Studies
Author(s) and Year
Aiken(1985)
Anderson and
Edwards (1986)
Azevedoetal. (2001)
Bockstael et al.
(1988)
Bockstael et al.
(1989)
Breffleetal. (1999)
Cameron (1988)
Cameron and
Huppert(1989)
Carson and Mitchell
(1993)
Carson etal. (1994)
Clonts and Malone
(1990)
Croke etal. (1986-
87)
DeZoysa(1995)
Desvousges et al.
(1987)
Hayes etal. (1992)
Herriges and Shogren
(1996)
Kite (2002)
Huang etal. (1997)
Hushak and Bielen
(1999)
Kaoru(1993)
Lant and Roberts
(1990)
Lant and Tobin
(1989)
Lichtkoppler and
Elaine (1999)
Lindsey(1994)
Lipton (2004)
Loomis(1996)
Loomis et al. (2000)
Lyke(1993)
Matthews et al.
(1999)
Olsen etal. (1991)
Opaluch etal. (1998)
Roberts and Leitch
(1997)
Roweetal. (1985)
Sanders etal. (1990)
Schulze etal. (1995)
Shrestha and
Alavalapati (2004)
Observatio
ns
1
1
5
1
2
2
1
1
4
2
3
9
2
12
2
2
2
2
2
1
3
9
1
8
1
1
2
2
2
3
1
1
1
4
2
2
State
CO
RI
IA
DC,MD,
VA
MD
WI
TX
CA
National
CA
AL
IL
OH
PA
RI
IA
MS
NC
OH, MI
MA
IA,IL
IA,IL
OH
MD
MD
WA
CO
WI
MN
ID, MT,
OR,WA
NY
MN, SD
CO
CO
MT
FL
Waterbody Type
all freshwater
salt
impoundment/marshes
Lake
estuary
estuary
estuary
Estuary
river/stream
multiple
estuary
river/stream
river/stream
river and lake
river/stream
estuary
Lake
river/stream
estuary
river/stream
salt
impoundment/marshes
river/stream
river/stream
multiple
estuary
estuary
river/stream
river/stream
lake
river/stream
river/stream
estuary
lake
river/stream
river/stream
river and lake
multiple
Type of Water Quality
Improvement
general water quality
general water quality
nutrients
general water quality;
phosphorus and nitrogen
general water quality
general water quality
general water quality
wildlife habitat
general water quality
DDTandPCBs
general water quality
general water quality
sediment and nutrients;
wildlife habitat
general water quality
general water quality
general water quality
general water quality
general water quality
general water quality
fecal conform
sediment
general water quality
PCBs and general water
quality
nutrients
general water quality
general water quality
general water quality
general water quality
phosphorus
wildlife habitat
general water quality
general water quality
general water quality
general water quality
hazardous pollutants
phosphorus and wildlife
habitat
Affected
Recreational Uses
Fishing
fishing and swimming
fishing and swimming
swimming, beach,
boating, fishing,
outings
Swimming
Fishing
Fishing
game fishing
boating; fishing;
swimming
Fishing
multiple uses
multiple uses
multiple uses
multiple uses
swimming; fishing
boating and fishing
multiple uses
Fishing
multiple uses
Fishing
boating, fishing, and
swimming
boating, fishing
all recreational uses
multiple uses
Boating
Fishing
multiple uses
Fishing
boating and fishing
Fishing
Shellfishing
multiple uses
boating, fishing, and
swimming
Swimming
boating, fishing, and
swimming
multiple uses
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
E-5
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix E: Meta-Analysis
Exhibit E-1: Selected Summary Information for Studies
Author(s) and Year
Stumborg et al.
(2001)
Sutherland and
Walsh (1985)
Viscusi et al. (2008)
Welle (1986)
Wey(1990)
Whitehead et al.
(2002)
Whitehead and
Groothuis(1992)
Whitehead et al.
(1995)
Whittington et al.
(1994)
Observatio
MS
2
1
2
6
2
1
3
2
1
State
WI
MT
National
MN
RI
NC
NC
NC
TX
Waterbody Type
lake
river and lake
river and lake
all freshwater
salt impoundment/
marshes
river/stream
river/stream
estuary
estuary
Type of Water Quality
Improvement
phosphorus
general water quality
general water quality
acid rain
general water quality
general water quality
sediment and nutrients
general water quality
heavy metals and
pesticides
Affected
Recreational Uses
multiple uses
Swimming
multiple uses
game fishing and
wildlife viewing
Other
fishing, boating,
swimming
multiple uses
boating, fishing, and
swimming
multiple uses
The Agency's review of the relevant economic literature showed that available surface water valuation studies
focus primarily on general water quality rather than specific pollutants or changes. Even in cases in which
specific pollutants are the primary policy issue, the stated preference surveys from which welfare estimates
are derived often characterize water quality changes only in general (i.e., non-pollutant specific) terms.
Hence, the associated welfare measures are conditioned on this general description. Of the 45 studies, 26
presented only WTP values for changes in general water quality (approximately 60 percent). Of the studies
that did address specific changes, eight specified nutrients and/or sediment, four addressed hazardous
pollutants including heavy metals and pesticides, three addressed wildlife habitat, one addressed acid
deposition, one addressed fecal coliform bacteria, one presented values for both changes in general water
quality and nutrient reductions, and one presented values for both changes in general water quality and
wildlife habitat. Preliminary model estimates showed no evidence that the type of pollutant considered had a
statistically significant influence on WTP across and within studies. For this reason, EPA used a standardized
scale to define both the baseline water quality and the water quality change valued in the original study.
Additional details are provided below.
From these 45 studies, the Agency compiled a data set for the meta-analysis of total WTP values. EPA
specified a regression model based on these data to estimate a range of total household benefits for surface
water and aquatic habitat improvements. General empirical methods follow those outlined by Johnston et al.
(2005), following standard approaches in the meta-analysis literature. The model and results are described in
the next section.
E.2 Total WTP Meta-Analysis Regression Model and Results
EPA estimated both trans-log and semi-log meta-regression models based on 115 WTP estimates for
improvements in water resources, derived from 45 original studies.83 These metadata, the model specification,
model results, and interpretation of the results are described in the following sections. EPA, however, notes
that only the trans-log model is used in the analysis of benefits of the proposed ELGs. The alternative
specification (semi-log) is presented for comparative purposes only.
83In its analysis of nonuse benefits for the final 316(b) Phase II rule, EPA also specified trans- and semi-log regression
models similar to the models estimated for Phase II discussed in this section. See Chapter A12, "Methods for Estimating
Non-use Benefits," in the Regional Analysis document for further details regarding both the log-log and semi-log
regression models estimated in EPA's analysis of nonuse benefits for the final Phase II rule (USEPA 2004c);
http://www.epa.gov/ost/316b/casestudv/final.htm).
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
E-6
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
In a frequently cited work, Glass (1976) characterizes meta-analysis as "the statistical analysis of a large
collection of results for individual studies for the purposes of integrating the findings. It provides a rigorous
alternative to the casual, narrative discussion of research studies which is commonly used to make some sense
of the rapidly expanding research literature" (p. 3; cited in Poe et al. 2001, p. 138). Meta-analysis is being
increasingly explored as a potential means to estimate resource values in cases where original targeted
research is impractical, or as a means to reveal systematic components of WTP (e.g., Smith and Osborne
1996; Santos 1998; Rosenberger and Loomis 2000a; Poe et al. 2001; Woodward and Wui 2001; Bateman and
Jones 2003). While the literature often urges caution in the use and interpretation of benefit transfers for direct
policy application (e.g., Desvousges et al. 1998; Poe et al. 2001; Navrud and Ready 2007), such methods are
"widely used in the United States by government agencies to facilitate benefit-cost analysis of public policies
and projects affecting natural resources" (Bergstrom and De Civita 1999). Transfers based on meta-analysis
are likewise common in both the United States and Canada (Bergstrom and De Civita 1999; Bergstrom and
Taylor 2006).
Depending on the suitability of available data, a meta-analysis can provide a superior alternative to the
calculation and use of a simple arithmetic mean WTP over the available observations, as it allows estimation
of the systematic influence of study, economic, and natural resource attributes on WTP (U.S. EPA, 2010b;
Rosenberger and Phipps 2007; Shrestha et al. 2007). The primary advantage of a regression-based (statistical)
approach is that it accounts for differences among study characteristics that may contribute to changes in
WTP, to the extent permitted by available data (Johnston et al. 2005; Rosenberger and Phipps 2007). An
additional advantage is that meta-analysis can reveal systematic factors influencing WTP, allowing analysts to
assess whether, for example, WTP estimates are (on average) sensitive to scope (Smith and Osborne 1996).
There is, however, some controversy regarding whether regression-based meta-analyses should be used for
direct benefit transfer. Many contemporary sources in the literature note the potential ability of regression-
based meta-analyses to generate benefit functions better able to adjust and forecast benefits at policy sites in
question, and either explicitly or implicitly favor the use of meta-analysis over alternative benefit transfer
approaches (e.g., Johnston et al. 2005; Bergstrom and Taylor 2006; Moeltner et al. 2007; Rosenberger and
Johnston 2007; Rosenberger and Phipps 2007; Shrestha et al. 2007). EPA (USEPA 2000b) characterizes
meta-analysis as "the most rigorous" benefit transfer method. In contrast, the EPA Science Advisory (2007)
Board's "Advisory on EPA's Issues in Valuing Mortality Risk Reduction" recommends against the use of
regression-based meta-analysis for VSL (value of statistical life) transfers, and other authors advise caution in
such uses (Navrud and Ready 2007). The primary disagreement is whether it is appropriate to use meta-
analysis results as a reduced form model to estimate benefits, and whether the empirical ability of meta-
analysis in many cases to generate benefit transfers with reduced transfer errors offsets the lack of an
underlying theoretical model to "calibrate" benefit estimates across studies (cf Smith and Pattanayak 2002;
Smith et al. 2002). While the Agency recognizes this ongoing controversy, it notes that there are a large
number of practitioners and publications supporting the use of regression-based meta-analysis for benefit
transfer. It also removes the element of subjective judgment associated with selecting a single study or value
for benefit transfer. Hence, the following model is presented as a means to provide a benefit function that
capitalizes on the substantial information available for existing water quality valuation studies,
notwithstanding potential concerns voiced by some regarding the use of meta-analyses for such purposes.
£.2.7 Metadata Total WTP Regression Model
Meta-analysis is largely an empirical, data-driven process, but one in which variable and model selection is
guided by theory (Bergstrom and Taylor 2006). Given a reliance on information available from the underlying
studies that comprise the metadata, meta-analysis models most often represent a middle ground between
model specifications that would be most theoretically appropriate and those specifications that are possible
given available data. Smith and Osborne (1996), Rosenberger and Loomis (2000a), Poe et al. (2001),
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-7
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
Bateman and Jones (2003), Dalhuisen et al. (2003), Johnston et al. (2005, 2006), Bergstrom and Taylor
(2006), Moeltner et al. (2007), and others provide insight into the mechanics of specifying and estimating
meta-equations in resource economics applications.
Past meta-analyses have incorporated a range of different statistical methods, with none universally accepted
as superior (Johnston et al. 2005). EPA followed recent work of Bateman and Jones (2003) and Johnston et al.
(2005) in applying a multilevel model specification to the metadata, to address potential correlation among
observations gathered from single studies. Also following prior work (e.g., Smith and Osborne 1996; Poe et
al. 2001), EPA applied the Huber-White robust variance estimation. As described by Smith and Osborne
(1996, p. 293), "this approach treats each study as the equivalent of a sample cluster with the potential for
heteroskedasticity... across clusters." Weighted models are avoided following the arguments of Bateman and
Jones (2003).84
To guide development of the model and variable specifications, EPA relied upon a set of general principles.
These principles are designed to help prevent excessive data manipulations and other factors that may lead to
misleading model results. The general principles include, all else being equal:
> Fewer and simpler data transformations are preferred to more extensive ones.
> In the absence of overriding theoretical considerations, continuous variables are generally
preferred to discrete variables derived from underlying continuous distributions.
> Models should attempt to capture elements of the scope and scale of resource changes.
> Models should distinguish WTP associated with different types of resources and resource uses,
particularly where relevant to the policy question at hand.
> Following the "weak structural utility theoretic" (WSUT) approach of Bergstrom and Taylor
(2006, p. 352), exogenous structural constraints are avoided to afford the flexibility necessary to
appropriately model empirical patterns that may not necessarily flow from an underlying
theoretical modeling structure. The dependent variable in the meta-analysis is the natural
logarithm of estimated household WTP for water quality improvements in aquatic habitat, as
reported in each original study. For this analysis, original study values were adjusted to 2007$
based on the relative change in Consumer Price Index (CPI) from the study year to 2007. Total
WTP over the sample ranged from $5.33 to $502.70, with a mean value of $83.09.
As noted above, two model specifications are estimated (cf Johnston et al. 2005). For the first specification,
all right-hand-side variables are linear, resulting in a standard semi-log functional form. The second
specification is identical to the semi-log model, except for the specification of the explanatory variable
measuring water quality change and baseline as natural logs. This results in a trans-log functional form, also
common in empirical applications (Johnston et al. 2001, 2005). Both the semi-log and trans-log models have
advantages related to (1) their fit to the data, (2) the intuitive results that are provided, and (3) their common
use in the empirical valuation and meta-analysis literature (e.g., Smith and Osborne 1996; Santos 1998;
Johnston et al. 2001, 2005, 2006). The trans-log model, however, has the additional structural advantage that
estimated WTP is necessarily zero when water quality change is also zero (cf. Johnston et al. 2001)—a
property suggested by theory that analysts may wish to weight against model fit considerations when
choosing a model for benefit estimation. While linear forms are also common in this literature (Rosenberger
84For comparison, models were also estimated using ordinary least squares (OLS) with robust variance estimation,
weighted least squares with robust variance estimation, and multilevel models with standard (non-robust) variance
estimation. None of these models outperformed the illustrated model in terms of overall model significance and fit, and
statistical significance of individual coefficients.
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-8
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
and Loomis 2000a, 2000b; Poe et al. 2001; Bateman and Jones 2003), specifications requiring more intensive
data transformations (e.g., Box-Cox, log-log) are less common.
As noted in the preceding section, the metadata include independent variables characterizing specific details
of the resource(s) valued such as the baseline resource conditions; the extent of resource improvements and
whether they occur in estuarine or freshwater; the geographic region and scale of resource improvements
(e.g., the number of waterbodies); resource characteristics (e.g., baseline conditions, the extent of water
quality change, and ecological services affected by resource improvements); characteristics of surveyed
populations (e.g., users, nonusers); and other specific details of each study. For ease of exposition, these
variables are categorized into those characterizing (1) study and methodology, (2) surveyed populations, (3)
geographic region and scale, and (4) resource improvements. Attributes included within each category are
summarized below.
Study and methodology variables characterize such features as:
> The year in which a study was conducted
> The payment vehicle and elicitation format (e.g., discrete choice versus open-ended, voluntary
versus non-voluntary, interview versus mail versus phone)
> WTP estimation methods and conventions (e.g., approaches to protest and outlier bids, use of
parametric versus non-parametric statistical methods, estimation of mean or median WTP, the
use of annual or lump-sum payments)
> Whether the original survey represented water quality changes using the WQI.
Surveyed populations variables characterize such features as:
> The average income of respondents
> Whether the survey specifically targeted nonusers.
Geographic region and scale variables characterize such features as:
> The number of waterbodies affected by the policy
> Whether the study considered water quality improvements in all waterbodies in a region
> The geographic area of the country in which the study was conducted.
Resource improvement variables characterize such features as:
> The extent of water quality change estimated as a difference between the baseline and post-
change water quality index
> Baseline water quality index
> Those studies for which recreational uses such as fishing are specifically noted in the survey
> Aquatic species affected by resource improvements (e.g., game fish and shellfish)
> Those studies identifying large increases in fish populations (i.e., greater than 50 percent).
Although the interpretation and calculation of most independent variables requires little explanation, a few
variables require additional detail. These include the variables characterizing surface water quality and its
measurement. Many (23) observations in the metadata characterize quality changes using variants of the WQI
(e.g., Mitchell and Carson 1989). This scale is linked to specific pollutant levels, which, in turn, are linked to
the presence of aquatic species and recreational uses. However, some observations provide water quality
measures using other, primarily descriptive, means that differ from the WQI.
To allow consistent comparisons of water quality change using a single scale, EPA mapped all water quality
measures to the original WQI developed by McClelland (1974). WQI values were therefore developed for
those studies that did not originally use this index. This scale was chosen for two reasons:
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-9
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
> WQI values are linked to specific pollutant levels, which, in turn, are linked to the presence of
aquatic species and suitability for particular recreational uses. Therefore, the WQI can be used to
link water quality changes from reduced nutrient and heavy metal connections to effects on
human uses and support for aquatic species habitat.
> A large number of the original studies in the metadata included WQI measures as "native"
components of the original surveys. Hence, for these studies, no additional transformations were
required.
While not all studies in the metadata included the WQI as a native survey component, in most cases the
descriptions of water quality (present in the studies that did not apply the WQI) rendered mapping of water
quality measures to the WQI straightforward. In cases where baseline and improved (or declined) water
quality was not defined by suitability for recreational activities (e.g., boating, fishing, and swimming) or
corresponding qualitative measures (e.g., poor, fair, good), EPA used descriptive information available from
studies (e.g., amount/indication of the presence of specific pollutants, historical decline of the quality of the
resource) to approximate the baseline level of water quality and the magnitude of the change.85 For studies
that valued discrete changes in the size of species populations, EPA characterized the baseline quality based
on the current presence and prevalence of the species at hand, and assumed population increases to
correspond to modest increases in water quality in order to be conservative.86 To account for the uncertainty
involved in mapping those studies that are not based on the WQI, EPA introduced the binary variable WQI,
which indicates those studies in which WQI measurements were an original component of the survey
instrument. This approach is based largely on the published methods of Johnston et al. (2005), drawn from
prior Agency work for the 316(b) Phase II Cooling Water Intake rule.
Variables incorporated in the final model are listed and described in Exhibit E-2.
85For example, a study by Huang et al. (1997) described current water quality as degraded from 1981 levels in terms of
reduced fish catches (60 percent) and reduced number of open shellfish beds (25 percent). However, because the water
resource was still supporting recreational fishery, the baseline water quality was set to "fishable" on the WQI.
86 For example, a study by Lyke (1993) describes the baseline conditions as follows: (1) "there are no naturally
reproducing lake trout in Lake Michigan; all lake trout found there are from hatcheries." (2) "Lake Superior stocks of
self-reproducing lake trout were much reduced, but not wiped out, and both natural and hatchery-raised lake trout are
found there." These baseline conditions correspond to the "game-fishable" level on the WQI. The study estimates WTP
for restoring natural populations of lake trout to the Wisconsin Great Lakes. Therefore, the expected change that would
occur within the "game-fishable" category is likely to be small.
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-10
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix E: Meta-Analysis
Exhibit E-2: Variables and Descriptive Statistics for the Total WTP Regression Model
Variable
ln_WTP
year_indx
discrete
volunt
mail
lump_sum
nonparam
quality_ch
lnquality_cha
WQI
non-reviewed
outlier_bids
median_WTP
income
nonusers
single river
single_lake°
Description
Natural log of WTP for specified resource improvements.
Year in which the study was conducted, converted to an index by
subtracting 1980.
Binary (dummy) variable indicating that WTP was estimated
using a discrete choice survey instrument.
Binary (dummy) variable indicating that WTP was estimated
using a payment vehicle described as voluntary as opposed to,
for example, property taxes.
Binary (dummy) variable indicating that the survey was
conducted through mail (default value for this dummy is a phone
survey).
Binary (dummy) variable indicating that payments were to occur
on something other than an annual basis over a long period of
time, such as property taxes. For example, some studies
specified that payments would occur over a five-year period.
Binary (dummy) variable indicating that WTP was estimated
using non-parametric methods.
The change in mean water quality, specified on the WQI
(McClelland 1974; Mitchell and Carson 1989). Defined as the
difference between baseline and post-compliance quality. Where
the original study (survey) did not use the WQI, EPA mapped
water quality descriptions to analogous levels on the WQI to
derive water quality change (see text).
Natural log of the change in mean water quality (quality ch),
specified on the WQI (McClelland 1974; Mitchell and Carson
1989).
Binary (dummy) variable indicating that the original survey
reported resource changes using a standard WQI.
Binary (dummy) variable indicating that the study was not
published in a peer-reviewed journal.
Binary (dummy) variable indicating that outlier bids were
excluded when estimating WTP.
Binary (dummy) variable indicating that the study reported
median, not mean, WTP.
Mean income of survey respondent, either as reported by the
original survey or calculated by EPA based on U.S. Census
Bureau averages for the original surveyed region.
Binary (dummy) variable indicating that the survey was
implemented over a population of nonusers (default category for
this dummy is a survey of any population that includes users).
Binary (dummy) variable indicating that resource change
explicitly took place over a single river (default is a change in an
estuary or that takes place on a national scale).
Binary (dummy) variable indicating that resource change
explicitly took place over a single lake (default is a change in an
estuary or that takes place on a national scale).
Units and
Measurement
Natural log of
dollars (Range:
2. 12 to 6.22)
Year index
(Range: 1 to 28)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
WQI units
(Range: 2.5 to 65)
WQI units
(Range: 0.92 to
4.17)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Dollars
(Range: 34,955 to
158,347)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Mean
(Std. Dev.)
4.42
(0.75)
9.68
(6.42)
0.28
(0.45)
0.05
(0.22)
0.48
(0.52)
0.13
(0.34)
0.50
(0.50)
21.3
(10.4)
2.91
(0.59)
0.34
(0.48)
0.32
(0.47)
0.94
(0.24)
0.04
(0.20)
5,7049.59
(13,946.64)
0.09
(0.29)
0.23
(0.42)
0.08
(0.28)
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
E-11
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix E: Meta-Analysis
Exhibit E-2: Variables and Descriptive Statistics for the Total WTP Regression Model
Variable
regional_fresh
multiple_river
salt_pond
num_riv_pond
mr
mp
allmult
nonspec
fish_use
fishplus
baseline
Inbase
Description
Binary (dummy) variable indicating that resource change
explicitly took place over an entire region such as a state (default
is a change in an estuary or a change that takes place on a
national scale).
Binary (dummy) variable indicating that resource change
explicitly took place over multiple rivers (default is a change in
an estuary or that takes place on a national scale).
Binary (dummy) variable indicating that resource change
explicitly took place over multiple salt impoundments (default is
a change in an estuary or that takes place on a national scale).
Number of rivers or salt impoundments affected by policy; if
unspecified num riv_pond = 0. (In the present data, only studies
addressing rivers and lakes specified >1 number of waterbodies.
All others specified either 1 waterbody, or the number was
unspecified.)
Binary (dummy) variable indicating that the survey included
respondents from more than one of the EPA regions.
Binary (dummy) variable indicating that the survey included
respondents from the Mountain Plain region.d
Binary (dummy) variable indicating that either all or multiple
aquatic species are affected by the resource change.
Binary (dummy) variable indicating that the study did not
specify what species would be affected by water quality
improvements.
Binary (dummy) variable identifying studies in which changes in
fishing use are specifically noted in the survey.
Binary (dummy) variable identifying studies in which a fish
population or harvest change of 50 percent or greater is reported
in the survey.
Baseline water quality, specified on the WQI.
Natural log of baseline water quality, specified on the WQI.
Units and
Measurement
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Number of
specified rivers or
impoundments
(Range: 0 to 15)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
Binary
(Range: 0 or 1)
WQI units (Range:
1.61 to 5.2)
WQI units (Range:
5 to 70)
Mean
(Std. Dev.)
0.36
(0.48)
0.06
(0.24)
0.03
(0.18)
1.01
(2.98)
0.08
(0.26)
0.02
(0.13)
0.18
(0.39)
0.42
(0.50)
0.56
(0.50)
0.08
(0.28)
39.79
(20.37)
3.46
(0.80)
a. The variable quality_ch is defined earlier in this table as the difference between baseline and post-compliance quality, specified on the WQI
(Mitchell and Carson 1989).
b. Examples of rivers and streams considered in the studies include the Columbia, Potomac, Elwha, Eagle, and Tar-Pamlico rivers.
c. Includes one study that focused on a segment of the Lake Erie shoreline.
d. The Mountain Plain region includes the following states: Colorado, Iowa, Kansas, Missouri, Montana, Nebraska, North Dakota, South Dakota,
Utah, Wyoming.
£.2.2 Total WTP Regression Model and Results
As noted above, EPA estimated the meta-analysis regression using a multilevel, random-effects specification.
This model follows the general approaches of Bateman and Jones (2003) and Johnston et al. (2005), among
others. Multilevel (or hierarchical) models may be estimated as either random-effects or random-coefficients
models, and are described in detail elsewhere (Singer 1998). The fundamental distinction between these and
classical linear models is the two-part modeling of the equation error to account for hierarchical data. Here,
the metadata are comprised of multiple observations per study, and there is a corresponding possibility of
correlated errors among observations that share a common study or author.
The common approach to modeling such potential correlation is to divide the residual variance of estimates
into two parts, a random error that is independently and identically distributed across all studies and for each
observation, and a random effect that represents systematic variation related to each study. The model is
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
E-12
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
estimated as a two-level hierarchy, with level one corresponding to WTP estimates (individual observations),
and level two corresponding to individual studies. The random effect may be interpreted as a deviation from
the mean equation intercept associated with individual studies (Bateman and Jones 2003). The model is
estimated using a maximum likelihood estimator, assuming that random effects show a multivariate normal
distribution. Following Bateman and Jones (2003), observations are unweighted. Covariances are obtained
using the Huber-White covariance estimator (Smith and Osborne 1996). Random-effects models such as the
multilevel model applied here are becoming increasingly standard in resource economics applications, and are
estimable using a variety of readily available software packages.
A Note on Model Specification
As noted above, EPA considered two functional forms in this analysis: semi-log and trans-log. In both cases,
the dependent variable is the natural log of estimated household WTP for surface water quality improvement.
For Model One, all right-hand-side variables are linear, resulting in a semi-log functional form common in
meta-analysis (e.g., Smith and Osborne 1996; Johnston et al. 2003). While linear forms are also common
(Rosenberger and Loomis 2000a, 2000b; Poe et al. 2001; Bateman and Jones 2003), the semi-log and trans-
log forms were chosen based on statistical performance and ability to capture curvature in the valuation
function, and because they allow independent variables to influence WTP in a multiplicative rather than
additive manner.
Model Two is a trans-log model, identical to the semi-log specification save for the inclusion of water quality
measures (baseline and quality_ch) as natural logarithms. This form—common in the hedonic modeling
literature (Johnston et al. 2001) and illustrated by Johnston et al. (2005) within the meta-analysis literature—
shares many advantages of the semi-log functional form, but also incorporates the desirable quality that WTP
is constrained to zero when quality change is also equal to zero.
Following standard econometric practice and the "weak structural utility theory" approach to meta-analysis
summarized above (Bergstrom and Taylor 2006), the final models are specified based on guidance from
theory and prior literature. For example, Arrow et al. (1993) made a fundamental distinction between discrete
choice and open-ended payment mechanisms (such as iterative bidding and payment cards). Hence, this
distinction is made in the final model (i.e., including the variable discrete_ch). Similarly, other "survey
methodology" variables in the model were chosen based on theoretical considerations and prior findings in
the literature (e.g., voluntary versus mandatory payment vehicles, parametric versus non-parametric,
treatment of protest and outlier bids, use of mean versus median WTP). Also included are variables
characterizing the scope and scale of the resource change, based on theoretical expectations that such factors
should be relevant to welfare estimates.
Few variables were excluded solely because of lack of statistical significance. Individual variables were only
excluded if they could not be shown to be statistically significant in any version of the model (restricted or
unrestricted), and there was no overriding rationale for retaining the variable in the model. For example,
variables distinguishing different types of discrete choice instruments (e.g., conjoint versus dichotomous
choice) added no significant explanatory power to the model (p = 0.58).
It is important to note that although empirical considerations play a role in model development, certain
variables were retained in the model for theoretical reasons, even if significance levels were low. Such
specification of meta-analysis models using a combination of theoretical guidance and empirical
considerations is standard in modeling efforts (Bergstrom and Taylor 2006).
Exhibit E-3 presents results of the total WTP regression model.
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-13
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix E: Meta-Analysis
Exhibit E-3: Estimated Multilevel Model Results for the Trans-log and Semi-log Total WTP
Regression Models: WTP for Aquatic Habitat Improvements
Variable
intercept
year indx
discrete
volunt
nonparam
income
WQI
outlier bids
single river
single lake
multiple river
regional fresh
salt_pond
num riv pond
mr
mp
nonusers
allmult
nonspec
quality ch (lnquality_ch in
trans-log model)
fish use
fishplus
baseline (Inbase in trans-log
model)
mail
lump sum
non reviewed
median wtp
-2 Log Likelihood
Trans-log Model
Parameter Estimate
5.7109 c
-0.08043 c
-0.1248
-1.3233 c
-0.6698C
2.698 x 10"6'
-0.3275
-0.8837C
-0.4279 c
-0.06316
-1.4752C
0.1588
0.9849C
0.1173C
-0.8846C
1.6337C
-0.4036C
___!,
^4042^
0.40653
___!,
__!,
0.02610
-0.2013
__,,
-0.27183
-0.5358C
133.9C
Standard Error
0.9352
0.01482
0.2230
0.1653
0.1434
3.9 x W'6
0.2692
O2855
O1412
O2386
0.3540
0.1505
0.3580
0. 02806
0.1832
0.2980
0.1314
0.1644
01731
0.1488
0.1291
0.1820
0.1183
0.1466
0.2387
0.1625
0.1875
Semi-log Model
Parameter Estimate
6.0946 c
-0.06707C
-0.1696
-1.3049C
-0.6892C
-1.16xlO"7
-0.363 la
lQ/78295
'&TT245
---68
-1.5054°
0.22193
1.1357°
0.1145C
-0.7932C
1.5168°
-0.445 lc
-0.4044C
^02988
0.03208C
-0.4480C
0.4017C
0.005205
-0.3073C
0.7188C
-0.3744C
I574675E
11 5.?
Standard Error
0.5127
0.01430
0.1735
0.1661
0.1164
3.238 x IV6
0.2096
O2164
O1222
0;2223
0.3184
0.1168
0.3142
0.02510
0.1343
0.2574
0.1301
0.1266
01351
0.006337
0.1034
0.1217
0.003739
0.1155
0.1873
0.1234
0.1898
Covariance Factors:
Study Level (ou)
Residual (o e)
0
0.1876C
0
0.1599C
a. Significant at the 0.10 level.
b. Significant at the 0.05 level.
c. Significant at the 0.01 level.
£.2.3 Interpretation of Total WTP Regression Analysis Results
Regression results reveal strong systematic elements influencing WTP. The analysis finds both statistically
significant and intuitive patterns that influence WTP for surface water quality improvements. In general, the
statistical fit of the three estimated equations is good; model results suggest a considerable systematic
component of WTP variation that allows forecasting of WTP based on site and study characteristics.
Likelihood ratio tests show that model variables are jointly significant at the p<0.01 level for the trans-log
model and the p < 0.05 level for the semi-log model. In both models, the majority of independent variables
are statistically significant at p<0.05, with most statistically significant at pO.Ol. Signs of significant
parameter estimates generally correspond with intuition, where prior expectations exist. As shown in Exhibit
E-3, the random effect is statistically insignificant (i.e., study level covariance factors are zero). Considering
these factors, the statistical performance of both models compare favorably to prior meta-analyses in the
valuation literature.
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
E-14
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
Despite differences in the functional form of the two models, statistical results are robust across models. In
most cases, coefficient magnitudes and standard errors vary to only a small degree. Measures of equation fit
are similar, and both models are significant at the p<0.05 level or better. Such results mirror those of Johnston
et al. (2005), whose earlier meta-analysis of WTP for water quality improvements finds a high degree of
robustness to changes in model specification.
The initial discussion emphasizes results of the trans-log model. Although policy implications of both model
specifications are nearly identical for moderate to relatively large water quality improvements (e.g., more than
5 percent increase in WQI), the trans-log model provides more accurate WTP estimates when the expected
water quality change is very small (e.g., less than 1 percent increase in WQI).
One of the primary means to assess the validity of benefit transfers—and the only one that may be applied in
cases wherein the true value for the study site is unknown—is value surface tests (Bergstrom and De Civita
1999). These tests involve assessments of ways in which "different factors may cause values to vary across
sites, providing guidance for adjustments needed to make a valid transfer of value estimates from the study
site(s) to the policy site" (Bergstrom and De Civita 1999). Following general approaches for such value
surface assessments—which generally involve comparisons of empirical patterns found via meta-analysis to
theoretical expectations or norms—the Agency concluded that most results of the estimated value surface
suggest an appropriate benefit function. Results of these value surface assessments are detailed in the
following sections.
Resource Improvement Effects
Seven variables characterize resource improvements and uses; most are of the expected sign. The coefficient
on the quality_ch variable is positive and statistically significant (p<0.01), indicating that larger water quality
improvements generate larger WTP. This is an important result, and indicates that WTP is sensitive to the
scope of water quality improvements. The estimated model showed that WTP values are not sensitive to the
baseline water quality from which water quality change would occur. The estimated parameter on the variable
baseline representing the baseline water quality from which water quality change would occur is not
statistically significant (p>0.1). This finding differs from that of Johnston et al. (2005), which shows that
WTP for marginal water quality improvements declines as baseline water quality improves. Here, the value is
positive but is only significant at p<0.17 in the semi-log model. In the trans-log model, this parameter is not
statistically significant. The reason for this result is unknown, but may be related to highly valued uses that
are associated with larger values on the WQI. For example, increases beginning at higher levels on the WQI
may cross thresholds allowing such highly valued uses as swimming and drinking, such that increases at these
high-quality levels may be valued more highly than otherwise similar changes at lower baselines, which may
not allow such uses. Such thresholds, or other non-convex preference patterns, may lead to unexpected results
for the baseline water quality variable (baseline).
Both models reveal that water quality changes associated with recreational fishing uses lead to a significant
decrease in total WTP values, compared to improvements that do not affect fishing. The variable fishjise
identifies those studies that specified effects of water quality improvements on recreational fishing (e.g.,
increase in catch rates). The associated parameter estimate is significant (p<0.05) and has the negative sign. In
contrast, the variable fishplus identifies those studies for which the associated survey identified particularly
large gains in fish populations or harvest rates (>50 percent). The positive and statistically significant result
(p<0.05) indicates that large gains in fish populations or harvests are associated with statistically significant
increases in total WTP. Results suggest that while water quality improvements targeting fishing uses may not
be valued particularly highly on average, very large resultant improvements in fish populations or harvests are
associated with increases in WTP, ceteris paribus.
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-15
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
The variables alljnult and nonspec indicate that water quality improvements affect multiple species
(alljnulf) or unspecified species (nonspec), respectively. The default category from which these variables
allow systematic variations in WTP is a focus on particular aquatic species affected by water quality (e.g.,
shellfish or game fish). The associated coefficients are negative, indicating that WTP is lower when a survey
instrument does not specify what aquatic species would be affected (nonspec) or when all or multiple species
are affected (alljnult). The latter finding seems to be counterintuitive at first. However, when a survey
instrument focuses on the effect of water quality on particular species, it is a likely indication that these
effects are of significant concern to the affected communities, which typically leads to a higher WTP. That is,
this result suggests that WTP is higher when water improvements can be shown to offer targeted benefits to
specific, and often high-profile, species groups—as opposed to cases in which improvements benefit an often
poorly characterized group of species.
Geographic Region and Scale Effects
Ten binary variables characterize geographic region and scale; seven are statistically significant at p<0.10.
The default category from which these variables allow systematic variations in WTP is an estuarine
waterbody. Also included in this default are a small number of observations addressing national level
improvements. Compared to this baseline, WTP associated with rivers is lower (single_river and
multiple _river both have negative and significant values). Single Jake has a negative value, but it is not
significant. WTP for water quality gains in salt impoundments (salt_pond) is higher than for estuaries
(p<0.05). This is not surprising since water quality gains in salt impoundments correspond to an increase in
the number of acres of shellfish beds.
Of particular importance for the general validity of empirical findings, the model results further suggest that
WTP is sensitive to the number ofwaterbodies under consideration and geographic scale of improvement
(regionalJresh). Of the waterbody categories distinguished above, both rivers and salt impoundments
allowed variation in numbers of affected waterbodies explicitly described by the survey. This variation is
captured by the variable num_riv_pond.S7 The associated parameter estimate is statistically significant
(p<0.01) and indicates that WTP increases with the number ofwaterbodies considered. The parameter
estimate on the regional Jresh variable is positive and significant (p<0.10) in the semi-log model, indicating
that large-scale regional water quality improvements lead to an increase in WTP. These results, combined
with the statistical significance of the water quality change variables noted above, suggest that WTP values
(in this case for water quality improvements) are strongly sensitive to scope, both in terms of the number of
waterbodies considered, geographic scale of improvement, and the magnitude of water quality change. In the
trans-log model, the regional Jresh variable is positive but not statistically significant (p>0.30).
Finally, the regional indicator variables mp and mult_reg are statistically significant at p<0.01, suggesting that
there are significant differences among WTP estimates from surveys in different geographical regions of the
United States. The parameter estimate on the mult_reg variable is negative, indicating that WTP for non-local
water quality improvements (e.g., out of state) are lower compared to in-state or local resource improvements.
This is consistent with prior findings that WTP for water quality improvements declines with the distance
from the resource (Bateman et al. 2006). The magnitude of the Pacific Mountain (mp) regional effect suggests
that spurious or otherwise unexplained effects (e.g., the effect of specific researchers who appear more than
once in the data) may drive their overall magnitude. For example, the size of the positive parameter estimate
associated with WTP in the Pacific Mountain dummy (mp) leads in many cases to relatively large increases in
WTP for Pacific Northwest policies. Hence, EPA believes that particular, spurious, or unexplained aspects of
studies from this region may have caused the associated parameter estimate to have a larger-than-expected
Technically, this variable is the sum of two interaction variables: (1) an interaction between multiple_river and the number of
waterbodies noted in the survey (0 if unspecified) and (2) an interaction between salt_pond and the number ofwaterbodies noted in
the survey (0 if unspecified).
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-16
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
influence on WTP. Although effects of regional dummy variables often escape simple, intuitive
characterization, EPA notes that they are often statistically significant in meta-analysis found in the valuation
literature. Similar issues are found by Johnston et al. (2005), for example.
Surveyed Populations Effects
Only two variables, nonusers and income, are used to characterize surveyed populations. In particular, the
nonusers variable is of substantial policy relevance. The negative and strongly significant (p<.01) parameter
estimate indicates that surveys of nonusers only, who by definition have only nonuse values for the resource
improvements in question (cf. Freeman 2003, p. 142), generate lower WTP values than surveys that include
users, who may have both use and nonuse values. Based on this statistically significant result, it is possible to
use this model to estimate nonuse values, interpreted as the mean WTP values estimated by surveys of
nonusers only. Such methods, however, may underestimate nonuse values of the general population, if the
nonuse values of users exceed those of nonusers (Whitehead and Blomquist 1991a,b).
The income parameter estimate is positive in the trans-log model, as expected, but is not statistically
significant. Such lack of statistical significance for income parameters is not uncommon in meta-analyses
found in the literature (e.g., Johnston et al. 2005).
Study and Methodology Effects
As often found in meta-analyses within the valuation and benefit transfer literature (Navrud and Ready 2007),
a variety of study and methodology effects can be shown to influence WTP for water quality improvements.
While expected, this does indicate that the methodological approach influences WTP, as argued by Arrow et
al. (1993). Of nine variables characterizing study and methodological effects, eight are statistically significant
at p<0.10. Among these is the year in which a study was conducted (yearjndx, a continuous variable), with
later studies associated with lower WTP. This is the expected result, as the focus of survey design over time
has often been on the reduction of survey biases that would otherwise result in an overstatement of WTP
(Arrow etal. 1993).
Model results reveal that voluntary (voluntary=\) payment vehicles (i.e., surveys that describe hypothetical
payments as voluntary) are associated with reduced WTP estimates. This result counters common intuition
and empirical findings that voluntary payment vehicles are associated with overstatements of true WTP
(Carson et al. 2000). The reason for this counter-intuitive finding is unknown, but may reflect an
unwillingness among respondents to offer large voluntary payments, given the fear that others will free-ride
(Johnston et al. 2005). Reduced WTP estimates are also associated with studies applying non-parametric
methods to WTP estimation (nonparam). Survey elicitation method does not have a strong effect in this
model; studies using discrete choice formats have lower WTP values, but this difference is not statistically
significant.
Smaller WTP estimates are associated with studies that eliminate or trim outlier bids when estimating WTP
(outlier_bids=\; p<0.01). Studies that report median WTP (median_WTP; p<0.01) have lower WTP values.
Lower WTP is associated with the use of the WQI in the original survey (WQI=\). This parameter is,
however, significant in the semi-log model only (p<0.1). As is the case with a variety of study design
variables, there is no necessary expectation with respect to the direction of this effect. Nonetheless, this
finding might suggest the capacity of such scales to clarify the specific magnitude and implications of water
quality change, and hence (perhaps) reduce methodological misspecification or symbolic biases that might act
to systematically inflate estimated WTP.
Survey format variables also have an effect on WTP, as might be expected. Mail has a negative and
statistically significant coefficient (p<0.01) in the semi-log model, compared to the default of telephone
surveys or interviews. This parameter is negative, but not statistically significant (p>0.17) in the trans-log
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-17
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix E: Meta-Analysis
model. It may be possible that the interview and telephone survey format results in larger WTP values either
because the respondents are better able to understand the valuation scenario, or because respondents may feel
pressure from interviewers to bias their WTP estimates upward. Finally, studies that ask respondents to report
an annual payment (as opposed to a lump_sum payment) have higher WTP estimates (p<0.05). This likely to
reflect the fact that annual payments are regarded as an infinite contribution and may reflect the respondent's
uncertainly regarding his future income and budget constraints.
£.2.4 Model Selection
To select the model for estimating benefits of water quality improvements from the regulation, EPA
calculated WTP values for a range of WQI changes using both semi-log and trans-log models. In all cases the
baseline WQI is set to 50, which approximates the average WQI value across all RF1 reaches in the United
States. EPA assigned values to other independent variables corresponding with theory, characteristics of the
water resource, and the policy context. Section 4 provides a complete list of values assigned to the remaining
independent regressors.
As shown in Exhibit E-4 both the semi-log and trans-log models yield similar WTP for water quality changes
greater than five points as measured by WQI. However, the semi-log model is not sensitive to very small
water quality changes (i.e., changes less than one point on the WQI index). Because the expected water
quality changes from the regulation are relatively small, the Agency selected the trans-log specification for
estimating benefits from the proposed ELGs, as a more conservative option.
Exhibit E-4: Comparison of WTP for Different Changes in WQI Based on Semi-log and
Trans-log Models
Model
Trans-log
Semi-log
20 Point
Change in
WQI
105.0186
89.0567
10 Point
Change in
WQI
79.2313
64.6167
5 Point
Change in
WQI
59.7760
55.0407
1 Point
Change in
WQI
31.0738
48.4122
0.50 Point
Change
in WQI
23.4436
47.6419
0.10 Point
Change
in WQI
12.1869
47.0345
0.01 Point
Change
in WQI
4.7796
46.8989
Semi-log Error Term = 0.1599.
Trans-log Error Term =0.1876.
E.3 Model Limitations
The validity and reliability of benefit transfer—including that based on meta-analysis—depends on a variety
of factors. While benefit transfer can provide valid measures of use and nonuse benefits, tests of its
performance have provided mixed results (e.g., Desvousges et al. 1998; Vandenberg et al. 2001; Smith et al.
2002; Shrestha et al. 2007). Nonetheless, benefit transfers are increasingly applied as a core component of
benefit cost analyses conducted by EPA and other government agencies (Bergstrom and De Civita 1999;
Rosenberger and Phipps 2007). Moreover, Smith et al. (2002, p. 134) argue that "nearly all benefit cost
analyses rely on benefit transfers, whether they acknowledge it or not." Given the increasing [or as Smith et
al. (2002) might argue, universal] use of benefit transfers, an increasing focus is on the empirical properties of
applied transfer methods and models.
Although the statistical performance of the model is good, EPA notes several limitations of the model. These
limitations stem largely from information available from the original studies, as well as degrees of freedom
and statistical significance. An important factor in any benefit transfer is the ability of the study site or
estimated valuation equation to approximate the resource and context under which benefit estimates are
desired. As is common, the meta-analysis model presented here provides a close but not perfect match to the
context in which values are desired. Although all of the studies used in the meta-analysis valued changes in
water quality improvements, many studies did not specify the cause of water quality impairment in the
baseline or focused on causes that are different from the pollutant of concern in the regulation (i.e., heavy
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
E-18
-------
Benefit and Cost Analysis for Proposed ELGs Appendix E: Meta-Analysis
metals and nutrients). Preliminary models, however, suggest no systematic patterns in WTP associated with
such factors, at least in the present metadata.
Additional limitations relate to the paucity of demographic variables available for inclusion in the model. The
only demographic variable incorporated in the analysis (income) was not statistically significant. Moreover,
other demographic variables are unavailable.
The estimated model is statistically significant and allows estimation of WTP based on study and site
characteristics. However, strictly speaking, model findings are relative to the specific case studies considered,
and must be viewed within the context of the 115-observation data set, with all the appropriate caveats.
Although this represents a fairly standard-to-large sample size for a meta-analysis in this context (the 45
studies in the analysis gather data from a total of 23,589 respondents), it is relatively small relative to other
statistical applications in resource and environmental economics. Model results are also subject to choices
regarding functional form and statistical approach, although many of the primary model effects are robust to
reasonable changes in functional form and/or statistical methods. The rationale for the specific functional
form chosen here (the semi-log form) is detailed above.
As in all cases, results of the meta-analysis are dependent on the sample of studies available for the given
resource change (Navrud and Ready 2007), and may be subject to various selection biases if the available
literature does not provide a representative, unbiased perspective on welfare estimates associated with
resource changes (Rosenberger and Johnston 2007). In this case, however, the Agency took various steps to
ameliorate such potential biases, including the incorporation of both peer-reviewed and gray literature to
avoid possible publication biases (Rosenberger and Johnston 2007), and the use of a comprehensive literature
review in the attempt to avoid—as much as possible—other types of selection biases.
The relatively large (positive) magnitude of the parameter estimate for the Pacific Mountain U.S. regional
dummy variable (mp) leads EPA to question the appropriate interpretation of this effect. While it is
theoretically possible that WTP for water quality changes is substantially higher in the Pacific Northwest
(e.g., people who live in this region are outdoor enthusiasts), the magnitude of the effect suggested by the
model seems unlikely from an intuitive perspective. As suggested above, it is possible that spurious,
unexplained factors influence the magnitude of this parameter in the present model. However, assessments of
preliminary model runs suggest that this effect is relatively robust given the present data and selection of
variables available. Nonetheless, EPA recommends that the magnitude of the predicted shift in WTP
associated with the Pacific Mountain region should be viewed with caution.
April 19, 2013 Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute E-19
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix F: Impacts of Pollutants on Aquatic Species
Appendix F: Impacts of Steam Electric Pollutants on Aquatic Species
Exhibit F-1: Common Coal-combustion Wastewater Pollutants (adapted from EPA, 2009a)
compound
Arsenic
BOD
Boron
Cadmium
Chlorides
Chromium
Copper
Iron
Lead
Manganese
Mercury
Nitrogen
pH
Phosphorus
Selenium
Potential environmental concern
Frequently observed in high concentrations in coal combustion wastewater; causes poisoning
of the liver in fish and developmental abnormalities; is associated with an increased risk of
cancer in humans in the liver and bladder.
Can cause fish kills because of a lack of available oxygen; increases the toxicity of other
pollutants, such as mercury. Has been associated with FGD wastewaters that use organic
acids for enhanced SO2 removal in the scrubber.
Frequently observed in high concentrations in coal combustion wastewater; leachate into
groundwater has exceeded state drinking water standards; human exposure to high
concentrations can cause nausea, vomiting, and diarrhea. Can be toxic to vegetation.
Elevated levels are characteristic of coal combustion wastewater-impacted systems;
organisms with elevated levels have exhibited tissue damage and organ abnormalities.
Sometimes observed at high concentrations in coal combustion wastewater (dependent on
FGD system practices); elevated levels observed in fish with liver and blood abnormalities.
Elevated levels have been observed in groundwater receiving coal combustion wastewater
leachate; invertebrates with elevated levels require more energy to support their metabolism
and therefore exhibit diminished growth.
Coal combustion wastewater can contain high levels; invertebrates with elevated levels
require more energy to support their metabolism and therefore exhibit diminished growth.
Leachate from impoundments has caused elevated concentrations in nearby surface water;
biota with elevated levels have exhibited sublethal effects including metabolic changes and
abnormalities of the liver and kidneys.
Concentrations in coal combustion wastewater are elevated initially, but lead settles out
quickly; leachate has caused groundwater to exceed state drinking water standards. Human
exposure to high concentrations of lead in drinking water can cause serious damage to the
brain, kidneys, nervous system, and red blood cells.
Coal combustion wastewater leachate has caused elevated concentrations in nearby
groundwater and surface water; biota with elevated levels have exhibited sublethal effects
including metabolic changes and abnormalities of the liver and kidneys.
Biota with elevated levels have exhibited sublethal effects including metabolic changes and
abnormalities of the liver and kidneys; can convert into methylmercury, increasing the
potential for bioaccumulation; human exposure at levels above the MCL for relatively short
periods of time can result in kidney damage.
Frequently observed at elevated levels in coal combustion wastewater; may cause
eutrophication of aquatic environments.
Acidic conditions are often observed in coal combustion wastewater; acidic conditions may
cause other coal combustion wastewater constituents to dissolve, increasing the fate and
transport potential of pollutants and increasing the potential for bioaccumulation in aquatic
organisms.
Frequently observed at elevated levels in coal combustion wastewater; may cause
eutrophication of aquatic environments.
Frequently observed at high concentrations in coal combustion wastewater; readily
bioaccumulates; elevated concentrations have caused fish kills and numerous sublethal
effects (e.g., increased metabolic rates, decreased growth rates, reproductive failure) to
aquatic and terrestrial organisms. Short term exposure at levels above the MCL can cause
hair and fingernail changes; damage to the peripheral nervous system; fatigue and irritability
in humans. Long term exposure can result in damage to the kidney, liver, and nervous and
circulatory systems.
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
F-1
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix F: Impacts of Pollutants on Aquatic Species
Exhibit F-1: Common Coal-combustion Wastewater Pollutants (adapted from EPA, 2009a)
Compound
Total Dissolved Solids
Potential Environmental Concern
High levels are frequently observed in coal combustion wastewater; elevated levels can be a
stress on aquatic organisms with potential toxic effects; elevated levels can have impacts on
agriculture & wetlands.
Zinc Frequently observed at elevated concentrations in coal combustion wastewater; biota with
elevated levels have exhibited sublethal effects such as requiring more energy to support
their metabolism and therefore exhibiting diminished growth, and abnormalities of the liver
and kidneys.
Exhibit F-2: T&E Species with Habitat Overlapping Waterbodies Affected by Steam Electric Plants
Species
Acipenser brevirostmm
Acipenser medirostris
Acipenser oxyrinchus desotoi
Acipenser oxyrinchus oxyrinchus
Alasmidonta heterodon
Alasmidonta raveneliana
Amblema neislerii
Amblyopsis rosae
Ambystoma bishopi
Ambystoma cingulatum
Ambystoma macrodactylum
Ambystoma tigrinum
Ammodramus savannamm floridanus
Anguispira picta
Antrobia culveri
Antrolana lira
Aphelocoma coerulescens
Athearnia anthonyi
Batrisodes texanus
Batrisodes venyivi
Boloria acrocnema
Brachylagus idahoensis
Brachyramphus marmoratus
Brychius hungerfordi
Bufo houstonensis
Cambarus aculabrum
Campeloma decampi
Campephilus principalis
Canis lupus
Canis rufus
Charadrius melodus
Cicindela dorsalis dorsalis
Cicindela nevadica lincolniana
Cicindela puritana
Cicurina baronia
Cicurina madia
Cicurina venii
Cicurina vespera
Species Group
Fishes
Fishes
Fishes
Fishes
Clams
Clams
Clams
Fishes
Amphibians
Amphibians
Amphibians
Amphibians
Birds
Snails
Snails
Crustaceans
Birds
Snails
Insects
Insects
Insects
Mammals
Birds
Insects
Amphibians
Crustaceans
Snails
Birds
Mammals
Mammals
Birds
Insects
Insects
Insects
Arachnids
Arachnids
Arachnids
Arachnids
Vulnerability
High
High
High
High
High
High
High
High
Moderate
Moderate
Moderate
Moderate
Low
Low
High
High
Low
High
Low
Low
Low
Low
Moderate
High
Moderate
High
High
Low
Low
Moderate
Moderate
Moderate
Moderate
Moderate
Low
Low
Low
Low
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
F-2
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix F: Impacts of Pollutants on Aquatic Species
Exhibit F-2: T&E Species with Habitat Overlapping Waterbodies Affected by Steam Electric Plants
Species
Corynorhinus (=Plecotus) townsendii ingens
Corynorhinus (=Plecotus) townsendii virginianus
Cryptobranchus alleganiensis
Cyprinella caerulea
Cyprogenia stegaria
Dendroica chrysoparia
Dendroica kirtlandii
Discus macclintocki
Dromus dramas
Drymarchon corais couperi
Elimia crenatella
Elliptic chipolaensis
Elliptic spinosa
Elliptic steinstansana
Elliptoideus sloatianus
Enhydra lutris nereis
Epioblasma brevidens
Epioblasma capsaeformis
Epioblasma florentina curtisii
Epioblasma florentina florentina
Epioblasma florentina walkeri
Epioblasma florentina walkeri (=E. walkeri)
Epioblasma metastriata
Epioblasma obliquata obliquata
Epioblasma obliquata perobliqua
Epioblasma othcaloogensis
Epioblasma torulosa rangiana
Epioblasma torulosa torulosa
Epioblasma turgidula
Etheostoma chermocki
Etheostoma chienense
Etheostoma etowahae
Etheostoma fonticola
Etheostoma nianguae
Etheostoma nuchale
Etheostoma phytophilum
Etheostoma scotti
Etheostoma sellare
Eubalaena glacialis
Eurycea nana
Fusconaia cor
Fusconaia cuneolus
Gambusia georgei
Gammarus acherondytes
Glaucomys sabrinus coloratus
Glaucomys sabrinus fuscus
Gopherus polyphemus
Graptemys flavimaculata
Graptemys oculifera
Species Group
Mammals
Mammals
Amphibians
Fishes
Clams
Birds
Birds
Snails
Clams
Reptiles
Snails
Clams
Clams
Clams
Clams
Mammals
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Fishes
Fishes
Fishes
Fishes
Fishes
Fishes
Fishes
Fishes
Fishes
Mammals
Amphibians
Clams
Clams
Fishes
Crustaceans
Mammals
Mammals
Reptiles
Reptiles
Reptiles
Vulnerability
Low
Low
High
High
High
Low
Low
Low
High
Low
High
High
High
High
High
Moderate
High
High
High
High
High
High
High
High
High
High
High
High
High
High
High
High
High
High
"High
"High
High
High
Low
High
High
High
High
Moderate
Low
Low
Low
High
High
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
F-3
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix F: Impacts of Pollutants on Aquatic Species
Exhibit F-2: T&E Species with Habitat Overlapping Waterbodies Affected by Steam Electric Plants
Species
Grus americana
Grus canadensis pulla
Hemistena lata
Heraclides aristodemus ponceanus
Herpmhtrus (=Felis) yagouaroundi cacomitli
Hesperia leonardus montana
Heterelmis comalensis
Juturnia kosteri
Lampsilis abrupta
Lampsilis altilis
Lampsilis higginsii
Lampsilis perovalis
Lampsilis subangulata
Lampsilis virescens
Lasmigona decorata
Leopardus (=Felis) pardalis
Leopardus (=Felis) wiedii
Leptodea leptodon
Leptonycteris nivalis
Leptoxis ampla
Leptoxis foremani
Leptoxis plicata
Leptoxis taeniata
Lepyrium showalteri
Lioplax cyclostomaformis
Lycaeides melissa samuelis
Lynx canadensis
Margaritifera hembeli
Medionidus acutissimus
Medionidus parvulus
Medionidus penicillatus
Medionidus simpsonianus
Mesodon clarki nantahala
Mesodon magazinensis
Microhexura montivaga
Microtus pennsylvanicus dukecampbelli
Mustela nigripes
Mycteria americana
Myotis grisescens
Neoleptoneta microps
Neonympha mitchellii francisci
Neonympha mitchellii mitchellii
Neotoma floridana smalli
Nicrophorus americanus
Notropis cahabae
Noturus crypticus
Noturus placidus
Obovaria retusa
Odocoileus virginianus clavium
Species Group
Birds
Birds
Clams
Insects
Mammals
Insects
Insects
Snails
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Mammals
Mammals
Clams
Mammals
Snails
Snails
Snails
Snails
Snails
Snails
Insects
Mammals
Clams
Clams
Clams
Clams
Clams
Snails
Snails
Arachnids
Mammals
Mammals
Birds
Mammals
Arachnids
Insects
Insects
Mammals
Insects
Fishes
Fishes
Fishes
Clams
Mammals
Vulnerability
Moderate
Moderate
High
Low
Low
Low
High
Low
High
High
High
High
High
High
High
Low
Low
High
Low
High
High
High
High
High
High
Low
Low
High
High
High
High
High
Low
Low
Low
Moderate
Low
Moderate
Moderate
Low
Low
Low
Low
Low
High
High
High
High
Moderate
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
F-4
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix F: Impacts of Pollutants on Aquatic Species
Exhibit F-2: T&E Species with Habitat Overlapping Waterbodies Affected by Steam Electric Plants
Species
Odocoileus virginianus leucums
Oncorhynchus clarkii stomias
Oncorhynchus clarkii stomias
Oncorhynchus kisutch
Orcimrs~orca
Orthalicus reses (not incl. nesodryas)
Palaemonias alabamae
Palaemonias ganteri
Panthera onca
Pegiasfabula
Percina antesella
Percina aurolineata
Percina rex
Percina tanasi
Peromyscus gossypinus allapaticola
Peromyscus polionotus ammobates
Peromyscus polionotus niveiventris
Peromyscus polionotus phasma
Phoebastria (=Diomedea) albatrus
Phoxinus cumberlandensis
Picoides borealis
Plethobasus cicatricosus
Plethobasus cooperianus
Plethodon nettingi
Pleurobema clava
Pleurobema collina
Pleurobema curium
Pleurobema decisum
Pleurobema furvum
Pleurobema georgianum
Pleurobema hanleyianum
Pleurobema marshalli
Pleurobema perovatum
Pleurobema plenum
Pleurobema pyriforme
Pleurobema taitianum
Pleurocera foremani
Polyborus plancus audubonii
Polygyriscus virginianus
Potamilus capax
Potamilus inflatus
Pseudemys alabamensis
Ptychobranchus greenii
Ptychocheilus lucius
Puma (=Felis) concolor coryi
Pyrgulopsis (=Marstonia) pachyta
Pyrgulopsis neomexicana
Pyrgulopsis ogmorhaphe
Pyrgulopsis roswellensis
Species Group
Mammals
Clams
Fishes
Fishes
Mammals
Snails
Crustaceans
Crustaceans
Mammals
Clams
Fishes
Fishes
Fishes
Fishes
Mammals
Mammals
Mammals
Mammals
Birds
Fishes
Birds
Clams
Clams
Amphibians
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Clams
Snails
Birds
Snails
Clams
Clams
Reptiles
Clams
Fishes
Mammals
Snails
Snails
Snails
Snails
Vulnerability
Moderate
High
High
High
Low
Low
Moderate
Moderate
Low
High
High
High
High
High
Low
Low
Low
Low
Low
High
Low
High
High
Low
High
High
High
High
High
High
High
High
High
High
"High
"High
High
Low
Low
High
High
High
High
High
Low
High
High
High
Low
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
F-5
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix F: Impacts of Pollutants on Aquatic Species
Exhibit F-2: T&E Species with Habitat Overlapping Waterbodies Affected by Steam Electric Plants
Species
Quadmla fragosa
Quadrula intermedia
Quadmla sparsa
Quadrula stapes
Rangifer tarandus caribou
Rhadine exilis
Rhadine infernalis
Rhadine persephone
Rhinichthys osculus thermalis
Salvelinus confluentus
Scaphirhynchus albus
Scaphirhynchus suttkusi
Sciurus niger cinereus
Somatochlora hineana
Sterna antillarum
Sternotherus depressus
Strix occidentalis caurina
Strix occidentalis lucida
Stygobromus (=Stygonectes) pecki
Stygoparnus comalensis
Succinea chittenangoensis
Sylvilagus palustris hefneri
Tartarocreagris texana
Texamaurops reddelli
Texella cokendolpheri
Texella reddelli
Texella reyesi
Toxolasma cylindrellus
Trichechus manatus
Triodopsis platysayoides
Tryonia alamosae
Tulotoma magnifica
Tympanuchus cupido attwateri
Typhlomolge rathbuni
Ursus americanus luteolus
Vermivora bachmanii
Villosa trabalis
Xyrauchen texanus
Zapus hudsonius preblei
Species Group
Clams
Clams
Clams
Clams
Mammals
Insects
Insects
Insects
Clams
Fishes
Fishes
Fishes
Mammals
Insects
Birds
Reptiles
Birds
Birds
Crustaceans
Insects
Snails
Mammals
Arachnids
Insects
Arachnids
Arachnids
Arachnids
Clams
Mammals
Snails
Snails
Snails
Birds
Amphibians
Mammals
Birds
Clams
Fishes
Mammals
Vulnerability
High
High
High
High
Low
Low
Low
Low
High
High
High
High
Low
High
High
High
Low
Low
High
Low
Low
High
Low
Low
Low
Low
Low
High
Low
Low
High
High
Low
High
Low
Moderate
High
High
Low
April 19, 2013
Internal Draft - Deliberative, Predecisional - Do not Quote, Cite, or Distribute
F-6
-------
Benefit and Cost Analysis for Proposed ELGs Appendix G: Sensitivity Analysis - Impoundment Failure Probability
Appendix G: Sensitivity of the Estimated Benefits from Avoided
Impoundment Failures to Failure Probability
To estimate the value of avoided impoundment failures, EPA estimated an overall average failure rate based
on data from a survey of impoundments conducted in support of regulations being developed by EPA's Office
of Resource Conservation and Recovery (ORCR; see U.S. EPA, 2010d). This analysis, presented in Chapter
7, applies a uniform probability of failure of 0.58 percent to all impoundments over the entire analysis period.
In practice, the probability of impoundment failure may depend on impoundment characteristics. To evaluate
the sensitivity of the impoundment failure benefits to differing assumptions, EPA used ORCR's survey data
to develop a statistical model of the probability of impoundment failure as a function of the impoundment
capacity.88 Equation G-l shows the estimated logistic regression model.
Equation G-l.
Where:
Where:
Probability of Failure(i) =
zit = ao + /?i x Incapacity j) + u;
Probability ofFailure(i)= the probability of impoundment /' failing
capacity\ = the capacity of impoundment /
ut = unobserved effect component specific to the universe of impoundments analyzed (varies with /')
Table G-l presents the specifications of the logistic regression model that best fit the survey data.
Table G-1. Logistic Regression Model
Parameter
a0
Pi
Coefficient
-7.0244
0.2904
Standard Error
0.7776
0.1212
P>|z|
0.017
0.000
95% Confidence Interval
0.0529
-8.5485
0.5279
-5.5002
Prob>Chi2 = 0.0166
Applying this model to the universe of 1,070 impoundments at steam electric plants subject to the ELGs and
data on impoundment capacities under the baseline and each regulatory scenario,89 EPA conducted this
analysis for five regulatory options: Options 1, 2, 3, 4, and 5. EPA estimated impoundment-specific failure
rates for the years 2019 through 2040. Under all scenarios (baseline and each analyzed regulatory option), the
probability of failure was assumed to be zero for years 2014 and 2018 due to integrity site assessments
conducted by EPA in 2009 through 2012, which are expected to prevent all failures for the first five years
after the recommended "action plan" to improve impoundment structures is completed (2014 through 2018).
While EPA tested other model specifications that included impoundment age as an explanatory variable, this variable
was not statistically significant in the best-fit model. Additionally, although other factors - such as the involvement of a
professional engineer, site-specific structural integrity inspections, statewide integrity inspection requirements,
impoundment height, and other factors - affect the probability of failure, EPA was unable to include these variables in
the model due to data limitations.
89 EPA used data on impoundment capacities from its 2010 Questionnaire for the Steam Electric Power Generating
Effluent Guidelines (the industry survey) to determine the baseline impoundment capacity and estimated changes in
capacity based on estimated reductions in flows of wastewater managed by an impoundment under each of the regulatory
options.
April 19, 2013
G-1
-------
Benefit and Cost Analysis for Proposed ELGs Appendix G: Sensitivity Analysis - Impoundment Failure Probability
The average annual probability of failure is 0.53 percent in the baseline, across all 1,070 impoundments at
steam electric plants (the probability of failure ranges between 0.04 percent and 2.75 percent, depending on
impoundment capacity). Changes due to the proposed ELGs reduce the average annual probability of failure
to 0.46 percent under Options 1 and 2, 0.42 percent under Option 3, and 0.19 percent under Options 4 and 5.
Table G-2 shows the results of this sensitivity analysis (in lower part of the table) as compared to results using
the uniform average probability of failure of 0.58 percent (in upper part of the table). As shown in the table,
the alternative approach to calculating the probability of failure for each impoundment results in higher
benefits than using the uniform average failure rate. For example, EPA estimates that Option 3 generates
annual avoided impoundment failure benefits valued at $367.9million, using a 3 percent discount rate ($306.6
million using a 7 percent discount rate), which is about 3.2 times the benefits estimated using a uniform
average failure rate.
Table G-2. Annual Benefits of Avoided Surface Impoundment Failures (millions; 2010 $)
Regulatory Option
3% Discount Rate
7% Discount Rate
Uniform Average Failure Rate (0.58 percent for all impoundments)
Option 1
Option 2
Option 3
Option 4
Option 5
$62.1
$62.1
$114.8
$295.1
$295.1
$52.2
$52.2
$95.9
$245.9
$245.9
Probability based pn Regression Results (Impoundment- and option-speciffc rate)
Option 1
Option 2
Option 3
Option 4
Option 5
$168.7
$168.7
$367.9
$918.0
$918.0
$142.2
$142.2
$306.6
$764.4
$764.4
a. Baseline value of total failure costs minus option value of total failure costs
Note that this analysis does not account for the effect of best management practices (BMPs) - including
integrity inspections and preventive maintenance - that are expected to further reduce the probability of
impoundment failures under the proposed ELGs. As discussed in Chapter 7, EPA does not have sufficient
information to accurately quantify and monetize the benefits of implementing BMPs. Preventing all future
impoundment failures would provide annual benefits estimated at up to $1,108.2 million (using 3 percent
discount and applying the regression equation described in this appendix).
EPA does not anticipate BMPs to fully realize such benefits, however. Inspections and other BMPs aim to
prevent future failures by identifying conditions that have contributed to past impoundment failures and
releases (e.g., slope instability, structural defects, seepage, overtopping, and inadequate management
practices; see NRC, 2002); they are not expected to be as effective at preventing impoundment failures caused
by unusual weather events or earthquakes. Further, data from the Steam Electric Industry Survey and 2010
CCR Impoundment Survey suggest that steam electric plants already implement inspections and monitoring
programs of varying scope and frequency (U.S. EPA, 2010e; U.S. EPA, 2012d), but field assessments
conducted subsequent to the 2010 CCR Impoundment Survey indicate that the existing inspection programs at
some plants failed to identify embankment erosion, seepage, and other conditions (U.S. EPA; 2012e).
April 19, 2013
G-2
-------
Benefit and Cost Analysis for Proposed ELGs
Appendix H: C-R Function for Analysis of Air-related Benefits
Appendix H: Concentration Response Function Used in the Analysis of
Air-Related Benefits
Most of the concentration response (C-R) functions relating criteria air pollutants to population incidence of
an adverse health effect are log-linear (or exponential) in form:
Equation H-l.
= a*e
ftc
where x is the ambient air pollutant concentration (PM25 or O3 in this analysis), y is the incidence of the
adverse health effect corresponding to x, ft is the coefficient of ambient concentration of the air pollutant
(describing the extent of change in y with a unit change in x), and a is the incidence of the adverse health
effect when there is no ambient air pollutant. Each epidemiological study provides b (an estimate of ft).
Let x0 denote the baseline (upper) level of the ambient air pollutant and x} denote the "control scenario"
(lower) level. In addition, let y0 denote the baseline incidence of the health effect (corresponding to the
baseline ambient pollutant level, x0) and j; denote the incidence after the rule is implemented, corresponding
to ambient pollutant level, x}. Equation H-l and the estimate, b, can be used to derive the following estimated
relationship between the absolute reduction in ambient air pollutant level, Ax = (x0 - Xj), and the
corresponding reduction in health effect incidence, Ay.
Equation H-2. ^v = (vn - V,) = V.
For this analysis, EPA used the values shown in Table H-l.
Table H-1: Summary of Studies and Concentration-Response Functions Used to Estimate
PM2.5-Related Benefits
Health Endpoint
Mortality, All
Causes
Study
Krewski et al.
(2009)
Location
116 U.S. cities
Age Range
30+
PM25
Coefficient
(Beta)
0.00583
Standard Error
0.00096
April 19, 2013
H-1
-------
Benefit and Cost Analysis for Proposed ELGs
Appendixl: Correlate d Benefits using Alternate SCC Values
Appendix I: CO2-Related Benefits using Alternate Social Cost of Carbon
Values
As discussed in Chapter 8, the Interagency Working Group on Social Cost of Carbon (IWGSCC, 2010)
developed social cost of carbon (SCC) values for three discount rates: 2.5 percent (average), 3 percent
(average and 95th percentile), and 5 percent (average).
In this appendix, we present the annualized benefits from reduced CO2 emissions using alternate SCC values.
Note that EPA used the SCC for the 3 percent discount rate (average) and 5 percent discount rate (average) to
estimate benefits presented in Chapter 8.
Table 1-1. Annualized Benefits from Reduced CO2 Emissions (Millions; 2010$)
Regulatory Option
Option 3
Option 4
5% Discount Rate,
Average
$8.8
$24.7
3% Discount Rate,
Average
$33.6
$94.6
2.5% Discount Rate,
Average
$52.7
$148.7
3% Discount Rate,
95th percentile
$102.5
$288.8
April 19, 2013
1-1
------- |