United States
Environmental Protection
Agency
Office Of Air Quality
Planning And Standards
Research Triangle Park, NC 27711
EPA-452/R-02-012
November 2002
FINAL REPORT
Air
Regulatory Impact Analysis of the
Proposed Reciprocating Internal
Combustion Engines NESHAP
Final Report
-------
-------
Regulatory Impact Analysis
of the Proposed
Reciprocating Internal Combustion Engines
NESHAP
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Air Quality Planning and Standards
Air Quality Strategies and Standards Division (C339-01);
Research Triangle Park, N.C. 27711
Final Report
November 2002
-------
Disclaimer
This report is issued by the Air Quality Standards & Strategies
Division of the Office of Air Quality Planning and Standards of
the U.S. Environmental Protection Agency (EPA). It presents
technical data on the National Emission Standard for Hazardous
Air Pollutants (NESHAP)for Reciprocating Internal Combustion
Engines, which is of interest to a limited number of readers. It
should be read in conjunction with the Background Information
Document (BID) for the NESHAP and other background material used
to develop the rule, which are located in the public docket for
the NESHAP proposed rulemaking. Copies of these reports and
other material supporting the rule are in Docket A-95-35 at EPA's
Air and Radiation Docket and Information Center, 1200
Pennsylvania Avenue, N.W.; Washington D.C. 20460. The EPA may
charge a reasonable fee for copying. Copies are also available
through the National Technical Information Services, 5285 Port
Royal Road, Springfield, VA 22161. Federal employees, current
contractors and grantees, and nonprofit organizations may obtain
copies from the Library Services Office (MD-35), U.S.
Environmental Protection Agency, Research Triangle Park, N.C.
27711; phone (919) 541-2777.
IV
-------
CONTENTS
EXECUTIVE SUMMARY ES-1
LIST OF FIGURES viii
LIST OF TABLES ix
Section Page
1.0 INTRODUCTION 1-1
1.1 Purpose 1-1
1.2 Legal History and Statutory Authority 1-2
1.3 Report Organization 1-3
2.0 NEED FOR REGULATION 2-1
2.1 Environmental Factors Which Necessitate Regulation 2-1
2.1.1 Air Emission Characterization 2-2
2.1.2 Harmful Effects of HAPs 2-5
2.2 Market Failure 2-8
2.3 Insufficient Political and Judicial Forces 2-10
2.4 Consequences of Regulatory Action 2-11
2.4.1 Consequences if EPA's Emission Reduction
Objectives Are Met 2-11
2.4.1.1 Regulatory Alternatives Considered 2-14
2.4.1.2 Alternative Regulatory Options Based on Risk ....2-14
2.4.1.3 Allocation of Resources 2-26
2.4.1.4 Emissions Reductions and Cost Impacts 2.26
2.4.1.5 Energy Impacts 2-29
2.4.1.6 State Regulation and New Source Review 2-29
2.4.1.7 Other Federal Programs 2-32
2.4.2 Consequences if EPA's Emission Reduction Objectives
Are Not Met 2-32
in
-------
3.0 PROFILE OF RICE UNITS AND TECHNOLOGIES 3-1
3.1 Engines Technologies 3-2
3.1.1 SI Two-Stroke Engines 3-3
3.1.2 SI Four-Stroke Engines 3-3
3.1.3 Compression Ignition Units 3-4
3.2 Emissions 3-5
3.3 Control Costs 3-5
3.4 Profile of RICE Units and Facilities in Inventory Database 3-15
3.4.1 Affected Units 3-15
3.4.2 Affected Facilities 3-16
3.5 Projected Growth of RICE 3-16
3.5.1 Growth Estimates by Industry 3-20
3.5.1.1 Mapping SIC Codes to NAICS Codes 3-20
3.5.1.2 Data Extrapolation to Proj ected National Unit
Estimates by Industry 3-21
3.5.2 Engineering Compliance Costs 3-25
3.5.2.1 Sample Industry Cost Calculation.
NAICS 211 3-25
3.5.2.2 National Engineering Compliance Costs 3-25
4.0 PROFILES OF AFFECTED INDUSTRIES 4-1
4.1 Crude Petroleum and Natural Gas (NAICS 211) 4-1
4.1.1 Introduction 4-2
4.1.2 Supply Side Characteristics 4-5
4.1.2.1 Production Processes 4-5
4.1.2.2 Types of Output 4-8
4.1.2.3 Major By-Products 4-10
4.1.2.4 Costs of Production 4-11
4.1.2.5 Imports and Domestic Capacity Utilization 4-11
4.1.3 Demand Side Characteristics 4-13
4.1.4 Organization of the Industry 4-15
4.1.5 Markets and Trends 4-19
4.2 Natural Gas Pipeline Industry 4-20
4.2.1 Introduction 4-20
4.2.2 Supply Side Characteristics 4-21
4.2.2.1 Service Description 4-22
4.2.2.2 Major By-Products 4-23
4.2.2.3 Costs of Production 4-23
4.2.2.4 Capacity Utilization 4-25
4.2.2.5 Imports 4-25
4.2.3 Demand Side Characteristics 4-26
4.2.4 Organization of the Industry 4-27
iv
-------
4.2.5 Market Trends 4-27
5.0 ECONOMIC IMPACT ANALYSIS 5-1
5.1 Economic Impact Methodology 5-2
5.1.1 Background on Economic Modeling Approaches 5-2
5.1.1.1 Modeling Dimension 1: Scope of Economic
Decisionmaking 5-2
5.1.1.2 Modeling Dimension 2: Interaction
Between Economic Sectors 5-4
5.1.2 Selected Modeling Approach for RICE Analysis 5-5
5.1.3 Directly Affected Markets 5-9
5.1.3.1 Market for Natural Gas 5-9
5.1.3.2 Market for Petroleum Products 5-10
5.1.3.3 Final Product and Service Markets 5-11
5.1.4 Indirectly Affected Markets 5-13
5.1.4.1 Market for Electricity 5-14
5.1.4.2 Market for Coal 5-14
5.1.4.3 Final Product and Service Markets 5-15
5.1.4.4 Impact on Residential Sector 5-15
5.1.4.5 Impact on Transportation Sector 5-15
5.2 Operationalizing the Economic Impact Model 5-16
5.2.1 Computer Model 5-18
5.2.2 Calculating Changes in Social Welfare 5-19
5.2.3 Supply and Demand Elasticities Used in the
Market Model 5-22
5.3 Economic Impact Estimates 5-24
5.4 Conclusions 5-31
6.0 IMPACTS ON FIRMS OWNING RICE UNITS 6-1
6.1 Identifying Small Businesses 6-2
6.2 Screening-Level Analysis 6-4
6.3 Analysis of Facility-Lev el and Parent-Level Data 6-4
6.4 Small Business Impacts 6-9
6.5 Assessment of Small Entity Screening 6-11
7.0 QUALITATIVE ASSESSMENT OF BENEFITS OF
EMISSION REDUCTIONS 7-1
7.1 Identification of Potential Benefit Categories 7-1
-------
7.2 Qualitative Description of Air Related Benefits 7-2
7.2.1 Benefits of Reducing HAP Emissions 7-2
7.2.1.1 Health Benefits of Reduction in HAP Emissions .... 7-3
7.2.1.2 Welfare Benefits of Reduction in HAP Emissions . . . 7-6
7.2.2 Benefits of Reducing Other Pollutants Due to HAP Controls . . 7-7
7.2.2.1 Benefits of Reduction in Carbon Monoxide
Emissions 7-7
7.2.2.2 Benefits of Reduced Nitrous Oxide Emissions 7-9
7.3 Lack of Approved Methods to Quantify Map Benefits 7-11
7.3.1 Evaluation of Alternative Regulatory Options Based
on Risk 7-12
7.4 Summary 7-12
.0 QUANTIFIED BENEFITS 8-1
8.1 Results in Brief 8-1
8.2 Introduction 8-2
8.3 Overview of Benefits Analysis Methodology 8-3
8.3.1 Methods for Estimating Benefits from Air Quality
Improvements 8-7
8.3.2 Quantifying Individual Health Effect Endpoints 8-10
8.3.2.1 Concentration-Response Functions for
Premature Mortality 8-14
8.3.3 Valuing Individual Health Effect Endpoints 8-21
8.3.3.1 Valuation of Reductions in Premature
Mortality Risk 8-24
8.3.3.2 Valuation of Reduction in Chronic Bronchitis 8-31
8.3.4 Methods for Describing Uncertainly 8-32
8.4 Derivation of Benefit Transfer Values for the RICE NESHAP 8-34
8.4.1 Ozone Benefit Transfer Values for Application to NOx
Emission Reductions 8-34
8.4.2 PM25 Benefit Transfer Values for Application to NOx
Emission Reductions 8-39
8.4.2 PM10 Benefit Transfer Values for Application to PM10
Emission Reductions 8-44
8.5 Application of Benefits Transfer Values to the RICE
NESHAP Rule 8-48
8.6 Limitation of the Analysis 8-49
8.6.1 Uncertainties and Assumptions 8-49
8.6.2 Unquantified Effects 8-51
vi
-------
8.7 Cost-Benefit Comparison 8-53
REFERENCES R-l
Appendix A: Economic Model of Markets Affected by the RICE MACT A-l
A.I Energy Markets Model A-l
A.I.I Supply Side Modeling A-2
A. 1.2 Demand Side Modeling A-2
A.2 Industrial and Commercial Markets A-3
A.2.1 Compute Percentage Change in Market Price A-4
A.2.2 Compute Percentage Change in Market Quantity A-4
A.3 With-Regulation market Equilibrium Determination A-4
A.4 Computing Social Costs A-5
A.4.1 Changes in Consumer Surplus A-5
A.4.2 Changes in Producer Surplus A-6
Appendix B: Economic Model Sensitivity Analyses B-l
B. 1 Energy Markets Model B-l
B.2 Final Product Market Elasticities B-2
B.3 Own and Cross Price Elasticities for Fuels B-2
B.4 Share of NAICS 211 Associated with Natural Gas and Petroleum . . . . B-3
Appendix C: Assumptions and Limitations of the Economic Model C-l
Appendix D: Summary of Studies of the Effects of Emissions of Hazardous Air Pollutants . D-l
vn
-------
LIST OF FIGURES
Number
3-1 Capacity Ranges for Engines in the Inventory Database 3-15
3-2 Characteristics of Engines in Inventory Database 3-16
5-1 Links Between Energy and Final Product Markets 5-7
5-2 Market Effects of Regulation-Induced Costs 5-10
5-3 Fuel Market Interactions with Facility-Level Production Decisions 5-13
5-4 Operationalizing the Estimation of Economic Impact 5-17
5-5 Changes in Economic Welfare with Regulation 5-21
6-1 Parent Size by Employment Range 6-6
6-2 Number of Parents by Sales Range 6-7
Vlll
-------
LIST OF TABLES
Number
2-1 HAP Emission Factors by Engine Configuration (Ibs/hour) 2-3
2-2 National Baseline HAP Emissions From RICE Units, 2005 2-6
2-3 Potential Health and Welfare Effects Associated with Exposure to
Hazardous Air Pollutants 2-7
2-4 Population of Existing RICE 2-12
2-5 Forecasted Population of New RICE 2-13
2-6 Summary of Regulatory Alternatives for RICE Subcategories 2-15
2-7 Dose-Response Assessment Values for HAP Reported Emitted by the
RICE Source Category 2-20
2-8 Summary of HAP Emission Reductions and Cost Impacts Associated
with the RICE NESHAP, 2005 2-28
3-1 HAP Emissions Factors by Engine Configuration (Ibs/hours) 3-5
3-2 Control Costs Associated with Model Engines 3-8
3-3 Control Costs Associated with Existing and New RICE 3-10
3-4 Costs of Monitoring for RICE Subcategories 3-11
3-5 Monitoring Option Applied to RICE Model Engine Categories 3-12
3-6 Total Annualized Control Cost for Affected Units 3-13
3-7 Cost Effectiveness for Each Model Engine Category 3-14
3-8 Number of Units with Assigned Model Numbers, the Number of Facilities
at Which They are Located, and the Average Number of Units per
Facility, by Industry in the Inventory Database 3-17
3-9 Population Estimates of Affected RICE Units, 2005 3-23
3-10 Affected RICE Population and Engineering Costs by
NAICS Code, 2005 3-24
3-11 Sample Cost Calculation: Estimating Compliance Costs for NAICS 211 ... 3-27
4-1 Crude Petroleum and Natural Gas Industries Likely to Be Affected by the
Regulation 4-3
4-2 Summary Statistics, Crude Oil and Natural Gas Extraction and Related
Industries 4-4
4-3 U.S. Supply of Crude Oil and Petroleum Products (10 barrels), 1998 4-9
4-4 U.S. Natural Gas Production, 1998 4-10
4-5 Costs of Production, Crude Oil and Natural Gas Extraction and Related
Industries 4-12
4-6 Estimated U.S. Oil and Gas Reserves, Annual Production, and
Imports, 1998 4-13
4-7 Size of Establishments and Value of Shipments, Crude Oil and Natural Gas
Extraction Industry (NAICS 211111), 1997 and 1992 4-16
4-8 Size of Establishments and Value of Shipments, Natural Gas Liquid
Extraction Industry (NAICS 211112), 1997 and 1992 4-17
ix
-------
4-9 Size of Establishments and Value of Shipments, Drilling Oil and Gas Wells
Industry, 1997 and 1992 4-18
4-10 Size of Establishments and Value of Shipments, Oil and Gas Support
Activities for Operations (NAICS 213112), 1997 4-19
4-11 Summary Statistics for the Natural Gas Pipeline Industry
(NAICS 4862), 1997 4-21
4-12 Summary Profile of Completed and Proposed Natural Gas Pipeline
Projects, 1996 to 2000 4-24
4-13 Energy Usage and Cost of Fuel, 1994-1998 4-25
4-14 Transmission Pipeline Capacity, Average Daily Flows, and Usage Rates,
1990 and 1997 4-26
4-15 Five Largest Natural Gas Pipeline Companies by System Mileage, 2000 . . . 4-28
5-1 Comparison of Modeling Approaches 5-3
5-2 Fuel Price Elasticities 5-19
5-3 Supply and Demand Elasticities 5-23
5-4 Supply and Demand Elasticities for Industrial and Commerical Sectors .... 5-25
5-5 Summary Table 5-26
5-6 Distribution of Social Cost 5-27
5-7 Market-Level Impacts 5-30
5-8 Impacts on the Number of New Engines Installed 5-33
6-1 Unit Counts and Percentages by Industry 6-3
6-2 Facility-Level and Parent-Level Data 6-5
6-3 Small Parent Companies 6-8
6-4 Summary Statistics for SBREFA Screening Analysis: Existing
Affected Small Entities 6-10
6-5 Profit Margins for Industry Sectors with Affected Small Businesses 6-12
7-1 Potential Health and Welfare Effects Associated with Exposure to
Hazardous Air Pollutants 7-4
8-1 Summary of Results: The Estimated PM and Ozone-Related Benefits
of the RICE NESHAP 8-2
8-2 Health Outcomes and Studies Included in the Analysis 8-12
8-3 Unit Values Used for Economic Valuation of Health Endpoints 8-23
8-4 Primary Sources of Uncertainty in the Benefit Analysis 8-35
8-5 Ozone $/ton Transfer Values for NOx Reductions Using Estimates from
the NOx SIP Call 8-39
8-6a Base Estimate of Annual Health Benefits Resulting From 50 Percent
RICE NOx Emission Reduction Scenario 8-42
8-6b Alternative Estimate of Annual Health Benefits Resulting From 50 Percent
RICE NOx Emission Reduction Scenario 8-43
8-7 Benefit Value Per Ton of NOx—Based on a 50% NOx Reduction at
RICE Units 8-44
-------
8-8a Base Estimate: Annual Health Benefits Resulting from 100
Percent RICE Direct PM Emission Reduction Scenario 8-46
8-8b Alternative Estimate: Annual Health Benefits Resulting from 100
Percent RICE Direct PM Emission Reduction Scenario 8-47
8-9 Benefit Value Per Ton of PM—Based on a 100% Reduction of PM10
at RICE Units 8-48
8-10 Benefits of the RICE NESHAP 8-49
8-11 Significant Uncertainties and Biases in Derivation of the Benefit
Transfer Values 8-50
8-12 Significant Uncertainties and Biases in Application of Benefit Transfer
Values to RICE NOx and PM Reductions 8-51
8-13 Unquantified Benefit Categories 8-52
8-14 Summary of Costs, Emission Reductions, and Quantifiable Benefits
by Engine Type 8-57
8-15 Annual Net Benefit of the RICE NESHAP in 2005 8-58
B-l Sensitivity Analysis: Elasticity of Supply in the Electricity Market B-2
B-2 Sensitivity Analysis: Supply and Demand Elasticities in the Industrial and
Commercial Markets ($106) B-3
B-3 Sensitivity Analysis: Own- and Cross- Price Elasticities Used to Model Fuel
Switching ($106) B-3
B-4 Sensitivity Analysis: Distribution of Affected Units in NAICS 211 Between
the Natural Gas and Petroleum Industries ($106) B-4
XI
-------
ACRONYMS AND ABBREVIATIONS
2SLB Two-Stroke Lean Burn
4SLB Four-Stroke Lean Burn
4SRB Four-Stroke Rich Burn
AIRS Airometric Information Retrieval System
A/F Air to fuel ratio
ASM Annual Survey of Manufacturers
BACT Best Available Control Technology
BCA Benefit Cost Analysis
CAA Clean Air Act Amendments of 1990
CD Criteria Document
C/E Cost Effectiveness
CH2O Formaldehyde
CI Compression Ignition
CNS Central Nervous System
CO Carbon monoxide
CRDM Climatological Regional Dispersion Model
CSR Cost to Sales Ratio
DOC Department of Commerce
EIA Department of Energy/Energy Information Administration
EPA Environmental Protection Agency
FACA Federal Advisory Committee Act
HAP Hazardous Air Pollutant
HEM Human Exposure Model
IARC International Agency for Research on Cancer
1C Internal Combustion
LAER Lowest Achievable Emission Rate
LEG Low Emission Combustion
LDCs Local Distribution Companies
LNG Liquefied Natural Gas
MACT Maximum Achievable Control Technology
MIR Maximum Individual Risk
mmBTU Million British Thermal Units
mmcf/d Million cubic feet per day
MRR monitoring, recordkeeping, and reporting
Mg Megagram
NAAQS National Ambient Air Quality Standard
NAICS North American Industry Classification System
NEMS National Energy Modeling System
NESHAP National Emission Standard for Hazardous Air Pollutants
NGLs Natural Gas Liquids
NSCR Non-Selective Catalytic Reduction
xn
-------
NSPS New Source Performance Standard
NOx Nitrogen Oxide
OMB Office of Management and Budget
OPEC Organization of the Petroleum Economic Community
OTAG Ozone Transport Assessment Group
PM Particulate Matter
ppbvd parts per billion by volume (dry)
ppmvd parts per million by volume (dry)
PSD Prevention of Significant Deterioration
RACT Reasonably Available Control Technology
RFA Regulatory Flexibility Act; also Regulatory Flexibility Analysis
RfC Reference-dose Concentration
RIA Regulatory Impact Analysis
RICE Reciprocating Internal Combustion Engine
SBREFA Small Business Regulatory Enforcement Fairness Act
SCR Selective Catalytic Reduction
SI Spark Ignition
SIC Standard Industrial Classification
SO2 Sulfur Dioxide
S-R Matrix Source Receptor Matrix
UAM-V Urban Airshed Model - Version V
URF Unit Risk Factor
U.S. United States
VOC Volatile Organic Compound
VSL Value of a Statistical Life
VSLY Value of a Statistical Life-Year
Xlll
-------
EXECUTIVE SUMMARY
This report summarizes the benefits and costs associated with the National Emissions
Standard for Hazardous Air Pollutants (NESHAP) for the Reciprocating Internal Combustion
Engines (RICE) source category. This source category includes spark ignition engines that
operate generally with natural gas and compression ignition engines that operate with diesel fuel,
and can be classified as two-stroke, or four-stroke engines. They are also classified by the
richness of the fuel mix: rich burn or lean burn. The affected RICE units operate in a variety of
markets and service industries. For instance, some are typically used along natural gas pipelines
to provide adequate pressure to transmit fuel through the pipeline. Others are also used to
provide power in a remote area of an operation in industries such as health services, energy
generation, oil and gas extraction, and quarrying of non-metallic minerals.
The proposed NESHAP for RICE will impact existing and new sources of RICE units
and is expected to reduce HAP emissions by 5,000 tons per year by the year 2005 due to controls
required to achieve the MACT floor—the minimum level of control mandated by the Clean Air
Act. The controls applied to RICE units will also achieve annual reductions of criteria
pollutants, including: 234,400 tons of carbon monoxide (CO) per year by 2005, and 167,900
tons of nitrogen oxides (NOx) per year, and 3,700 tons of particulate matter (PM10).
The total social cost of these HAP reductions is $255 million (1998$) in the 5th year after
implementation. This cost is spread across more than 25 different manufacturing and service
industries, which results in minimal changes in prices and production levels in most affected
industries. However, because natural gas engines are a large portion of the controlled units, the
natural gas market (including fuel usage for energy generation, as well as the extraction,
processing, and transmission industries for natural gas) has a larger share of the regulatory
ES-1
-------
burden associated with this rule. Natural gas prices are expected to rise by about 0.3 percent,
which is greater than for other affected industries, but which is considered a modest change in
comparison to historical price changes. Prices in other energy generation markets, such as oil,
coal and electricity do not change substantially, although a modest amount of fuel switching
from natural gas to electricity or coal is anticipated.
A screening of the impacts on firms owning RICE units was conducted for firms who
own existing RICE units. In our database of approximately 26,800 existing engines, we
determined that about 3,300 units could be affected by the existing source MACT. We were able
to identify the ownership of 889 of these engines. Using the subset of 889 units, we determined
these engines operate at 385 facilities owned by 84 parent firms. Of these firms, 13 were defined
as small entities. None of these small firms are anticipated to have compliance costs associated
with the existing source MACT that exceed three percent of firm revenues and only two small
firms have impacts between one and three percent. The average profit margin in the primary
affected industries is approximately five percent. Given that none of the small entities evaluated
in our subset have impacts that exceed the five percent profit margin, and only 16 percent may
have impacts greater than one percent of total revenues, we conclude that this proposed action
will not have a significant impact on a substantial number of small entities.
For new sources, it can be reasonably assumed that the investment decision to purchase a
new engine may be slightly altered as a result of the regulation. In fact, for the entire population
of affected engines (approximately 20,000 new engines over a 5-year period), only 5 fewer
engines (0.02 percent) may be purchased due to market responses to the regulation. It is not
possible, however, to determine future investment decisions at the small entities in the affected
industries, so we cannot link these 5 engines to any one firm (small or large). Overall, it is very
unlikely that a substantial number of small firms who may consider purchasing a new engine will
be significantly impacted because the decision to purchase new engines is not altered to a large
extent.
Although the proposed rule will not have a significant impact on a substantial number of
small entities, we nonetheless have tried to reduce the impact of this rule on small entities. In
this proposed rule, we are applying the minimum level of control (i.e., the MACT floor), and the
minimum level of monitoring, record keeping, and reporting to affected sources allowed by the
ES-2
-------
CAA. In addition, RICE units with capacities under 500 hp and those that operate as
emergency/temporary units are not covered by the rule. This provision is expected to reduce the
level of small entity impacts.
The HAPs that are reduced as a result of implementing the RICE NESHAP will produce
a variety of benefits, some of which include: the reduction in the incidence of cancer to exposed
populations, neurotoxicity, irritation, and crop or plant damage. The rule will also produce
benefits associated with reductions in CO, PM10, and NOx emissions. Human health effects
associated with exposure to CO include cardiovascular system and central nervous system
effects, which are directly related to reduced oxygen content of blood and which can result in
modification of visual perception, hearing, motor and sensorimotor performance, vigilance, and
cognitive ability. Human health effects associated with PM and NOx include respiratory
problems, such as chronic bronchitis, asthma, or even death.
Although the rule will achieve reductions in HAPs, CO, PM10 and NOx, the benefit
analysis presented in this RIA is only able to place a dollar value on the benefits associated with
the health effects of PM10 and NOx (as it transforms into PM), and the health effects of NOx as it
transforms into ozone.
We use two approaches (referred to as Base and Alternative Estimates) to provide
benefits in terms of health effects and in monetary terms. While there is a substantial difference
in the specific estimates, both approaches show that the RICE MACT may provide benefits to
public health, whether expressed as health improvements or as economic benefits. These include
prolonging lives, reducing cases of chronic bronchitis and hospital admissions, and reducing
thousands of cases in other indicators of adverse health effects, such as work loss days, restricted
activity days, and days with asthma attacks. In addition, there are a number of health and
environmental effects which we were unable to quantify or monetize. These effects, denoted by
"B" are additive to the both the Base and Alternative estimates of benefits. Also, in determining
the monetary value of the effects, we use two different discount rates to provide a present value
of the benefit estimates. We adopt a 3 percent discount rate to reflect reliance on a "social rate
of time preference" discounting concept, as recommended by EPA's Guidelines for Preparing
Economic Analyses (EPA, 2000b). We also calculate benefits using a 7 percent discount rate
consistent with an "opportunity cost of capital" concept to reflect the time value of resources
ES-3
-------
directed to meet regulatory requirements, as recommended by OMB Circular A-94 (OMB,
1992). In this analysis, the benefit estimates are not significantly affected by the choice of
discount rate. The Base Estimate of monetized benefits of the PM10 and NOx health effects in
1998 dollars are $280 million + B (using a 3 percent discount rate), or $265 million + B (using a
7 percent discount rate). The Alternative Estimate totals $40 million + B (using a 3 percent
discount rate), or $45 million + B (using a 7 percent discount rate).
The Base Estimate of benefits reflects the use of peer-reviewed methodologies developed
for earlier risk and benefit-cost assessments related to the Clean Air Act, such as the regulatory
assessments of the Heavy Duty Diesel and Tier II Rules and the Section 812 Report to Congress.
The Alternative Estimate explores important aspects of the key elements underlying estimates of
the benefits of reducing NOx emissions, specifically focusing on estimation and valuation of
mortality risk reduction and valuation of chronic bronchitis. The Alternative Estimate of
mortality reduction relies on recent scientific studies finding an association between increased
mortality and short-term exposure to particulate matter over days to weeks, while the Base
Estimate relies on a recent reanalysis of earlier studies that associate long-term exposure to fine
particles with increased mortality. The Alternative Estimate differs in the following ways: it
explicitly omits any impact of long-term exposure on premature mortality, it uses different data
on valuation and makes adjustments relating to the health status and potential longevity of the
populations most likely affected by PM. It also uses a cost-of-illness method to value reductions
in cases of chronic bronchitis while the Base estimate is based on individual's willingness to pay
to avoid a case of chronic bronchitis.
Given the lack of approved methods to value HAPs and CO, the benefits estimates
provided must be considered with all other non-monetized benefits and information on costs,
economic impacts, and legal requirements to understand the full impact of the rule on society.
The tables below summarize the regulatory impacts of the RICE NESHAP, including:
total social costs, economic impacts, small business impacts, quantifiable benefits, and net
benefits (i.e., benefits minus costs). Approximately 90 percent of the total benefits ($255 million
under the Base Estimate, and $35 million under the Alternative Estimate) are associated with
NOx reductions from the 4SRB subcategory for new and existing engines. Approximately 10
percent of the total benefits ($25 million under the Base Estimate, and $5 million under the
ES-4
-------
Alternative Estimate) are associated with the PM reductions from the compression ignition
engine subcategory at new sources.
Table ES-1. Summary of Regulatory Impacts of the RICE NESHAP
Summary of Social Costs (millions 1998$)a:
Natural Gas Market $ 35
Mining Sector $ 20
Construction Sector $ 10
Chemicals $ 20
Energy Use Sectors:
Commercial Sector $ 70
Residential Sector $ 40
Transportation Sector $ 15
Other Industrial Sectors (23 industries) $ 45
Total Social Costs $255
Economic Impacts:
Change in Natural Gas Prices
Change in Prices in Other Industries
Change in New Engine Purchases
Small Business Impacts:
Firms with costs above 1% of revenues
Firms with costs above 3% of revenues
Total Benefits (millions 1998$)a:
Base Estimate
Using 3% Discount Rate
Using 7% Discount Rate
Alternative Estimate
Using 3% Discount Rate
Using 7% Discount Rate
0.30%
0.00% to 0.05%
0.02% (5 out of 20,000 engines)
2
0
$280 + unquantified benefits
$265 + unquantified benefits
$40 + unquantified benefits
$45 + unquantified benefits
Costs and benefit values are rounded to the nearest $5 million.
ES-5
-------
Table ES-2. Summary of Costs, Emission Reductions, and Quantifiable Benefits,
by Engine Type
Quantifiable Annual
Type of
Engine
2SLB-New
4SLB-New
4SRB-
Existing
4SRB-New
CI-New
Total
Total
Annualized
Cost (million
$/yr in 2005)
$3
$64
$37
$47
$96
$255
Emission Reductions"
HAP
250
4,035
230
215
305
5,035
(tons/yr
CO
2,025
36,240
98,040
91,820
6,320
234,445
in 2005)
NOx
0
0
69,900
98,000
0
167,900
PM
0
0
0
0
3,700
3,700
Monetized Benefits'5' c (million
Base
B!
B3
$105
$100
$150
$140
$254
$280
$265
$/yr in
Estimate
+ B5
+ B6
+ B9
+ BIO
-B13
+ B
+ B
2005)
Alternative
Estimate
B2
B4
$15+B7
$15+B8
$20 + Bn
$25 + B12
$5+B14
$40 + B
$45 +B
For the calculation of PM-related benefits, total NOx reductions are multiplied by the appropriate benefit per ton value
presented in Table 8-7. For the calculation of ozone-related benefits, NOx reductions are multiplied by 5/12 to account for
ozone season months and 0.74 to account for Eastern States in the ozone analysis. The resulting ozone-related NOx
reductions are multiplied by $28 per ton. Ozone-related benefits are summed together with PM-related benefits to derive
total benefits of NOx reductions. All benefits values are rounded to the nearest $5 million.
Benefits of HAP and CO emission reductions are not quantified in this analysis and, therefore, are not presented in this table.
The quantifiable benefits are from emission reductions of NOx and PM only. For notational purposes, unquantified benefits
are represented with a "B" for monetary benefits. A detailed listing of unquantified NOx, PM, and HAP related health
effects is provided in Table 8-13.
Results reflect the use of two different discount rates; a 3% rate which is recommended by EPA's Guidelines for Preparing
Economic Analyses (EPA, 2000b), and 7% which is recommended by OMB Circular A-94 (OMB, 1992).
ES-6
-------
Table ES-3. Annual Net Benefits of the RICE NESHAP in 2005
Social Costs'5
Social Benefits"cd:
HAP-related benefits
CO-related benefits
Ozone- and PM-related welfare benefits
Ozone- and PM-related health benefits:
Base Estimate
-Using 3% Discount Rate
-Using 7% Discount Rate
Alternative Estimate
-Using 3% Discount Rate
-Using 7% Discount Rate
Net Benefits (Benefits - Costs)0 d:
Base Estimate
-Using 3% Discount Rate
-Using 7% Discount Rate
Alternative Estimate
-Using 3% Discount Rate
-Using 7% Discount Rate
Million 1998$a
$255
Not monetized
Not monetized
Not monetized
$280 + B
$265 + B
$40+ B
$45+B
$25+B
$10 + B
-$215+B
-$210+ B
All costs and benefits are rounded to the nearest $5 million. Thus, figures presented in this chapter may not exactly equal
benefit and cost numbers presented in earlier sections of the chapter.
Note that costs are the total costs of reducing all pollutants, including HAPs and CO, as well as NOx and PM10. Benefits in
this table are associated only with PM and NOx reductions.
Not all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit categories that have
not been quantified and monetized are listed in Table 8-13. B is the sum of all unquantified benefits and disbenefits.
Monetized benefits are presented using two different discount rates. Results calculated using 3 percent discount rate are
recommended by EPA's Guidelines for Preparing Economic Analyses (EPA, 2000b). Results calculated using 7 percent
discount rate are recommended by OMB Circular A-94 (OMB, 1992).
ES-7
-------
1.0 INTRODUCTION
The regulation under analysis in this report, which is being proposed under Section 112
of the Clean Air Act Amendments of 1990 (CAA), is the National Emission Standard for
Hazardous Air Pollutants (NESHAP)for Reciprocating Internal Combustion Engines (RICE).
This emission standard would regulate the emissions of certain hazardous air pollutants (HAPs)
from certain internal combustion engines. The RICE industry group includes any facility
engaged in the use of internal combustion engines to produce power for the production or
transmission of final goods in their operating process. This report analyzes the impact that
regulatory action is likely to have on the industries affected by the rule, and on society as a
whole. Included in this chapter is a summary of the purpose of this regulatory impact analysis
(RIA), the statutory history which preceded this regulation, and a description of the content of
this report. This report should be read in conjunction with other background documents and
supporting analyses, such the determination of the MACT floor memorandum, the memorandum
of baseline emissions of HAPs, and the detailed analyses of engineering costs and national
impacts. All of these documents are located in the public docket.
1.1 PURPOSE
The President issued Executive Order 12866 on October 4, 1993. It requires EPA to
prepare RIAs for all "economically significant" regulatory actions. The criteria set forth in
Section 1 of the Order for determining whether a regulation is economically significant are that
the rule: (1) is likely to have an annual effect on the economy of $100 million or more, or
1-1
-------
adversely and materially affect a sector of the economy, productivity, competition, jobs, the
environment, public health or safety, or state, local, or tribal governments or communities; (2) is
likely to create a serious inconsistency or otherwise interfere with an action taken or planned by
another agency; (3) is likely to materially alter the budgetary impact of entitlements, grants, user
fees, or loan programs or the rights and obligation of recipients thereof; or (4) is likely to raise
novel legal or policy issues arising out of legal mandates, the President's priorities, or the
principles set forth in the Executive Order. The EPA has determined that the RICE NESHAP is
a "significant" rule because it will have an annual effect on the economy of more than
$100 million, and is therefore subject to the requirements of Executive Order 12866. Along with
requiring an assessment of benefits and costs, E.O. 12866 specifies that EPA, to the extent
allowed by the CAA and court orders, demonstrate (1) that the benefits of the NESHAP
regulation will outweigh the costs and (2) that the maximum level of net benefits (including
potential economic, environmental, public health and safety and other advantages; distributive
impacts; and equity) will be reached. EPA has chosen a single regulatory option for evaluation
in this RIA. Benefits and costs are quantified to the greatest extent allowed by available data.
As stipulated in E.O. 12866, in deciding whether and how to regulate, EPA is required to assess
all costs and benefits of available regulatory alternatives, including the alternative of not
regulating. Accordingly, the cost benefit analysis in this report is measured against the baseline,
which represents industry and societal conditions in the absence of regulation.
1.2 LEGAL HISTORY AND STATUTORY AUTHORITY
The RICE NESHAP will require sources to achieve emission limits reflecting the
application of the maximum achievable control technology (MACT), consistent with
sections 112(d) of the CAA. This section provides a brief history of Section 112 of the Act and
background regarding the definition of source categories and emission points for Section 112
standards.
Section 112 of the Act provides a list of 189 HAPs and directs the EPA to develop rules
to control HAP emissions. The CAA requires that the rules be established for categories of
sources of the emissions, rather than being set by pollutant. In addition, the CAA establishes
specific criteria for establishing a minimum level of control and criteria to be considered in
1-2
-------
evaluating control options more stringent than the minimum control level. Assessment and
control of any remaining unacceptable health or environmental risk is to occur 8 years after the
rules are promulgated.
For the subject NESHAP, EPA chose regulatory options based on control options on an
emission point basis. The RICE NESHAP regulates emissions of all HAPs emitted from all
emission points at both new and existing RICE sources. An emission point is defined as a point
within a facility that operates one or more internal combustion engine(s) which emits one or
more HAPs. For RICE units, there is only one emission point for each engine—end-of-pipe
emissions after combustion of a fuel source (typically natural gas).
1.3 REPORT ORGANIZATION
Chapter 2 presents information on the need for a regulation of RICE units. This meets
the Executive Order 12866 requirement for EPA to promulgate only regulations that are required
by law, are necessary to interpret the law, or are necessary due to a compelling public need, such
as material failures of private markets to protect or improve the health and safety of the public,
the environment, or the well-being of the public. We present the market conditions which
necessitate regulatory action, and provide a characterization of the air emissions associated with
RICE units, and the significance of the environmental problem which EPA intends to address
through the regulation.
Chapter 3 provides a profile of RICE units and the control techniques which were
considered for the standard. We then present the a summary of regulatory compliance costs
(including the engineering costs associated with the control techniques and monitoring,
reporting, and record keeping costs) along with the issues and assumptions upon which the
estimates were based.
Chapter 4 provides economic profiles of the industries that operate RICE units, which
provides a characterization of the affected industries and presents background data necessary to
estimate total social costs of the regulation. Chapter 5 describes the methodology used to
estimate the economic effects of the regulation including, predicted price, output, and
employment impacts which reflect upon the quantification of the social costs of the regulatory
option. We also present a discussion of how this rule may influence purchase decisions for new
1-3
-------
engines. Chapter 6 then uses the estimated costs and economic impacts to present a screening
analysis of firm-level impacts on small and large firms owning RICE units.
Chapter 7 provides a qualitative description of the benefits from several of the pollutants
reduced as a result of regulatory action (including, the HAPs of concern - formaldehyde,
acetaldehyde, acrolein, and methanol—carbon monoxide, and nitrous oxides). As explained in
this chapter, due to data limitations some benefits cannot be quantified in terms of dollar value
and therefore we cannot provide a full presentation of monetized benefits for the purpose of
comparing with costs.
Chapter 8 provides a quantitative assessment of a portion of the benefits which are
identified in Chapter 7; namely, only those benefits associated with health effects of NOx
exposures. The methodology used to arrive at these estimates is outlined and any uncertainties
and limitations are identified. The quantitative benefits of NOx health effects are then compared
with total social costs, recognizing that a large portion of the benefits are not represented in the
benefit-cost comparison (including all benefits associated with HAP reductions, CO reductions,
and the welfare effects of NOx).
1-4
-------
2.0 NEED FOR REGULATION
One of the concerns about potential threats to human health and the environment from
internal combustion engines is the emission of HAPs. Health risks from emissions of HAPs into
the air include increases in potential cancer incidences in the nasal cavity, trachea, and the
respiratory system in general and other toxic effects. This chapter discusses the need for and
consequences of regulating of HAP emissions from RICE.
Section 2.1 presents the conditions of market failure which necessitate government
intervention. Section 2.2 identifies the insufficiency of political and judicial forces to control the
release of toxic air pollutants from internal combustion engines. Section 2.3 provides a
characterization of the HAP and other pollutant emissions from RICE, and a summary of the
health and welfare risks of these pollutants. Lastly, Section 2.4 identifies the consequences of
regulating versus the option of not regulating.
2.1 ENVIRONMENTAL FACTORS WHICH NECESSITATE REGULATION
Regulation of RICE units addresses of the adverse health effects caused by human
exposure to HAP emissions. This section characterizes the emissions attributable to RICE and
summarizes the adverse health effects associated with human exposure to HAP emissions.
2-1
-------
2.1.1 Air Emission Characterization
The HAP emissions from RICE units are all organic HAPs as are in section 112(b) of the
CAA. HAP emissions from RICE are primarily composed of formaldehyde, acetaldehyde,
acrolein, and methanol. The different HAPs emitted have different toxicities, and there are some
variations in the concentrations of individual HAPs and the emission release characteristics of
different emission points.
Baseline emissions from RICE were estimated using information gathered during a
Federal Advisory Committee Act (FACA) process for several source categories of combustion
units (Alpha Gamma, 2002a) and provided by vendors of RICE units in response to information
collection requests and questionnaires sent out under section 114 of the CAA. For the purpose
of calculating baseline emissions and emission reductions, HAP emission factors were calculated
for each potentially affected new and existing engine type (spark-ignition two-stroke lean burn
(SI2SLB), spark-ignition four-stroke lean burn (SI4SLB), spark-ignition four-stroke rich burn
(SI4SRB), and compression-ignition (CI) engines1). These factors were estimated from test data
contained in the Inventory Database for engines rated at greater than 500 hp, operating at all
loads. The total HAP emission factor was calculated by summing the average emission factors
for formaldehyde, acetaldehyde, acrolein, and methanol in terms of Ib of HAP per hour of engine
operation. Table 2-1 contains the HAP emissions factors for each engine configuration in
pounds per hour. Emissions are greatest for 2SLB engines, which, on average, emit 0.962 Ibs.
per hour of HAPs, and least for CI engines, which emit 0.0359 Ibs. per hour.
'Unless otherwise noted, 2SLB, 4SLB, and 4SRB are used in the remainder of this section to denote spark-ignition
engine categories. Compression-ignition engines are referred to as CI throughout the section regardless of the
number of engine strokes per cycle. Characteristics of these four RICE design categories are discussed in more
detail in Section 3.1.
2-2
-------
Table 2-1. HAP Emissions Factors by Engine Configuration (lbs/hour)a
Engine Configuration Emissions Factor (Ibs/hour)
2SLB 0.962
4SLB 0.887
4SRB 0.0707
_ CI _ 0.0359 _
a The HAP emissions factors presented are the sum of the factors for formaldehyde, acetaldehyde, acrolein, and
methanol.
This value was then converted to an annual HAP emission factor in terms of tons of HAP
per year for each of the four engine types (2SLB, 4SLB, 4SRB, and CI) using the following
equation:
EFHAP () * 6,500
(2.1) HAP Emission Factor (— ) = - —
2,000
ton
where EFHAP is the total HAP emissions factor in pounds per hour, 6,500 is the estimated average
number of hours of operation per year for engines in the Inventory Database, and 2,000 is the
conversion factor between pounds and tons.
Total baseline emissions were estimated for 2005, which was the year chosen for
quantitative analysis of the costs and benefits of the RICE NESHAP. Baseline emissions were
calculated by multiplying the HAP emission factor generated by applying equation (2.1) for each
engine type by the number of engines of that type projected to be subject to the rule in 2005,
adjusting for the proportion of each engine type expected to be controlled in the absence of the
rule and their level of control. For those engines that are currently controlling formaldehyde
emissions or would control them in the future even in the absence of the RICE NESHAP, it was
assumed that the same percent reduction achieved for formaldehyde is being achieved for all
HAPs. For instance, approximately 27 percent of 4SRB are currently using NSCR to achieve 75
percent reductions in formaldehyde emissions. Therefore, it was assumed that these engines are
20
-j
-------
also achieving 75 percent reductions in all HAPs. To calculate baseline emissions for each
engine type, the following relationship was used:
BaselineHAP
Emissions
(2.2) "•»
where EFHAP is the value calculated for that engine type using equation (2.1), Y is the proportion
of engines estimated to be uncontrolled in the baseline, N is the number of engines subject to the
RICE NESHAP, and • *is the percent reduction in formaldehyde emissions achieved for those
engines that are controlled in the baseline.
Based on a memorandum discussing the distribution of major and area sources of RICE
units (Alpha Gamma, 200la), we anticipate that about 60 percent of existing and future
stationary RICE units will be located at area sources. This is because most RICE engines or
groups of RICE engines are not major sources of HAP emissions by themselves, but may be
major because they are co-located at major HAP sites. Because area sources are not covered by
the NESHAP, engines located at area sources will not incur any compliance costs associated
with the RICE NESHAP. Thus, only 40 percent of the existing 4SRB engines that are above 500
hp and are not backup/emergency units (the only existing engines that receive costs under the
rule) and 40 percent of all RICE projected to be added in the future (above 500 hp that are not
backup/emergency units) are expected to be subject to the proposed rule.
For example, for existing 4SRB engines, EFHAP = 0.0707 * 6,500/2,000 = 0.2298, Y is
0.73, N is equal to 4,573 * 0.4 (to adjust for the proportion of engines located at major sources),
and • *is 75 (the values of Y, N, and • *for other affected engine types are provided later in this
section of the report in Tables 2-5 and 2-6). Thus, the estimated level of baseline HAP
emissions from existing 4SRB RICE that are subject to the rule is equal to 0.2298 * 0.73 * 4,573
* 0.4 + 0.2298 * 0.27 * 0.25 * 4,573 * 0.4, or 335 tons per year.
Table 2-2 presents the estimated annual baseline HAP emissions from RICE subject to
the NESHAP for each type of new and existing engine. Although all existing RICE located at
major sources are subject to the rule, the only existing engines that will be required to meet
2-4
-------
emissions limits under the NESHAP are 4SRB. For the other three potentially affected
subcategories, the MACT floor is considered to be no control. Because an above-the-floor
option was considered to have excessive costs, existing 2SLB, 4SLB, and CI engines will be
subject only to the MACT floor and are not required to add emission control or monitoring
equipment. Baseline HAP emissions from existing sources are 27,489 tons per year. As
mentioned above, 4SRB are the only subcategory directly affected by the rule, representing
about 50 percent of baseline emissions from existing RICE, however, approximately only 3
percent are expected to be located at major sources and apply controls. Baseline HAP emissions
from new sources are expected to have reached 19,200 tons per year by 2005. Unlike existing
sources, all new sources subject to the rule are required to control HAP emissions. As described
above, baseline emissions take into account the current estimated level of emissions control,
based on questionnaire responses submitted by vendors and users of RICE units. As a result,
baseline HAP and other pollutant emissions reflect the level of control that would be achieved in
the absence of the rule.
2.1.2 Harmful Effects of HAP s
Exposure to HAPs has been associated with a variety of adverse health effects. Direct
exposure to HAPs can occur through inhalation, soil ingestion, the food chain, and dermal
contact. Health effects associated with HAP emissions are addressed in this NESHAP. In
general, many HAPs are classified as possible, probable, or known human carcinogens, which
can result in pain and suffering of individuals associated with leukemia or other cancers and
possible death. Other HAPs have not been classified as human carcinogens, but have non-
carcinogenic toxic effects. Exposure to these pollutants will also result in adverse health and
welfare impacts to human populations.
2-5
-------
Table 2-2. National Baseline HAP Emissions from RICE Units, 2005
Baseline HAP Emissions Baseline HAP Emissions
from All RICE Sources" from Major Sources
Type of Engine (tons/yr) (tons/yr)
Existing Engines:
2SLB Clean Gaseous Fuel
4SLB Clean Gaseous Fuel
4SRB Clean Gaseous Fuel
Compression Ignition
Subtotal
New Engines:
2SLB Clean Gaseous Fuel
4SLB Clean Gaseous Fuel
4SRB Clean Gaseous Fuel
Compression Ignition
Subtotal
Total
13,888
11,729
838
1,034
27,489
1,565
15,685
785
1,165
19,200
46,689a
5,555
4,692
335
414
10,996
626
6,274
314
466
7,680
18,676
a This includes emissions from both major and area sources.
Table 2-3 lists the possible effects from exposure to HAP emissions. EPA has devised a
system, which was adapted from one developed by the International Agency for Research on
Cancer (IARC), for classifying chemicals based on the weight-of-evidence (EPA, 1987). Of the
HAPs reduced from this proposed regulation, formaldehyde and acetaldehyde are classified as
group B, or probable human carcinogens. This means that there is evidence to support that the
chemical may cause an increased risk of cancer in humans. Formaldehyde and acetaldehyde are
a concern to the EPA because long term exposure to these chemicals have been known to cause
lung and nasal cancer in animals and probably humans.
2-6
-------
Table 2-3. Potential Health and Welfare Effects Associated with
Exposure to Hazardous Air Pollutants
Effect Type
Effect Category
Effect End-Point
Health Mortality
Chronic Morbidity
Acute Morbidity
Welfare Materials Damage
Aesthetic
Agriculture
Ecosystem Structure
Carcinogenicity
Genotoxicity
Non-Cancer lethality
Neurotoxicity
Immunotoxicity
Pulmonary function decrement
Liver damage
Gastrointestinal toxicity
Kidney damage
Cardiovascular impairment
Hematopoietic (Blood disorders)
Reproductive/Developmental toxicity
Pulmonary function decrement
Dermal irritation
Eye irritation
Corrosion/Deterioration
Unpleasant odors
Transportation safety concerns
Yield reductions/Foliar injury
Biomass decrease
Species richness decline
Species diversity decline
Community size decrease
Organism lifespan decrease
Trophic web shortening
2-7
-------
The remaining HAPs reduced by the rule are noncarcinogens. Though they do not cause
cancer, they are considered hazardous because of the other significant adverse health effects with
which they are associated, such as problems with the central nervous system, irritation of the
skin, eyes, or respiratory tract, and many other effects. These adverse effects are discussed in
more detail in Chapter 7 of this RIA.
The rule will also produce benefits associated with reductions in CO and NOx.
Emissions of CO and NOx have been associated with a variety of health impacts. Human health
effects associated with exposure to CO include cardiovascular system and central nervous
system (CNS) effects, which are directly related to reduced oxygen content of blood and which
can result in modification of visual perception, hearing, motor and sensorimotor performance,
vigilance, and cognitive ability.
Emissions of NOx can irritate the lungs and lower resistance to respiratory infection
(such as influenza). NOx, together with VOCs, are precursors to the formation of tropospheric
ozone. It is exposure to ozone that is responsible for adverse respiratory impacts, including
coughing and difficulty in breathing. Repeated exposure to elevated concentrations of ozone
over long periods of time may also lead to chronic, structural damage to the lungs. Paniculate
matter (PM) can also be formed from NOx emissions. Scientific studies have linked PM (alone
or in combination with other air pollutants) with a series of health effects. These health effects
include premature death and increased hospital admissions and emergency room visits, increased
respiratory symptoms and disease, decreased lung function, and alterations in lung tissue and
structure and in respiratory tract defense mechanisms. Children, the elderly, and people with
cardiopulmonary disease, such as asthma, are most at risk from the health effects of ozone and
PM. NOx emissions are also an important precursor to acid rain and may affect both terrestrial
and aquatic ecosystems. Atmospheric deposition of nitrogen leads to excess nutrient enrichment
problems ("eutrophication") in the Chesapeake Bay and several nationally important estuaries
along the East and Gulf Coasts. Nitrogen dioxide and airborne nitrate also contribute to pollutant
haze, which impairs visibility and can reduce residential property values and the value placed on
scenic views.
2.2 MARKET FAILURE
2-8
-------
The U.S. Office of Management and Budget (OMB) directs regulatory agencies to
demonstrate the need for a major rule (OMB, 1992). The RIA must show that a market failure
exists and that it cannot be resolved by measures other than Federal regulation. Market failures
are categorized by OMB as externalities, natural monopolies, or inadequate information. The
operators of RICE units participate in highly competitive industries, thus the natural monopoly
condition does not exist; nor does the condition of inadequate information due to the highly
organized nature of the affected industries. They do, however, create a negative externality from
the effects of the air pollution generated from RICE units. This means that, in the absence of
government regulation, the decisions of generators of air pollution do not fully reflect the costs
associated with that pollution. For a user of an internal combustion engine, air pollution from
the engine is a product or by-product that can be disposed of cheaply by venting it to the
atmosphere. Left to their own devices, many users of these engines treat air as a free good and
do not fully "internalize" the damage caused by toxic emissions. This damage is born by
society, and the receptors (the people who are adversely affected by the pollution) are not able to
collect compensation to offset their costs. They cannot collect compensation because the
adverse effects, like increased risks of morbidity and mortality, are non-market goods, that is,
goods that are not explicitly and routinely traded in organized free markets.
HAP emissions represent an externality in that operations that use RICE impose costs on
others outside of the marketplace. In the case of this type of negative externality, the market
price of goods and services does not reflect the costs, borne by receptors of the HAPs, generated
by the use of these engines. Government regulation, therefore, can be used to improve the
situation. For example, the NESHAP will require certain types of internal combustion engines to
reduce the quantity of HAPs that are emitted. With the NESHAP in effect, the cost that affected
industries must incur to produce products or services that require RICE as an input will more
closely approximate the full social costs of production. The more the costs of pollution are
internalized by the users of RICE, the greater the improvement in the way the market functions.
In the long run, affected industries will be forced to increase the prices of their products and
services in order to cover total production costs (including the internalized pollution costs that
result from the RICE NESHAP). As market prices rise to better reflect the costs to society
imposed by the use of RICE, consumers will reduce their demand for the affected products and
2-9
-------
services accordingly. As a result of the behavioral changes by consumers and producers, fewer
products and services will be provided to the market. The reduction in output will tend to reduce
emissions from RICE, which provides benefits to society, but it will also impose costs on
producers and consumers.
2.3 INSUFFICIENT POLITICAL AND JUDICIAL FORCES
There are a variety of reasons why many emission sources, in EPA's judgment, should be
subject to reasonably uniform national standards. The principal reasons are:
Air pollution crosses jurisdictional lines.
•• The people who breathe the air pollution travel freely, sometimes coming in
contact with air pollution outside their home jurisdiction.
Harmful effects of air pollution detract from the nation's health and welfare
regardless of whether the air pollution and harmful effects are localized.
Uniform national standards, unlike potentially piecemeal local standards, are not
likely to create artificial incentives or artificial disincentives for economic
development in any particular locality.
One uniform set of requirements and procedures can reduce paperwork and
frustration for firms that must comply with emission regulations across the
country.
Because RICE units are typically a small component to a larger operation or production
process, and because they are located in a wide variety of manufacturing and service industries,
it would be too costly for individuals or small groups to organize and obtain the political or
judicial force to reduce the level of air pollution from these sources.
2-10
-------
2.4 CONSEQUENCES OF REGULATORY ACTION
To address the health and welfare concern from the emission of HAPs, the proposed rule
reduces emissions at "major" sources of RICE HAP emissions (i.e., those that emit more than 10
tons of any one HAP or more than 25 tons of a combination of HAPs). Although the rule does
not apply to all RICE units that emit HAPs, it will reduce the magnitude of the negative
externality that exists in the affected industries. Below we provide an assessment of the
consequences of the attainment of EPA emission reduction objectives, and the likely
consequences if these objectives are not met.
2.4.1 Consequences if EPA 's Emission Reduction Objectives are Met
The EPA collected information and identified four subcategories (or types) of RICE units
in operation today that are potentially affected by the RICE NESHAP, including:
•• Spark-Ignition, Clean Gaseous Fuel 2-Stroke Lean Burn Engines (2SLB),
Spark-Ignition, Clean Gaseous Fuel 4-Stroke Lean Burn Engines (4SLB),
Spark-Ignition, Clean Gaseous Fuel 4-Stroke Rich Burn Engines (4SRB), and
•• Compression Ignition Engines (CI).
Table 2-4 and 2-5 present the population of existing and new sources of RICE units (Alpha
Gamma, 2002a), broken into the total number of engines in each model category and the number
that will be directly affected (i.e., incur compliance costs). These population estimates are based
on data contained in the Inventory Database and information provided by the EPA Office of
Mobile Sources (now the Office of Transportation and Air Quality) regarding estimated five year
sales volume for engines, which was derived from the Power Systems Research database, and
confidential sales projection information provided to EPA by engine manufacturers.
2-11
-------
Table 2-4. Population of Existing RICE3
Engine Subcategory
2SLB Clean Gaseous Fuel
4SLB Clean Gaseous
Fuel"
4SRB Clean Gaseous
Fuel6
Compression Ignition
HP Range"
500-1,000
1,000-5,000
5,000-10,000
Total
500-1,000
1,000-5,000
5,000-10,000
Total
500-1,000
1,000-5,000
5,000-10,000
Total
500-1,000
1,000-5,000
5,000-10,000
Total
Total Number of
Engines
1,412
2,726
305
4,444
866
3,095
188
4,149
3,353
1,215
5
4,573
5,312
3,541
None
8,853
Number of Affected
Engines0
0
0
0
0
0
0
0
0
1,341
486
2
1,829
0
0
0
0
Source: Alpha Gamma Technologies, Inc.; Memorandum to Sims Roy, U.S. EPA; National Impacts Associated
with Reciprocating Internal Combustion Engines; June, 2002a.
a The presented population excludes RICE that are used as emergency power units or that are less than 500 HP.
b There are no existing RICE greater than 10,000 HP.
0 The only existing RICE affected by the proposed rule are 4SRB engines located at major sources. The number of
affected engines was rounded to the nearest integer in this table for presentation purposes, but fractional engines
were used in calculations.
d 3 percent of existing 4SLB clean gaseous fuel RICE are controlled with a CO oxidation catalyst.
e 27 percent of existing 4SRB clean gaseous fuel RICE are controlled with NSCR.
2-12
-------
Table 2-5. Forecasted Population of New RICE, 2005a
Engine Subcategory
2SLB Clean
Gaseous Fuel
4SLB Clean
Gaseous Fueld
4SRB Clean
Gaseous Fuel6
Compression
Ignition
HP Rangeb
500-1,000
1,000-5,000
5,000-10,000
Total
500-1,000
1,000-5,000
5,000-10,000
Total
500-1,000
1,000-5,000
5,000-10,000
Total
500-1,000
1,000-5,000
5,000-10,000
Total
Total New RICE
Projected to be
Added by 2005
500
None
None
500
2,124
3,412
12
5,548
1,858
2,417
8
4,283
5,987
3,991
0
9,978
Affected New
RICE, 2005C
200
0
0
200
850
1,365
5
2,219
743
967
3
1,713
2,395
1,596
0
3,991
Source: Alpha Gamma Technologies, Inc.; Memorandum to Sims Roy, U.S. EPA; National Impacts Associated
with Reciprocating Internal Combustion Engines; June, 2002a.
a The forecasted population of new RICE are assumed for stationary applications not including emergency power
units.
b It is predicted that no RICE greater than 10,000 HP will be sold during the next five years.
0 The only existing RICE affected by the proposed rule are 4SRB engines located at major sources. The number of
affected engines was rounded to the nearest integer in this table for presentation purposes, but fractional engines
were used in calculations.
d It is predicted that 3 percent of new 4SLB clean gaseous fuel RICE will be controlled with a CO oxidation
catalyst in the absence of this regulation.
e It is predicted that 27 percent of new 4SRB clean gaseous fuel RICE will be controlled with NSCR in the
absence of this regulation.
2-13
-------
2.4.1.1 Regulatory Alternatives Considered.
Based on information in our database, we determined the MACT floor for new and
existing sources. For existing sources, the MACT floor (defined in the CAA as the average
control level achieved by the top 12 percent of similar sources) identifies controls on 4SRB
subcategory only, whereas all uncontrolled new sources in each of the five subcategories will be
required to control to the new source MACT floor levels (defined in the CAA as the best
available control achieved in the subcategory).
Table 2-6 presents the regulatory alternatives considered for this proposal. The first
regulatory alternative represents the MACT floor level of performance for engine subcategories.
The second regulatory alternative, a more stringent, above-the-floor alternative, was also
evaluated. The above-the-floor alternative was developed to introduce an alternative which
results in higher HAP emission reductions compared to the MACT floor performance levels.
However, EPA determined that the incremental costs associated with the above-the-floor MACT
options (with cost per ton over $300,000 for some subcategories) were excessive and are not
evaluated in this analysis.
2.4.1.2 Alternative Regulatory Options Based on Risk
We have made every effort in developing this proposal to minimize the cost to the
regulated community and allow maximum flexibility in compliance options consistent with our
statutory obligations. We recognize, however, that the proposal may still require some facilities
to take costly steps to further control emissions even though those emissions may not result in
exposures which could pose an excess individual lifetime cancer risk greater than one in one
million or which exceed thresholds determined to provide an ample margin of safety for
protecting public health and the environment from the effects of hazardous air pollutants. We
are, therefore, specifically soliciting comment on whether there are further ways to structure the
proposed rule to focus on the facilities which pose significant risks and avoid the imposition of
high costs on facilities that pose little risk to public health and the environment.
2-14
-------
Table 2-6. Summary of Regulatory Alternatives for RICE Subcategories
Engine Subcategory
Regulatory
Alternative
Requirement
Performance Level3
to
Existing Sources: MACT Floor
2SLB Clean Gaseous Fuel
Ab ove-the-Fl oor
4SLB Clean Gaseous Fuel
MACT Floor
Ab ove-the-Fl oor
4SRB Clean Gaseous Fuel MACT Floor
Compression Ignition
Digester/Landfill
MACT Floor
Ab ove-the-Fl oor
MACT Floor
Above-the-Floor
New Sources:
2SLB Clean Gaseous Fuel MACT Floor
4SLB Clean Gaseous Fuel MACT Floor
No control
Oxidation catalyst
No control
Oxidation catalyst
NSCR with a required formaldehyde
control efficiency and a formaldehyde
outlet concentration limit
No control
Oxidation catalyst
No control
Catalytic control with pretreatment
system*3
Oxidation catalyst with a required CO
control efficiency and a formaldehyde
outlet concentration limit
Oxidation catalyst with a required CO
control efficiency and a formaldehyde
outlet concentration limit
Equipment standard
Equipment standard
75 percent CH2O efficiency
or emission limitation of 350
ppbvd CH2O
Equipment standard
Equipment standard
60 percent CO efficiency or
emission limitation 17 ppmvd
CH2O
93 percent CO efficiency or
emission limitation of 14
ppmvd CH2O
-------
Table 2-6. Summary of Regulatory Alternatives for RICE Subcategories (continued)
Engine Subcategory Regulatory
Alternative
Requirement
Performance Levela
4SRB Clean Gaseous
Fuel
Digester/Landfill
MACT Floor
Compression Ignition MACT Floor
MACT Floor
NSCR with a required formaldehyde
control efficiency and a formaldehyde
outlet concentration limit
Oxidation catalyst with a required CO
control efficiency and a formaldehyde
outlet concentration limit
No control
75 percent CH2O efficiency or
emission limitation of 350
ppbvd CH2O
70 percent CO efficiency or
emission limitation of 580
ppbvd CH2O
Source: Alpha Gamma Technologies, Inc.; Memorandum to Sims Roy, U.S. EPA; National Impacts Associated with Reciprocating Internal Combustion
Engines; June, 2002a.
All concentrations must be corrected to 15 percent oxygen, dry basis.
-------
Representatives of the plywood and composite wood products industry provided EPA
with descriptions of three mechanisms that they believed could be used to implement more cost-
effective reductions in risk. The docket for today's proposed rule contains "white papers"
prepared by industry that outline their proposed approaches (see docket number A-95-35, Item
#II-D-9). These approaches could be effective in focusing regulatory controls on facilities that
pose significant risks and avoiding the imposition of high costs on facilities that pose little risk to
public health or the environment, and we are seeking public comment on the utility of each of
these approaches with respect to this rule.
Applicability Cutoffs for Threshold Pollutants Under Section 1 U(d}(4} of the CAA. The
first approach is an "applicability cutoff for threshold pollutants that is based on EPA's
authority under CAA section 112(d)(4) to establish standards for HAP which are "threshold
pollutants." A "threshold pollutant" is one for which there is a concentration or dose below
which adverse effects are not expected to occur over a lifetime of exposure. For such pollutants,
section 112(d)(4) allows EPA to consider the threshold level, with an ample margin of safety,
when establishing emission standards. Specifically, section 112(d)(4) allows EPA to establish
emission standards that are not based upon the maximum achievable control technology
(MACT) specified under section 112(d)(2) for pollutants for which a health threshold has been
established. Such standards may be less stringent than MACT. Historically, EPA has
interpreted section 112(d)(4) to allow categories of sources that emit only threshold pollutants to
avoid further regulation if those emissions result in ambient levels that do not exceed the
threshold, with an ample margin of safety.2
A different interpretation would allow us to exempt individual facilities within a source
category that meet the section 112(d)(4) requirements. There are three potential scenarios under
this interpretation of the section 112(d)(4) provision. One scenario would allow an exemption
for individual facilities that emit only threshold pollutants and can demonstrate that their
emissions of threshold pollutants would not result in air concentrations above the threshold
levels, with an ample margin of safety, even if the category is otherwise subject to MACT. A
second scenario would allow the section 112(d)(4) provision to be applied to both threshold and
2See 63 FR 18503, 18765 (April 15, 1998) (Pulp and Paper INESHAP).
2-17
-------
non-threshold pollutants, using the 1 in a million cancer risk level for decisionmaking for non-
threshold pollutants. A third scenario would allow a section 112(d)(4) exemption at a facility
that emits both threshold and non-threshold pollutants. For those emission points where only
threshold pollutants are emitted and where emissions of the threshold pollutants would not result
in air concentrations above the threshold levels, with an ample margin of safety, those emission
points could be exempt from the MACT standard. The MACT standard would still apply to the
non-threshold emissions from the source. For this third scenario, emission points that emit a
combination of threshold and non-threshold pollutants that are co-controlled by MACT would
still be subject to the MACT level of control. However, any threshold HAP eligible for
exemption under section 112(d)(4) that are controlled by control devices different from those
controlling non-threshold HAP would be able to use the exemption, and the facility would still
be subject to the parts of the standard that control non-threshold pollutants or that control both
threshold and non-threshold pollutants.
Estimation of hazard quotients and hazard indices. Under the section 112(d)(4)
approach, EPA would have to determine that emissions of each of the threshold pollutants
emitted by RICE sources at the facility do not exceed the threshold levels, with an ample margin
of safety. The common approach for evaluating the potential hazard of a threshold air pollutant
is to calculate a "hazard quotient" by dividing the pollutant's inhalation exposure concentration
(often assumed to be equivalent to its estimated concentration in air at a location where people
could be exposed) by the pollutant's inhalation Reference Concentration (RfC). An RfC is
defined as an estimate (with uncertainty spanning perhaps an order of magnitude) of a
continuous inhalation exposure that, over a lifetime, likely would not result in the occurrence of
adverse health effects in humans, including sensitive individuals. The EPA typically establishes
an RfC by applying uncertainty factors to the critical toxic effect derived from the lowest- or no-
observed-adverse-effect level of a pollutant (EPA, 1994). A hazard quotient less than one means
that the exposure concentration of the pollutant is less than the RfC, and, therefore, presumed to
be without appreciable risk of adverse health effects. A hazard quotient greater than one means
that the exposure concentration of the pollutant is greater than the RfC. Further, EPA guidance
for assessing exposures to mixtures of threshold pollutants recommends calculating a "hazard
index" by summing the individual hazard quotients for those pollutants in the mixture that affect
2-18
-------
the same target organ or system by the same mechanism (EPA, 2000d). Hazard index (HI)
values would be interpreted similarly to hazard quotients; values below one would generally be
considered to be without appreciable risk of adverse health effects, and values above one would
generally be cause for concern.
For the determinations discussed herein, EPA would generally plan to use RfC values
contained in EPA's toxicology database, the Integrated Risk Information System (IRIS). When a
pollutant does not have an approved RfC in IRIS, or when a pollutant is a carcinogen, EPA
would have to determine whether a threshold exists based upon the availability of specific data
on the pollutant's mode or mechanism of action, potentially using a health threshold value from
an alternative source, such as the Agency for Toxic Substances and Disease Registry (ATSDR)
or the California Environmental Protection Agency (CalEPA). Table 2-7 provides RfC's, as well
as unit risk estimates, for the HAP emitted by facilities in the RICE source category. A unit risk
estimate is defined as the upper-bound excess lifetime cancer risk estimated to result from
continuous exposure to an agent at a concentration of 1 • g/m3 in air.
To establish an applicability cutoff under section 112(d)(4), EPA would need to define
ambient air exposure concentration limits for any threshold pollutants involved.
There are several factors to consider when establishing such concentrations. First, we
would need to ensure that the concentrations that would be established would protect public
health with an ample margin of safety. As discussed above, the approach EPA commonly uses
when evaluating the potential hazard of a threshold air pollutant is to calculate the pollutant's
hazard quotient, which is the exposure concentration divided by the RfC.
2-19
-------
Table 2-7. Dose-Response Assessment Values for HAP Reported Emitted
by the RICE Source Category.
Chemical Name
Acetaldehyde
Acrolein
Formaldehyde
Methanol
CAS No.
75-07-0
107-02-8
50-00-0
67-56-1
Reference
Concentration" (mg/m3)
9.0E-03 (IRIS)
2.0E-05 (IRIS)
9.8E-03 (ATSDR)
4.0E+00 (CAL)
Unit Risk Estimate15
(l/(ug/m3))
2.2E-06 (IRIS)
1.3E-05 (IRIS)
Sources:
IRIS = EPA Integrated Risk Information System (http://www.epa.gov/iris/subst/index.html).
ATSDR = U.S. Agency for Toxic Substances and Disease Registry (http://www.atsdr.cdc.gov/mrls.html).
CAL = California Office of Environmental Health Hazard Assessment
(http: //www. oehha. ca. go v/air/chronic_rels/AllChrels.html).
HEAST = EPA Health Effects Assessment Summary Tables (EPA, 1997b).
a Reference Concentration: An estimate (with uncertainty spanning perhaps an order of magnitude) of a continuous inhalation
exposure to the human population (including sensitive subgroups which include children, asthmatics and the elderly) that is
likely to be without an appreciable risk of deleterious effects during a lifetime. It can be derived from various types of
human or animal data, with uncertainty factors generally applied to reflect limitations of the data used.
b Unit Risk Estimate: The upper-bound excess lifetime cancer risk estimated to result from continuous exposure to an agent at
a concentration of 1 • g/m3 in air. The interpretation of the Unit Risk Estimate would be as follows: if the Unit Risk
Estimate = 1.5 x 10-6 per • g/m3, 1.5 excess tumors are expected to develop per 1,000,000 people if exposed daily for a
lifetime to 1 • g of the chemical in 1 cubic meter of air. Unit Risk Estimates are considered upper bound estimates, meaning
they represent a plausible upper limit to the true value. (Note that this is usually not a true statistical confidence limit.) The
true risk is likely to be less, but could be greater.
EPA's "Supplementary Guidance for Conducting Health Risk Assessment of Chemical
Mixtures" (EPA, 2000f) suggests that the noncancer health effects associated with a mixture of
pollutants ideally are assessed by considering the pollutants' common mechanisms of toxicity.
The guidance also suggests, however, that when exposures to mixtures of pollutants are being
evaluated, the risk assessor may calculate a hazard index (HI). The recommended method is to
calculate multiple hazard indices for each exposure route of interest, and for a single specific
toxic effect or toxicity to a single target organ. The default approach recommended by the
guidance is to sum the hazard quotients for those pollutants that induce the same toxic effect or
affect the same target organ. A mixture is then assessed by several His, each representing one
toxic effect or target organ. The guidance notes that the pollutants included in the HI calculation
are any pollutants that show the effect being assessed, regardless of the critical effect upon which
2-20
-------
the RfC is based. The guidance cautions that if the target organ or toxic effect for which the HI
is calculated is different from the RfC's critical effect, then the RfC for that chemical will be an
overestimate, that is, the resultant HI potentially may be overprotective. Conversely, since the
calculation of an HI does not account for the fact that the potency of a mixture of HAP can be
more potent than the sum of the individual HAP potencies, an HI may potentially be
underprotective.
Options for establishing a hazard index limit. One consideration in establishing a hazard
index limit is whether the analysis considers the total ambient air concentrations of all the
emitted HAP to which the public is exposed.3 There are at least several options for establishing
a hazard index limit for the section 112(d)(4) analysis that reflect, to varying degrees, public
exposure.
One option is to allow the hazard index posed by all threshold HAP emitted from RICE
sources at the facility to be no greater than one. This approach is protective if no additional
threshold HAP exposures would be anticipated from other sources in the vicinity of the facility
or through other routes of exposure (e.g., through ingestion).
A second option is to adopt a "default percentage" approach, whereby the hazard index
limit of the HAP emitted by the facility is set at some percentage of one (e.g., 20 percent or 0.2).
This approach recognizes the fact that the facility in question is only one of many sources of
threshold HAP to which people are typically exposed every day. Because noncancer risk
assessment is predicated on total exposure or dose, and because risk assessments to focus only
on an individual source, establishing a hazard index limit of 0.2 would account for an
assumption that 20 percent of an individual's total exposure is from that individual source. For
the purposes of this discussion, we will call all sources of HAP, other than the facility in
question, "background" sources. If the facility is allowed to emit HAP such that its own impacts
could result in HI values of one, total exposures to threshold HAP in the vicinity of the facility
could be substantially greater than one due to background sources, and this would not be
protective of public health, since only HI values below one are considered to be without
appreciable risk of adverse health effects. Thus, setting the hazard index limit for the facility at
3Senate Debate on Conference Report (October 27, 1990), reprinted in "A Legislative History of the Clean Air Act
Amendments of 1990," Comm. Print S. Prt. 103-38 (1993) ("Legis. Hist.") at 868.
2-21
-------
some default percentage of one will provide a buffer which would help to ensure that total
exposures to threshold HAP near the facility (i.e., in combination with exposures due to
background sources) will generally not exceed one, and can generally be considered to be
without appreciable risk of adverse health effects.
A third option is to use available data (from scientific literature or EPA studies, for
example) to determine background concentrations of HAP, possibly on a national or regional
basis. These data would be used to estimate the exposures to HAP from non-RICE sources in
the vicinity of an individual facility. For example, the EPA's National-scale Air Toxics
Assessment (NATA) (EPA, 2002c) and ATSDR's Toxicological Profiles (ATSDR, 2002)
contain information about background concentrations of some HAP in the atmosphere and other
media. The combined exposures from RICE sources and from other sources (as determined from
the literature or studies) would then not be allowed to exceed a hazard index limit of one.
A fourth option is to allow facilities to estimate or measure their own facility-specific
background HAP concentrations for use in their analysis.
Tiered analytical approach for predicting exposure. Establishing that a facility meets the
cutoffs established under section 112(d)(4) will necessarily involve combining estimates of
pollutant emissions with air dispersion modeling to predict exposures. The EPA envisions that
we would promote a tiered analytical approach for these determinations. A tiered analysis
involves making successive refinements in modeling methodologies and input data to derive
successively less conservative, more realistic estimates of pollutant concentrations in air and
estimates of risk.
As a first tier of analysis, EPA could develop a series of simple look-up tables based on
the results of air dispersion modeling conducted using conservative input assumptions. By
specifying a limited number of input parameters, such as stack height, distance to property line,
and emission rate, a facility could use these look-up tables to determine easily whether the
emissions from their sources might cause a hazard index limit to be exceeded.
A facility that does not pass this initial conservative screening analysis could implement
increasingly more site-specific but more resource-intensive tiers of analysis using EPA-approved
modeling procedures, in an attempt to demonstrate that exposure to emissions from the facility
does not exceed the hazard index limit. The EPA's guidance could provide the basis for
2-22
-------
conducting such a tiered analysis (EPA, 1992c). It is also possible that ambient monitoring data
could be used to supplement or supplant the tiered modeling approach described above. It is
envisioned that the appropriate monitoring to support such a determination could be extensive.
Accounting for dose-response relationships. In the past, EPA routinely treated
carcinogens as non-threshold pollutants. The EPA recognizes that advances in risk assessment
science and policy may affect the way EPA differentiates between threshold and nonthreshold
HAP. The EPA's draft Guidelines for Carcinogen Risk Assessment (EPA, 1999b) suggest that
carcinogens be assigned non-linear dose-response relationships where data warrant. Moreover,
it is possible that dose-response curves for some pollutants may reach zero risk at a dose greater
than zero, creating a threshold for carcinogenic effects. It is possible that future evaluations of
the carcinogens emitted by this source category would determine that one or more of the
carcinogens in the category is a threshold carcinogen or is a carcinogen that exhibits a non-linear
dose-response relationship but does not have a threshold.
The dose-response assessments for formaldehyde and acetaldehyde are currently
undergoing revision by the EPA. As part of this revision effort, EPA is evaluating formaldehyde
and acetaldehyde as potential non-linear carcinogens. The revised dose-response assessments
will be subject to review by the EPA Science Advisory Board, followed by full consensus
review, before adoption into the EPA Integrated Risk Information System (IRIS). At this time,
EPA estimates that the consensus review will be completed by the end of 2003. The revision of
the dose-response assessments could affect the potency factors of these HAP, as well as their
status as threshold or nonthreshold pollutants. At this time, the outcome is not known. In
addition to the current reassessment by EPA, there have been several reassessments of the
toxicity and carcinogenicity of formaldehyde in recent years, including work by the World
Health Organization and the Canadian Ministry of Health.
If the section 112(d)(4) approach were adopted, the rulemaking would likely indicate that
the requirements of the rule do not apply to any source that demonstrates, based on a tiered
approach that includes EPA-approved modeling of the affected source's emissions, that the
anticipated HAP exposures do not exceed the specified hazard index limit.
2.4.1.2.1 Subcateaorv Delisting Under Section 1 U(c}(9}(E} of the CAA
2-23
-------
EPA is authorized to establish categories and subcategories of sources, as appropriate,
pursuant to CAA section 112(c)(l), in order to facilitate the development of MACT standards
consistent with section 112 of the CAA. Further, section 112(c)(9)(B) allows EPA to delete a
category (or subcategory) from the list of major sources for which MACT standards are to be
developed when the following can be demonstrated: 1) in the case of carcinogenic pollutants,
that "no source in the category . . . emits [carcinogenic] air pollutants in quantities which may
cause a lifetime risk of cancer greater than one in one million to the individual in the population
who is most exposed to emissions of such pollutants from the source"; 2) in the case of
pollutants that cause adverse noncancer health effects, that "emissions from no source in the
category or subcategory . . . exceed a level which is adequate to protect public health with an
ample margin of safety"; and 3) in the case of pollutants that cause adverse environmental
effects, that "no adverse environmental effect will result from emissions from any source."
Given these authorities and the suggestions from the white paper prepared by industry
representatives (see docket number A-95-35, Item # II-D-9), EPA is considering whether it
would be possible to establish a subcategory of facilities within the larger RICE category that
would meet the risk-based criteria for delisting. Such criteria would likely include the same
requirements as described previously for the second scenario under the section 112(d)(4)
approach, whereby a facility would be in the low-risk subcategory if its emissions of threshold
pollutants do not exceed the HI limits and if its emissions of non-threshold pollutants do not
exceed a cancer risk level of 10"6.
Since each facility in such a subcategory would be a low-risk facility (i.e., if each met
these criteria), the subcategory could be delisted in accordance with section 112(c)(9), thereby
limiting the costs and impacts of the proposed MACT rule to only those facilities that do not
qualify for subcategorization and delisting. EPA estimates that the maximum potential effect of
this approach would be the same as that of applying the section 112(d)(4) approach that allows
exemption of facilities emitting threshold and non-threshold pollutants if exemption criteria are
met.
Facilities seeking to be included in the delisted subcategory would be responsible for
providing all data required to determine whether they are eligible for inclusion. Facilities that
2-24
-------
could not demonstrate that they are eligible to be included in the low-risk subcategory would be
subject to MACT and possible future residual risk standards.
Establishing that a facility qualifies for the low-risk subcategory under section 112(c)(9)
will necessarily involve combining estimates of pollutant emissions with air dispersion modeling
to predict exposures. The EPA envisions that we would employ the same tiered analytical
approach described earlier in the section 112(d)(4) discussion for these determinations.
Another approach under section 112(c)(9) would be to define a subcategory of facilities
within the RICE source category based upon technological differences, such as differences in
production rate, emission vent flow rates, overall facility size, emissions characteristics,
processes, or air pollution control device viability. If it could then be determined that each
source in this technologically-defined subcategory presents a low risk to the surrounding
community, the subcategory could then be delisted in accordance with section 112(c)(9).
If this section 112(c)(9) approach were adopted, the rulemaking would likely indicate
that the rule does not apply to any source that demonstrates that it belongs in a subcategory
which has been delisted under section 112(c)(9).
2.4.1.3 Allocation of Resources.
One of the consequences of the proposed rule is that there will be improved allocation of
society's resources associated with RICE. The negative externality of treating air pollution as a
free good results in production costs that are less than the optimal level to society (a level that
would incorporate the costs associated with the air pollution). Thus, the output levels in the
affected industries that utilize RICE units also exceed the optimal level to society. With this
rule, the costs of the harmful effects of the processes that use these engines will be internalized
by the producers. This, in turn, will affect consumers' purchasing decisions. To the extent these
newly-internalized costs are then passed along to the end users of products from industries that
utilize RICE units in their production process, and to the extent that these end users are free to
buy as much or as little of these products as they wish, they will purchase less (relative to their
purchases of other competing services). If this same process of internalizing negative
externalities occurs throughout all of the affected industries, an economically optimal situation is
approached. This is the situation in which the marginal cost of resources devoted to productions
2-25
-------
of products that utilize RICE during production processes equals the marginal value of the
products to the end users of these products. Although there are uncertainties in this progression
of impacts, in the aggregate and in the long run, the NESHAP will move society toward this
economically optimal situation.
2.4.1.4 Emissions Reductions and Cost Impacts.
The environmental impact of the rule includes the reduction of HAP, CO, NOx, and PM
emissions and are presented relative to the baseline, which represents the level of control in the
absence of the proposed rule. The estimates include the impacts of applying control to:
(1) existing RICE units and (2) additional RICE units that are expected to begin operation by
2005. Thus, the overall estimates represent annual impacts occurring in 2005. Under the
proposed rule, it is estimated that the emissions of HAP from RICE units would be reduced by
about 5,000 tons per year (approximately 200 tons per year from existing sources and 4,800 tons
per year from new sources), emissions of CO would be reduced by 234,400 tons per year,
emissions of NOx would be reduced by 167,900 tons per year, and directly emitted PM will be
reduced by approximately 3,700 tons per year. Emission levels of other air pollutants (VOC)
were not quantified.
The cost impact of the rule includes the capital cost of new control equipment, the
associated operation and maintenance cost, and the cost of monitoring, recordkeeping, and
reporting. Under the proposed rule, it has been determined that oxidation catalysts, such as CO
oxidation catalyst and non-selective catalytic reduction (NSCR), are applicable controls for the
reduction of HAP from RICE. Cost impacts include the total capital investment of new
oxidation catalyst or NSCR equipment, the cost of energy (utilities) required to operate the
control equipment, operation and maintenance costs, and the cost of monitoring, reporting, and
record keeping. For 2SLB and 4SLB burn clean gaseous fuel engines, and compression ignition
engines, the annualized monitoring costs ranged from $5,959/year to $58,800/year. For 4SRB
clean gaseous fuel engines, the annualized monitoring costs ranged from $6,496/year to
$21,618/year.
Total control costs and total annual control costs for affected RICE units are presented in
Table 2-8. For the MACT floor for existing 4SRB clean gaseous fuel engines, the estimated
2-26
-------
total capital investment is $68.4 million and the total annualized cost is $38.1 million (1998
dollars). For the MACT floor for new sources, the estimated total capital investment is $372.2
million and the total annualized cost is $215.6 million for new sources projected to enter by
2005. Overall, the total annualized compliance costs in 2005 across both new and existing
sources are estimated to be $253.7 million.
Considering total annualized capital costs, monitoring, reporting, and record keeping
costs at all affected sources along with behavioral responses in the affected markets (see Section
5 for further discussion of the economic model), this proposed rule has estimated total social
costs of approximately $253.7 million in the 5th year after implementation. The estimated social
costs differs only very slightly from the estimated engineering compliance costs (excluding
behavioral adjustments) in this case (about $20,000 less) because the resulting price changes in
each affected market are so small that there is little behavioral response by consumers and
producers.
2.4.1.5 Energy Impacts.
Energy impacts associated with this regulation would be due to additional energy
consumption that the proposed regulation would require by installing and operating control
equipment. The only energy requirement for the operation of the control technologies is due to a
small increase in fuel consumption resulting from back pressure caused by the control system.
This energy impact is however considered minimal in comparison to cost of other impacts, and is
therefore considered negligible.
2-27
-------
Table 2-8. Summary of HAP Emission Reductions and Cost Impacts
Associated with the RICE NESHAP, 2005
to
to
oo
Existing
Engine
Subcategory
2SLB Clean
Gaseous Fuel
4SLB Clean
Gaseous Fuel
4SRB Clean
Gaseous Fuel
Compression
Ignition
Total:
Baseline
Emissions
(ton/yr)
5,555
4,692
335
414
10,996
HAP Emission
Reduction
(ton/yr)
0
0
230
0
230
Total Capital
Investment
($l,000)a
0
0
68,370
0
68,370
Total
Annualized
Cost ($l,000)a
0
0
38,125
0
38,125
Baseline
Emissions
(ton/yr)
626
6,274
314
466
7,680
New
HAP Emission
Reduction
(ton/yr)b
250
4,034
216
302
4,802
Total Capital
Investment
($l,000)a
5,846
109,468
91,098
165,752
372,164
Total
Annualized
Cost ($l,000)a
3,122
65,774
47,853
98,852
215,601
Total capital investment and total annual costs include the cost of monitoring. Monitoring costs were calculated based on selecting option 1 for 2SLB and 4SLB clean
gaseous fuel, and for compression ignition engines greater than or equal to 5000 HP, and selecting option 3 for 2SLB and 4SLB clean gaseous fuel, and compression
ignition engines between 500 HP and 5000 HP. Monitoring costs were calculated based on selecting option 5 for 4SPJ3 clean gaseous fuel engines greater than or equal to
5000 HP, and selecting option 6 for 4SRB clean gaseous fuel engines between 500 HP and 5000 HP.
-------
2.4.1.6 State Regulation and New Source Review.
Many RICE emit significant quantities of NOx and sometimes CO. States in the
Northeast U.S. and to a lesser extent in other parts of the U.S. have required that reasonably
available control technology (RACT) be installed on many existing engines for control of NOx.
These RACT controls vary from state to state. In some cases RACT NOx controls require the
use of ignition enhancement or ignition retard which achieves a NOx reduction of about 10 to 15
percent. In other cases, RACT NOx control may be low emission combustion (LEG) technology
which can reduce NOx emissions by 80 to 90 percent. Finally, in other cases, selective catalytic
reduction (SCR) and NSCR technologies have been installed to meet RACT requirements. SCR
and NSCR can reduce NOx emissions by 90 percent. Existing 4SRB RICE have already added
any required NOx or CO controls needed to meet state, local or federal requirements. A new
engine going into the Northeast U.S. or any area where RACT is currently required would be
expected to control NOx to similar levels as existing engines are currently required.
Existing 2SLB, 4SLB, and CI are not required to install MACT controls. Under the
provisions of the NOx SIP call, however, large (> 2500HP and/or 1 ton/day NOx emissions) new
2SLB, 4SLB, and CI engines will have to reduce NOx emissions potentially beyond the RACT
level in the NOx SIP call region (21 Eastern U.S. States and the District of Columbia) by 2007.
The NOx SIP call is a rulemaking meant to help the Northeastern states meet the 1-hour ozone
National Ambient Air Quality Standard (NAAQS). To estimate the potential impact of the RICE
MACT rule in the states affected by the NOx SIP call, queries on the RICE Inventory Database
were performed to determine the number of engines, size, and controls applied to each type of
engine in these states. Information from the Database indicates that selective catalytic reduction
(SCR) is being applied to two CI engines. Catalytic reduction, including oxidation catalysts and
NSCR, is being applied to a total of 30 engines in the database (14 4SRB and 16 CI). There are
additional engines with existing controls, but none of these controls are considered applicable
techniques for reducing HAP from RICE (Alpha-Gamma, 2002b).
The installation of groups of new engines or even one large new engine may trigger new
source review (NSR) in a non-attainment area for NOx or CO, or prevention of significant
deterioration (PSD) in an attainment area for NOx or CO, because of the magnitude of
uncontrolled emissions of NOx or CO emissions. In such cases lowest achievable emission
2-29
-------
reduction (LAER) technology or best available control technology (BACT) would have to be
installed. The NSCR technology for 4SRB engines can reduce NOx by 90 percent and selective
catalytic reduction (SCR) technology can also reduce NOx by similar amounts. Since NSCR
will achieve the MACT standard and also the NOx and CO standards, no additional impacts are
expected for this type of engine for existing and new engines as a result of the RICE MACT.
For new 2SLB, 4SLB, and CI engines, it would be expected that RACT NOx controls may be
required. No additional CO controls would be required since oxidation catalyst systems also
reduce CO in addition to HAPs. It is also expected that some of the larger engines that can
trigger NSR/PSD review will have to add NOx controls such as SCR in addition to controls
required by the RICE MACT oxidation catalyst systems. We expect these cases to be limited in
number.
No existing control technologies are in place specifically to address the reduction of
HAPs from RICE. There are several existing control techniques designed to reduce other
emissions from RICE that could potentially reduce HAP emissions. However, EPA has
determined that, among existing add-on controls, controls that involve oxidation are the most
likely to reduce HAP emissions from RICE. For rich burn engines, the only currently known
applicable technology is NSCR. The only known applicable technology for lean burn engines is
the use of oxidation catalysts. There are three other control technologies that could potentially
reduce HAP emissions from RICE: air injection, particulate traps, and catalyzed diesel
particulate filters. However, the effectiveness of HAP reduction has not been demonstrated for
any of these technologies. No other current control device is considered to be applicable for
HAP emission reductions from RICE.
For those engines that have installed or will install NSCR or oxidation catalysts to meet
restrictions on NOx or other emissions, HAP emissions are reduced incidentally. This has been
taken into account in calculating baseline emissions and the incremental emission reductions that
will be achieved by the RICE NESHAP. Searches of EPA's RACT/BACT/LAER Clearinghouse
(RBLC), California's BACT Clearinghouse, and the RICE Inventory Database were conducted
to estimate the number of existing RICE that are equipped with these controls. In addition,
several state environmental agencies, EPA regions, and catalyst vendors were contacted to gather
more information. The search revealed very few installations of oxidation catalysts. Based on
2-30
-------
searches of EPA's RBLC database, only five facilities permitted in the last three years have
stationary RICE equipped with an oxidation catalyst. The states and EPA regions contacted
indicated there were very few or zero facilities in their areas that are known to use oxidation
catalysts. A catalyst vendor contacted by EPA indicated that 4,000 catalysts have been installed
on stationary RICE since 1985. This vendor projects 200 catalyst installations per year, with
approximately 60 percent being oxidation catalysts and the other 40 percent NSCR. Estimates
based on information regarding existing engines in the Inventory Database indicate that 27
percent of existing 4SRB are equipped with NSCR, 3 percent of existing 4SLB are using
oxidation catalysts, and no existing 2SLB or CI engines were identified as using either (Alpha-
Gamma, 2002b). Based on the information gathered, EPA estimates that 27 percent of existing
and new 4SRB, 3 percent of existing and new 4SLB, and 0 percent of existing and new 2SLB
and CI RICE would be controlled in the absence of this NESHAP.
2.4.1.7 Other Federal Programs.
No other Federal programs are known except as discussed in 2.4.1.5.
2.4.2 Consequences if EPA 's Emission Reduction Objectives are Not Met
The most obvious consequence of failure to meet EPA's emission reduction objectives
would be emissions reductions and benefits that are not as large as is projected in this report.
However, costs are not likely to be as large either. Whether it is noncompliance from ignorance
or error, or from willful intent, or simply slow compliance due to owners and/or operators
exercising legal delays, poor compliance can save some producers money. Unless states respond
by allocating more resources into enforcement, then poor compliance could bring with it smaller
aggregate nationwide control costs. EPA has not included an allowance for poor compliance in
its estimates of emissions reductions, due to the fact that poor compliance is unlikely. Also, if
the emission control devices degraded rapidly over time or in some other way did not function as
expected, there could be a misallocation of resources. This situation is very unlikely, given that
the NESHAP is based on demonstrated technology.
2-31
-------
3.0 PROFILE OF RICE UNITS AND TECHNOLOGIES
EPA identified 26,832 engines located at commercial, industrial, and government
facilities based on information contained in the EPA Inventory Database V.4—Internal
Combustion (1C) Engines (referred to as the Inventory Database). The list of engines in this
database was itself developed from information in the Aerometric Information Retrieval System
(AIRS) and Ozone Transport Assessment Group (OTAG) databases and state and local permit
records. As part of the Industrial Combustion Coordinated Rulemaking (ICCR) FACA process,
industry and environmental stakeholders reviewed the engines units in the EPA Inventory
Database. These stakeholders contributed to the Inventory Database by identifying and
including omitted units. From this initial population of 26,832 engines, there were 10,118
engines that were excluded from further analysis because they were either less than 500 hp or
used to supply emergency/backup power or both. These engines are not covered by the proposed
regulation. Of the 16,714 remaining engines in the Inventory Database that are potentially
affected by the rule, 2,645 units had sufficient information to assign model numbers (e.g., fuel
type, engine configuration, horsepower). These 2,645 units were linked to 834 existing
facilities. These engines are primarily in either the oil and gas extraction industry or the natural
gas transmission industry. Because the only existing RICE units affected by the rules are
SI4SRB, most of the engines in the database would not have any control costs. Only 889 of the
2,645 existing engines in the database with sufficient information to assign a model number are
expected to incur control costs. However, the database is assumed to be representative of the
industries where new engines will be added in the future. This section provides background
information on RICE technologies, the units and facilities in the Inventory Database, and engines
3-1
-------
population estimates. Also included is a discussion of pollutants associated with these units and
the cost of installing control technologies.
As mentioned in Section 2, EPA anticipates that about 60 percent of existing and future
stationary RICE units are currently or will be located at area sources (Alpha Gamma, 200la).
This is because most RICE engines or groups of RICE engines are not major HAP emission
sources by themselves, but may be major because they are co-located at major HAP sites.
Because area sources are not covered by the NESHAP, engines located at area sources will not
incur any compliance costs associated with the RICE NESHAP. Thus, only 40 percent of the
existing SI4SRB engines (the only existing engines with costs under the rule) and 40 percent of
all RICE projected to be added in the future (that are above 500 hp and are not
backup/emergency units) are expected to be directly affected by the proposed rule.
3.1 ENGINES TECHNOLOGIES
The 1C engines affected by the regulation are of four design categories as discussed in
Section 1: SI2SLB, SI4SLB, and SI4SRB, and CI.1 In an 1C engine, a mixture of air and fuel is
burned in engine cylinders. A series of pistons and a crankshaft convert the energy of the
expanding gases into mechanical work. Apart from the method of ignition, SI or CI, and the
number of strokes, two or four, engines are differentiated by their air-to-fuel (A/F) ratio. As
defined by the Gas Research Institute (GRI, 2000), the relative proportions of air and fuel are
expressed as the mass of air to that of fuel and is called the A/F ratio. The A/F ratio is called
"stoichiometric" if the mixture contains the minimum amount of air that supplies sufficient
oxygen to complete combustion of the fuel. Rich burn engines operate near the fuel-air
stoichiometric limit with excess oxygen levels less than 4 percent. Lean burn engines operate
with significantly higher excess oxygen levels (GRI, 2000). The majority of the information
contained in this section is from the Gas Research Institute's publication, "Engine Design,
Operation, and Control in the Natural Gas Industry" (GRI, 2000).
'Unless otherwise noted, 2SLB, 4SLB, and 4SRB are used in the remainder of this section to denote spark-ignition
engine categories. Compression-ignition engines are referred to as CI throughout the section regardless of the
number of engine strokes per cycle.
O O
3-2
-------
3.1.1 ,57 Two-Stroke Engines
A two-stroke engine completes the power cycle in one revolution of the crankshaft. The
crankshaft in an 1C engine is attached to the pistons. When the pistons move up and down, the
crankshaft turns and converts the reciprocating motion of the pistons into rotary motion. The
first stroke begins with the piston at the top of the cylinder. At this time, the engine's
combustion chamber contains a compressed mixture of fuel and air. The mixture is ignited by a
spark that causes a sudden increase in temperature and pressure that forces the piston downward,
transferring power to the crankshaft. As the piston travels downward, air and exhaust ports are
uncovered, allowing combustion gases to exit and fresh air to enter. During the second stroke,
the air and exhaust ports close and fuel is injected into the cylinder. As the piston returns to its
starting position, the upward motion compresses the fuel and air mixture. When the piston
reaches the top of the cylinder, the compressed fuel and air mixture is ignited again and the cycle
begins again.
Because fresh air is used to clear combustion gases from the cylinder, two-stroke engines
operate with an A/F ratio greater than stoichiometric and are, therefore, all of the "lean-burn"
design type. A/F ratios for 2SLB engines range between 20:1 and 60:1. Their exhaust
temperatures are normally between 550 and 800°F. All 2SLB engines are direct-injected (i.e.,
fuel is injected directly into the cylinder) (GRI, 2000).
3.1.2 ,57 Four-Stroke Engines
A four-stroke engine completes the power cycle in two revolutions of the crankshaft.
The first stroke is the intake stroke during which the intake valve opens and the exhaust valve
closes. The downward motion of the piston draws air (direct injected) or a mixture of air and
fuel (premixed) into the cylinder. During the second stroke, the intake valve closes, and the fuel
is injected (direct injected) into the cylinder as the piston moves upward to compress the air and
fuel mixture. As the piston finishes its upward stroke, a spark ignites the mixture, causing a
sudden increase in temperature and pressure. The increased pressure drives the piston downward
(i.e., the third stroke), delivering power to the crankshaft. During the fourth stroke, the exhaust
valve opens and the piston moves upwards to force the exhaust gases out of the cylinder. The
regulation will affect two types of spark ignition, four-stroke engines: 4SLB and 4SRB.
-------
Four-Stroke Lean Burn. Compared to the 2SLB engine, the 4SLB engine reduces the
presence of high fuel concentration and temperature gradients in the cylinder by mixing the air
and fuel during the second stroke. Compared to a 4SRB engine, the increased A/F ratio in 4SLB
engines reduces combustion and exhaust temperatures. A/F ratios for this engine configuration
are similar to those of 2SLB engines.
Four-Stroke Rich Burn. 4SRB engines have A/F ratios near stoichiometric, meaning that
in these engines the proportion of fuel relative to air is greater than in lean-burn engines. All
turbo-charged engines that do not introduce fresh air to sweep combustion gases out of the
cylinder after ignition are 4SRB engines (GRI, 2000). A/F ratios for these engines typically
range between 16:1 and 20:1. Exhaust temperature is higher in rich-burn engines than in lean-
burn engines.
3.1.3 Compression Ignition Units
CI units almost always operate as lean burn engines. They can be configured as either
2SLB or 4SLB; the distinction is that CI engines are fueled by distillate fuel oil (diesel oil), not
by natural gas. Fuel consumption is an important determinant in the type of emissions from
these units; combustion of natural gas and combustion of diesel oil may each have separate types
and proportions of emissions. Because of this difference in fuel consumption, the type of control
equipment, and thus cost, varies from natural gas-fueled units, even if those using diesel are of
the same engine configuration and horsepower (hp).
5-4
-------
3.2 EMISSIONS
The proposed regulation aims to reduce HAP emissions. HAPs of concern include
formaldehyde, acetaldehyde, acrolein, and methanol. Without the regulation, annual HAP
emissions from sources subject to the RICE NESHAP are estimated to be 18,700 tons each year
by 2005. The proposed regulation will decrease emissions from existing sources by
approximately 200 tons per year and emissions from new sources by about 4,800 tons per year
by 2005. Estimation of baseline emissions and emission reductions is described further in
Section 2.
Emissions factors differ substantially between engine configurations. Table 3-1 contains
the HAP emissions factors for each engine configuration in pounds per hour. Emissions are
greatest for 2SLB engines, which, on average, emit 0.962 Ibs. per hour of HAPs, and least for CI
engines, which emit 0.0359 Ibs. per hour. In estimating the emission factors, test data from the
Emissions Database from engines rated at greater than 500 hp, operating at all loads, were used.
Table 3-1. HAP Emissions Factors by Engine Configuration (lbs/hour)a
Engine Configuration Emissions Factor (Ibs/hour)
2SLB 0.962
4SLB 0.887
4SRB 0.0707
CI 0.0359
Source: Alpha Gamma Technologies, Inc.; Memorandum to Sims Roy, U.S. EPA; National Impacts Associated with
Reciprocating Internal Combustion Engines; January, 2002a.
a The HAP emissions factors presented are the sum of the factors for formaldehyde, acetaldehyde, acrolein, and
methanol.
3.3 CONTROL COSTS
The primary method identified by EPA for controlling emissions from 2SLB, 4SLB, and
CI engines is the use of oxidation catalyst systems. However, few existing 2SLB, 4SLB, and CI
engines currently use these systems to control their emissions. Less than 1 percent of 2SLB and
CI engines are controlled, and only about 3 percent of 4SLB engines are controlled. All of these
numbers are below the criteria for a MACT floor in each subcategory, so the MACT floor in
3-5
-------
these categories was considered to be no control. An above-the-floor MACT option of requiring
oxidation catalyst systems was considered for these subcategories of engines, but it was
determined that the incremental cost of this alternative would be excessive (EPA, 2000a).
Unlike the situation for the other engine configurations, the average of the top 12 percent
of existing 4SRB stationary RICE sources control emissions. The method typically used to
control emissions from 4SRB engines is known as non-selective catalytic reduction (NSCR).
Because the average of the top 12 percent of existing engines in this category are controlled, the
MACT floor for existing 4SRB engines is considered to be the level of HAP emissions reduction
achieved by using NSCR systems. Although the percentage of existing 2SLB, 4SLB, and CI
engines that are controlled with oxidation catalyst systems is not high enough to mandate a
MACT floor requiring control for existing units, there are stationary RICE units operating with
these systems in each of these subcategories. Therefore, the MACT floor for new sources in
these subcategories is defined as the level of HAP emissions control achieved using oxidation
catalyst systems. For new 4SRB engines, the MACT floor is the same as for existing engines.
The required control for new 4SRB engines is the level of HAP emissions reduction achieved
using NSCR systems (EPA, 2000).
Each unit in the Inventory Database was grouped into one of 12 categories, or model
types, based on its engine configuration, horsepower, and fuel type. For each of those model
types, the annualized cost of installing pollution control equipment to achieve the floor level of
control and the associated administrative, operating, monitoring, and maintenance costs for that
equipment were estimated based on information collected from catalyst vendors. First, the total
direct and indirect capital costs were estimated as follows. Data on equipment costs (EC) for
oxidation catalysts and NSCR for 26 model engines were collected from Engelhard Corporation
and Miratech Corporation (the two firms surveyed that provided cost estimates). Because these
costs did not include instrumentation, tax, freight, or installation, purchased equipment costs
(PEC) were calculated as 118 percent of EC. Direct installation costs (DIG) were then estimated
as 30 percent of PEC. The direct capital costs are equal to PEC plus DIG. The indirect capital
costs were estimated to be 31 percent of PEC to account for indirect installation costs (e.g.,
engineering, construction and field expenses, contractor fees, start-up, a performance test, and
contingencies). Thus, total capital costs (TCC) are estimated to equal about 1.9 times as much as
3-6
-------
the equipment costs, i.e., TCC = EC(1.18)(1.3) + EC(1.18)(1.31) = EC(1.9) (Alpha Gamma,
2001b).
To calculate the annualized control costs for each model engine, the direct and indirect
annualized costs were calculated. Direct annual costs (DCC) were calculated as $71.30 plus
$5/hp for maintenance based on information from vendors. Indirect annualized costs were
estimated as 60 percent of maintenance costs for overhead plus 4 percent of TCC for property
tax, insurance, and administrative charges plus the annualized capital costs based on an interest
rate of 7 percent amortized over 10 years (annualized cost = 1(^1+1^—TCC, where /' is the
(i+i)n-i
interest rate and n is the equipment life). The annualized direct and indirect costs were then
summed to estimate total annualized compliance costs (Alpha Gamma, 200Ib).
For example, the 600 hp Clark RA6 2SLB has a control equipment cost of $7,000
according to the vendor providing the information. The total estimated capital cost to control
emissions from this engine model is then 1.9 times $7,000, or $13,300. Annualizing this capital
cost over 10 years at 7 percent yields an annualized capital cost of $1,894. Annual maintenance
costs for this engine are $71.30 plus $5 times 600 hp, which comes to $3,071. Overhead on the
maintenance costs are 60 percent of $3,071, or $1,843. Finally, annual costs for tax, insurance,
and administrative charges are estimated to be 4 percent of the total capital costs ($13,300),
which is approximately $532. Overall, annualized control costs for this type of engine are
estimated to be $7,339. Table 3-2 presents the annualized control costs estimated for each of the
engine models with available information.
The average annualized control cost per hp was then calculated for 2SLB, 4SLB, 4SRB,
and CI engines by averaging the estimated annualized control cost per hp across three to five
sample engines in each category, as shown in Table 3-2. Based on the engines included in the
sample, the average annualized control cost is approximately $12/hp for 2SLB, $11/hp for 4SLB,
$14/hp for 4SRB, and $11/hp for CI engines (Alpha Gamma, 2002a).
5-7
-------
Table 3-2. Control Costs Associated with Model Engines
Model Engines
Clark RA6
Cooper Bessemer GMV10
Cooper Bessemer GMV10TC
Cooper Bessemer 10V250
Worthington ML20
2SLB Average:
Caterpillar 3 5 12
Caterpillar 3 5 12
Waukesha 7042 GL
Cooper Bessemer LSV16G
4SLB Average:
Waukesha F3 52 1GSI
Waukesha 7042 G
Waukesha L7042 GSI
4SRB Average:
Detroit 16V71
Caterpillar D3 99
Detroit 12V92
Cummins KTA50
Detroit 16V149
CI Average:
Capital Control
Cost per Model
HP Engine
Rating ($)
600
1100
1350
3800
7500
1000
1220
1478
5200
738
1024
1478
510
750
818
1850
1965
13,299
27,072
30,777
72,003
121,112
14,344
21,325
28,497
84,352
27,833
32,012
40,690
12,102
11,399
13,964
31,775
22,399
Annual
Control Cost
per Model
Engine
($/yr)
7,339
13,851
16,527
43,646
82,202
10,730
13,763
17,135
57,098
11,094
14,144
19,532
6,401
8,193
9,205
20,709
19,919
Capital
Control Cost
per Model
Engine
($ per HP)
22
25
23
19
16
21
14
17
19
16
17
38
31
28
32
24
15
17
17
11
17
Annual Control
Cost
per Model
Engine
($ per HP/yr)
12
13
12
11
11
12
11
11
12
11
11
15
14
13
14
13
11
11
11
10
11
Source: Alpha-Gamma Technologies, Inc.; Memorandum to Sims Roy, U.S. EPA; National Impacts Associated with
Reciprocating Internal Combustion Engines; January, 2002a.
These estimated costs per hp were then used to estimate the annualized control costs for
each of the twelve model engine categories (see Table 3-3). For each model engine, the costs
were calculated by multiplying the average cost per hp for the appropriate engine configuration
by the midpoint of the horsepower range for that model. For instance, the estimated annualized
control cost for a 2SLB engine between 500 and 1,000 hp is 750 hp * $12/hp, which is equal to
$9,000.
-------
In addition to the annualized control costs for RICE, there are monitoring costs
associated with the proposed rule. Costs for several monitoring options were developed for each
of the engine subcategories. The most appropriate method of monitoring was selected for each
of the twelve model engine categories based on cost-effectiveness considerations and the
potential emissions that could result from poorly performing emission controls. Tables 3-4 and
3-5 present the estimated annualized costs of monitoring for each of the options considered and
the option chosen for each model engine category, respectively.
The total annualized compliance costs and monitoring costs calculated for each engine
model were used to estimate costs per engine for each of the 12 model unit categories. The total
annualized cost of control and monitoring for these units ranges between $14,209 and $148,800.
Table 3-6 lists the model types, characteristics, and total costs for each of the 12 unit categories.
All affected engines that have capacities between 500 and 1,000 hp have estimated costs less
than $17,000 per year. Affected engines that have capacities between 1,000 and 5,000 hp have
control and monitoring costs between $38,959 and $48,496 per year. Affected engines with
capacities greater than 5,000 hp have annualized control and monitoring costs greater than
$125,000 per year. Based on the proportion of each model number identified in the Inventory
Database, the mean cost expected per affected new engine is $34,366 and the median is $38,959.
The unit-level cost elements were then summed to determine costs at the facility- and parent
firm-levels.
5-9
-------
Table 3-3. Control Costs Associated with Existing and New RICE
Engine Subcategory
Existing Engines'1
4SRB Stationary RICE
New Engines'1
2SLB Stationary RICE
4SLB Stationary RICE
4SRB Stationary RICE
CI Stationary RICE
HP Range3
500-1,000
1,000-5,000
5,000-10,000
500-1,000
1,000-5,000
5,000-10,000
500-1,000
1,000-5,000
5,000-10,000
500-1,000
1,000-5,000
5,000-10,000
500-1,000
1,000-5,000
5,000-10,000
Total # Engines
Affected
(2005)"
3,353
1,215
5
500
0
0
2,124
3,412
12
1,858
2,417
8
5,987
3,991
0
Average
HP
750
3000
7,500
750
3000
7,500
750
3000
7,500
750
3,000
7,500
750
3,000
7,500
Control Cost
per Engine0
($/engine)
24,000
96,000
240,000
15,750
63,000e
157,500e
12,750
51,000
127,500
24,000
96,000
240,000
12,750
51,000
127,500d
Annualized
Control Cost
per Engine0
($/yr)
10,500
42,000
105,000
9,000
36,000e
90,000e
8,250
33,000
82,500
10,500
42,000
105,000
8,250
33,000
82,500d
Source: Alpha-Gamma Technologies, Inc.; Memorandum to Sims Roy, U.S. EPA; National Impacts Associated with
Reciprocating Internal Combustion Engines; January, 2002a.
a There are no existing stationary RICE greater than 10,000 HP, and the presented population excludes emergency power units
and engines 500 HP or less.
b Control costs are calculated using the average HP for the HP range in question, multiplied times the average control cost in
$ per HP, obtained from Table 3-2.
c The only engines affected are those existing 4SRB and new RICE that are or will be located at major sources. The number of
affected sources was rounded to the nearest integer in this table for presentation, but fractional engines were used in
calculations.
d It was estimated that 3 percent of 4SLB and 27 percent of 4SRB engines would be controlled in the absence of the regulation
(no 2SLB or CI engines are projected to be controlled). These engines would not incur control costs under the RICE
NESHAP.
e These values are the estimated annualized control costs that would be incurred if any units in these subcategories were to
comply with the RICE NESHAP. However, there are projected to be no new engines in these subcategories by 2005.
3-10
-------
Table 3-4. Costs of Monitoring for RICE Subcategories
Engine Subcategory
2SLB Stationary RICE
4SLB Stationary RICE
4SRB Stationary RICE
CI Stationary RICE
Monitoring Option3
Option 1
Option 2
Option 3
Option 4
Option 1
Option 2
Option 3
Option 4
Option 5
Option 6
Option 1
Option 2
Option 3
Option 4
Monitoring Capital Cost
($/engine)
208,900
5,699
13,479
13,479
208,900
5,699
13,479
13,479
5,699
5,699
208,900
5,699
13,479
13,479
Total Annualized
Monitoring Cost
($/engine)
58,800
21,618
5,959
3,938
58,800
21,618
5,959
3,938
21,618
6,496
58,800
21,618
5,959
3,938
Source: Alpha-Gamma Technologies, Inc.; Memorandum to Sims Roy, U.S. EPA; National Impacts Associated with
Reciprocating Internal Combustion Engines; January, 2002a.
a Monitoring costs are independent of engine horsepower.
Option 1: CEMforCO.
Option 2: Semi-annual stack testing for CO using Method 10A and continuous parametric monitoring (catalyst pressure drop
and temperature).
Option 3: Quarterly stack testing using portable CO monitor (ASTM D6522-00) and continuous parametric monitoring
(catalyst pressure and temperature).
Option 4: Initial stack testing using portable CO monitor (ASTM D6522-00) and continuous parametric monitoring (catalyst
pressure and temperature).
Option 5: Annual stack testing for formaldehyde (FTIR or CARB 430) and continuous parametric monitoring (catalyst
pressure and temperature).
Option 6: Initial stack testing for formaldehyde (FTIR or CARB 430) and continuous parametric monitoring (catalyst
pressure and temperature).
3-11
-------
Table 3-5. Monitoring Option Applied to RICE Model Engine Categories
Engine Subcategory
2SLB Stationary RICE
4SLB Stationary RICE
4SRB Stationary RICE
CI Stationary RICE
HP Range
500-1,000
1,000-5,000
5,000-10,000
500-1,000
1,000-5,000
5,000-10,000
500-1,000
1,000-5,000
5,000-10,000
500-1,000
1,000-5,000
5,000-10,000
Monitoring
Option
Selected
Option 3
Option 3
Option 1
Option 3
Option 3
Option 1
Option 6
Option 6
Option 5
Option 3
Option 3
Option 1
Monitoring
Capital Cost
($/engine)
13,479
13,479
208,900
13,479
13,479a
208,900a
5,699
5,699
5,699
13,479
13,479
208,900a
Total Annualized
Monitoring Cost
($/engine)
5,959
5,959
58,800
5,959
5,959a
58,800a
6,496
6,496
21,618
5,959
5,959
58,800a
These values are the estimated monitoring costs that would be incurred if any units in these subcategories were to comply
with the RICE NESHAP. However, there are projected to be no new engines in these subcategories by 2005.
Option 1: CEMforCO.
Option 2: Semi-annual stack testing for CO using Method 10A and continuous parametric monitoring (catalyst pressure drop
and temperature).
Option 3: Quarterly stack testing using portable CO monitor (ASTM D6522-00) and continuous parametric monitoring
(catalyst pressure and temperature).
Option 4: Initial stack testing using portable CO monitor (ASTM D6522-00) and continuous parametric monitoring (catalyst
pressure and temperature).
Option 5: Annual stack testing for formaldehyde (FTIR or CARB 430) and continuous parametric monitoring (catalyst
pressure and temperature).
Option 6: Initial stack testing for formaldehyde (FTIR or CARB 430) and continuous parametric monitoring (catalyst
pressure and temperature).
3-12
-------
Table 3-6. Total Annualized Control Cost for Affected Units
Model
Number
1
2
3
4
5
6
7
8
9
10
11
12
Engine
Configuration
2SLB
2SLB
2SLB
4SLB
4SLB
4SLB
4SRB
4SRB
4SRB
CI
CI
CI
Fuel Type
Natural gas
Natural gas
Natural gas
Natural gas
Natural gas
Natural gas
Natural gas
Natural gas
Natural gas
Diesel
Diesel
Diesel
Hp Range
500 to 1,000
1,000 to 5,000
5,000 to 10,000
500 to 1,000
1,000 to 5,000
5,000 to 10,000
500 to 1,000
1,000 to 5,000
5,000 to 10,000
500 to 1,000
1,000 to 5,000
5,000 to 10,000
Annualized
Control Cost
$9,000
$36,000
$90,000
$8,250
$33,000
$82,500
$10,500
$42,000
$105,000
$8,250
$33,000
$82,500
Annual
Monitoring Average Total
Cost Annualized Cost
$5,959
$5,959
$58,800
$5,959
$5,959
$58,800
$6,496
$6,496
$21,618
$5,959
$5,959
$58,800
$14,959
$41,959
$148,800
$14,209
$38,959
$141,300
$16,996
$48,496
$126,618
$14,209
$38,959
$141,300
-------
Because the baseline emissions per engine, percentage reduction in emissions that will be
achieved under the proposed rule, and the annualized control cost differ between engine models,
the cost-effectiveness of HAP reductions will also differ between engine model categories.
Table 3-7 presents estimates of the cost-effectiveness for each RICE model engine category
affected by the RICE NESHAP. Controlling emissions from 4SLB is the most cost-
effectiveness, whereas reducing emissions from CI engines is the least cost-effective. In each
subcategory, emission reductions are achieved at the lowest cost per ton of HAP in the 1,000 to
5,000 hp engine size range.
Table 3-7. Cost Effectiveness for Each Model Engine Category
Total Cost per HAP Emission Reduction Cost Effectiveness
Engine($/year) per Engine (ton/year) ($/ton)
New 2SLB
500-1,000 HP
1,000-5,000 HP
5,000-10,000 HP
New 4SLB
500-1,000 HP
1,000-5,000 HP
5,000-10,000 HP
New and Existing 4SRB
500-1,000 HP
1,000-5,000 HP
5,000-10,000 HP
NewCI
500-1,000 HP
1,000-5,000 HP
5,000-10,000 HP
14,959
41,959
148,800
14,209
38,959
141,300
16,996
48,496
126,618
14,209
38,959
141,300
0.71
2.84
7.11
1.08
4.31
10.77
0.23
0.93
2.33
0.05
0.18
0.45
21,039
14,754
20,928
13,189
9,040a
13,115a
72,807
51,937
54,241
314,674
215,697
312,924a
Source: Calculations by Alpha-Gamma Technologies based on information contained in Alpha-Gamma Technologies, Inc.;
Memorandum to Sims Roy, U.S. EPA; National Impacts Associated with Reciprocating Internal Combustion Engines;
January, 2002a.
a These values are the estimated cost-effectiveness that would be achieved if any of these units were to comply with the RICE
NESHAP. However, there are projected to be no new engines in these subcategories by 2005.
3.4 PROFILE OF RICE UNITS AND FACILITIES IN INVENTORY DATABASE
3.4.1 Affected Units
3-14
-------
Engines in the Inventory Database range in capacity from 500 to 8,000 hp. Despite the
presence of units with horsepower capacity of 5,000 or more, the vast majority of units are less
than 1,500 hp (see Figure 3-1). About 80 percent of the Inventory units, 2,088 engines, have
capacities less than 1,500 hp. More than half of those engines have less than 1,000 hp. Only
557 units are greater than 1,500 hp.
About two-thirds of the units in the Inventory Database are described as lean-burn units
(see Figure 3-2). All of the rich-burn units are four-stroke; the lean-burn units are split fairly
evenly between two-stroke and four-stroke configurations. Also, 95 percent of the units use
natural gas for fuel (only about 5 percent are CI units).
1,400
1
c
M-
o
E
3
500 to 999 1,000 to
1,499
1,500 to
1,999
2,000 to
2,499
2,500 to
2,999
>3,000
Capacity Range (hp)
Figure 3-1. Capacity Ranges for Engines in the Inventory Database
3-15
-------
Engine Configuration
Fuel Type
2SLB
4SRB
34%
Natural
Gas
95%
4SLB
35%
Figure 3-2. Characteristics of Engines in Inventory Database
3.4.2 Affected Facilities
The 2,645 units in the Inventory Database for which sufficient identifying information is
available are located at 834 facilities. Table 3-8 presents the distribution of units and facilities
by industry grouping. Most of the Inventory Database units are concentrated in two industries:
oil and gas extraction and pipeline transportation. These units are for the most part located at
compression stations on natural gas pipelines or at oil and gas fields and plants. The only other
industries with relatively sizable numbers of units at the three-digitNAICS code level are the
mining (except oil and gas) industry (NAICS 212), hospitals (NAICS 622), and electric utilities
(NAICS 221).
3.5 PROJECTED GROWTH OF RICE
The Agency estimates that, without the rule, the United States will have 20,309 new
RICE engines with horsepower greater than 500 (that are not used as backup/emergency units)
by 2005 (see Table 2-5). These estimates are based on the expected growth in the number of
engines in each of the 12 model categories listed in Table 3-9. All growth estimates are based on
information provided by the EPA Office of Mobile Sources (now the Office of Transportation
3-16
-------
Table 3-8. Number of Units With Assigned Model Numbers, the Number of Facilities
at Which They are Located, and the Average Number of Units per Facility,
by Industry in the Inventory Database"
NAICS
112
211
212
221
234
311
312
322
324
325
326
327
331
421
441
486
488
524
531
541
562
611
622
922
Unknown
Total
Industry Description
Animal Production
Oil and Gas Extraction
Mining (Except Oil and Gas)
Utilities
Heavy Construction
Food Manufacturing
Beverage and Tobacco Product Manufacturing
Paper Manufacturing
Petroleum and Coal Products Manufacturing
Chemical Manufacturing
Plastics and Rubber Products Manufacturing
Nonmetallic Mineral Product Manufacturing
Primary Metal Manufacturing
Wholesale Trade, Durable Goods
Motor Vehicle and Parts Dealers
Pipeline Transportation
Support Activities for Transportation
Insurance Carriers and Related Activities
Real Estate
Professional, Scientific, and Technical Services
Waste management and Remediation Services
Educational Services
Hospitals
Justice, Public Order, and Safety Activities
Industry Classification Unknown
Source: Industrial Combustion Coordinated Rulemaking (ICCR). 1998.
Number of
Units
1
1,148
33
35
1
15
9
1
11
16
1
1
3
1
4
1,282
1
5
1
13
2
1
36
4
20
2,645
Number of
Facilities
1
312
28
15
1
4
1
1
7
4
1
1
1
1
1
424
1
3
1
1
1
1
20
1
2
834
Data/Information Submitted to the
Average
Number of
Units Per
Facility
1.0
3.7
1.2
2.3
1.0
3.8
9.0
1.0
1.6
4.0
1.0
1.0
3.0
1.0
4.0
3.0
1.0
1.7
1.0
13.0
2.0
1.0
1.8
4.0
10.0
3.1
Coordinating
Committee at the Final Meeting of the Industrial Combustion Coordinated Rulemaking Federal Advisory Committee.
EPA Docket Numbers A-94-63, II-K-4b2 through -4b5. Research Triangle Park, North Carolina. September 16-17.
Although there are a total of 26,832 engines in the Inventory Database, only 2,645 of these units are potentially affected by
the rule (i.e., they are greater than 500 hp and are not emergency/backup units) and have enough information to assign a
model number. These are the units in the Inventory Database that serve as the basis for assigning compliance costs by
industry.
3-17
-------
Table 3-9. Population Estimates of Affected RICE Units, 2005a
Model
Number
1
2
3
4
5
6
7
8
9
OJ
5
11
12
Total
Engine
Configuration
2SLB
2SLB
2SLB
4SLB
4SLB
4SLB
4SRB
4SRB
4SRB
CI
CI
CI
Units in
Inventory
Database with
Model #"
259
500
57
170
608
37
650
238
1
63
60
2
2,645
Total
Existing
Affected
Units0
0
0
0
0
0
0
1,341
486
2
0
0
0
1,829
Existing 5-year Growth in
Affected Total Affected
Uncontrolled 5-year Growth in Uncontrolled Total Affected
Units Affected Units Unitsd Units (2,005)
0
0
0
0
0
0
979
355
1
0
0
0
1,335
200
0
0
850
1,365
5
743
967
o
J
2,395
1,596
0
8,124
200
0
0
824
1,324
5
542
706
2
2,395
1,596
0
7,594
200
0
0
850
1,365
5
2,084
1,453
5
2,395
1,596
0
9,953
Source: Industrial Combustion Coordinated Rulemaking (ICCR). 1998. Data/Information Submitted to the Coordinating Committee at the Final Meeting of the
Industrial Combustion Coordinated Rulemaking Federal Advisory Committee. EPA Docket Numbers A-94-63, II-K-4b2 through -4b5. Research Triangle Park,
North Carolina. September 16-17.
a The only RICE directly affected by the rule are those existing 4SRB and new RICE that are or will be located at major sources. The number of units was rounded to the
nearest integer in this table for presentation, but fractional engines were used in calculations.
b Only the engines in the Inventory Database with sufficient information to assign a model number were used in estimating costs by industry.
c The only existing engines affected by this rule are 4SRB engines, some of which are already controlled in the absence of this rule. Monitoring costs due to the rule
apply to all of the 4SRB engines, even those already controlled.
d It is assumed that 27 percent of new 4SRB and 3 percent of new 4SLB engines would be controlled in the absence of this regulation. Therefore, the costs of controls for
these engines are not included in the total cost of the regulation. However, the monitoring costs incurred by all of these engines due to the rule are included in
calculating the total cost.
-------
and Air Quality) regarding estimated five year sales volume for engines, which was derived from
the Power Systems Research database, and confidential sales projection information provided to
EPA by engine manufacturers. However, not all of these engines will be affected by the RICE
NESHAP because it only applies to RICE located at major sources. The percentage of sources
that are major in the natural gas prime mover (60 percent), crude petroleum and natural gas (33
percent), and electric services (100 percent) sectors were estimated by obtaining information
from representative industry organizations (Alpha Gamma, 200la). Estimates for the percentage
of engines owned by the Department of Defense that are located at major sources (31 percent)
were obtained from a representative of the Naval Facilities Engineering Service Center and EPA
assumed that only 25 percent of all other engines would be located at major sources (Alpha
Gamma, 200la).
EPA calculated the overall percentage of existing engines at major sources based on the
percentage of existing engines owned by each of these five segments (Department of Defense,
13 percent; natural gas prime movers, 25 percent; crude petroleum and natural gas, 33 percent;
electric services, 5 percent; and other miscellaneous, 24 percent) and the percentage of those
existing engines estimated to be major sources. Using this method, the percentage of RICE
located at major sources is estimated to be approximately 40 percent (Alpha Gamma, 200la).
Based on an assumption that the proportion of existing engines located at major sources is a
good approximation for the percentage of future engines that will be located at major sources,
EPA assumed that only 40 percent of RICE engines subject to the proposed rule that will be
installed in the future will incur compliance costs.
Thus, the Agency estimates that the U.S. will have 8,124 new 1C engines with
horsepower greater than 500 by the end of 2005 that will be affected by the rule (see Table 2-5)
based on the assumption that 40 percent of new RICE would be located at major sources. Table
3-9 lists several unit counts: units in the Inventory Database with assigned model numbers,
existing affected units, and projected unit growth over 5 years. The latter two categories are also
broken out by the total number of units and the number of units that would have been controlled
regardless of the rule.
Existing 2SLB engines (model numbers 1, 2, and 3) are not affected by the rule. As new
2SLB units come online, however, they will be required to install the requisite control equipment
3-19
-------
and operators will have to adhere to monitoring requirements. It is estimated that 200 new 2SLB
engines of greater than 500 hp will have come into operation at major sources by the end of
2005, none of which are expected to be greater than 1,000 hp.
Existing 4SLB engines (model numbers 4, 5, and 6) are also not affected by this rule. In
the absence of this rule, it is expected that 3 percent of new units would come online controlled
in the future based on the percentage of units currently controlled (Alpha Gamma, 2002a).
Therefore, only the remaining 97 percent of units located at major sources (2,152 of 2,219 units)
will have control costs associated with the rule. The cost of controlling the additional remaining
3 percent was not included in the rule's cost because it would have been borne by industry
regardless of the rule; the rule will not affect those business decisions. However, all 2,219 new
4SLB engines located at major sources will incur monitoring costs. It is expected that very few
of these units will be greater than 5,000 hp.
The only existing engines that are affected by the rule are 4SRB engines (model numbers
7, 8, and 9). Those engines that are located at major sources and not already controlled, 1,335
units, will have to install control equipment. All existing 4SRB engines located at major sources
(1,829 units) must comply with the monitoring component of the rule. For new sources, the
Agency estimates that 27 percent (463 units) would come online controlled without the rule
based on the current population of 4SRB engines (Alpha Gamma, 2000). Thus, control costs for
these units are not included in the total cost of the rule. However, all 1,713 units projected to
enter into operation at major sources by the end of 2005 will incur monitoring costs. Most
existing units are less than 1,000 hp, but the majority of new units are expected to be between
1,000 and 5,000 hp.
Similar to 2SLB and 4SLB engines, only new CI engines (model numbers 10, 11, and 12)
will be affected by this rule. Existing CI engines do not have to add any controls. None of these
engines are projected to be controlled in the absence of regulation. Therefore, all 3,991 units
estimated to enter into operation at major sources by the end of 2005 will be subject to both
control and monitoring costs under the regulation. About 60 percent of these units are expected
to be under 1,000 hp; no units are expected to be greater than 5,000 hp.
3-20
-------
3.5.1 Grow th Estimates by Industry
Although growth estimates by engine configuration and horsepower are available,
estimates of the growth in the number of units by industry are not. To assess the distribution of
the engines estimated to be operating in 2005 across industries, it was assumed that the
distribution of each model engine number across industries for the units in the Inventory
Database with assigned model numbers is representative of the distribution of future units across
industries. This distribution was then used to estimate the number of affected engines that would
be added in each industry by 2005.
3.5.1.1 Mapping SIC Codes to NAICS Codes
Although the economic analysis was originally conducted based on SIC-level costs, the
SIC information included with affected unit and facility records in the Inventory Database was
later complemented with the appropriate NAICS code to reflect the change in industry
classification that has occurred in recent years. The original 4-digit SIC codes for these units
and facilities were mapped to corresponding 3-digit NAICS code (3-digit NAICS codes are the
functional equivalent of 2-digit SIC codes, the highest level of detail often shown in economic
analyses). The 1997 NAICS and 1987 SIC Correspondence Tables prepared by the Bureau of
the Census were used to determine the matching NAICS codes.2 The process of mapping SIC
codes to NAICS codes was relatively straightforward because, although there are 2,645 RICE
units in the Inventory Database with sufficient information to assign model engine numbers,
three 4-digit SIC codes accounted for more than 91 percent of the units:
1,268 units in SIC 4922 ("Natural Gas Transmission") were mapped to NAICS
486 ("Pipeline Transportation").
601 units in SIC 1321 ("Natural Gas Liquids") were mapped to NAICS 211 ("Oil
and Gas Extraction").
• 543 units in SIC 1311 ("Crude Petroleum and Natural Gas") were mapped to
NAICS 211 ("Oil and Gas Extraction").
2The 1997 NAICS and 1987 SIC Correspondence Tables can be viewed on the Bureau of the Census website at
http://www.census.gov/epcd/www/naicstab.htm.
3-21
-------
Overall, there were 47 different 4-digit SIC codes in the database, with all of them having well-
defined corresponding 3-digit NAICS codes. There were no instances where a 4-digit SIC code
was divided into two separate NAICS codes. Thus, the assignment of costs at the NAICS level
yields very similar costs by industry to those achieved using SIC codes (as well as very similar
results), but is consistent with the recent movement towards using NAICS codes in regulatory
analyses.
3.5.1.2 Data Extrapolation to Projected National Unit Estimates by Industry
The Inventory Database contains information on type of engine (e.g., 2SLB, 4SLB,
4SRB, CI), engine size (hp), and SIC code, among other data. As discussed above, a column
containing the 3-digit NAICS code was added by mapping SIC codes to their corresponding
NAICS classifications. To develop national economic impact estimates by industry based on the
subset of units with sufficient data included in the Inventory Database, national unit population
estimates (Alpha Gamma, 2002a) for both existing and new sources in 2005 were used.
However, these estimates were provided for 12 model engines (defined by engine type and size),
not by industry. Therefore, the industry classification of units in the Inventory Database was
used to estimate the distribution of the RICE population estimates across industries.
The projected distribution of engines by industry was based on the current distribution in
the Inventory Database. For example, it was estimated that 500 units of engine model 1 (2SLB,
500 to 1,000 hp) will be added by 2005 (Alpha Gamma, 2002a), with 200 units located at major
sources. There are 259 units identified as model 1 in the Inventory Database. Therefore, for
each model 1 unit that is included in the database for a particular industry, it was assumed that
1.931 model 1 units (i.e., 500/259) would be added in that industry by 2005. In other words, it
was assumed that the current distribution of each model engine across industries, as reported in
the Inventory Database, is representative of the future distribution of each model engine category
across industries. For instance, the database included 122 model 1 engines in NAICS 486, 131
in NAICS 211, 2 in NAICS 311, and 4 in NAICS 541. Therefore, the projected distribution of
the 500 model 1 engines projected to be added by 2005 was approximately 235.6 in NAICS 486,
253.0 in NAICS 211, 3.9 in NAICS 311, and 7.7 in NAICS 541. It was assumed that 40 percent
3-22
-------
of the engines in each NAICS code would be located at major sources and would be subject to
the rule.
NAICS codes 211 and 486 represent over 91 percent of the units in the Inventory
Database, but only 60 percent of the estimated affected population in 2005. This is due to the
large increase in CI units projected and the extremely small share of CI units that are in these
two NAICS codes based on the Inventory Database. For example, there are 63 engines that are
model 10 (CI, 500 to 1,000 hp) in the database, but only 1 (1.6 percent) is in NAICS 211 and 3
(4.8 percent) are in NAICS 486. It was projected that a total of 2,395 affected model 10 engines
will be added by 2005 (24 percent of total affected engines) (Alpha Gamma, 2002a), but very
few are projected to be in NAICS codes 211 or 486. Overall, 49 percent of new affected units
are projected to be CI units (3,991 CI units/8,124 total projected units) with NAICS codes 211
and 486 accounting for only 0.8 percent and 4.8 percent, respectively.
The total number of affected units estimated to exist in 2005 by industry is presented in
Table 3-10. The third column lists the number of units in the Inventory Database with assigned
model numbers (the units that served as the basis for cost estimates by industry). The fourth
column presents the estimated population of affected engines projected by industry for 2005.
3-23
-------
Table 3-10. Affected RICE Population and Engineering Costs by NAICS Code, 2005
NAICS
112
211
212
221
234
311
312
322
324
325
326
327
331
421
441
486
488
524
531
541
562
611
622
922
Unknown
Total
Industry Description
Animal Production
Oil and Gas Extraction
Mining (Except Oil and Gas)
Utilities
Heavy Construction
Food Manufacturing
Beverage and Tobacco Product
Manufacturing
Paper Manufacturing
Petroleum and Coal Products
Manufacturing
Chemical Manufacturing
Plastics and Rubber Products
Manufacturing
Nonmetallic Mineral Product
Manufacturing
Primary Metal Manufacturing
Wholesale Trade, Durable Goods
Motor Vehicle and Parts Dealers
Pipeline Transportation
Support Activities for Transportation
Insurance Carriers and Related Activities
Real Estate
Professional, Scientific, and Technical
Services
Waste management and Remediation
Services
Educational Services
Hospitals
Justice, Public Order, and Safety
Activities
Industry Classification Unknown
Source: Industrial Combustion Coordinated Rulemaking (ICCR).
Number of Units in
Inventory Database
with Model #a
1
1,148
33
35
1
15
9
1
11
16
1
1
3
1
4
1,282
1
5
1
13
2
1
36
4
20
2.645
Estimated 2005
Affected
Populationb
3
2,875
1,032
859
—
63
31
27
148
173
27
38
7
38
13
3,110
3
86
38
9
53
27
1,163
129
3
9.953
1998. Data/Information Submitted to the Coordinating
Annualized
Engineering
Costs (1998$)
45,411
71,102,348
20,401,095
25,707,611
—
1,971,951
629,936
1,036,633
2,811,969
4,469,266
1,036,633
540,111
255,691
540,111
181,645
80,076,833
45,411
3,200,721
540,111
273,032
2,073,266
1,036,633
26,397,114
3,153,487
45,411
247.572.429
I Committee at the
Final Meeting of the Industrial Combustion Coordinated Rulemaking Federal Advisory Committee. EPA Docket Numbers A-94-63,
II-K-4b2 through -4b5. Research Triangle Park, North Carolina. September 16-17.
Although there are a total of 26,832 engines in the Inventory Database, only 2,645 of these units are potentially affected by
the rule (i.e., they are greater than 500 hp and are not emergency/backup units) and have enough information to assign a
model number. These are the units in the Inventory Database that serve as the basis for assigning compliance costs by
industry.
3.5.2 Engineering Compliance Costs
3-24
-------
Based on the projected distribution of each model engine type across industries, total
annualized costs were estimated by multiplying the projected number of affected engines in each
model engine category by the annualized compliance cost per engine for that model engine type.
This calculation was performed for each industry as follows:
f
12 12
(3.1)
9S COW • CON • UNC • UNC •
+ AFF * ACC
where TACCj is the total annualized compliance cost for industry j (there are 25 industry
categories in the model), i = 1,...,12 represents the model engine categories, n^ is the number of
engines of model type i used in industry j that are included in the Inventory Database and have
sufficient information available to assign them a model number, AFFCONi is the number of
affected engines of model type i projected to exist in 2005 that would be controlled in the
absence of the RICE NESHAP, ACCCONi represents the annualized compliance cost for a single
engine of model type i that would be controlled in the absence of the RICE NESHAP3, and
AFFuNc,; and ACC^c; are the measures for RICE that would be uncontrolled in the absence of
the NESHAP corresponding to AFFCONi and ACCCONi. As an example of the calculation of total
annualized costs for an industry, the calculations used in estimating the total annualized costs of
the RICE NESHAP for NAICS 211 are described below.
3.5.2.1 Sample Industry Cost Calculation: NAICS 211
RICE in the Inventory Database that were identified as being used in SIC codes 1311
(Oil and Gas Extraction) and 1321 (Natural Gas Liquids) were mapped into NAICS 211. In the
Inventory Database, there are 1,148 units identified that were placed in this NAICS code. They
are distributed among model engine types as shown in Table 3-11 (column 2). Compliance costs
3It was estimated that 0 percent of 2SLB, 3 percent of 4SLB, 27 percent of 4SRB, and 0 percent of CI engines would
be controlled in the absence of the RICE NESHAP (Alpha Gamma, 2002a). The engines that would be
controlled in the absence of the NESHAP still have compliance costs associated with the rule because they are
subject to monitoring requirements.
3-25
-------
forNAICS 211 were estimated by applying equation (3.1) to the data contained in columns 1
through 4 of Table 3-11.
For example, the total annualized compliance cost for NAICS 211 to upgrade model 1
25
engines was calculated as follows. ForNAICS 211, n1211 = 131 and 5Xj =259. Because there
j=i
are projected to be no model 1 engines that would be controlled in the absence of this regulation,
AFFCON j is equal to zero. For model 1 engines that would be uncontrolled in the baseline, the
annualized cost per engine, ACCl5 was estimated to be $14,959 (Alpha Gamma, 2002a). The
total number of affected model 1 engines that would be uncontrolled in the baseline, AFF^^ is
estimated to be 200 (see Table 2-5). Thus, the cost to NAICS 211 of controlling model 1
engines, TACCU11, is equal to 131/259*[200*$14,959+0*$5,959], or $1,513,227.
Using similar calculations for each model engine type and summing across all 12 model
engine types yields the total projected cost to NAICS 211. That total is estimated to be
$71,102,348, as reported in Tables 3-10 and 3-11.
3.5.2.2 National Engineering Compliance Costs
Based on the projections in Table 3-10 of the affected RICE population, the engineering
control costs of this regulation would be $247.6 million in 2005. These costs are inputs into the
market model used in Section 5 to estimate the changes in price and quantity taking place in each
affected market as a result of the regulation as well as the social costs of the rule. The
magnitude and distribution of the regulatory costs' impact on the economy depend on the
relative size of the impact on individual markets (relative shift of the market supply curves) and
the behavioral responses of producers and consumers in each market (as measured by the
elasticity of supply and the elasticity of demand). To the extent that the projections by engine
model are inaccurate, the Inventory Database is not representative of the current distribution of
engines, and/or the distribution of future affected engines across industries will differ from the
current distribution, the actual costs experienced across industries may differ from those
projected. In addition, there are costs for reporting and record keeping totalling $6.1 million that
are not included in the economic model.
3-26
-------
Table 3-11. Sample Cost Calculation: Estimating Compliance Costs for NAICS 211
Engine
Model
(0
1
2
3
4
5
6
7
8
9
10
11
12
Total
Engines in
Inventory
Database
(NAICS 21 1/
Total)
("i.21/' "y)
131/259
257/500
6/57
66/170
184/608
11/37
349/650
142/238
1/1
1/63
0/60
0/2
1,148/2,645
Projected Number of
Affected Engines
(2005)a (Uncontrolled/
Controlled in Baseline)
(Ar r ujjQi/ Ar r CON,I)
200/0
0/0
0/0
824/25
1,324/41
5/0
1,522/563
1,061/392
4/1
2,395/0
1,596/0
0/0
8,930/1,023
Cost Per Affected
Engine (Uncontrolled/
Controlled in Baseline)
(ACCujjcyACCcoN.i)
$14,959/$5,959
$41,959/$5,959
$148,800/$58,800
$14,209/$5,959
$38,959/$5,959
$141,300/$58,800
$16,996/$6,496
$48,496/$6,496
$126,618/$21,618
$14,209/$5,949
$38,959/$5,949
$141,300/$58,800
NA
Projected Cost
for NAICS 211
by Model Engine
Category
(TACCU11)
$1,513,227
$0
$0
$4,605,127
$15,682,396
$198,107
$15,848,536
$32,209,416
$505,430
$540,111
$0
$0
$71,102,348
Note: The number of engines has been rounded to the nearest integer for presentation. However, fractional engines were
used in calculations. Thus, applying equation (3.1) using the values in columns 1 through 4 may not yield the exact
cost presented in column 5 due to rounding.
3-27
-------
4.0 PROFILES OF AFFECTED INDUSTRIES
This section contains profiles of the industries most directly affected by the proposed
regulation of RICE units. Most existing engines that would be subject to the regulation are
concentrated in two industries, oil and natural gas extraction (NAICS 211) and natural gas
pipeline transportation (NAICS 4862). Together, they account for over 90 percent of the engines
identified by EPA in the Inventory Database that would fall under this rule. (The remaining
units are spread across various industries, most notably mining, hospitals, and various
manufacturing industries, such as food manufacturing and chemical manufacturing.) Most new
engines that would be affected by this regulation are also projected to be in these industries.
The oil and natural gas industry is divided into five distinct sectors: (1) exploration,
(2) production, (3) transportation, (4) refining, and (5) marketing. The NESHAP considers
controls on the use of RICE units, which are used in this industry primarily to power
compressors used for crude oil and natural gas extraction and natural gas pipeline transportation.
Therefore, this section contains background information on the oil and natural gas extraction
industry and the natural gas transmission industry to help inform the regulatory process.
4.1 CRUDE PETROLEUM AND NATURAL GAS (NAICS 211)
The crude petroleum and natural gas industry encompasses the oil and gas extraction
process from the exploration for oil and natural gas deposits through the transportation of the
product from the production site. The primary products of this industry are natural gas, natural
gas liquids, and crude petroleum.
4-1
-------
4.1.1 Introduction
The U.S. is home to half of the major oil and gas companies operating around the globe.
Although small firms account for nearly 45 percent of U.S. crude oil and natural gas output, the
domestic oil and gas industry is dominated by 20 integrated petroleum and natural gas refiners
and producers, such as Exxon Mobil, BP Amoco, and Chevron (Lillis, 1998). Despite the
presence of many large global players, the industry experiences a more turbulent business cycle
than most other major U.S. industries. Because oil is an international commodity, the U.S.
production of crude oil is affected by the world crude oil price, the price of alternative fuels, and
existing regulations. Domestic oil production has been falling in recent years. Total U.S. crude
oil production is expected to fall to 5.78 million barrels per day in 2000, the lowest annual U.S.
crude oil output since 1950 (EIA, 2000a). Because the industry imports 60 percent of the crude
oil used as an input into refineries, it is susceptible to fluctuations in crude oil output and prices,
which may be influenced by the Organization of Petroleum Exporting Countries (OPEC).1
In contrast, natural gas markets in the U.S. are competitive and relatively stable.
Domestic natural gas production has been on an upward trend since the mid-1980s. Almost all
natural gas used in the U.S. comes from domestic and Canadian sources.
There are four sub- or related industries to NAICS 211 (see Table 4-1):
•• NAICS 211111: Crude petroluem and natural gas extraction. Firms in this
industry are primarily engaged in (1) the exploration, development and/or the
production of petroleum or natural gas from wells in which the hydrocarbons will
initially flow or can be produced using normal pumping techniques, or (2) the
production of crude petroleum from surface shales or tar sands or from reservoirs
in which the hydrocarbons are semisolids. Establishments in this industry operate
oil and gas wells on their own account or for others on a contract or fee basis.
•• NAICS 211112: Natural gas liquid (NGL) extraction. Firms in this industry are
primarily engaged in the recovery of liquid hydrocarbons from oil and gas field
'OPEC is a cartel consisting of most of the world's largest petroleum-producing countries that attempts to increase
the profits of member countries.
4-2
-------
gases. Establishments primarily engaged in sulfur recovery from natural gas are
included in this industry.
•• NAICS 213111: Drilling oil and gas wells. Firms in this industry are primarily
engaged in drilling oil and gas wells for others on a contract or fee basis. This
industry includes contractors that specialize in spudding in, drilling in, redrilling,
and directional drilling.
NAICS 213112: Support activities for oil and gas operations. Firms in this
industry perform oil and gas field services (except contract drilling) for others, on
a contract or fee basis. Services included are exploration (except geophysical
surveying and mapping); excavating slush pits and cellars; grading and building
foundations at well locations; well surveying; running, cutting, and pulling
casings, tubes, and rods; cementing wells; shooting wells; perforating well
casings; acidizing and chemically treating wells; and cleaning out, bailing, and
swabbing wells.
Table 4-1. Crude Petroleum and Natural Gas Industries Likely to Be Affected
by the Regulation
NAICS Description
211111 Crude Petroleum and Natural Gas Extraction
211112 Natural Gas Liquid Extraction
213111 Drilling Oil and Gas Wells
213112 Support Activities for Oil and Gas Operations
In 1997, more than 6,800 crude oil and natural gas extraction companies (NAICS
211111) generated $75 billion in revenues (see Table 4-2). Revenues for 1997 were
approximately 5 percent higher than revenues in 1992, although the number of companies and
employees declined 11.5 and 42.5 percent, respectively.
-------
Table 4-2. Summary Statistics, Crude Oil and Natural Gas Extraction
and Related Industries
NAICS
211111
211112
213111
213112
Number of
Industry Companies
Crude Oil and
Natural Gas
Extraction
1992
1997
Natural Gas Liquid
Extraction
1992
1997
Drilling Oil and Gas
Wells
1992
1997
Support Activities for
Oil and Gas
Operations
1997
7,688
6,802
108
89
1,698
1,371
6,385
Number of Revenues
Establishments ($1997 103)
9,391
7,781
591
529
2,125
1,638
7,068
71,622,600
75,162,580
26,979,200
24,828,503
3,552,707
7,317,963
11,547,563
Employees
174,300
100,308
12,000
10,549
47,700
53,865
106,339
Sources: U.S. Department of Commerce, Bureau of the Census. 1999a.
Washington, DC: U.S. Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 1995a.
Washington, DC: U.S. Department of Commerce.
1997Economic Census, Mining Industry Series.
1992 Census of Mineral Industries, Industry Series.
Table 4-2 shows the NGL extraction industry (NAICS 211112) experienced a decline in
the number of companies, establishments, and employees between 1992 and 1997. The
industry's revenues declined nearly 8.0 percent during this time, from $27 billion per year to
$24.8 billion per year.
Revenues for NAICS 213111, drilling oil and gas wells, more than doubled between
1992 and 1997. In 1992, the industry employed 47,700 employees at 1,698 companies and
generated $3.6 billion in annual revenues. By the end of 1997, the industry's annual revenues
4-4
-------
were $7.3 billion, a 106 percent improvement. Although the total number of companies and
establishments decreased from 1992 levels, industry employment increased 13 percent to 53,685.
The recent transition from the Standard Industrial Classification (SIC) system to the
North American Industrial Classification System (NAICS) changed how some industries are
organized for information collection purposes and thus how certain economic census data are
aggregated. Some SIC codes were combined, others were separated, and some activities were
classified under one NAICS code and the remaining activities classified under another. The
support activities for oil and gas operations is an example of an industry that was reclassified.
Under NAICS, SIC 1382, Oil and Gas Exploration Services, and SIC 1389, Oil and Gas Services
Not Elsewhere Classified, were combined. The geophysical surveying and mapping services
portion of SIC 1382 was reclassified and grouped into NAICS 54136. The adjustments to SIC
1382/89 have made comparison between the 1992 and 1997 economic censes difficult at this
time. The U.S. Census Bureau has yet to publish a comparison report. Thus, for NAICS 213112
only 1997 census data are presented. For that year, nearly 6,400 companies operated under
NAICS 213112, employing more than 100,000 people and generating $11.5 billion in revenues.
4.1.2 Supply Side Characteristics
Characterizing the supply side of the industry involves describing the production
processes, the types of output, major by-products, costs of production, and capacity utilization.
4-5
-------
4.1.2.1 Production Processes
Domestic production occurs within the contiguous 48 states, Alaska, and at offshore
facilities. There are four major stages in oil and gas extraction: exploration, well development,
production, and site abandonment (EPA, 1999d). Exploration is the search for rock formations
associated with oil and/or natural gas deposits. Nearly all oil and natural gas deposits are located
in sedimentary rock. Certain geological clues, such as porous rock with an overlying layer of
low-permeability rock, help guide exploration companies to a possible source of hydrocarbons.
While exploring a potential site, the firm conducts geophysical prospecting and exploratory
drilling.
After an economically viable field is located, the well development process begins. Well
holes, or well bores, are drilled to a depth of between 1,000 and 30,000 feet, with an average
depth of about 5,500 feet (EPA, 1999d). The drilling procedure is the same for both onshore and
offshore sites. A steel or diamond drill bit, which may be anywhere between 4 inches and 3 feet
in diameter, is used to chip off rock to increase the depth of the hole. The drill bit is connected
to the rock by several pieces of hardened pipe known collectively as the drill string. As the hole
is drilled, casing is placed in the well to stabilize the hole and prevent caving. Drilling fluid is
pumped down through the center of the drill string to lubricate the equipment. The fluid returns
to the surface through the space between the drill string and the rock formation or casing. Once
the well has been drilled, rigging, derricks, and other production equipment are installed.
Onshore fields are equipped with a pad and roads; ships, floating structures, or a fixed platform
are procured for offshore fields.
Production is the process of extracting hydrocarbons through the well and separating
saleable components from water and silt. Oil and natural gas are naturally occurring co-
products, and most production sites produce a combination of oil and gas; however, some wells
produce little natural gas, while others may produce only natural gas. Once the hydrocarbons are
brought to the surface, they are separated into a spectrum of products. Natural gas is separated
from crude oil by passing the hydrocarbons through one or two decreasing pressure chambers.
Crude oil is always delivered to a refinery for processing and excess water is removed, at which
point the oil is about 98 percent pure, a purity sufficient for storage or transport to a refinery
(EPA, 1999b). Natural gas may be processed at the field or at a natural gas processing plant to
4-6
-------
remove impurities. The primary extracted streams and recovered products associated with the
oil and natural gas industry include crude oil, natural gas, condensate, and produced water. The
products are briefly described below.
Crude oil can be classified as paraffinic, naphthenic, or intermediate. Paraffmic (or
heavy) crude is used as an input to the manufacture of lube oils and kerosene. Naphthenic (or
light) crude is used as an input to the manufacture of gasoline and asphalt. Intermediate crudes
are those that do not fit into either category. The classification of crude oil is determined by a
gravity measure developed by the American Petroleum Institute (API). API gravity is a weight
per unit volume measure of a hydrocarbon liquid. A heavy crude is one with an API gravity of
20° or less, and a light crude, which flows freely at atmospheric temperature, usually has an API
gravity in the range of the high 30s to the low 40s (EPA, 1999c).
Natural gas is a mixture of hydrocarbons and varying quantities of nonhydrocarbons that
exist either in gaseous phase or in solution with crude oil from underground reservoirs. Natural
gas may be classified as either wet or dry gas. Wet gas is unprocessed or partially processed
natural gas produced from a reservoir that contains condensable hydrocarbons. Dry gas is
natural gas whose water content has been reduced through dehydration, or natural gas that
contains little or no commercially recoverable liquid hydrocarbons.
Condensates are hydrocarbons that are in a gaseous state under reservoir conditions
(prior to production), but which become liquid during the production process. Condensates have
an API gravity in the 50° to 120° range (EPA, 1999c). According to historical data, Condensates
account for about 4.5 to 5 percent of total crude oil production.
Produced water is recovered from a production well or is separated from the extracted
hydrocarbon streams. More than 90 percent of produced water is reinjected into the well for
disposal and to enhance production by providing increased pressure during extraction. The
remainder is released into surface water or disposed of as waste.
In addition to the products discussed above, other various hydrocarbons may be
recovered through the processing of the extracted streams. These hydrocarbons include mixed
natural gas liquids, natural gasoline, propane, butane, and liquefied petroleum gas.
Natural gas is conditioned using a dehydration and a sweetening process, which removes
hydrogen sulfide and carbon dioxide, so that it is of high enough quality to pass through
4-7
-------
transmission systems. The gas may be conditioned at the field or at one of the 623 operating
gas-processing facilities located in gas-producing states, such as Texas, Louisiana, Oklahoma,
and Wyoming. These plants also produce the nation's NGLs, propane and butane (NGSA et al.,
2000c).
Site abandonment occurs when a site lacks the potential to produce economic quantities
of natural gas or when a production well is no longer economically viable. The well(s) are
plugged using long cement plugs and steel plated caps, and supporting production equipment is
disassembled and moved offsite.
4.1.2.2 Types of Output
The oil and gas industry's principal products are crude oil, natural gas, and NGLs (see
Tables 4-3 and 4-4). Refineries process crude oil into several petroleum products. These
products include motor gasoline (40 percent of crude oil); diesel and home heating oil
(20 percent); jet fuels (10 percent); waxes, asphalts, and other nonfuel products (5 percent);
feedstocks for the petrochemical industry (3 percent); and other lesser products (EIA, 1999a).
Natural gas is produced from either oil wells (known as "associated gas") or wells that
are drilled for the primary objective of obtaining natural gas (known as "nonassociated gas") (see
Table 4-4). Methane is the predominant component of natural gas (about 85 percent), but ethane
(about 10 percent), propane, and butane are also significant components (see Table 4-3).
Propane and butane, the heavier components of natural gas, exist as liquids when cooled and
compressed. These latter two components are usually separated and processed as natural gas
liquids (EPA, 1999d). A small amount of the natural gas produced is consumed as fuel by the
engines used in extracting and transporting the gas, and the remainder is transported through
pipelines for use by residential, commercial, industrial, and electric utility users.
4-8
-------
Table 4-3. U.S. Supply of Crude Oil and Petroleum Products (103 barrels), 1998
Commodity
Crude Oil
Natural Gas Liquids
Ethane/Ethylene
Propane/Propylene
Normal Butane/Butylene
Isobutane/Isobutylene
Other
Other Liquids
Finished Petroleum Products
Finished Motor Gasoline
Finished Aviation Gasoline
Jet Fuel
Kerosene
Distillate Fuel Oil
Residual Fuel Oil
Naptha
Other Oils
Special Napthas
Lubricants
Waxes
Petroleum Coke
Asphalt and Road Oil
Still Gas
Miscellaneous Products
Total
Field
Production
2,281,919
642,202
221,675
187,369
54,093
66,179
112,886
69,477
69,427
69,427
3,063,025
Refinery
Production
245,918
11,444
200,815
29,333
4,326
5,970,090
2,880,521
7,118
556,834
27,848
1,249,881
277,957
89,176
78,858
24,263
67,263
8,355
260,061
181,910
239,539
20,506
6,216,008
Imports
3,177,584
82,081
6,230
50,146
8,612
5,675
11,418
211,266
437,515
113,606
43
45,143
466
76,618
100,537
22,388
61,554
2,671
3,327
613
263
10,183
103
3,908,446
Source: Energy Information Administration. 1999b. Petroleum Supply Annual 1998, Volume I. Washington, DC:
U.S. Department of Energy.
4-9
-------
Table 4-4. U.S. Natural Gas Production, 1998
Gross Withdrawals Production (10 cubic feet)
From Gas Wells 17,558,621
From Oil Wells 6,365,612
Less Losses and Repressuring 5,216,477
Total 18,707,756
Source: Energy Information Administration. 1999b. Natural Gas Annual 1998. Washington, DC: U.S.
Department of Energy.
4.1.2.3 Major By-Products
In addition to the various products of the oil and natural gas extraction process described
above, there are some additional by-products generated during the extraction process. Oil and
natural gas are composed of widely varying constituents and proportions depending on the site of
extraction. The removal and separation of individual hydrocarbons during processing is possible
because of the differing physical properties of the various components. Each component has a
distinctive weight, boiling point, vapor pressure, and other characteristics, making separation
relatively simple. Most natural gas is processed to separate hydrocarbon liquids that are more
valuable as separate products, such as ethane, propane, butane, isobutane, and natural gasoline.
Natural gas may also include water, hydrogen sulfide, carbon dioxide, nitrogen, helium, or other
diluents/contaminants. The water present is either recovered from the well or separated from the
hydrocarbon streams being extracted. More than 90 percent of the produced water is reinjected
into the well to increase pressure during extraction. If hydrogen sulfide, which is poisonous and
corrosive, is present, it is removed and further processed to recover elemental sulfur for
commercial sale. In addition, processing facilities may remove carbon dioxide to prevent
corrosion and to use for injection into the well to increase pressure and enhance oil recovery,
recover helium for commercial sale, and may remove nitrogen to increase the heating value of
the gas (NGSA et al., 2000c). Finally, the engines that provide pumping action at wells and push
crude oil and natural gas through pipes to processing plants, refineries, and storage locations
4-10
-------
produce HAPs. HAPs produced in engines include formaldehyde, acetaldehyde, acrolein, and
methanol.
4.1.2.4 Costs of Production
The 42 percent decrease in the number of people employed by the crude oil and natural
gas extraction industry between 1992 and 1997 was matched by a corresponding 40 percent
decrease in the industry's annual payroll (see Table 4-5). During the same period, industry
outlays for supplies, such as equipment and other supplies, increased over 32 percent, and capital
expenditures nearly doubled. Automation, mergers, and corporate downsizing have made this
industry less labor-intensive (Lillis, 1998).
Unlike the crude oil and gas extraction industry, the NGL extraction industry's payroll
increased over 6 percent even though total industry employment declined 12 percent. The
industry's expenditures on capital projects, such as investments in fields, production facilities,
and other investments, increased 11.4 percent between 1992 and 1997. The cost of supplies did,
however, decrease 13 percent from $23.3 billion in 1992 to $20.3 billion in 1997.
Employment increased in NAICS 213111, Drilling Oil and Gas Wells. In 1992, the
industry employed 47,700 people, increasing 13 percent to 53,685 in 1997. During a period
where industry revenues increased over 100 percent, the industry's payroll increased 41 percent
and the cost of supplies increased 182 percent.
4.1.2.5 Imports and Domestic Capacity Utilization
Domestic annual oil and gas production is a small percentage of total U.S. reserves. In
1998, oil producers extracted approximately 1.5 percent of the nation's proven crude oil reserves
(see Table 4-6). A slightly lesser percentage of natural gas was extracted (1.4 percent), and an
even smaller percentage of NGLs was extracted (0.9 percent). The U.S. produces approximately
40 percent (2,281 million barrels) of its annual crude oil consumption, importing the remainder
of its crude oil from Canada, Latin America, Africa, and the Middle East (3,178 million barrels).
Approximately 17 percent (3,152 billion cubic feet) of U.S. natural gas supply is imported. Most
imported natural gas originates in Canadian fields in the Rocky Mountains and off the coast of
Nova Scotia and New Brunswick.
4-11
-------
174,300 $8,331,849
100,308 $4,968,722
12,000
10,549
$509,272
$541,593
47,700 $1,358,784
53,865 $1,918,086
106,339 $3,628,416
$16,547,510 $10,860,260
$21,908,191 $21,117,850
Table 4-5. Costs of Production, Crude Oil and Natural Gas Extraction and Related
Industries
Cost of Supplies Used, Capital
Payroll Purchased Machinery Expenditures
NAICS Industry Employees ($1997 103) Installed, Etc. ($1997 103) ($1997103)
211111 Crude Oil and
Natural Gas
Extraction
1992
1997
211112 Natural Gas Liquid
Extraction
1992
1997
213111 Drilling Oil and
Gas Wells
1992
1997
213112 Support Activities
for Oil and Gas
Operations
1997
$23,382,770
$20,359,528
$3,076,039
$609,302
$678,479
$1,344,509 $286,509
$7,317,963 $2,209,300
$1,165,018
Sources: U.S. Department of Commerce, Bureau of the Census. 1999a. 1997 Economic Census, Mining, Industry Series.
Washington, DC: U.S. Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 1995a. 1992 Census of Mineral Industries, Industry Series.
Washington, DC: U.S. Department of Commerce.
4-12
-------
Table 4-6. Estimated U.S. Oil and Gas Reserves, Annual Production, and Imports, 1998
Category
Crude Oil (106 barrels)
Natural Gas (109 cubic feet)
Natural Gas Liquids (106 barrels)
Reserves
152,453
1,330,930
26,792
Annual
Production
2,281
18,708
246
Imports
3,178
3,152
NA
Sources: Energy Information Administration. 1999d. U.S. Crude Oil, Natural Gas, and Natural Gas Liquids Reserves 1998
Annual Report. Washington, DC: U.S. Department of Energy.
Energy Information Administration. 1999b. Petroleum Supply Annual 1998, Volume I. WashingtonDC: U.S.
Department of Energy.
4.1.3 Demand Side Characteristics
Characterizing the demand side of the industry involves describing product
characteristics. Crude oil, or unrefined petroleum, is a complex mixture of hydrocarbons that is
the most important of the primary fossil fuels. Refined petroleum products are used for
petrochemicals, lubrication, heating, and fuel. Petrochemicals derived from crude oil are the
source of chemical products such as solvents, paints, plastics, synthetic rubber and fibers, soaps
and cleansing agents, waxes, jellies, and fertilizers. Petroleum products also fuel the engines of
automobiles, airplanes, ships, tractors, trucks, and rockets. Other applications include fuel for
electric power generation, lubricants for machines, heating, and asphalt (Berger and Anderson,
1978). Because the market for crude oil is global and its price influenced by OPEC, slight
increases in the cost of producing crude oil in the U.S. will have little effect on the prices of
products that use crude oil as an intermediate good. Production cost increases are likely to be
absorbed mainly by the producer, with little of the increased cost passed along to consumers.
Natural gas is a colorless, flammable gaseous hydrocarbon consisting for the most part of
methane and ethane. Natural gas is used by residential, commercial, industrial, and electric
utility users. Total consumption of natural gas in the U.S. was 21,262 billion cubic feet in 1998.
Industrial consumers accounted for the largest share of this total, consuming 8,686 billion cubic
feet, while residential, commercial, and electric utility consumption was 4,520 billion cubic feet,
3,005 billion cubic feet, and 3,258 billion cubic feet, respectively. The remainder of U.S.
4-13
-------
consumption was by natural gas producers in their plants and on their gas pipelines. The largest
single application for natural gas is as a domestic or industrial fuel. Natural gas is also becoming
increasingly important for generating electricity. Although these are the primary uses, other
specialized applications have emerged over the years, such as a nonpolluting fuel for buses and
other motor vehicles. Carbon black, a pigment made by burning natural gas with little air and
collecting the resulting soot, is an important ingredient in dyes, inks, and rubber compounding
operations. Also, much of the world's ammonia is manufactured from natural gas; ammonia is
used either directly or indirectly in urea, hydrogen cyanide, nitric acid, and fertilizers (Tussing
and Tippee, 1995).
The primary substitutes for oil and natural gas are coal, electricity, and each other.
Consumers of these energy products are expected to respond to changes in the relative prices
between these four energy markets by changing the proportions of these fuels they consume. For
example, if the price of natural gas were to increase relative to other fuels, then it is likely that
consumers would substitute oil, coal, and electricity for natural gas. This effect of changing
prices is commonly referred to as fuel-switching. The extent to which consumers change their
fuel usage depends on such factors as the availability of alternative fuels and the capital
requirements involved. If they own equipment that can run on multiple fuels, then it may be
relatively easy to switch fuel usage as prices change. However, if existing capital cannot easily
be modified to run on an alternative fuel, then it is less likely for a consumer to change fuels in
the short run. If the relative price of the fuel currently in use remains elevated in the long run,
some additional consumers will switch fuels as they replace existing capital with new capital
capable of using relatively cheaper fuels. For example, if the price of natural gas were to
increase greatly relative to the price of electricity for residential consumers, most consumers are
unlikely to replace their natural gas furnaces immediately due to the high cost of doing so.
However, new construction would be less likely to include natural gas furnaces, and if the price
of natural gas were to remain relatively high compared with electricity in the long run,
residential consumers would be more likely to replace their natural gas furnaces with electric
heat pumps as their existing furnaces wear out.
4.1.4 Organization of the Industry
4-14
-------
Many oil and gas firms are merging to remain competitive in both the global and
domestic marketplaces. By merging with their peers, these companies may reduce operating
expenses and reap greater economies of scale than they would otherwise. Recent mergers, such
as BP Amoco and Exxon Mobil, have reduced the number of companies and facilities operating
in the U.S. Currently, there are 20 domestic major oil and gas companies, and only 40 major
global companies in the world (Conces, 2000). Most U.S. oil and gas firms are concentrated in
states with significant oil and gas reserves, such as Texas, Louisiana, California, Oklahoma, and
Alaska.
Tables 4-7 through 4-10 present the number of facilities and value of shipments by
facility employee count for each of the four industries. In 1997, 6,802 oil and gas extraction
companies operated 7,781 facilities, an average of 1.14 facilities per company (see Table 4-7).
Facilities with more than 100 employees produced more than 55 percent of the industry's value
of shipments. Although the number of companies and the number of facilities operating in 1992
were both greater then than in 1997, the distribution of shipment values by employee size was
similar to that of 1992.
Facilities employing fewer than 50 people in the NGLs extraction industry accounted for
64 percent, or $15.8 billion, of the industry's total value of shipments in 1997 (see Table 4-8).
487 of the industry's 529 facilities are in that employment category. This also means that a
relatively small number of larger facilities produce 36 percent of the industry's annual output, in
terms of dollar value. The number of facilities with zero to four employees and the number with
50 or more employees decreased during the 5-year period, accounting for most of the
10.5 percent decline in the number of facilities from 1992 to 1997. The average number of
facilities per company was 5.5 and 5.9 in 1992 and 1997, respectively.
4-15
-------
Table 4-7. Size of Establishments and Value of Shipments, Crude Oil and Natural Gas
Extraction Industry (NAICS 211111), 1997 and 1992
Average Number of
Employees in Facility
0 to 4 employees
5 to 9 employees
10 to 19 employees
20 to 49 employees
50 to 99 employees
100 to 249 employees
250 to 499 employees
500 to 999 employees
1,000 to 2,499 employees
2,500 or more employees
Total
Number of
Facilities
5,249
1,161
661
412
132
105
40
14
5
2
7,781
1997
Value of
Shipments
($1997 103)
$5,810,925
$3,924,929
$4,843,634
$10,538,529
$8,646,336
$41,318,227
$75,162,580
1992
Number of
Facilities
6184
1402
790
523
203
154
68
46
18
3
9,391
Value of
Shipments
($1997 103)
$5,378,330
$3,592,560
$4,504,830
$8,820,100
$5,942,130
$11,289,730
$8,135,850
$14,693,630
$9,265,530
D
$71,622,600
D = undisclosed.
Sums do not add to totals due to independent rounding.
Sources: U.S. Department of Commerce, Bureau of the Census. 1999a. 1997 Economic Census, Mining, Industry Series:
Crude Petroleum and Natural Gas Extraction. EC97N-2111A. Washington, DC: U.S. Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 1995a. 1992 Census,ofMineral Industries, Industry Series:
Crude Petroleum and Natural Gas. MIC92-I-13A. Washington, DC: U.S. Department of Commerce.
4-16
-------
Table 4-8. Size of Establishments and Value of Shipments, Natural Gas Liquid Extraction
Industry (NAICS 211112), 1997 and 1992
Average Number of
Employees in Facility
0 to 4 employees
5 to 9 employees
10 to 19 employees
20 to 49 employees
50 to 99 employees
100 to 249 employees
250 to 499 employees
500 to 999 employees
1,000 to 2,499 employees
2,500 or more employees
Total
Number of
Facilities
143
101
122
121
35
3
3
1
0
0
529
1997
Value of
Shipments
($1997 103)
$1,407,192
$1,611,156
$4,982,941
$7,828,439
$5,430,448
D
D
D
—
—
$24,828,503
1992
Number of
Facilities
190
92
112
145
36
14
2
0
0
0
591
Value of
Shipments
($1997 103)
$2,668,000
$1,786,862
$5,240,927
$10,287,200
$4,789,849
$2,205,819
D
—
—
—
$26,979,200
D = undisclosed.
Sums do not add to totals due to independent rounding.
Sources: U.S. Department of Commerce, Bureau of the Census. 1999b. 1997 Economic Census, Mining, Industry Series:
Natural Gas Liquid Extraction. EC97N-2111b. Washington, DC: U.S. Department of Commerce.
U. S. Department of Commerce, Bureau of the Census. 1995b. 1992 Census of Mineral Industries, Industry Series:
Natural Gas Liquids. MIC92-I-13B. Washington, DC: U.S. Department of Commerce.
As mentioned earlier, the oil and gas well drilling industry's 1997 value of shipments
were 106 percent larger than 1992' s value of shipments. However, the number of companies
primarily involved in this industry declined by 327 over 5 years, and 487 facilities closed during
the same period (see Table 4-9). The distribution of the number of facilities by employment size
shifted towards those that employed 20 or more people. In 1997, those facilities earned two-
thirds of the industry's revenues.
4-17
-------
Table 4-9. Size of Establishments and Value of Shipments, Drilling Oil and Gas Wells
Industry (NAICS 213111), 1997 and 1992
Average Number of
Employees in Facility
0 to 4 employees
5 to 9 employees
10 to 19 employees
20 to 49 employees
50 to 99 employees
100 to 249 employees
250 to 499 employees
500 to 999 employees
1,000 to 2,499 employees
2,500 or more employees
Total
Number of
Facilities
825
215
197
200
95
75
10
14
6
1
1,638
1997
Value of
Shipments
($1997 103)
$107,828
$231,522
$254,782
$1,008,375
$785,804
$1,069,895
$435,178
$1,574,139
D
D
$7,317,963
1992
Number of
Facilities
1,110
321
244
233
120
70
19
5
3
—
2,125
Value of
Shipments
($1997 103)
$254,586
$182,711
$256,767
$572,819
$605,931
$816,004
$528,108
$97,254
$238,427
—
$3,552,707
D = undisclosed.
Sums do not add to totals due to independent rounding.
Sources: U.S. Department of Commerce, Bureau of the Census. 1999c. 1997 Economic Census, Mining, Industry Series:
Drilling Oil and Gas Wells. EC97N-2131A. Washington, DC: U.S. Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 1995c. 1992 Census of Mineral Industries, Industry Series:
Oil and Gas Field Services. MIC92-I-13C. Washington, DC: U.S. Department of Commerce.
In 1997, 6,385 companies operated 7,068 oil and gas support activities facilities, an
average of 1.1 facilities per company. The Inventory Database includes 1,599 facilities in
NAICS 21. Most facilities employed four or fewer employees; however, those facilities with 20
or more employees accounted for the majority of the industry's revenues (see Table 4-10).
4-18
-------
Table 4-10. Size of Establishments and Value of Shipments, Support Activities for Oil and
Gas Operations (NAICS 213112), 1997
Average Number of Employees at
Facility
0 to 4 employees
5 to 9 employees
10 to 19 employees
20 to 49 employees
50 to 99 employees
100 to 249 employees
250 to 499 employees
500 to 999 employees
1,000 to 2,499 employees
2,500 or more employees
Total
1997
Number of Facilities
4,122
1,143
835
629
211
84
21
13
9
1
7,068
Value of Shipments
($1997 103)
$706,396
$571,745
$904,356
$1,460,920
$1,480,904
$1,175,766
$754,377
$1,755,689
D
D
$11,547,563
D = undisclosed.
Sums do not add to totals due to independent rounding.
Source: U.S. Department of Commerce, Bureau of the Census. 1999d. 1997 Economic Census, Mining, Industry Series:
Support Activities for Oil and Gas Operations. EC97N-2131B. Washington, DC: U.S. Department of Commerce.
4.1.5 Marke ts and Trends
Between 1990 and 1998, crude oil consumption increased 1.4 percent per year, and
natural gas consumption increased 2.0 percent per year. The increase in natural gas consumption
came mostly at the expense of coal consumption (EPA, 1999d). The Energy Information
Administration (EIA) anticipates that natural gas consumption will continue to grow at a similar
rate through the year 2020 to 32 trillion cubic feet/year. Prices are expected to grow steadily,
increasing overall by about 0.6 percent annually (EIA, 1999a). They also expect crude oil
consumption to grow at an annual rate of less than 1 percent over the same period (EIA, 1999a).
For ease of comparison, the quantities used for all energy markets modeled for this analysis are
defined in terms of quadrillions of Btus and prices are defined as dollars per million Btus. In
4-19
-------
2005, the year used for this analysis, the EIA (2000c) projects 24.57 quadrillion Btus of natural
gas will be consumed at an average price of $4.23/million Btus, and 41.21 quadrillion Btus of
petroleum products will be consumed at an average price of $8.22/million Btus.
4.2 NATURAL GAS PIPELINE INDUSTRY
The natural gas pipeline industry (NAICS 4862) comprises establishments primarily
engaged in the pipeline transportation of natural gas from processing plants to local distribution
systems. Also included in this industry are natural gas storage facilities, such as depleted gas
fields and aquifers.
4.2.1 Introduction
The natural gas industry can be divided into three segments, or links: production,
transmission, and distribution. Natural gas pipeline companies are the second link, performing
the vital function of linking gas producers with the local distribution companies and their
customers. Pipelines transmit natural gas from gas fields or processing plants through high
compression steel pipe to their customers. By the end of 1998, there were more than 300,000
miles of transmission lines (OPS, 2000).
The interstate pipeline companies that linked the producing and consuming markets
functioned mainly as resellers or merchants of gas until about the 1980s. Rather than acting as
common carriers (i.e., providers only of transportation), pipelines typically bought and resold the
gas to a distribution company or to some other downstream pipelines that would later resell the
gas to distributers. Today, virtually all pipelines are common carriers, transporting gas owned
by other firms instead of wholesaling or reselling natural gas (Tussing and Tippee, 1995).
According to the U.S. Bureau of the Census, the natural gas pipeline industry's revenues
totaled $19.6 billion in 1997. Pipeline companies operated 1,450 facilities and employed 35,789
people (see Table 4-11). The Inventory Database contains 1,401 facilities in NAICS 4862, so
the majority of pipeline companies are included. The industry's annual payroll is nearly
$1.9 billion.
4-20
-------
Table 4-11. Summary Statistics for the Natural Gas Pipeline Industry (NAICS 4862), 1997
Establishments 1,450
Revenue ($103) $19,626,833
Annual Payroll ($103) $1,870,950
Paid Employees 35,789
Source: U.S. Department of Commerce, Bureau of the Census. 2000. 1997 Economic Census, Transportation and
Warehousing: Geographic Area Series. EC97T48A-US. Washington, DC: Government Printing Office.
The recent transition from the SIC system to the NAICS changed how some industries
are organized for information collection purposes and thus how certain economic census data are
aggregated. Some SIC codes were combined, others were separated, and some activities were
classified under one NAICS code and the remaining activities classified under another. The
natural gas transmission (pipelines) industry is an example of an industry code that was
reclassified. Under NAICS, SIC 4922, natural gas transmission (pipelines), and a portion of SIC
4923, natural gas distribution, were combined. The adjustments have made comparison between
the 1992 and 1997 economic censes difficult at this time. The U.S. Census Bureau has yet to
publish a comparison report. Thus, for this industry only 1997 census data are presented.
4.2.2 Supply Side Characteristics
Characterizing the supply side involves describing services provided by the industry, by-
products, the costs of production, and capacity utilization.
4-21
-------
4.2.2.1 Service Description
Natural gas is delivered from gas processing plants and fields to distributers via a
nationwide network of over 300,000 miles of transmission pipelines (NGSA et al., 2000a). The
majority of pipelines are composed of steel pipes that measure from 20 to 42 inches in diameter
and operate 24 hours a day. Natural gas enters pipelines at gas fields, storage facilities, or gas
processing plants and is "pushed" through the pipe to the city gate or interconnections, the point
at which distribution companies receive the gas. Pipeline operators use sophisticated computer
and mechanical equipment to monitor the safety and efficiency of the network.
Reciprocating internal combustion engines compress and provide the pushing force
needed to maintain the flow of gas through the pipeline. When natural gas is transmitted, it is
compressed to reduce the volume of gas and to maintain pushing pressure. The gas pressure in
pipelines is usually between 300 and 1,300 psi, but lesser and higher pressures may be used. To
maintain compression and keep the gas moving, compressor stations are located every 50 to
100 miles along the pipeline. Most compressors are large reciprocating engines powered by a
small portion of the natural gas being transmitted through the pipeline.
There are over 8,000 gas compressing stations along U.S. gas pipelines, each equipped
with one or more engines. The combined output capability of U.S. compressor engines is over
20 million horsepower (NGSA et al., 2000a). Nearly 5,000 engines have individual output
capabilities from 500 to over 8,000 horsepower. The replacement cost of this subset of larger
engines is estimated by the Gas Research Institute to be $18 billion (Whelan, 1998).
Before or after natural gas is delivered to a distribution company, it may be stored in an
underground facility. Underground storage facilities are most often depleted oil and/or gas
fields, aquifers, or salt caverns. Natural gas storage allows distribution and pipeline companies
to serve their customers more reliably by withdrawing more gas from storage during peak-use
periods and reduces the time needed to respond to increased gas demand (NGSA et al., 2000b).
In this way, storage guarantees continuous service, even when production or pipeline
transportation services are interrupted.
4.2.2.2 Major By-Products
4-22
-------
There are no major by-products of the natural gas pipeline industry itself. However, the
engines that provide pumping action at plants and push crude oil and natural gas through
pipelines to customers and storage facilities produce HAPs. As noted previously, HAPs
produced in engines include formaldehyde, acetaldehyde, acrolein, and methanol.
4.2.2.3 Costs of Production
Between 1996 and 2000, pipeline firms committed over $14 billion to 177 expansion and
new construction projects. These projects added over 15,000 miles and 36,178 million cubic feet
per day (MMcf/d) capacity to the transmission pipeline system. Because there are compression
stations about every 50 to 100 miles along gas pipelines, the addition of 15,000 miles of pipeline
implies that 150 to 300 compression stations were added. There are varying numbers of engines
at different stations, but the average is three engines per compression station in the Inventory
Database. Thus, approximately 450 to 900 new engines were added along pipelines over the
period 1996 through 2000. Table 4-12 summarizes the investments made in pipeline projects
during the past 5 years. Building new pipelines is more expensive than expanding existing
pipelines. For the period covered in the table, the average cost per project mile was $862,000.
However, the costs for pipeline expansions averaged $542,000, or 29 cents per cubic foot of
capacity added. New pipelines averaged $1,157,000 per mile at 48 cents per cubic foot of
capacity.
Pipelines must pay for the natural gas that is consumed to power the compressor engines.
The amount consumed and the price paid have fluctuated in recent years. In 1998, pipelines
consumed 635,477 MMcf of gas, paying, on average, $2.01 per 1,000 cubic feet. Thus, firms
spent approximately $1.28 billion in 1998 for the fueling of RICE units used on pipelines.
Pipelines used less natural gas in 1998 than in previous years; the price paid for that gas
fluctuated between $1.49 and $2.29 between 1994 and 1997 (see Table 4-13). For companies
that transmit natural gas through their own pipelines the cost of the natural gas consumed is
considered a business expense.
4-23
-------
Table 4-12. Summary Profile of Completed and Proposed Natural Gas Pipeline Projects, 1996 to 2000
Year
1996
1997
1998
1999
2000
Total
Number
of
Projects
26
42
54
36
19
177
System
Mileage
1,029
3,124
3,388
3,753
4,364
15,660
All
New
Capacity
(Mmcf/d)
2,574
6,542
11,060
8,205
7,795
36,178
Type Projects
Project
Costs
($106)
$552
$1,397
$2,861
$3,135
$6,339
$14,285
Average Cost
per Mile
($103)
$448
$415
$1,257
$727
$1,450
$862
Costs per
Cubic Foot
Capacity
(cents)
21
21
30
37
81
39
New Pipelines
Average
Cost per
Mile ($103)
$983
$554
$1,301
$805
$1,455
$1,157
Costs per
Cubic Foot
Capacity
(cents)
17
22
31
46
91
48
Expansions
Average
Cost per
Mile ($103)
$288
$360
$622
$527
$940
$542
Costs per
Cubic Foot
Capacity
(cents)
27
21
22
31
57
29
Note: Sums may not add to totals because of independent rounding.
Source: Energy Information Administration. 1999a. Natural Gas 1998: Issues and Trends. Washington, DC: U.S. Department of Energy.
-------
Table 4-13. Energy Usage and Cost of Fuel, 1994-1998
Year
1994
1995
1996
1997
1998
Pipeline Fuel (MMcf)
685,362
700,335
711,446
751,470
635,477
Average Price
($ per 1,000 cubic feet)
1.70
1.49
2.27
2.29
2.01
Source: Energy Information Administration. 1999b. Natural Gas Annual 1998. Washington, DC: US Department of Energy.
4.2.2.4 Capacity Utilization
During the past 15 years, interstate pipeline capacity has increased significantly. In
1990, the transmission pipeline system's capacity was 74,158 Mmcf/day (see Table 4-14). By
the end of 1997, capacity reached 85,847 Mmcf/day, an increase of approximately 16 percent.
The system's usage, however, has increased at a faster rate than capacity. The average daily
flow was 60,286 Mmcf/day in 1997, a 22 percent increase over 1990's rates. Currently, the
system operates at approximately 72 percent of capacity.
4.2.2.5 Imports
Approximately 17 percent of the U.S. natural gas supply is imported, primarily from
Canadian fields. In many economic analyses, the imported supply is treated separately from the
domestic supply because of the difference in the impact of domestic regulation. However, it is
assumed that the imported gas will still be subject to control costs when it is transported through
pipelines in the U.S. Thus, the imported supply is not differentiated because the regulation will
affect it in a similar manner to domestically supplied gas since they use the same distribution
method.
4-25
-------
Table 4-14. Transmission Pipeline Capacity, Average Daily Flows,
and Usage Rates, 1990 and 1997
Capacity (Mmcf per day)
Average Flow (Mmcf per
day)
Usage Rate (percent)
1990
74,158
49,584
68
1997
85,847
60,286
72
Percent Change
16
22
4
Source: Energy Information Administration. 1999a. Natural Gas 1998: Issues and Trends. Washington, DC: US Department
of Energy.
4.2.3 Demand Side Characteristics
Most pipeline customers are local distribution companies that deliver natural gas from
pipelines to local customers. Many large gas users will buy from marketers and enter into
special delivery contracts with pipelines. However, local distribution companies (LDCs) serve
most residential, commercial, and light industrial customers. LDCs also use compressor engines
to pump natural gas to and from storage facilities and through the gas lines in their service area.
While economic considerations strongly favor pipeline transportation of natural gas,
liquified natural gas (LNG) emerged during the 1970s as a transportation option for markets
inaccessible to pipelines or where pipelines are not economically feasible. Thus, LNG is a
substitute for natural gas transmission via pipelines. LNG is natural gas that has been liquified
by lowering its temperature. LNG takes up about 1/600 of the space gaseous natural gas takes
up, making transportation by ship possible. However, virtually all of the natural gas consumed
in the U.S. reaches its consumer market via pipelines because of the relatively high expense of
transporting LNG and its volatility. Most markets that receive LNG are located far from
pipelines or production facilities, such as Japan (the world's largest LNG importer), Spain,
France, and Korea (Tussing and Tippee, 1995).
4-26
-------
4.2.4 Organization of the Industry
Much like other energy-related industries, the natural gas pipeline industry is dominated
by large investor-owned corporations. Smaller companies are few because of the real estate,
capital, and operating costs associated with constructing and maintaining pipelines (Tussing and
Tippee, 1995). Many of the large corporations are merging to remain competitive as the industry
adjusts to restructuring and increased levels of competition. Increasingly, new pipelines are built
by partnerships: groups of energy-related companies share capital costs through joint ventures
and strategic alliances (EIA, 1999a). Ranked by system mileage, the largest pipeline companies
in the U.S. are El Paso Energy (which recently merged with Southern Natural Gas Co.), Enron,
Williams Cos., Coastal Corp., and Duke Energy (see Table 4-15). El Paso Energy and Coastal
intend to merge in mid-2000.
4.2.5 Marke ts and Trends
During the past decade, interstate pipeline capacity has increased 16 percent. Many
existing pipelines underwent expansion projects, and 15 new interstate pipelines were
constructed. In 1999 and 2000, proposals for pipeline expansions and additions called for a
$9.5 billion investment, an increase of 16.0 billion cubic feet per day of capacity (EIA, 1999a).
The EIA (1999a), a unit of the Department of Energy, expects natural gas consumption to
grow steadily, with demand forecasted to reach 32 trillion cubic feet by 2020. The expected
increase in natural gas demand has significant implications for the natural gas pipeline system.
The EIA (1999a) expects the interregional pipeline system, a network that connects the
lower 48 states and the Canadian provinces, to grow at an annual rate of 0.7 percent between
2001 and 2020. However, natural gas consumption is expected to grow at more than twice that
annual rate, 1.8 percent, over that same period. The majority of the growth in consumption is
expected to be fueled by the electric generation sector. According to the EIA, a key issue is what
kinds of infrastructure changes will be required to meet this demand and what the financial and
environmental costs will be of expanding the pipeline network.
4-27
-------
Table 4-15. Five Largest Natural Gas Pipeline Companies by System Mileage, 2000
Company
Headquarters
Sales Employment Miles of
($1999106) (1999) Pipeline
El Paso Energy Corporation Houston, TX
Incl. El Paso Natural Gas Co.
Southern Natural Gas Co.
Tennessee Gas Pipe Line Co.
Enron Corporation Houston, TX
Incl. Northern Border Pipe Line Co.
Northern Natural Gas Co.
Transwestern Pipeline Co.
Williams Companies, Inc. Tulsa, OK
Incl. Transcontinental Gas Pipe Line
Northwest Pipe Line Co.
Texas Gas Pipe Line Co.
The Coastal Corporation Houston, TX
Incl. ANR Pipeline Co.
Colorado Interstate Gas Co.
Duke Energy Corporation Charlotte, NC
Incl. Panhandle Eastern Pipeline Co.
Algonquin Gas Transmission Co.
Texas Eastern Transmission Co.
$5,782
$40,112
3,593
,197
$21,742
4,700 40,200
17,800 32,000
21,011 27,000
13,000 18,000
21,000 11,500
Sources: Heil, Scott F., Ed. 1998. Ward's Business Directory of U.S. Private and Public Companies 1998, Volume 5. Detroit,
MI: Gale Research Inc.
Sales, employment, and system mileage: Hoover's Incorporated. 2000. Hoover's Company Profiles. Austin, TX:
Hoover's Incorporated, .
4-28
-------
5.0 ECONOMIC IMPACT ANALYSIS
The proposed rule to control emissions of HAPs from RICE will affect many U.S.
industries because these engines are primarily used as inputs in extracting and transporting fuels
(oil and natural gas). Therefore, the proposed regulations will increase the cost of producing
these fuels and will lead to an increase in energy costs to industrial, commercial, and residential
customers. In addition to the effect on energy prices, many industrial facilities use RICE as part
of their production process and will face direct control costs on these engines. The response of
producers to these additional costs determines the economic impacts of the regulation.
Specifically, the cost of the regulation may induce some owners to change their current operating
rates or even to close their operations (either the entire facility or individual product lines).
These choices affect, and in turn are affected by, the market prices for fuels and the market
prices in the final product markets. This section describes the methodology, data, and model
used to estimate the economic impacts of the proposed regulation for the year 2005 and provides
the economic analysis results
5.1 ECONOMIC IMPACT METHODOLOGY
This section summarizes the Agency's approach to modeling the responses of fuel
markets to the imposition of the proposed regulation. In conducting an economic analysis and
determining the economic impacts, the analyst should recognize the alternatives available to
each producer in response to the regulation and the context of these choices. The Agency
evaluated the economic impacts of this NESHAP using a market-based approach that gives
5-1
-------
producers the choice of whether to continue producing these products and, if so, to determine the
optimal level consistent with market signals.
The Agency's approach is soundly based on standard microeconomic theory, employs a
comparative statics approach, and assumes certainty in relevant markets. Supply curves were
developed for each energy market (see Appendix A), and prices and quantities were determined
in perfectly competitive markets for each fuel market and each final product and service market.
5.1.1 Background on Economic Modeling Approaches
In general, the economic analysis methodology needs to allow EPA to consider the
effects of the different regulatory alternatives. Several types of economic impact modeling
approaches have been developed to support regulatory development. These approaches can be
viewed as varying along two modeling dimensions:
•• the scope of economic decision making accounted for in the model and
•• the scope of interaction between different segments of the economy.
Each of these dimensions was considered in determining the approach for this study. The
advantages and disadvantages of different modeling approaches are discussed below.
5.1.1.1 Modeling Dimension 1: Scope of Economic Decision making
Models incorporating different levels of economic decision making can generally be
categorized as with behavior responses and without behavior responses (accounting approach).
Table 5-1 provides a brief comparison of the two approaches. The nonbehavioral approach
essentially holds fixed all interactions between facility production and market forces. It assumes
that firms absorb all control costs and consumers do not face any of the costs of regulation.
5-2
-------
Table 5-1. Comparison of Modeling Approaches
EIA With Behavioral Responses
• Incorporates control costs into production function
• Includes change in quantity produced
• Includes change in market price
• Estimates impacts for
• • affected producers
• • unaffected producers
• • consumers
• • foreign trade
EIA Without Behavioral Responses
• Assumes firm absorbs all control costs
• Typically uses discounted cash flow analysis to evaluate burden of control costs
• Includes depreciation schedules and corporate tax implications
• Does not adjust for changes in market price
• Does not adjust for changes in plant production
Typically, engineering control costs are weighted by the number of affected units to develop
"engineering" estimates of the total annualized costs. These costs are then compared to
company or industry sales to determine the regulation's impact.
In contrast, the behavioral approach is grounded in economic theory related to producer
and consumer behavior in response to changes in market conditions. Owners of affected
facilities are economic agents that can, and presumably will, make adjustments such as changing
production rates or altering input mixes that will generally affect the market environment in
which they operate. As producers change their behavior in response to regulation, consumers are
typically faced with changes in prices that cause them to alter the quantity that they are willing to
purchase. In essence, this approach models the expected reallocation of society's resources in
response to a regulation. The changes in price and production from the market-level impacts are
used to estimate the distribution of social costs between consumers and producers.
5O
-------
5.1.1.2 Modeling Dimension 2: Interaction Between Economic Sectors
Because of the large number of markets potentially affected by the regulation on RICE,
an issue arises concerning the level of sectoral interaction to model. In the broadest sense, all
markets are directly or indirectly linked in the economy; thus, all commodities and markets are
to some extent affected by the regulation. For example, controls on RICE may indirectly affect
almost all markets for goods and services to some extent because the cost of fuel (an input in the
provision of most goods and services) is likely to increase with the regulation in effect.
However, the impact of rising fuel prices will differ greatly between different markets depending
on how important fuel is as an input in that market.
The appropriate level of market interactions to be included in the EIA is determined by
the scope of the regulation across industries and the ability of affected firms to pass along the
regulatory costs in the form of higher prices. Alternative approaches for modeling interactions
between economic sectors can generally be divided into three groups:
Partial equilibrium model: Individual markets are modeled in isolation. The only
factor affecting the market is the cost of the regulation on facilities in the industry
being modeled.
General equilibrium model: All sectors of the economy are modeled together.
General equilibrium models operationalize neoclassical microeconomic theory by
modeling not only the direct effects of control costs, but also potential input
substitution effects, changes in production levels associated with changes in
market prices across all sectors, and the associated changes in welfare
economywide. A disadvantage of general equilibrium modeling is that substantial
time and resources are required to develop a new model or tailor an existing
model for analyzing regulatory alternatives.
Multiple-market partial equilibrium model: A subset of related markets are
modeled together, with intersectoral linkages explicitly specified. To account for
the relationships and links between different markets without employing a full
general equilibrium model, analysts can use an integrated partial equilibrium
model. The multiple-market partial equilibrium approach represents an
5-4
-------
intermediate step between a simple, single-market partial equilibrium approach
and a full general equilibrium approach. This approach involves identifying and
modeling the most significant subset of market interactions using an integrated
partial equilibrium framework. In effect, the modeling technique is to link a
series of standard partial equilibrium models by specifying the interactions
between supply functions and then solving for prices and quantities across all
markets simultaneously. In instances where separate markets are closely related
and there are strong interconnections, there are significant advantages to
estimating market adjustments in different markets simultaneously using an
integrated market modeling approach.
5.1.2 Selected Modeling Approach for RICE Analysis
To conduct the analysis for the RICE MACT, the Agency used a market modeling
approach that incorporates behavioral responses in a multiple-market partial equilibrium model
as described above. This approach allows for a more realistic assessment of the distribution of
impacts across different groups than the nonbehavioral approach, which may be especially
important in accurately assessing the impacts of a significant rule affecting numerous industries.
Because of the size and complexity of this regulation, it is important to use a behavioral model to
examine the distribution of costs across society. Because the regulations on RICE affect energy
costs, an input into many production processes, complex market interactions need to be captured
to provide an accurate picture of the distribution of regulatory costs. Because of the large
number of affected industries under this MACT, an appropriate model should include multiple
markets and the interactions between them. Multiple-market partial equilibrium analysis
provides a manageable approach to incorporate interactions between energy markets and product
markets into the economic analysis to accurately estimate the regulation's impact.
The model used for this analysis includes industrial (manufacturing), commercial,
residential, transportation, and energy markets affected by the controls placed on engines. The
5-5
-------
industrial and commercial sectors are divided into 24 final product and service markets.1 The
energy markets are divided into natural gas, petroleum products, coal, and electricity.
Figure 5-1 presents an overview of the key market linkages included in the economic
impact model we propose to use for analyzing the RICE MACT. The analysis' emphasis is on
the energy supply chain, including the extraction and transportation of natural gas and
petroleum, the generation of electricity, and the consumption of energy by producers of final
products and services. The industries most directly affected by the RICE MACT are those
involved in extracting and transporting natural gas. However, changes in the equilibrium price
and quantity of natural gas affect all of the other energy markets. As shown in Figure 5-1,
wholesale electricity generators consume natural gas, petroleum products, and coal to generate
electricity that is then used to produce final products and services. In addition, many final
product markets use natural gas and petroleum products directly as an input into their production
process. This analysis explicitly models the linkages between these market segments.
RICE are used to extract and transport natural gas and petroleum products used by a wide
range of industrial, commercial, residential, and transportation sectors in the U.S. economy. As
a result, control costs associated with the proposed regulation will directly affect the cost of
extraction and transportation of natural gas and petroleum products using RICE to
generate compression and
•• using RICE directly as part of a production process, such as for rock crushing in
the mining sector.
'These markets are defined at the two- and three-digit NAICS code level. This allows for a fairly disaggregated
examination of the regulation's impact on producers. However, if the costs of the regulation are concentrated on
a particular subset of one of these markets, then treating the cost as if it fell evenly on the entire NAICS code
may underestimate the impacts on the subset of producers that are affected by the regulation.
5-6
-------
era
•s
re
E
5'
W
re
P^-
re
re
re
era
=
a
s
5'
o
a
c
n
re
P^-
VI
Fuel Markets
Supply Demand
Exogenous Endogenous
Oil
Gas
Coal
2-i dema
demand
i demand
Z-i demand
coal
V V
Electricity
Electricity Market
Energy Consumption
Industry A
Btu ^ Manufacturing
Production ^ Process
I Reeulatorv 1
- Regulatory -
Costs
Industry B
Industry C
demand
Intermediate or
Final Product Markets
Supply
Endogenous
product A
supply
Demand
Exogenous
Product A
-------
There are several categories of RICE, as described in Section 2. The categories
that fall under the proposed regulation are spark ignition 2SLB, spark ignition
4SLB, spark ignition 4SRB, and CI RICE.2 Most industries that use engines use
multiple categories. 2SLB, 4SLB, and 4SRB engines are all used primarily in
either oil and gas extraction or on natural gas pipelines. They are also distributed
across many other industrial and commercial SIC codes, although in relatively
small numbers. The CI engines in the Inventory Database fall mainly in the
hospital services industry and in other commercial businesses.
In addition to the direct impact of control costs on entities installing new RICE and
existing entities using 4SRB, indirect impacts are passed along the energy supply chain through
changes in prices. For example, production costs will increase for mining companies using
RICE as a result of the direct control costs on RICE as well as the resulting increase in the price
of natural gas and electricity used as energy inputs in the production process.
Also included in the impact model is feedback of changes in output in the final product
markets into the demand for Btus in the fuel markets. The change in facility output is
determined by the size of the Btu cost increase (typically variable cost per output), the facility's
production function (slope of facility-level supply curve), and the characteristics of the facility's
downstream market (other market suppliers and market demanders). For example, if consumers'
demand for a final product is not very sensitive to price, then producers can pass the majority of
the cost of the regulation through to consumers and the facility output may not change
appreciably. However, if only a small proportion of market output is produced at facilities
affected by the regulation, then competition will prevent the affected facilities from raising their
prices significantly.
One possible feedback pathway that this analysis does not plan on modeling is technical
changes in the manufacturing process. For example, if the cost of Btus increases, a facility may
use measures to increase manufacturing efficiency or capture waste heat. Facilities could also
possibly change the input mix that they use, substituting other inputs for fuel. These facility-
2Although CI engines can be either 2SLB or 4SLB, these two categories have been combined for this analysis, and
the acronyms 2SLB and 4SLB are reserved for spark ignition engines of these configurations.
5-8
-------
level responses will also act to reduce pollution, but including these responses is beyond the
scope of this analysis.
The intermarket linkages connecting the fuel markets and final product markets are
described in the sections below.
5.1.3 Directly Affected Markets
Markets where RICE are used as an input to production are considered to be directly
affected. Producers using engines will be required to add costly controls to any new engines that
they acquire and to existing 4SRB engines. They also must incur monitoring costs to ensure that
the controls are working properly. Therefore, the regulation will increase their production costs
and cause these directly affected firms to reduce the quantity that they are willing to supply at
any given price.
5.1.3.1 Market for Natural Gas
Because the majority of RICE are used in either extracting oil and natural gas or
transporting natural gas, the energy market most directly affected by the proposed regulations is
the natural gas industry. Because it will be more costly to produce natural gas under the new
regulations, firms involved in producing natural gas are expected to supply less gas to the market
at any given price than they did prior to the new rule. These decreases at the facility level will
lead to a decrease in industry supply. The magnitude of the upward shift in the supply curve and
the price elasticities of supply and demand are the two factors that determine the impacts on the
natural gas market. Because 25 percent of 4SRB and 3 percent of 4SLB engines are projected to
be controlled in the absence of the proposed regulation, these engines are considered to be
unaffected by the regulation. Figure 5-2 illustrates the shifts in the supply curves for a
representative energy market.
5-9
-------
Qn Q10 Quantity
(a) Producers bearing control
costs (affected)
Q20 Q21 Quantity
(b) Producers bearing no control
costs (unaffected)
Qio
Qn
^20
Q20
Q2i
STO
STI
QTO
Qn
QT1 QTO Quantity
(c) Total Market
market price without regulation
market price with regulation
supply function for affected firms without regulation
supply function for affected firms with regulation
quantity sold for affected firms without regulation
quantity sold for affected firms with regulation
supply function for unaffected firms both with and without regulation
quantity sold for unaffected firms without regulation
quantity sold for unaffected firms with regulation
total market supply function without regulation
total market supply function with regulation
total market quantity sold without regulation
total market quantity sold with regulation
Figure 5-2. Market Effects of Regulation-Induced Costs
5.1.3.2 Market for Petroleum Products
The market for petroleum products is also included in the economic impact model for
RICE. For petroleum products, a single composite product is used to model market adjustment.
A composite product was used in this market because engines are used in the extraction of crude
petroleum; as a result, the increased production costs were not assigned to specific end products,
such as fuel oil #2 or reformulated gasoline. This will tend to understate the impacts for
petroleum products where extraction costs as a percentage of production costs are greater than
average and overstate impacts for products where extraction costs as a percentage of production
costs are less than average.
Control costs associated with RICE will increase the cost of petroleum extraction. The
cost impacts are assumed to be distributed over all domestically consumed petroleum products.
This is because it is assumed that affected units will be distributed across all firms involved in
5-10
-------
the production of these products. The supply curve for petroleum products will shift upward by
the proportional increase in total production costs caused by the control costs on RICE.
5.1.3.3 Final Product and Service Markets
Final product and service markets are also directly affected by the regulation. Many
manufacturing facilities use engines in their production processes. Commercial entities use
engines as generators, especially in the health services field. In addition to the direct costs of the
regulation, final product and service markets are indirectly affected through price increases in the
energy markets.
Directly affected producers of final products and services are segmented into industrial
and commercial sectors defined at the two- and three-digit NAICS code level. A partial
equilibrium analysis was conducted to model the supply and demand for final product and
service markets. Changes in production levels and fuel switching due to the regulation's impact
on the price of Btus were then linked back into the energy markets.
Impact on the Final Product and Service Markets. The impact of the regulation on
manufacturers in this sector is modeled as an increase in the cost of Btus used in the production
process. In this context, Btus refer to the generic energy requirements that are used to generate
process heat, process steam, or shaft power. Compliance costs associated with the regulation
will increase the cost of Btu production in the manufacturing sectors. The cost of Btu production
for industry increases due to both direct control costs on engines owned by manufacturers and
increases in the price of fuels. Because Btus are an input into the production process, these price
increases lead to an upward shift in the facility (and industry) supply curves as shown in Figure
5-2, leading to a change in the equilibrium market price and quantity.
The changes in equilibrium supply and demand in each final product and service market
are modeled to estimate the regulation's impact on each manufacturing sector. In a perfectly
competitive market, the point where supply equals demand determines the market price and
quantity, so market price and quantity are determined by solving the model for the price where
the quantity supplied and the quantity demanded are equal. The size of the regulation-induced
shifts in the supply curve are a function of the total direct control costs associated with new
engines and existing 4SRB engines and the indirect fuel costs (determined by the change in fuel
5-11
-------
price and intensity of use) in each final product and service market. The proportional shift in the
supply curve is determined by the ratio of total control costs (both direct and indirect) to
production costs.
This impact on the price of Btus facing industrial users feeds back to the fuel market in
two ways (see Figure 5-3). The first is through the company's input decision concerning the
fuel(s) that will be used for its manufacturing process. As the cost of Btus increases, firms may
switch fuels and/or change production processes to increase energy efficiency and reduce the
number of Btus required per unit of output. Fuel switching impacts are modeled using cross-
price elasticities of demand between energy sources. For example, a cross-price elasticity of
demand between natural gas and electricity of 0.5 implies that a 1 percent increase in the price of
electricity will lead to a 0.5 percent increase in the demand for natural gas. Own-price
elasticities of demand are used to estimate the change in the use of fuel by demanders. For
example, a demand elasticity of-0.175 for electricity implies that a 1 percent increase in the
price of electricity will lead to a 0.175 percent decrease in the quantity of electricity demanded.
The second feedback pathway to the energy markets is through the facility's change in
output. Because Btus are an input into the production process, price increases lead to an upward
shift in the facility supply curves (not modeled individually). This leads to an upward shift in the
industry supply curve when the shifts at the facility level are aggregated across facilities. A shift
in the industry supply curve leads to a change in the equilibrium market price and quantity. In a
perfectly competitive market, the point where supply equals demand determines the market price
and quantity. The Agency assumes constant returns to scale in production so that the percentage
change in Btus consumed by manufacturers equals the percentage change in the equilibrium
market quantity in each final product and service market.
5-12
-------
Compliance Costs
A $/Btu
Fuel
Markets
$/Btu
Btu Production
Decision
A $/Btu
Production
Decision
Output
Market
A Fuel Use
A Output
Figure 5-3. Fuel Market Interactions with Facility-Level Production Decisions
The Agency assumed that the demand curves for final products and services in all
manufacturing sectors are unchanged by the regulation. However, because the demand function
quantifies the change in quantity demanded in response to a change in price, the baseline demand
conditions are important in determining the regulation's impact. The key demand parameters
will be the elasticities of demand with respect to changes in the price of final products. For these
markets, a "reasonable" range of elasticity values is assigned based on estimates from similar
commodities. Because price changes are anticipated to be small, the point elasticities at the
original price and quantity should be applicable throughout the relevant range of prices and
quantities examined in this model.
5.1.4 Indirectly Affected Markets
In addition to the many markets that are directly affected by the regulation on RICE,
some markets feel the regulation's impacts despite having no direct costs resulting from the
regulation. Firms in these markets generally face changes in the price of energy that affect their
production decisions.
5-13
-------
5.1.4.1 Market for Electricity
Although EPA assumed that there are no direct impacts on the production of electricity
because engines are not commonly used by utilities to generate power, the market for electricity
will still be indirectly affected through changes in fuel prices. Electricity generators are
extremely large consumers of coal and natural gas as well as petroleum products to a lesser
extent. These fuels are used to generate electricity, so as the prices of fuels rise, there is a
decrease in the amount of electricity that producers are willing to supply. This impact feeds
back into the fuel markets as utilities reduce their purchases of fuels. In addition to the decrease
in supply due to the regulation, an increase in demand is expected as fuel consumers switch from
natural gas and petroleum to electricity. Therefore, it is ambiguous whether equilibrium quantity
will rise or fall. The price elasticities of supply and demand are the important factors influencing
the size of the impacts and whether quantity will increase or decrease.
5.1 A.2 Market for Coal
The coal market is not directly affected by the regulation, but it is included in the market
model. Although engines are not commonly used in the production or transportation of coal, the
supply of coal will be affected by the price of energy used in coal production, and the demand
for coal by utility generators and manufacturers will be affected through changes in the relative
price of alternative (noncoal) energy sources such as natural gas and petroleum products. The
demand for coal from the industrial, transportation and, residential sectors will increase as
consumers switch away from the fuels that face increases in price due to controls. The demand
for coal from electric utilities may either increase or decrease depending on whether the
equilibrium quantity of electricity rises or falls as a result of the regulation.
5-14
-------
5.1.4.3 Final Product and Service Markets
Some final product markets do not include any engines and are therefore not directly
affected by the RICE MACT. However, these markets will still be affected indirectly due to the
changes in energy prices that they will face following the regulation. There will be a tendency
for these users to shift away from natural gas and petroleum products and towards electricity and
coal.
5.1.4.4 Impact on Residential Sector
The residential sector does not bear any direct costs associated with the regulation
because they do not own RICE. However, they bear indirect costs due to price increases. The
residential sector is a significant consumer of electricity, natural gas, and petroleum products
used for heating, cooling, and lighting, as well as many other end uses. The change in the
quantity of energy demanded by these consumers in response to changes in energy prices is
modeled as a single demand curve parameterized by demand elasticities for residential
consumers obtained from the literature. Once again, it is expected that in addition to a decrease
in the total amount of energy consumed, there will be reallocation across fuels consumed.
5.1.4.5 Impact on Transportation Sector
The transportation sector does not face any direct costs due to the regulation because
RICE are not typically used in this sector. The main fuels used in this market are petroleum
products. The change in the quantity of energy demanded by these consumers in response to
changes in prices is modeled as a single demand curve parameterized by demand elasticities for
this sector from the literature. The major impact on this market is an increase in the price of a
key input causing a reduction in output. There may also be some fuel switching in this sector
towards electricity and coal.
5-15
-------
5.2 OPERATIONALIZING THE ECONOMIC IMPACT MODEL
Figure 5-4 illustrates the linkages used to operationalize the estimation of economic
impacts associated with the compliance costs. Compliance costs placed on existing 4SRB and
new RICE shift the supply curve for natural gas and petroleum because RICE are used in the
extraction and transportation of these fuels. Adjustments in the natural gas and petroleum
energy markets determine the share of the cost increases that producers (natural gas and
petroleum companies) and consumers (electricity utilities, product manufacturers, commercial
business, and residential households) bear. There are also some relatively small compliance
costs on the electricity market from the very few affected engines used in this market.
Increased fuel costs for electricity generators will decrease the supply of electricity. The
new equilibrium price and quantity in the electricity market will determine the distribution of
impacts between producers (electricity generators) and consumers (product manufacturers,
commercial businesses, and residential households). Changes in wholesale electricity
generators' demand for input fuels (due to changes in the market quantity of electricity) feed
back into the natural gas and petroleum markets as utilities change the allocation of fuels used as
inputs.
Manufacturers experience supply curve shifts due to control costs on affected engines
they operate and increased prices for natural gas, petroleum, and electricity. The share of these
costs borne by producers (manufacturers) and consumers is determined by the new equilibrium
price and quantity in the final product markets. Changes in manufacturers' Btu demands due to
fuel switching and changes in production levels feed back into the electricity, natural gas, and
petroleum markets. Adjustments in price and quantity in all energy and final product markets
occur simultaneously. A computer model was used to numerically simulate market adjustments
by iterating over commodity prices until equilibrium is reached (i.e., until the quantity supplied
equals the quantity demanded in all markets being modeled). Using the results provided by the
model, economic impacts of the regulation (changes in consumer and producer surplus) were
estimated for all sectors of the economy being modeled.
5-16
-------
era
•s
re
o
•a
=
5°
era
re
o
a, »
^ o
n
o
=
o
5
5
•a
o
U
Fuel Markets
Assume
Supply
Gas
Coal
Model
Demand
2-i demand
oil
2-1 demand
NO
i demand
Electricity Market
Electricity
i demand
Energy Consumption
Final Product Markets
Fuel Prices
Industry A
Btu
Production
Manufacturing
Process
Product Supply
Product Price
Regulatory Costs
(A Fuel Prices)
Fuel Prices
Industry B
Fuel Prices
Industry Z
Fuel Prices
Commercial Businesses
Fuel Prices
Residential Households
A Production Process
(Fuel Switching)
AQ
P = market price of final
output
Q = quantity sold of final
output
A Production Levels
-------
5.2.1 Computer Model
The computer model comprises a series of computer spreadsheet modules. The modules
integrate the engineering cost inputs and the market-level adjustment parameters to estimate the
regulation's impact on the price and quantity in each market being analyzed. At the heart of the
model is a market-clearing algorithm that compares the total quantity supplied to the total
quantity demanded for each market commodity.
Forecast prices and production levels for 2005 are used to calibrate the baseline scenario
(without regulation) for the model. Then, the compliance costs associated with the regulation are
introduced as a "shock" to the system, and the supply and demand for market commodities are
allowed to adjust to account for the increased production costs resulting from the regulation.
Using an iterative process, if the supply does not equal demand in all markets, a new set of prices
is "called out" and sent back to producers and consumers to "ask" what quantities they would
supply and demand based on these new prices. This technique is referred to as an auctioneer
approach because new prices are continually called out until an equilibrium set of prices is
determined (i.e., where supply equals demand for all markets).
Supply and demand quantities are computed at each price iteration. The market supply
for each energy and final product market is obtained by using a mathematical specification of the
supply function, and the key parameter is the point elasticity of supply at the baseline condition.
The demand curves for the energy markets are the sum of demand responses across all
markets. For example, the demand for natural gas is the sum of the demand for the electricity
industry, all manufacturing sectors, the commercial sector, and the residential sector. The
demand for electricity is the sum of the demand for the manufacturing sectors, the commercial
sector, and the residential sector. The demand for energy in the manufacturing sectors is a
derived demand calculated using baseline energy usage and changes associated with fuel
switching and changes in production levels.
The demand for final products in the two- and three-digit NAICS code manufacturing
sectors is obtained by using a mathematical specification of the demand function. Similarly, the
energy demand in the commercial and residential sectors is obtained through mathematical
specification of the demand functions (see Appendix A).
5-18
-------
EPA modeled fuel switching using secondary data developed by the U.S. Department of
Energy for the National Energy Modeling System (NEMS). Table 5-2 contains fuel price
elasticities of demand for electricity, natural gas, petroleum products, and coal. The diagonal
elements in the table represent own-price elasticities. For example, the table indicates that for
steam coal, a 1 percent change in the price of coal will lead to a 0.499 percent decrease in the use
of coal. The off diagonal elements are cross-price elasticities and indicate fuel switching
propensities. For example, for steam coal, the second column indicates that a 1 percent increase
in the price of coal will lead to a 0.061 percent increase in the use of natural gas.
Table 5-2. Fuel Price Elasticities
Own and Cross Elasticities in 2015
Inputs
Electricity
Natural Gas
Steam Coal
Residual
Distillate
Electricity
-0.074
0.496
0.021
0.236
0.247
Natural Gas
0.092
-0.229
0.061
0.036
0.002
Coal
0.605
1.087
-0.499
0.650
0.578
Residual
0.080
0.346
0.151
-0.587
0.044
Distillate
0.017
0.014
0.023
0.012
-0.055
Source: U.S. Department of Energy, Energy Information Administration (EIA). January 1998. Model Documentation Report:
Industrial Sector Demand Module of the National Energy Modeling System. DOE/EIA-M064(98). Washington, DC:
U.S. Department of Energy.
5.2.2 Calculating Changes in Social Welfare
The RICE MACT will impact almost every sector of the economy either directly through
control costs or indirectly through changes in the price of energy and final products. For
example, a share of control costs that originate in the energy markets is passed through the final
product markets and borne by both the producers and consumers of final products. To estimate
the total change in social welfare without double-counting impacts across the linked partial
equilibrium markets being modeled, EPA quantified social welfare changes for the following
categories:
5-19
-------
•• change in producer surplus in the energy markets,
change in producer surplus in the final product markets,
•• change in consumer surplus in the final product markets, and
•• change in consumer surplus in the residential, commercial, and transportation
energy markets.
Figure 5-5 illustrates the change in producer and consumer surplus in the intermediate
energy market and the final product markets. For example, assume a simple world with only one
energy market, wholesale electricity, and one final product market, pulp and paper. If the
regulation increased the cost of generating wholesale electricity, then part of the cost of the
regulation will be borne by the electricity producers as decreased producer surplus, and part of
the costs will be passed on to the pulp and paper manufacturers. In Figure 5-5(a), the pulp and
paper manufacturers are the consumers of electricity, so the change in consumer surplus is
displayed. This change in consumer surplus in the energy market is captured by the final
product market (because the consumer is the pulp and paper industry in this case), where it is
split between consumer surplus and producer surplus in those markets. Figure 5-5(b) shows the
change in producer surplus in the energy market, where B represents an increase in producer
surplus and C represents a decrease.
As shown in Figures 5-5(c) and 5-5(d), the cost affects the pulp and paper industry by
shifting up the supply curve in the pulp and paper market. These higher electricity prices
therefore lead to costs in the pulp and paper industry that are distributed between producers and
consumers of paper products in the form of lower producer surplus and lower consumer surplus.
Note that the change in consumer surplus in the intermediate energy market must equal the total
change in consumer and producer surplus in the final product market. Thus, to avoid double-
5-20
-------
(a) Change in Consumer
Surplus in the Energy
Market
(b) Change in Producer Surplus
in the Energy Market
s'
(c) Change in Consumer
Surplus in Final Product
Markets
(d) Change in Producer Surplus
in Final Product Markets
Figure 5-5. Changes in Economic Welfare with Regulation
5-21
-------
counting, the change in consumer surplus in the intermediate energy market was not quantified;
instead the total change in social welfare was calculated as
(5.1) Change in Social Welfare = • • PSE + • • PSF + • • CSF + • • CSR
where
• PSE = change in producer surplus in the energy markets,
• PSF = change in producer surplus in the final product markets,
• CSF = change in consumer surplus in the final product markets, and
• CSR = change in consumer surplus in the commercial, residential, and transportation
energy markets.
Appendix A contains the mathematical algorithms used to calculate the change in producer and
consumer surplus in the appropriate intermediate and final product markets.
The engineering control costs presented in Section 2.3 are inputs (regulatory "shocks") in
the market model approach. The magnitude and distribution of the regulatory costs' impact on
the economy depend on the relative size of the impact on individual markets (relative shift of the
market supply curves) and the behavioral responses of producers and consumers in each market
(measured by the price elasticities of supply and demand).
5.2.3 Supply and Demand Elasticities Used in the Market Model
The market model incorporates behavioral changes based on the price elasticities of
supply and demand. The price elasticities used to estimate the economic impacts presented in
Section 5.3 are given in Table 5-3. Because most of the direct cost impacts fall on engines
involved in the production of natural gas, the price elasticity of supply in the natural gas market
is one of the most important factors influencing the size and distribution of the economic impacts
associated with the RICE regulation. The supply elasticities in all of the other energy markets
also have a significant impact on the results. However, estimates of the elasticity of supply for
electric power were unavailable. This is in part because, under traditional regulation, the electric
utility industry had a mandate to serve all its customers. In addition, utilities' rates were
regulated and were based on allowing them to earn a market rate of return. As a result, the
5-22
-------
Table 5-3. Supply and Demand Elasticities
Elasticity of Demand
Energy Sectors
Electricity
Natural gas
Petroleum
Coal
Elasticity of
Supply
0.75
0.41b
0.58b
1.0C
Manufacturing
Derived demand
Derived demand
Derived demand
Derived demand
Commercial"
-0.24
-0.47
-0.28
-0.28
Transportation3
-0.24
-0.47
-0.28
-0.28
Residential3
-0.23
-0.26
-0.28
-0.28
a Energy Information Administration. 2000. "Issues in Midterm Analysis and Forecasting 1999—Table 1."
. As obtained on May 8, 2000.
b Dahl, Carol, and Thomas E. Dugan. 1996. U.S. Energy Product Supply Elasticities: A Survey and Application to the U.S. Oil
Market. Resource and Energy Economics 18:243-263.
c Zimmerman, M.B. 1977. "Modeling Depletion in the Mineral Industry: The Case of Coal." The Bell Journal of Economics
8(2):41-65.
market concept of supply elasticity was not the driving force in utilities' capital investment
decisions. However, wholesale market deregulation was initiated by the Energy Policy Act of
1992 and most states have begun to address the issue of retail deregulation. The overall trend is
clearly toward deregulation of retail electric markets and the movement is gaining momentum.
In future years, the market for electric power will probably look more like a typical competitive
industry because of deregulation. To operationalize the model, a supply elasticity of 0.75 was
assumed for the electricity market based on an assumption that the supply of electricity is fairly
inelastic in the short run.
In contrast, many studies have been conducted on the elasticity of demand for electricity,
and it is generally agreed that, in the short run, the demand for electricity is relatively inelastic.
Most residential, commercial, and industrial electricity consumers do not significantly adjust
short-run behavior in response to changes in the price of electricity. The elasticity of demand for
electricity is primarily driven by long-run decisions regarding equipment efficiency and fuel
substitution.
Additional elasticity of demand parameters for the residential, commercial, and
transportation sectors were obtained from the Energy Information Administration by fuel type
(natural gas, petroleum, coal). The demand elasticities also have a very significant impact on the
5-23
-------
model results. The elasticities of demand for energy are not provided for manufacturing because
the model calculates the derived demand from this sector for each of the energy markets modeled
based on the estimated output from these markets. In effect, adjustments in the final product
markets due to changes in production levels and fuel switching are used to estimate changes in
energy demand, eliminating the need for demand elasticity parameters in the energy markets.
Supply and demand elasticities for goods and services produced in the industrial and commercial
markets are reported in Table 5-4. Appendix B contains a sensitivity analysis for the key supply
and demand elasticity assumptions.
5.3 ECONOMIC IMPACT ESTIMATES
This study used a market model to estimate total changes in social welfare and to
investigate the distribution of impacts between consumers and producers. In addition, producer
impacts are distributed across industries within the energy and manufacturing sectors.
Table 5-5 summarizes the economic impact estimates. The total change in social welfare
in 2005 is estimated to be $247.55 million. This estimate includes market adjustments in final
product markets and fuel switching adjustments in the manufacturing sector in response to
changes in relative prices. For comparison, the baseline engineering costs and social costs
without fuel switching are also presented in Table 5-5. Social welfare losses in the model with
fuel switching adjustments are $0.02 million less than the estimated baseline engineering costs as
a result of behavior changes by producers and consumers that reflect lower cost alternatives.
5-24
-------
Table 5-4. Supply and Demand Elasticities for Industrial and Commercial Sectors
NAICS Description
Industrial Sectors
1 1 Agricultural Sector
21 Other Mining Sector
23 Construction Sector
311 Food
312 Beverage and Tobacco Products
313 Textile Mills
3 14 Textile Product Mills
315 Apparel
316 Leather and Allied Products
321 Wood Products
322 Paper
323 Printing and Related Support
325 Chemicals
326 Plastics and Rubber Products
327 Nonmetallic Mineral Products
331 Primary Metals
332 Fabricated Metal Products
333 Machinery
334 Computer and Electronic Products
335 Electrical Equip., Appliances, and
Components
336 Transportation Equipment
337 Furniture and Related Products
339 Miscellaneous
Commercial Sector (NAICS 42-45;51-56;61-72)
Supply3
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
Demandb
-1.80
-0.30
-1.00
-1.00
-1.30
-1.50
-1.50
-1.10
-1.20
-1.00
-1.50
-1.80
-1.80
-1.80
-1.00
-1.00
-0.20
-0.50
-0.30
-0.50
-0.50
-1.80
-0.60
-1.00
a Assumed supply elasticity. Sensitivity analysis of this assumption is presented in Appendix B.
b Source: Personal communication from Larry Sorrels, EPA to Mike Gallaher, RTI. August 15, 2000. Qualitative Market
Assessment—PM NAAQS.
5-25
-------
Table 5-5. Summary Table
Change in Social Welfare
(Millions of $1998)
Engineering control costs 247.57
Social costs with market adjustments 247.56
Social costs with market adjustments and fuel 247.55
switching
Total reporting and record keeping costs 6.15
Total social costs 253.73
Table 5-6 presents the distribution of economic impacts between producers and
consumers and shows the distribution of impacts across sectors/markets. The market analysis
estimates that consumers will bear a burden of $125.4 million in 2005 (51 percent of the total
social cost) because of the increased price of energy, the increased prices of final products, and
the smaller quantities of energy and final products generally available. Producer surplus is
projected to decrease by $122.1 million in 2005 (49 percent of the total social cost) as a result of
the direct control costs, higher energy costs, and reductions in output with the majority of the
producer surplus losses logically falling on natural gas producers because the rule applies to
engines that are primarily used in natural gas production. The costs to natural gas producers are
approximately 29 percent of the total producer surplus loss or 14 percent of the total social cost
of the regulation. Producer surplus also falls in the petroleum products market and in each of the
final product markets. However, there are energy markets in which producer surplus actually
increases as a result of the regulation. In particular, both the electricity and coal markets
experience increases in producer surplus. Like natural gas producers, the producers of electricity
and coal also face higher input costs due to increases in the price of oil and natural gas.
However, the increase in input costs is much less for these producers than the increase in costs
applied to natural gas and oil producers. As a result, the supply curve shifts less for electricity
and coal than for natural gas and petroleum products, and the price does not increase as much.
The fact that the prices of electricity and coal increase less than those of natural gas and
5-26
-------
Table 5-6. Distribution of Social Costs
Change in:
Sectors/Markets
Energy Markets
Producer Surplus Consumer Surplus Social Welfare
Petroleum (NAICS 3241 1, 4861)
Natural gas (NAICS
21111,4862,2212)
Electricity (NAICS 2211 1, 221 122, 221 121)
Coal (NAICS 2121)
Subtotal
NAICS Code
Industrial Sector
11
21
23
311
312
313
314
315
316
321
322
323
325
326
327
331
332
333
334
335
336
337
339
Description
Agricultural Sector
Other Mining Sector
Construction Sector
Food
Beverage and Tobacco Products
Textiles Mills
Textile Product Mills
Apparel
Leather and Allied Products
Wood Products
Paper
Printing and Related Support
Chemicals
Plastics and Rubber Products
Nonmetalic Mineral Products
Primary Metals
Fabricated Metal Products
Machinery
Computer and Electronic Products
Electrical Equipment, Appliances, and
Transportation Equipment
Furniture and Related Products
Miscellaneous
Industrial Sector Subtotal
Commercial Sector
Residential Sector
Transportation Sector
Subtotal
-$6.0
-$35.2
$3.2
$0.3
-$38.3
-$1.6
-$6.0
-$6.3
-$3.4
-$0.6
-$0.5
-$0.1
-$0.1
-$0.0
-$0.3
-$3.5
-$0.3
-$12.6
-$1.5
-$2.0
-$3.9
-$0.4
-$0.3
-$0.2
-$0.2
-$0.7
-$0.2
-$0.1
-$44.7
-$39.1
NA
NA
-$83.8
NA
NA
NA
NA
NA
-$0.7
-$15.0
-$4.7
-$2.5
-$0.3
-$0.3
-$0.1
-$0.1
-$0.0
-$0.3
-$1.7
-$0.1
-$5.2
-$0.6
-$1.5
-$2.9
-$1.4
-$0.5
-$0.5
-$0.3
-$1.0
-$0.1
-$0.2
-$39.9
-$29.3
-$40.0
-$16.2
-$125.4
NA
NA
NA
NA
NA
-$2.3
-$21.0
-$11.1
-$5.9
-$1.0
-$0.8
-$0.2
-$0.2
-$0.0
-$0.6
-$5.2
-$0.4
-$17.8
-$2.1
-$3.5
-$6.7
-$1.8
-$0.8
-$0.6
-$0.4
-$1.7
-$0.2
-$0.3
-$84.6
-$68.4
-$40.0
-$16.2
-$209.2
petroleum cause electricity and coal to become more attractive to energy consumers because
they have become relatively less expensive energy sources following the regulation despite their
5-27
-------
increase in price. This leads to an increase in the demand for electricity and coal as some
consumers switch their fuel usage to consume a smaller proportion of natural gas and petroleum
products and a larger proportion of electricity and coal due to the changing incentives facing
them as relative prices of energy products change. Consumers change their consumption until
the energy markets once again reach equilibrium at new levels of price and output. The increase
in demand for electricity and coal resulting from fuel switching by energy users outweighs the
increase in input costs and leads to increases in producer surplus in these two markets.
The total welfare loss for the industrial sectors affected by the rule is estimated to total
approximately $39.9 million for consumers and $44.7 million for producers in the aggregate, but
product prices and output do not show substantial changes. This may occur because in
comparison to the projected energy expenditures in these industries (estimated to be $180 billion
in 1998 [EIA, 2000]), the cost of this rule to producers as a percentage of their energy
expenditures is only 0.06 percent. Also, the total value of shipments for the affected industrial
sectors was $5.0 trillion in 1998, so the cost to consumers of these products as a percentage of
spending on the outputs from these industries is less than 0.01 percent.
The cost to residential consumers of energy is estimated to be $40.0 million. This cost
represents 0.04 percent of the projected annual residential energy expenditures of $111 billion
(EIA, 2000). The commercial sector also experiences a large portion of the total social cost with
an impact to this sector estimated at $68.4 million. For the commercial sector, energy
expenditures are projected to be $92 billion (EIA, 2000c). Therefore, the regulatory burden
associated with the RICE MACT is estimated as 0.07 percent of total energy expenditures by the
commercial sector. The cost to transportation consumers is estimated by the economic model to
be $16.2 million. This cost represents approximately 0.01 percent of energy expenditures for the
transportation sector ($16.2 million/$241 billion [EIA, 2000c]).
The equilibrium changes in price and quantity in the energy markets are presented in
Table 5-7. In both the petroleum and natural gas markets, output decreases and price increases
in response to the direct control costs. These control costs increase the cost of producing these
products and decrease the supply, resulting in producer surplus losses of $6.0 million and $35.2
million, respectively. The impacts are greater in the natural gas market because that is where the
majority of the affected engines operate. Even with the relatively large cost in the natural gas
5-28
-------
market, natural gas prices are estimated to increase by only 0.101 percent, while the impacts in
the other energy markets are expected to be much smaller as shown in Table 5-7. This increase
in the price of natural gas is reasonable given the engineering cost impact on the natural gas
market, which is estimated to be 0.132 percent of the initial price, and the increased cost of fuel
as an input into producing natural gas for consumption. The total cost impact of these two
effects is 0.135 percent of the initial market price of natural gas. The market price is expected to
increase by less than the increase in engineering costs and input fuel costs because the economic
model allows producers and consumers to change their behavior in response to price changes.
As price increases, consumers reduce the quantity that they are willing to purchase. Therefore, if
producers attempted to simply increase the price of their product by the full amount that their
costs increased, then there would be a surplus of natural gas because consumers would not be
willing to continue purchasing the initial quantity at a higher price. Producers would then
respond by lowering prices until a new equilibrium is reached to avoid holding excess inventory.
The market for petroleum products faces a similar situation. The engineering costs entering the
economic model are estimated to be 0.005 percent of the initial price. Adding in the increased
costs of energy used in the production of petroleum products, the total cost impact is about 0.007
percent of initial market price, whereas the model results indicate a 0.005 percent increase in the
price of petroleum products after taking behavioral responses into account.
In the electricity market, both price and quantity increase slightly (by 0.022 percent and
0.001 percent, respectively), which implies that, although the supply in this market decreases,
there is an increase in demand that is larger than the decrease in supply and which leads to a
minimal increase in equilibrium quantity. This is presumably due to consumers changing their
fuel usage in response to higher prices for natural gas and petroleum. In the petroleum products,
natural gas, and electricity markets, the change in price is larger in magnitude than the change in
5-29
-------
Table 5-7. Market-Level Impacts
Percent Change
Sectors/Markets
Energy Markets
Petroleum (NAICS 3241 1, 4861)
Natural gas (NAICS
21111,4862,2212)
Electricity (NAICS 221 1 1, 221 122, 221 121)
Coal (NAICS 2121)
NAICS Code
Industrial Sectors
11
21
23
311
312
313
314
315
316
321
322
323
325
326
327
331
332
333
334
335
336
337
339
Commercial Sector
Description
Agricultural Sector
Other Mining Sector
Construction Sector
Food
Beverage and Tobacco Products
Textiles Mills
Textile Product Mills
Apparel
Leather and Allied Products
Wood Products
Paper
Printing and Related Support
Chemicals
Plastics and Rubber Products
Nonmetalic Mineral Products
Primary Metals
Fabricated Metal Products
Machinery
Computer and Electronic Products
Electrical Equipment, Appliances, and
Transportation Equipment
Furniture and Related Products
Miscellaneous
Price
0.005%
0.101%
0.022%
0.001%
0.000%
0.020%
0.001%
0.001%
0.000%
0.000%
0.000%
0.000%
0.000%
0.000%
0.001%
0.000%
0.001%
0.000%
0.002%
0.001%
0.001%
0.000%
0.000%
0.000%
0.000%
0.000%
0.000%
0.000%
Quantity
-0.001%
-0.0140%
0.001%
0.001%
-0.001%
-0.006%
-0.001%
-0.001%
0.000%
-0.001%
0.000%
0.000%
0.000%
0.000%
-0.001%
0.000%
-0.002%
-0.001%
-0.002%
-0.001%
0.000%
0.000%
0.000%
0.000%
0.000%
0.000%
0.000%
0.000%
quantity because demand is more inelastic than supply in these markets, meaning that quantity is
relatively unresponsive to changes in price. Price and quantity both increase in the coal market
also (by 0.001 percent for both price and quantity), again because of a positive demand shift that
outweighs any negative supply shift resulting from an increase in the energy input costs for coal
production. Demand from utilities and other consumers is increasing due to switching towards
coal usage as well as the increase in output of electricity. Because the primary users of coal are
5-30
-------
electricity producers and much of the electricity produced in the U.S. is produced at coal burning
plants, an increase in the equilibrium quantity of electricity will lead to an increase in the derived
demand for coal from the utilities.
Table 5-7 also provides the percentage change in price and quantity for the
manufacturing final product markets. The regulation increases the price of final products in all
markets and decreases the quantity. The final product markets behave similarly to the petroleum
and natural gas markets. In each case, the estimated increase in price is less than the engineering
costs facing that particular product market. In general, the changes in price and quality are very
small. Only one market has a change in price or quantity greater than or equal to 0.02 percent.
That market is mining and the other mining sector (NAICS 21), which has an estimated increase
in price of 0.02 percent and an estimated decrease in quantity of 0.006 percent.
Although the impacts on price and quantity in the final product markets are estimated to
be small, one possible effect of modeling market impacts at the two- and three-digit NAICS code
level is that there may potentially be fuel-intensive industries within the larger NAICS code
definition that are affected more significantly than the average for that NAICS code. Thus, the
changes in price and quantity should be interpreted as an average for the whole NAICS code, not
necessarily for each disaggregated industry within that NAICS code.
These results have some uncertainty associated with them due to assumptions that are
made to operationalize the model. A full discussion of these uncertainties is provided in
Appendix C.
5.4 CONCLUSIONS
The total social cost estimated using the market analysis is $253.73 million in the year
2005. The economic impact from the market analysis is $0.02 million less than the estimated
baseline engineering costs because the market model accounts for behavioral changes of
producers and consumers. Although the rule affects engines that are primarily used in the
natural gas industry, the natural gas producers incur only 14 percent of the total social cost of the
regulation. The burden is spread across numerous markets because the price of energy increases
slightly as a result of the regulation, which increases the cost of production for all markets that
use energy as part of their production process.
5-31
-------
The market model estimates that the regulation will increase the cost of producing
petroleum products and natural gas, leading to decreases in the quantity of these products
produced and increases in their prices. Because of fuel switching away from natural gas and
petroleum and towards electricity and coal taking place, both the electricity and coal markets
have increases in demand that outweigh any reduction in supply caused by an increase in input
prices. The market analysis also indicates that the impacts of the regulation will be borne
primarily by natural gas producers and consumers in the manufacturing, commercial, and
residential sectors. The manufacturing markets that are most affected are the other mining sector
(NAICS 21), food (NAICS 311), chemicals (NAICS 325), and construction (NAICS 23)
markets.
Because of the minimal changes in price and quantity estimated for most of the affected
markets, EPA expects that there would be no discernable impact on international trade.
Although an increase in the price of U.S. products relative to those of foreign producers is
expected to decrease exports and increase imports, the changes in price due to the RICE MACT
are generally too small to significantly influence trade patterns. In addition, the market facing
the largest increase in price is the natural gas market, but imports of natural gas are essentially
limited to Canadian gas, which will also be subject to at least some of the costs of the regulation
as it is transported through pipelines in the U.S. There may also be a small decrease in
employment, but because the impact of the regulation is spread across so many industries and the
decreases in market quantities are so small, it is unlikely that any particular industry will face a
significant decrease in employment.
Because of the decrease in the quantity of natural gas and petroleum products projected
due to the RICE MACT, as well as the decrease in output in the final product markets, it is
expected that fewer new engines will be installed than in the absence of the regulation.
Table 5-8 shows the regulation's estimated impact on the number of new engines installed based
on a constant number of engines being added per unit of output in each affected market. The
manufacturing markets category is the sum of engines used in all 24 manufacturing markets
included in this analysis. However, the changes in quantity projected in each of these markets
were so small that none of the manufacturing markets were projected to have any reduction in
the number of new engines installed. The category labeled "other" contains all of the engines in
5-32
-------
the commercial market. Because the quantity of output was assumed unchanged in these
markets, it is assumed that the number of engines demanded in these sectors will also remain
constant. Because the percentage changes in price and quantity are so small, the estimated
impact on the number of engines is extremely small. According to the economic model,
approximately 2 fewer engines (0.01 percent of the projected total) will be installed due to the
regulation because of reductions in output in the natural gas and manufacturing markets.
Table 5-8. Impacts on the Number of New Engines Installed
New Engines
Baseline
With Regulation
Natural gas market
Petroleum products market
Manufacturing, mining, and agricultural markets
Commercial markets
Total
11,581
1,602
3,405
3,721
20,309
11,579
1,602
3,405
3,721
20,307
5-33
-------
6.0 IMPACTS ON FIRMS OWNING RICE UNITS
The regulatory costs imposed on domestic producers to reduce air emissions from
internal combustion engines will have a direct impact on owners of the affected facilities. Firms
or individuals that own the facilities with internal combustion engines are legal business entities
that have the capacity to conduct business transactions and make business decisions that affect
the facility. The legal and financial responsibility for compliance with a regulatory action
ultimately rests with these owners, who must bear the financial consequences of their decisions.
Environmental regulations, such as the proposed internal combustion engine standard, affect
both large and small entities (businesses or governments), but small entities may have special
problems in complying with such regulations.
The Regulatory Flexibility Act (RFA) of 1980 requires that special consideration be
given to small entities affected by federal regulation. Specifically, the RFA requires determining
whether a regulation will significantly affect a substantial number of small entities or cause a
disproportionate burden on small entities in comparison with large companies. In 1996, the
Small Business Regulatory Enforcement Fairness Act (SBREFA) was passed, which further
amended the RFA by expanding judicial review of agencies' compliance with the RFA and by
expanding small entity review of EPA rulemaking.
This analysis assesses the potential impacts of the standard on small entities. To make
this assessment, the costs of the regulation are, to the extent possible, mapped to firm-level data
(or government-level data) and proportional cost effects are estimated for each identified firm (or
government). Then, the focus is placed on small firms and the question of whether there are a
6-1
-------
substantial number with a large regulatory cost-to-sales impact. The control costs under the
MACT floor are used to estimate cost-to-sales ratios (CSRs).
6.1 IDENTIFYING SMALL BUSINESSES
To support the economic impact analysis of the proposed regulation, EPA identified
26,832 engines located at commercial, industrial, and government facilities. The population of
engines was developed from the EPA Industrial Combustion Coordinated Rulemaking (ICCR)
Inventory Database version 4.1.l The list of engines contained in these databases was developed
from information in the AIRS and OTAG databases, state and local permit records, and the
combustion source ICR conducted by the Agency. Industry and environmental stakeholders
reviewed the units contained in these databases as part of the ICCR FACA process. In addition,
stakeholders contributed to the databases by identifying and including omitted units.
Information was extracted from the ICCR databases to support the engines NESHAP. This
modified database containing information on only engines is referred to as the Inventory
Database.
From this initial population of 26,832 engines, 10,118 engines were excluded because the
proposed regulation will not cover engines smaller than 500 hp or engines used to supply
emergency/backup power. Table 6-1 provides the remaining population of 16,714 engines,
broken out by industry SIC code, the format in which the database was originally constructed.
Although data used in the economic model was later converted to NAICS, the data presented in
this table is by SIC code because there was insufficient data to map units without control costs to
the appropriate NAICS code.
Because it is not possible to project specific companies or government organizations that
will purchase new engines in the future, the small business screening analysis for the RICE
MACT is based on the evaluation of existing owners of engines. It is assumed that the existing
'The ICCR Inventory Database contains data for boilers, process heaters, incinerators, landfill gas flares, turbines,
and internal combustion engines.
6-2
-------
Table 6-1. Unit Counts and Percentages by Industry
Industry (SIC)
Agriculture (01 -09)
Mining (10-12, 14)
Petroleum & Natural Gas
Exploration (13)
Construction (15-17)
Manufacturing (20-39)
Utility Services (40-48)
Electricity & Gas Services
(49)
Wholesale Trade (50-51)
Retail Trade (52-59)
Finance, Real Estate, &
Insurance (60-67)
Services (70-89)
Government (90-98)
Not Elsewhere Classified (99)
Unknown
Total
Subset Mapped with
Control Costs
Number of
Units
1
33
1,145
1
57
9
1,306
1
4
6
50
4
0
28
2,645
Percentage of
Total Units
0.04
1.25
43.29
0.04
2.16
0.34
49.38
0.04
0.15
0.23
1.91
0.15
1.07
Inventory
Number
of Units
8
663
6,191
84
1,547
241
6,371
171
26
84
331
387
41
670
16,714
Database
Percentage
of Total
Units
0.05
3.97
37.04
0.50
9.26
1.44
38.12
1.02
0.16
0.50
1.98
2.32
0.25
4.01
size and ownership distribution of engines in the Inventory Database is representative of the
future growth in new engines. The remainder of this section presents cost and sales information
on small companies and government organizations that own existing engines.
6.2 SCREENING-LEVEL ANALYSIS
To conduct the small entity analysis, unit model numbers (Alpha Gamma Technologies,
Inc., 2000) were linked to individual units (engines) at affected facilities so that parent
6-3
-------
companies' aggregate control costs could be compared to company sales. Of the 16,714 affected
units in the Inventory Database, 2,645 units had sufficient information to assign model numbers.
Table 6-1 compares the unit counts and percentage of units by industry for the total Inventory
Database population and the subset of units used in the small entity analysis.
As indicated in Table 6-1, the subset of units used in the small entity analysis is fairly
representative of the population in the Inventory Database because the percentage of units in
each SIC code is similar to the percentage in the Inventory Database for most industries.
Petroleum & Natural Gas Exploration (NAICS 211) and Electricity & Gas Services (SIC
49/NAICS 221/486) account for the majority of units in both the Inventory Database and subset
populations.
6.3 ANALYSIS OF FACILITY-LEVEL AND PARENT-LEVEL DATA
The 2,645 units in the Inventory Database with full information were linked to 834
existing facilities. As shown in Table 6-2, these 834 facilities are owned by 153 parent
companies.
Employment and sales are typically used as measures of business size. Employment,
sales, and tax revenue data (when applicable) were collected for 141 of the 153 parent
companies.2 Sales and employment information was unavailable for 12 parent companies.
Figure 6-1 shows the distribution of employees by parent company. Employment for parent
companies ranges from 5 to 96,650 employees. Fifty-eight of the firms have fewer than 500
employees, and seven companies have more than 25,000 employees.
2Total annualized cost is compared to tax revenue to assess the relative impact on local governments.
6-4
-------
Table 6-2. Facility-Level and Parent-Level Data
NAICS
112
211
212
221
234
311
312
322
324
325
326
327
331
421
441
486
488
524
531
541
562
611
622
922
Unknown
Total
Industry Description
Animal Production
Oil and Gas Extraction
Mining (Except Oil and Gas)
Utilities
Heavy Construction
Food Manufacturing
Beverage and Tobacco Product Manufacturing
Paper Manufacturing
Petroleum and Coal Products Manufacturing
Chemical Manufacturing
Plastics and Rubber Products Manufacturing
Nonmetallic Mineral Product Manufacturing
Primary Metal Manufacturing
Wholesale Trade, Durable Goods
Motor Vehicle and Parts Dealers
Pipeline Transportation
Support Activities for Transportation
Insurance Carriers and Related Activities
Real Estate
Professional, Scientific, and Technical Services
Waste Management and Remediation Services
Educational Services
Hospitals
Justice, Public Order, and Safety Activities
Industry Classification Unknown
Number of
Facilities
1
312
28
15
1
4
1
1
7
4
1
1
1
1
1
424
1
3
1
1
1
1
20
1
2
834
Number of
Parent
Companies
1
37
16
9
1
4
0
1
5
3
2
0
1
0
1
48
1
3
1
0
0
1
17
1
153
Average
Number of
Facilities per
Parent
Company
1.0
8.4
1.8
1.7
1.0
1.0
1.0
1.4
1.3
0.5
1.0
1.0
8.8
1.0
1.0
1.0
1.0
1.2
1.0
Source: Industrial Combustion Coordinated Rulemaking (ICCR). 1998. Data/Information Submitted to the Coordinating
Committee at the Final Meeting of the Industrial Combustion Coordinated Rulemaking Federal Advisory
Committee. EPA Docket Numbers A-94-63, II-K-4b2 through -4b5. Research Triangle Park, North Carolina.
September 16-17.
6-5
-------
^n
1 -5
E
S. 20 -
•5
l_ -1C
0) l3
.a
E 10
3 IU
z
c
n
<25 25 to 49 50 to 99 100 to 500 to 1,000 to 5,000 to >25,000
499 999 4,999 24,999
Parent Employment
Figure 6-1. Parent Size by Employment Range
Includes 141 parent companies for which data are available.
Sales provide another measure of business size. Figure 6-2 presents the sales distribution
for affected parent companies. The median sales figure for affected companies is $300 million,
and the average sales figure is $4.7 billion (excluding the federal government). As shown in
Figure 6-2, the distribution of firm sales is fairly evenly distributed, but approximately two-
thirds of all parent companies have sales greater than $100 million. These figures include all
sales associated with the parent company, not just facilities affected by the regulation (i.e.,
facilities with internal combustion engines).
6-6
-------
oc;
1
§ ?n
TO ^
D.
"o 15
-° 10
E 10
3
2 5
<5
5 to 9 10 to 49 50 to 99 100 to 500 to 1,000 to 5, 000 to 10, 000 to >25,000
499 999 4,999 9,999 24,999
Parent Sales ($106)
Figure 6-2. Number of Parents by Sales Range
Includes 141 parent companies for which data are available.
Based on Small Business Administration guidelines (SBA, 1999), 47 entities were
identified as small. Small businesses by business type are presented in Table 6-3.3 The oil and
gas extraction industry and the mining industry each have 14 small companies. Seven small
companies are in the utilities industry and 5 are in pipeline transportation. The remaining small
businesses are distributed across seven different three-digit NAICS code groupings. Also, six
cities are classified as small governments because they have fewer than 50,000 residents, based
on guidelines established by EO 12875.
3Small business guidelines typically define small businesses based on employment, and the threshold varies from
industry to industry. For example, in the paints and allied products industry, a business with fewer than 500
employees is considered a small business; whereas in the industrial gases industry, a business with fewer than
1,000 employees is considered small. However, for a few industries, usually services, sales are used as the
criterion. For example, in the veterinary hospital industry, companies with less than $5 million in annual sales
are defined as small businesses.
6-7
-------
Table 6-3. Small Parent Companies
NAICS
112
211
212
221
234
311
312
322
324
325
326
327
331
421
441
486
488
524
531
541
562
611
622
922
Unknown
Total
Industry Description
Animal Production
Oil and Gas Extraction
Mining (Except Oil and Gas)
Utilities
Heavy Construction
Food Manufacturing
Beverage and Tobacco Product Manufacturing
Paper Manufacturing
Petroleum and Coal Products Manufacturing
Chemical Manufacturing
Plastics and Rubber Products Manufacturing
Nonmetallic Mineral Product Manufacturing
Primary Metal Manufacturing
Wholesale Trade, Durable Goods
Motor Vehicle and Parts Dealers
Pipeline Transportation
Support Activities for Transportation
Insurance Carriers and Related Activities
Real Estate
Professional, Scientific, and Technical Services
Waste management and Remediation Services
Educational Services
Hospitals
Justice, Public Order, and Safety Activities
Industry Classification Unknown
Number of
Facilities
1
312
28
15
1
4
1
1
7
4
1
1
1
1
1
424
1
3
1
1
1
1
20
1
2
834
Number of
Parent
Companies
1
37
16
9
1
4
0
1
5
3
2
0
1
0
1
48
1
3
1
0
0
1
17
1
153
Number of
Small
Companies
0
14
14
7
1
2
0
1
2
1
0
0
0
0
0
5
0
0
0
0
0
0
0
0
47
Source: Industrial Combustion Coordinated Rulemaking (ICCR). 1998. Data/Information Submitted to the Coordinating
Committee at the Final Meeting of the Industrial Combustion Coordinated Rulemaking Federal Advisory
Committee. EPA Docket Numbers A-94-63, II-K-4b2 through -4b5. Research Triangle Park, North Carolina.
September 16-17.
6.4 SMALL BUSINESS IMPACTS
-------
Although there are a total of 47 small entities identified in the Inventory Database, only
13 of them own 4SRB engines. As mentioned in previous sections, the only existing engines
affected by the rule are 4SRB units, while all other types of engines will only have requirements
on new engines rather than existing units. These small entities own a total of 39 4SRB units at
21 facilities. The impacts on the affected entities in the Inventory Database are summarized in
Table 6-4 assuming that each of the 39 4SRB units are located at major sources. This is an upper
bound cost scenario because only 40 percent of all RICE units are estimated to be at major
sources, and therefore subject to the rule. Based on this percentage, only about 16 of the 39
4SRB units identified at facilities owned by small businesses would be located at major sources.
It is reasonable to expect that the percentage of facilities owned by small businesses that are
major sources would be lower than the average for the whole source category, so even fewer
existing 4SRB owned by small businesses may be affected. However, because it is unknown
which facilities are major sources and which are area sources, it was assumed that all existing
4SRB owned by small businesses are located at major sources and subject to the rule to provide
a conservative estimate of the small business impacts. Even under this scenario, there are no
small firms that have compliance costs above 3 percent of firm revenues and only two small
firms owning 4SRB engines that have impacts between 1 and 3 percent of revenues. In addition
to twelve small firms with 4SRB engines, there is one small government in the Inventory
Database affected by this rule. The costs to this city are approximately $3 per capita annually
assuming their engine is affected by the rule, less than 0.01 percent of median household income.
6-9
-------
Table 6-4. Summary Statistics for SBREFA Screening Analysis:
Existing Affected Small Entities
Total Number of Small Entities 13a
Average Annual Compliance Cost ($106/yr)b $120,067
Small Entities with Sales/Revenue Datab Number Share
Compliance Costs < 1% of sales 10 83.3%
Compliance Costs between 1 and 3% of sales 2 16.7%
Compliance Costs > 3% of sales 0 0.0%
Total 12 100.0%
Compliance Cost-to-Sales Ratios Descriptive Statisticsb
Average 0.73%
Median 0.58%
Minimum 0.06%
Maximum 2.27%
a One of these is a small city for which no sales were available.
b Assumes no market responses (i.e., price and output adjustments) by regulated entities and that all of these entities are
classified as major sources (upper bound cost scenario).
Based on this subset of the existing engines population, the regulation will affect no
small entities owning RICE at a CSR greater than 3 percent, while approximately 4 percent
(2/47) of small entities owning RICE greater than 500 hp will have compliance costs between 1
and 3 percent of sales under an upper bound cost scenario. The total existing population of
engines with greater than 500 hp that are not backup units is estimated to be 22,018 (Alpha
Gamma, 2002a). Assuming the same breakdown of large and small company ownership of
engines in the total population of existing engines as in the subset with parent company
information identified, the Agency expects that approximately 17 small entities in the existing
population of RICE owners would have CSRs between 1 and 3 percent under an upper bound
cost scenario where all RICE owned by small entities are located at major sources.
In addition, because many small entities owning RICE will not be affected because of the
exclusion of engines with less than 500 hp, the percentage of all small companies owning RICE
that are affected by this regulation is even smaller. Based on the proportion of engines in the
Inventory Database that are greater than 500 hp and are not backup units (16,714/26,832, or 62.3
6-10
-------
percent) and assuming that small companies own the same proportion of small engines (less than
500 hp) as they do of engines greater than 500 hp, the Agency estimates that 628 small
companies own RICE. Of all small companies owning RICE, 2.7 percent (17/628) are expected
to have CSRs between 1 and 3 percent under an upper bound cost scenario. If the percentage of
RICE owned by small companies that are located at major sources is the same as the engine
population overall (40 percent), only about 1.1 percent of small companies owning RICE would
be expected to have CSRs greater than 1 percent.
6.5 ASSESSMENT OF SMALL ENTITY SCREENING
As outlined above, this regulation will affect only a very small percentage of small
entities owning RICE units. To determine whether the impacts on existing small entities are
significant, typical profit margins in the affected industries were considered. The engines
included in the database are owned and operated in more than 25 different industries, but the
majority of the small businesses affected by the proposed regulation are in the oil and gas
extraction; mining and quarrying; and electric, gas, and sanitary services sectors (see Table 6-3).
As shown in Table 6-5, the average profit margin for these sectors is approximately 5 percent.
Table 6-5 also shows the profit margins for the other industry sectors with affected small entities.
All profit margins of industry sectors with affected small businesses are above 2 percent. Based
on this median profit margin data, it seems reasonable to review the number of small firms with
CSRs above 3 percent in screening for significant impacts.
This analysis shows that none of the small entities in the Inventory Database have
impacts greater than 5 percent and only two small firms have impacts between 1 and 3 percent
even under an upper bound cost scenario. Based on the low number of affected small firms, the
fact that no small firms have CSRs between 3 and 5 percent, and the fact that industry profit
margins average 5 percent, this analysis concludes that this proposed regulation will not have a
significant impact on a substantial number of existing small entities.
For new sources, it can be reasonably assumed that the investment decision to purchase a
new engine may be slightly altered as a result of the regulation. For the entire population of
affected engines projected to exist in 2005, the economic model predicts 2 fewer engines (0.01
percent of the projected total in the absence of the regulation) will be purchased because of
6-11
-------
Table 6-5. Profit Margins for Industry Sectors with Affected Small Businesses
NAICS
212
211
212
234
311
322
325
221/486
Industry Description
Metal Mining
Oil & Gas Extraction
Mining & Quarrying of Nonmetallic
Minerals, Except Fuels
Heavy Construction
Food & Kindred Products
Paper & Allied Products
Chemicals & Allied Products
Electric, Gas, & Sanitary Services
Median Profit Margin
5.1%
4.6%
2.1%
3.5%
3.6%
3.3%
2.7%
7.5%
Source: Dun & Bradstreet 1997. Industry Norms & Key Business Ratios. Desktop Edition 1996-97. Murray Hill, NJ: Dun &
Bradstreet, Inc.
market responses to the regulation. Specifically, the slight declines in output in industries that
use RICE leads to a small decrease in the number of engines needed to produce that output. It is
not feasible, however, to determine future investment decisions at the small entities in the
affected industries, so EPA cannot link these 2 engines to any one firm (small or large). Overall,
it is very unlikely that a substantial number of small firms who may consider purchasing a new
engine will be significantly affected because the decision to purchase new engines is not altered
to a large extent. In addition, the rule is likely to increase profits at the many small firms owning
RICE that are not affected by the rule by increasing their revenues due to the estimated increase
in prices in the energy markets and final product markets.
Although this proposed rule will not have a significant economic impact on a substantial
number of small entities, EPA nonetheless has tried to reduce the impact of this rule on small
entities. In this proposed rule, the Agency is applying the minimum level of control (i.e., the
MACT floor) and the minimum level of monitoring, recordkeeping, and reporting to affected
sources allowed by the CAA. In addition, as mentioned earlier in this report, new RICE units
with capacities under 500 hp and those that operate as emergency/temporary units are not
covered by this proposed rule. This provision should reduce the level of small entity impacts.
EPA continues to be interested in the potential impacts of the proposed rule on small entities and
welcomes comments on issues related to such impacts.
6-12
-------
6-13
-------
7.0 QUALITATIVE ASSESSMENT OF BENEFITS
OF EMISSION REDUCTIONS
The emission reductions achieved by this environmental regulation will provide benefits
to society by improving environmental quality. In this chapter, and the following chapter,
information is provided on the types and levels of social benefits anticipated from the RICE
NESHAP. This chapter discusses the health and welfare effects associated with the HAPs and
other pollutants emitted by RICE. The following chapter places a monetary value of a portion of
the benefits that are described here.
In general, the reduction of HAP emissions resulting from the regulation will reduce
human and environmental exposure to these pollutants and thus, reduce potential adverse health
and welfare effects. This chapter provides a general discussion of the various components of
total benefits that may be gained from a reduction in HAPs through this NESHAP. The rule will
also achieve reductions of carbon monoxide (CO), nitrous oxides (NOx), and to a lesser extent
volatile organic compounds (VOCs) and particulate matter (PM). HAP benefits are presented
separately from the benefits associated with other pollutant reductions.
7.1 IDENTIFICATION OF POTENTIAL BENEFIT CATEGORIES
The benefit categories associated with the emission reductions predicted for this
regulation can be broadly categorized as those benefits which are attributable to reduced
exposure to HAPs, and those attributable to reduced exposure to other pollutants. Some of the
HAPs associated with this regulation have been classified as probable human carcinogens. As a
result, a potential benefit of the proposed regulation is a reduction in the risk of lung and
nasopharyngeal cancer illness and possibly mortality. Other benefit categories include: reduced
7-1
-------
incidence of neurological effects and irritants associated with exposure to noncarcinogenic
HAPs, and reduced incidence of cardiovascular and central nervous system problems associated
with CO, and mortality and other morbidity effects associated with NOx (or with NOx as it
transforms into PM). In addition to health impacts occurring as a result of reductions in HAP
and other pollutant emissions, there are welfare impacts which can also be identified. In general,
welfare impacts include effects on crops and other plant life, materials damage, soiling, and
acidification of estuaries. Each category is discussed separately in the following section.
7.2 QUALITATIVE DESCRIPTION OF AIR RELATED BENEFITS
The health and welfare benefits of HAPs, CO and NOx reductions are summarized
separately in the discussions below. Appendix D also provides greater detail from the
epidemiological, animal, and occupational studies that have been conducted for the HAP
pollutants. Note that because the level of emission reductions of VOCs and PM are relatively
small, we do not provide a description of potential benefits of these pollutants in this chapter
(except to the extent that NOx can become PM once it is in the ambient air and result in adverse
effects as a PM particle).
7.2.1 Benefits of Reducing HAP Emissions
According to baseline emission estimates, this source category currently emits
approximately 27,489 tons per year of HAPs at existing sources and it is estimated that by the
year 2005, new RICE sources will emit 3,840 tons per year of HAPs. This totals 31,329 tons
annually at all RICE sources. The regulation will reduce approximately 5,000 tons of emissions
of formaldehyde, acetaldehyde, acrolein, and methanol at new and existing sources by 2005.
Human exposure to HAPs may occur directly through inhalation or indirectly through
ingestion of food or water contaminated by HAPs or through dermal exposure. HAPs may also
enter terrestrial and aquatic ecosystems through atmospheric deposition. HAPs can be deposited
on vegetation and soil through wet or dry deposition. HAPs may also enter the aquatic
environment from the atmosphere via gas exchange between surface water and the ambient air,
wet or dry deposition of particulate HAPs and particles to which HAPs adsorb, and wet or dry
deposition to watersheds with subsequent leaching or runoff to bodies of water (EPA,1992a).
7-2
-------
This analysis is focused only on the air quality benefits of HAP reduction. A summary of the
range of potential physical health and welfare effects categories that may be associated with
HAP emissions is provided in Table 7-1. As noted in the table, exposure to HAPs can lead to a
variety of acute and chronic health impacts as well as welfare impacts.
7.2.1.1 Health Benefits of Reduction in HAP Emissions.
The HAP emissions reductions achieved by this rule are expected to reduce exposure to
ambient concentrations of formaldehyde, acetaldehyde, and methanol, which will reduce a
variety of adverse health effects considering both cancer and noncancer endpoints.
Formaldehyde and acetaldehyde are classified as probable human carcinogens, according to the
Integrated Risk Information System (IRIS), an EPA system for reviewing, classifying, and listing
chemicals by cancer risk (EPA, 2000c). These HAPs are a concern to EPA because long term
exposure to these chemicals have been linked with cases of lung and nasopharyngeal cancer
deaths in humans in an occupational setting. Therefore, a reduction in human exposure to
formaldehyde and acetaldehyde could lead to a decrease in cancer risk and ultimately to a
decrease in cancer illness and mortality.
The remaining species of HAP emitted by RICE, methanol, has not been shown to cause
cancer. However, exposure to this pollutant may still result in adverse health impacts to human
and non-human populations. In general, noncancer health effects can be grouped into the
following broad categories: genotoxicity, developmental toxicity, reproductive toxicity,
systemic toxicity, and irritation. Genotoxicity is a broad term that usually refers to a chemical
that has the ability to damage DNA or the chromosomes. Developmental toxicity refers to
-------
Table 7-1. Potential Health and Welfare Effects Associated with
Exposure to Hazardous Air Pollutants
Effect Type
Effect Category
Effect End-Point
Citation
Health Mortality
Chronic Morbidity
Acute Morbidity
Carcinogenicity
Genotoxicity
Non-Cancer lethality
Neurotoxicity
Immunotoxicity
Pulmonary function decrement
Liver damage
Gastrointestinal toxicity
Kidney damage
Cardiovascular impairment
Hematopoietic (Blood disorders)
Reproductive/Developmental toxicity
Pulmonary function decrement
Dermal irritation
Eye irritation
EPA (1990), Graham et al. (1989)
Graham etal. (1989)
Voorhees et al. (1989)
All morbidity end-points obtained
from Graham et al. (1989), Voorhees
etal. (1989), Cote etal. (1988)
-------
Table 7-1. Potential Health and Welfare Effects Associated with
Exposure to Hazardous Air Pollutants (continued)
Effect Type Effect Category
Effect End-Point
Citation
Welfare Materials Damage
Aesthetic
Agriculture
Ecosystem Structure
Corrosion/Deterioration
Unpleasant odors
Transportation safety concerns
Yield reductions/Foliar injury
Biomass decrease
Species richness decline
Species diversity decline
Community size decrease
Organism lifespan decrease
Trophic web shortening
NAS (1975)
Stern et al. (1973)
Weinstein and Birk (1989)
Source: Mathtech, 1992
-------
adverse effects on a developing organism that may result from exposure prior to conception,
during prenatal development, or postnatally to the time of sexual maturation. Adverse
developmental effects may be detected at any point in the life span of the organism.
Reproductive toxicity refers to the harmful effects of HAP exposure on fertility, gestation, or
offspring, caused by exposure of either parent to a substance. Systemic toxicity affects a portion
of the body other than the site of entry. Irritation, for the purpose of this document, refers to any
effect which results in irritation of the eyes, skin, and respiratory tract (EPA, 1992a). For
methanol, IRIS does not present summary data on inhalation effects. IRIS does provide detailed
summaries of studies of the effects from oral doses of methanol, but they are not summarized for
the purposes of this RIA.
For the HAPs covered by the RICE NESHAP, evidence on the potential toxicity of the
pollutants varies. However, given sufficient exposure conditions, each of these HAPs has the
potential to elicit adverse health or environmental effects in the exposed populations. It can be
expected that emission reductions achieved through the subject NESHAP will decrease the
incidence of these adverse health effects.
7.2.1.2 Welfare Benefits of Reduction in HAP Emissions.
The welfare effects of exposure to HAPs have received less attention from analysts than
the health effects. However, this situation is changing, especially with respect to the effects of
toxic substances on ecosystems. Over the past ten years, ecotoxicologists have started to build
models of ecological systems which focus on interrelationships in function, the dynamics of
stress, and the adaptive potential for recovery. This perspective is reflected in Table 7-1 where
the end-points associated with ecosystem functions describe structural attributes rather than
species specific responses to HAP exposure. This is consistent with the observation that chronic
sub-lethal exposures may affect the normal functioning of individual species in ways that make it
less than competitive and therefore more susceptible to a variety of factors including disease,
insect attack, and decreases in habitat quality (EPA, 1991). All of these factors may contribute
to an overall change in the structure (i.e., composition) and function of the ecosystem.
The adverse, non-human biological effects of HAP emissions include ecosystem,
recreational, and commercial fishery impacts. Atmospheric deposition of HAPs directly to land
7-6
-------
may affect terrestrial ecosystems. Atmospheric deposition of HAPs also contributes to adverse
aquatic ecosystem effects. This not only has adverse implications for individual wildlife species
and ecosystems as a whole, but also the humans who may ingest contaminated fish and
waterfowl. In general, HAP emission reductions achieved through the RICE NESHAP should
reduce the associated adverse environmental impacts.
7.2.2 Benefits of Reducing Other Pollutants Due to HAP Controls
As is mentioned above, controls that will be required on RICE to reduce HAPs will also
reduce emissions of other pollutants, namely: CO, NOx, VOCs, and PM. The adverse effects
from CO, NOx, and PM emissions are presented below, but because emission reductions of
VOCs are small in magnitude, the effects from these pollutants are not discussed in this analysis.
7.2.2.1 Benefits of Reduction in Carbon Monoxide Emissions.
The EPA Staff Paper for carbon monoxide (CO) provides a summary of the health effects
information pertinent to the NAAQS for CO (EPA, 2000e). The Staff Paper concludes that
human health effects associated with exposure to CO include cardiovascular system and central
nervous system (CNS) effects. In addition, consideration is given to combined exposure to CO,
other pollutants, drugs, and the influence of environmental factors. Cardiovascular effects of CO
are directly related to reduced oxygen content of blood, resulting in tissue hypoxia (i.e., oxygen
starvation). Most healthy individuals have mechanisms (e.g., increased blood flow, blood vessel
dilation) which compensate for this reduction in tissue oxygen, although the effect of reduced
maximal exercise capacity has been reported in some healthy persons. Several other medical
conditions such as occlusive vascular disease, chronic obstructive lung disease, and anemia can
increase susceptibility to potential adverse effects of CO during exercise. Effects of CO on the
CNS involve both behavioral and physiological changes. These include modification of visual
perception, hearing, motor and sensorimotor performance, vigilance, and cognitive ability.
Although acute poisoning induced by CO can be lethal and is probably the best known
health endpoint of CO, this only occurs at very high concentrations of CO (greater than 100 ppm,
hourly average). In the ambient air, exposures to lower-levels of CO predominate and at these
levels the best documented adverse health endpoint in human subjects is the decrease in time to
7-7
-------
onset of reproducible exercise-induced chest pain. Results of some human exposure studies and
reports of workers routinely exposed to combustion products provide support for recent
epidemiology research suggesting day-to-day variations in ambient CO concentrations are
related to cardiovascular hospital admissions and daily mortality, especially for individuals over
65 years of age (EPA, 1999a). Uncertainties about the association between these health
endpoints and ambient CO and the relative influence of indoor vs. outdoor CO have not been
resolved and will require further research.
There are certain people who are more "at risk" to CO exposures. Individuals with
preexisting illness or cardiovascular diseases which limit oxygen absorption or oxygen transport
to body tissues would be somewhat more susceptible to the effects of CO. Very little data are
available demonstrating human health effects in healthy individuals caused by or associated with
exposures to low CO concentrations. Decrements in maximal exercise duration and performance
in healthy individuals have been reported, however, these decrements are small and likely to
affect only athletes in competition. No effects were seen in healthy individuals during
submaximal exercise, representing more typical daily activities. Most recent evidence of CNS
effects induced by exposure to CO indicates that behavioral impairments in healthy individuals
should not be expected until CO levels are well above what would be caused by typical ambient
air levels of CO (EPA, 1999a). Also, evidence of CO-induced fetal toxicity or of interactions
with high altitudes, drugs, other pollutants, or other environmental stresses remains uncertain or
suggests that effects of concern will occur in healthy individuals only with exposure to very high
levels of CO. The Staff Paper concludes that newer health effects evidence published since the
last NAAQS review supports the current EPA standards for CO and does not currently support a
need for more stringent standards.
7.2.2.2 Benefits of Reduced Nitrous Oxide Emissions.
Emissions of NOx produce a wide variety of health and welfare effects (EPA,1999e).
Nitrogen dioxide can irritate the lungs and lower resistance to respiratory infection (such as
influenza). NOx emissions are an important precursor to acid rain and may affect both terrestrial
and aquatic ecosystems. Atmospheric deposition of nitrogen leads to excess nutrient enrichment
problems ("eutrophication") in the Chesapeake Bay and several nationally important estuaries
7-8
-------
along the East and Gulf Coasts. Eutrophication can produce multiple adverse effects on water
quality and the aquatic environment, including increased algal blooms, excessive phytoplankton
growth, and low or no dissolved oxygen in bottom waters. Eutrophication also reduces sunlight,
causing losses in submerged aquatic vegetation critical for healthy estuarine ecosystems.
Deposition of nitrogen-containing compounds also affects terrestrial ecosystems. Nitrogen
fertilization can alter growth patterns and change the balance of species in an ecosystem.
Nitrogen dioxide and airborne nitrate also contribute to pollutant haze, which impairs
visibility and can reduce residential property values and the value placed on scenic views.
NOx in combination with volatile organic compounds (VOC) also serve as precursors to
ozone. Based on a large number of recent studies, EPA has identified several key health effects
caused when people are exposed to elevated levels of ozone. Short-term exposures (1 to 3
hours) to high ambient ozone concentrations have been linked to increased hospital admissions
and emergency room visits for respiratory problems. Repeated exposure to ozone can also make
people more susceptible to respiratory infection and lung inflammation and can aggravate
preexisting respiratory disease, such as asthma. Prolonged exposure to ozone can cause repeated
inflammation of the lung, impairment of lung defense mechanisms, and irreversible changes in
lung structure, which could lead to premature aging of the lungs and/or chronic respiratory
illnesses such as emphysema, chronic bronchitis, and chronic asthma.
Children are at most risk from ozone exposure because they typically are active outside
playing and exercising, during the summer when ozone levels are highest. Further, children are
more at risk than adults from ozone exposure because their respiratory systems are still
developing. Adults who are outdoors and moderately active during the summer months, such as
construction workers and other outdoor workers, also are among those most at risk. These
individuals, as well as people with respiratory illnesses such as asthma, especially children with
asthma, can experience reduced lung function and increased respiratory symptoms, such as chest
pain and cough, when exposed to relatively low ozone levels during periods of moderate
exertion. In addition to human health effects, ozone adversely affects crop yield, vegetation and
forest growth, and the durability of materials. Ozone causes noticeable foliar damage in many
crops, trees, and ornamental plants (i.e., grass, flowers, shrubs, and trees) and causes reduced
growth in plants.
7-9
-------
Particulate matter (PM) can also be formed from NOx emissions. Secondary PM is
formed in the atmosphere through a number of physical and chemical processes that transform
gases such as sulfur dioxide, NOx, and VOC into particles. Scientific studies have liked PM
(alone or in combination with other air pollutants) with a series of health effects (see Chapter 8
for a detailed discussion of studies used to evaluate health impacts of PM emissions). Coarse
particles can accumulate in the respiratory system and aggravate health problems such as asthma.
Fine particles penetrate deeply into the lungs and are more likely than coarse particles to
contribute to a number of the health effects. These health effects include premature death and
increased hospital admissions and emergency room visits, increased respiratory symptoms and
disease, decreased lung function, and alterations in lung tissue and structure and in respiratory
tract defense mechanisms. Children, the elderly, and people with cardiopulmonary disease, such
as asthma, are most at risk from these health effects.
PM also causes a number of adverse effects on the environment. Fine PM is the major
cause of reduced visibility in parts of the United States, including many of our national parks and
wilderness areas. Other environmental impacts occur when particles deposit onto soil, plants,
water, or materials. For example, particles containing nitrogen and sulfur that deposit onto land
or water bodies may change the nutrient balance and acidity of those environments, leading to
changes in species composition and buffering capacity.
Particles that are deposited directly onto leaves of plants can, depending on their
chemical composition, corrode leaf surfaces or interfere with plant metabolism. Finally, PM
causes soiling and erosion damage to materials.
Thus, reducing the emissions of NOx from RICE can help to improve some of the effects
mentioned above, either those directly related to NOx emissions, or the effects of ozone and PM
resulting from the combination of NOx with other pollutants.
7.3 LACK OF APPROVED METHODS TO QUANTIFY HAP BENEFITS
The primary effect associated with the HAPs that are controlled with the proposed rule is
the incidence of cancer. In previous analyses of the benefits of reductions in HAPs, EPA has
quantified and monetized the benefits of reduced incidences of cancer (EPA, 1992b, 1995). In
7-10
-------
some cases, EPA has also quantified (but not monetized) reductions in the number of people
exposed to non-cancer HAP risks above no-effect levels (EPA, 1995).
Monetization of the benefits of reductions in cancer incidences requires several important
inputs, including central estimates of cancer risks, estimates of exposure to carcinogenic HAPs,
and estimates of the value of an avoided case of cancer (fatal and non-fatal). In the above
referenced analyses, EPA relied on unit risk factors (URF) developed through risk assessment
procedures. The unit risk factor is a quantitative estimate of the carcinogenic potency of a
pollutant, often expressed as the probability of contracting cancer from a 70 year lifetime
continuous exposure to a concentration of one • g/m3 of a pollutant. These UKFs are designed to
be conservative, and as such, are more likely to represent the high end of the distribution of risk
rather than a best or most likely estimate of risk.
In a typical analysis of the expected health benefits of a regulation (see for example the
Heavy Duty Engine/Diesel Fuel Rule's Regulatory Impact Analysis; EPA, 2000d), health effects
are estimated by applying changes in pollutant concentrations to best estimates of risk obtained
from epidemiological studies. As the purpose of a benefit analysis is to describe the benefits
most likely to occur from a reduction in pollution, use of high-end, conservative risk estimates
will lead to a biased estimate of the expected benefits of the regulation. For this reason, we will
not attempt to quantify the health benefits of reductions in HAPs unless best estimates of risks
are available. While we used high-end risk estimates in past analyses, recent advice from the
EPA Science Advisory Board and internal methods reviews have suggested that we avoid using
high-end estimates in current analyses. EPA is working with the Science Advisory Board to
develop better methods for analyzing the benefits of reductions in HAPs. However the methods
to conduct a risk analysis of HAP reductions produces high-end estimates of benefits due to
assumptions required in such analyses. While we used high-end risk estimates in past analyses,
recent advice from the EPA Science Advisory Board (SAB) and internal methods reviews have
suggested that we avoid using high-end estimates in current analyses. EPA is working with the
SAB to develop better methods for analyzing the benefits of reductions in HAPs. While not
appropriate as part of a primary estimate of benefits, to estimate the potential baseline risks
posed by the RICE source category and the potential impact of applicability cutoffs discussed in
Section 3 of this RIA, EPA performed a "rough" risk assessment, described below. There are
7-11
-------
large uncertainties regarding all components of the risk quantification step, including location of
emission reductions, emission estimates, air concentrations, exposure levels and dose-response
relationships. However, if these uncertainties are properly identified and characterized, it is
possible to provide estimates of the reduction in inhalation cancer incidence associated with this
rule. It is important to keep in mind that these estimates will only cover a very limited portion of
the potential HAP effects of the rule, as they exclude non-inhalation based cancer risks and non-
cancer health effects.
7.3.1 Evaluation of Alternative Regulatory Options Based on Risk
For the RICE source category, four HAP make up the majority of the total HAP. Those
four HAP are methanol, formaldehyde, acetaldehyde, and acrolein. Three of these,
acetaldehyde, acrolein, and formaldehyde, are included in the HAP listed for the EPA's Urban
Air Toxics Program.
The HAP emitted by RICE facilities do not appear on EPA's published lists of
compounds believed to be persistent and bioaccumulative.
Two of the HAP, acetaldehyde and formaldehyde, are considered to be non-threshold
carcinogens, and cancer potency values are reported for them in IRIS. Acrolein and methanol
are not carcinogens, but are considered to be threshold pollutants, and inhalation reference
concentrations are reported for them in IRIS and by the California Environmental Protection
Agency (CalEPA), respectively.
To estimate the potential baseline risks posed by the RICE source category, EPA
performed a crude risk analysis of the RICE source category that focused only on cancer risks.
The results of the analysis are based on approaches for estimating cancer incidence that carry
significant assumptions, uncertainties, and limitations. Based on the assessment, if this proposed
rule is implemented at all affected RICE facilities, annual cancer incidence is estimated to be
reduced on the order often cases/year. Due to the uncertainties associated with the analysis,
annual cancer incidence could be higher or lower than these estimates. (Details of this
assessment are available in the docket.)
7.4 SUMMARY
7-12
-------
The HAPs that are reduced as a result of implementing the RICE NESHAP will produce
a variety of benefits, some of which include: the reduction in the incidence of cancer to exposed
populations, neurotoxicity, irritation, and crop or plant damage. The rule will also produce
benefits associated with reductions in CO and NOx emissions. Human health effects associated
with exposure to CO include cardiovascular system and CNS effects, which are directly related
to reduced oxygen content of blood and which can result in modification of visual perception,
hearing, motor and sensorimotor performance, vigilance, and cognitive ability. Human health
effects associated with NOx include respiratory problems, such as chronic bronchitis, asthma, or
even death from complications from PM concentrations created from NOx emissions. Based on
this information and the level of reductions anticipated from the RICE NESHAP, the benefits of
the rule will be substantial.
7-13
-------
8.0 QUANTIFIED BENEFITS
8.1 RESULTS IN BRIEF
In this section, we calculate monetary benefits for the reductions in ambient PM
concentrations resulting from the NOx and PM emission reductions expected from the RICE
NESHAP. Benefits related to ozone, PM10 and PM2 5 reductions are calculated using a benefit
transfer approach which uses dollar per ton values generated from air quality analyses of NOx
and PM emission reductions at RICE facilities. We have used two approaches (Base and
Alternative) to provide source benefit estimates from which the benefit transfer values are
derived. These approaches differ in their treatment of estimation and valuation of mortality risk
reductions and in the valuation of cases of chronic bronchitis. Total benefits (in 1998$) from
RICE NOx and PM emission reductions at major sources are presented in Table 8-1.
This benefit analysis does not quantify all potential benefits or disbenefits associated with
NOx and PM reductions. This analysis also does not quantify the benefits associated with
reductions in hazardous air pollutants. The magnitude of the unquantified benefits associated
with omitted categories and pollutants, such as avoided cancer cases, damage to ecosystems, or
materials damage to industrial equipment and national monuments, is not known. However, to
the extent that unquantified benefits exceed unquantified disbenefits, the estimated benefits
presented above will be an underestimate of actual benefits. There are many other sources of
uncertainty in the estimates of quantified benefits. These sources of uncertainty, along with the
methods for estimating monetized benefits for the RICE NESHAP and a more detailed analysis
of the results are presented below.
3-1
-------
Table 8-1. Summary of Results:
The Estimated PM and Ozone-Related Benefits of the RICE NESHAP
Total Benefits3 b
Estimation Method (millions 1998$)
Base Estimate:
Using a 3% discount rate $280 + B
Using a 7% discount rate $265 + B
Alternative Estimate:
Using a 3% discount rate $40 + B
Using a 7% discount rate $45 + B
a Benefits of HAP and CO emission reductions are not quantified in this analysis and, therefore, are not presented in this table.
The quantifiable benefits are from emission reductions of NOx and PM only. For notational purposes, unquantified benefits
are indicated with a "B" to represent additional monetary benefits. A detailed listing of unquantified NOx, PM, and HAP
related health effects is provided in Table 8-13.
b Results reflect the use of two different discount rates; a 3% rate which is recommended by EPA's Guidelines for Preparing
Economic Analyses (EPA, 2000b), and 7% which is recommended by OMB Circular A-94 (OMB, 1992).
8.2 INTRODUCTION
This chapter presents the methods used to estimate the monetary benefits of the
reductions in NOx and PM emissions associated with RICE NESHAP controls. Results are
presented for the emission controls described in Chapter 2. The benefits that result from the rule
include both the Base impacts from application of control technologies or changes in operations
and processes, and the secondary effects of the controls. The regulation induced reductions in
PM and NOx emissions will result in changes in the physical damages associated with exposure
to elevated ambient concentrations of the criteria pollutants, PM and ozone. These damages
include changes in both human health and welfare effects categories.
The remainder of this chapter provides the following:
Subsection 3 provides an overview of the benefits methodology.
•• Subsection 4 discusses methods for estimating the NOx and direct PM transfer
values used as inputs to the benefits analysis.
8-2
-------
•• Subsection 5 provides estimates of health and welfare benefits associated with
NESHAP controls based on the benefit transfer values and emission reductions.
•• Subsection 6 discusses potential benefit categories that are not quantified due to
data and/or methodological limitations, and provides a list of analytical
uncertainties, limitations, and biases.
•• Subsection 7 presents the net benefits (benefits minus costs) of the RICE
NESHAP.
8.3 OVERVIEW OF BENEFITS ANALYSIS METHODOLOGY
This section documents the general approach used to estimate benefits resulting from
emissions reductions from RICE sources. We follow the basic methodology described in the
Regulatory Impact Analysis of the Heavy Duty Engine/Diesel Fuel rule [hereafter referred to as
theHDDRIA] (EPA, 2000d).
On September 26, 2002, the National Academy of Sciences (NAS) released a report on
its review of the Agency's methodology for analyzing the health benefits of measures taken to
reduce air pollution. The report focused on EPA's approach for estimating the health benefits of
regulations designed to reduce concentrations of airborne particulate matter (PM).
In its report, the NAS said that EPA has generally used a reasonable framework for
analyzing the health benefits of PM-control measures. It recommended, however, that the
Agency take a number of steps to improve its benefits analysis. In particular, the NAS stated
that the Agency should:
include benefits estimates for a range of regulatory options;
estimate benefits for intervals, such as every five years, rather than a single year;
•• clearly state the project baseline statistics used in estimating health benefits,
including those for air emissions, air quality, and health outcomes;
•• examine whether implementation of proposed regulations might cause unintended
impacts on human health or the environment;
when appropriate, use data from non-U.S. studies to broaden age ranges to which
current estimates apply and to include more types of relevant health outcomes;
-------
•• begin to move the assessment of uncertainties from its ancillary analyses into its
Base analyses by conducting probabilistic, multiple-source uncertainty analyses.
This assessment should be based on available data and expert judgment.
Although the NAS made a number of recommendations for improvement in EPA's
approach, it found that the studies selected by EPA for use in its benefits analysis were generally
reasonable choices. In particular, the NAS agreed with EPA's decision to use cohort studies to
derive benefits estimates. It also concluded that the Agency's selection of the American Cancer
Society (ACS) study for the evaluation of PM-related premature mortality was reasonable,
although it noted the publication of new cohort studies that should be evaluated by the Agency.
Several of the NAS recommendations addressed the issue of uncertainty and how the
Agency can better analyze and communicate the uncertainties associated with its benefits
assessments. In particular, the Committee expressed concern about the Agency's reliance on a
single value from its analysis and suggested that EPA develop a probabilistic approach for
analyzing the health benefits of proposed regulatory actions. The Agency agrees with this
suggestion and is working to develop such an approach for use in future rulemakings.
In this RIA, the Agency has used an interim approach that shows the impact of several
important alternative assumptions about the estimation and valuation of reductions in premature
mortality and chronic bronchitis. This approach, which was developed in the context of the
Agency's Clear Skies analysis, provides an alternative estimate of health benefits using the time
series studies in place of cohort studies, as well as alternative valuation methods for mortality
and chronic bronchitis risk reductions.
The analysis that follows evaluates the benefits of the RICE NESHAP across four
subcategories of control. Only one subcategory will have controls on existing RICE units. For
new sources, estimated emission reductions will occur in all subcategories at sources that
become operational by 2005. Based on a memo discussing the distribution of major and area
sources of RICE units (Alpha-Gamma, 200 la), we anticipate that at least 60 percent of the
stationary RICE in operation in 2005 will be located at area sources which are not affected by
this regulatory action. Therefore, this analysis presents the benefits of emission reductions
8-4
-------
occurring at major sources only (i.e., for the 40 percent of the total estimate of emissions at all
RICE units in 2005).
The location of new sources is not known. Based on 1996 emissions inventory data, we
find NOx emissions from RICE sources to occur throughout the U.S. As such, we also expect
the operation of new RICE units in 2005 to be spread across the country. Due to the limitations
in availability of data on location of emission reductions from specific RICE sources, this
benefits analysis is based on benefit transfer, rather than on modeling of changes in air quality
and health effects from the location specific emissions reductions achieved under the RICE
NESHAP. Although the NESHAP regulation is expected to result in reductions in emissions of
many hazardous air pollutants as well as NOx and PM, benefit transfer values are generated for
only NOx and PM due to limitations in availability of transfer values, concentration-response
functions, or air quality and exposure models. For this analysis, we focus on directly emitted
PM, and NOx in its role as a precursor in the formation of ambient ozone and particulate matter.
Other potential impacts of PM and NOx reductions not quantified in this analysis, as well as
potential impacts of HAP reductions are described in Chapter 7.
The general term "benefits" refers to any and all outcomes of the regulation that
contribute to an enhanced level of social welfare. In this case, the term "benefits" refers to the
dollar value associated with all the expected positive impacts of the regulation, that is, all
regulatory outcomes that lead to higher social welfare. If the benefits are associated with market
goods and services, the monetary value of the benefits is approximated by the sum of the
predicted changes in consumer (and producer) "surplus." These "surplus" measures are standard
and widely accepted measures in the field of applied welfare economics, and reflect the degree
of well-being enjoyed by people given different levels of goods and prices. If the benefits are
non-market benefits (such as the risk reductions associated with environmental quality
improvements), however, other methods of measuring benefits must be used. In contrast to
market goods, non-market goods such as environmental quality improvements are public goods,
whose benefits are shared by many people. The total value of such a good is the sum of the
dollar amounts that all those who benefit are willing to pay.
Given the current limitations on availability of data on facility-specific emission
reductions, we have selected benefit transfer as the most appropriate methodology for this
8-5
-------
benefits analysis. Benefit transfer is the process of applying quantified benefits derived for a
study scenario to a policy scenario for which quantified benefits are desired. This is particularly
useful when time or data constraints do not allow for direct and complete quantification of
benefits. The benefit transfer value is typically expressed as dollar or health effect benefits per
ton of emissions reduced. The PM value per ton can be determined by examining the direct
health impacts of changes in ambient PM. To estimate the value per ton of NOx reduced, we
need estimates of the value per ton of NOx as an ozone precursor and as a PM precursor. We
apply two different approaches to benefit transfer for PM and ozone related benefits, due to
differences in availability of data and models. Our approach to benefit transfer for PM related
benefits is to generate a benefits analysis for an emissions control scenario similar to the RICE
NESHAP scenario, calculate a dollar per ton estimate based on this analysis, and apply that
estimate to the emissions reductions expected to result from the NESHAP controls. Our
approach for ozone-related benefits is to use a dollar per ton estimate generated from a previous
ozone related benefits analyses of NOx reductions from utility and industrial combustion
sources. The difference in approach for ozone and PM benefits is due to the fact that a suitable
PM air quality model is available, while a suitable ozone model is not.
Development of a benefit transfer value for each criteria pollutant requires selection of an
existing set of air quality modeling results that, to the extent possible, parallels the air quality
modeling that would be conducted for the current policy if the data and resources were available.
This requires review of the magnitude, type, and geographic distribution of emissions reductions
used in the air quality analyses, the regions of analysis, and the ambient pollutants modeled in
the analyses. Once an existing set of air quality modeling results has been selected, two pieces
of information need to be extracted from the results: (1) changes in ambient concentrations of
the pollutant, i.e., ozone and (2) reductions in precursor emissions of the pollutant of interest,
i.e., NOx. These data, along with the set of concentration-response functions and valuation
functions, constitute the input set for the benefit transfer value function. The benefit transfer
function for pollutant i is specified as:
8-6
-------
The numerator in the transfer value formula is total monetary benefits, which is
determined by applying economic valuation functions to changes in incidences of health and
welfare endpoints and summing over all endpoints. Changes in incidences of health and welfare
endpoints are calculated by applying epidemiological concentration-response functions to the
changes in ambient concentrations of the pollutant.
Using the estimated benefit transfer values, national benefits for PM and NOx reductions
can be obtained using the following formula:
(8.2) TotalBenefits=TVozone •ANoxozone +TVPM25 • ANOxPM +TVdirectPM • AdirectPM
where TVozone is the transfer value for ozone, TVPM is the transfer value for PM, TVdirectPM is the
transfer value for directly emitted PM, • NOxozone is the change in NOx ozone precursor
emissions, • NOxPM is the change in NOx PM precursor emissions, and • directPM is the change
in direct PM emissions. The relevant NOx emission changes for ozone formation are those
occurring during the summer
(TransferValue)t =——:—:—-—-L- — ozone season, while those for
(Emission Re auctions)t
PM formation are year round.
8.3.1 Methods for Estimating Benefits from Air Quality Improvements
Environmental and health economists have a number of methods for estimating the
economic value of improvements in (or deterioration of) environmental quality. The method
used in any given situation depends on the nature of the effect and the kinds of data, time, and
3-7
-------
resources that are available for investigation and analysis. This section provides an overview of
the methods we selected to monetize the benefits included in this RIA.
We note at the outset that EPA rarely has the time or resources to perform extensive new
research in the form of evaluating the response in human health effects from specific changes in
the concentration of pollutants, or by issuing surveys to collect data of individual's willingness
to pay for a particular rule's given change in air quality, which is needed to fully measure the
economic benefits of individual rulemakings. As a result, our estimates are based on the best
available methods of benefit transfer from epidemiological studies and studies of the economic
value of reducing certain health and welfare effects. Benefit transfer is the science and art of
adapting Base benefits research on concentration-response functions and measures of the value
individuals place on an improvement in a given health effect to the scenarios evaluated for a
particular regulation. Thus, we strive to obtain the most accurate measure of benefits for the
environmental quality change under analysis given availability of current, peer reviewed
research and literature. Where appropriate, adjustments are made for the sociodemographic and
economic characteristics of the affected population, and other factors in order to improve the
accuracy and robustness of benefits estimates.
In general, economists tend to view an individual's willingness-to-pay (WTP) for an
improvement in environmental quality as the most complete and appropriate measure of the
value of an environmental or health risk reduction. An individual's willingness-to-accept
(WTA) compensation for not receiving the improvement is also a valid measure. Willingness to
pay and Willingness to accept are comparable measures when the change in environmental
quality is small and there are reasonably close substitutes available. However, WTP is generally
considered to be a more readily available and conservative (i.e. more likely to underestimate than
overestimate) measure of benefits. Adoption of WTP as the measure of value implies that the
value of environmental quality improvements is dependent on the individual preferences of the
affected population and that the existing distribution of income (ability to pay) is appropriate.
For many goods, WTP can be observed by examining actual market transactions. For
example, if a gallon of bottled drinking water sells for one dollar, it can be observed that at least
some persons are willing to pay one dollar for such water. For goods not exchanged in the
market, such as most environmental "goods," valuation is not as straightforward. Nevertheless,
-------
a value may be inferred from observed behavior, such as sales and prices of products that result
in similar effects or risk reductions, (e.g., non-toxic cleaners or bike helmets). Alternatively,
surveys may be used in an attempt to directly elicit WTP for an environmental improvement.
One distinction in environmental benefits estimation is between "use values" and "non-
use values." Although no general agreement exists among economists on a precise distinction
between the two, the general nature of the difference is clear. Use values are those aspects of
environmental quality that affect an individual's welfare more or less directly. These effects
include changes in product prices, quality, and availability, changes in the quality of outdoor
recreation and outdoor aesthetics, changes in health or life expectancy, and the costs of actions
taken to avoid negative effects of environmental quality changes.
Non-use values are those for which an individual is willing to pay for reasons that do not
relate to the direct use or enjoyment of any environmental benefit, but might relate to existence
values and bequest values. Non-use values are not traded, directly or indirectly, in markets. For
this reason, the measurement of non-use values has proved to be significantly more difficult than
the measurement of use values. The air quality changes produced by this NESHAP cause
changes in both use and non-use values, but the monetary benefit estimates are almost
exclusively for use values.
More frequently than not, the economic benefits from environmental quality changes are
not traded in markets, so direct measurement techniques can not be used. Avoided cost methods
are ways to estimate the costs of pollution by using the expenditures made necessary by
pollution damage. For example, if buildings must be cleaned or painted more frequently as
levels of PM increase, then the appropriately calculated increment of these costs is a reasonable
lower bound estimate (under most conditions) of true economic benefits when PM levels are
reduced. Avoided costs methods are used to estimate some of the health-related benefits related
to morbidity, such as hospital admissions (see the FtDD RIA for a detailed discussion of methods
to value benefit categories).
Indirect market methods can also be used to infer the benefits of pollution reduction. The
most important application of this technique for our analysis is the calculation of the value of a
statistical life for use in the estimate of benefits from mortality reductions. There exists no
market where changes in the probability of death are directly exchanged. However, people make
8-9
-------
decisions about occupation, precautionary behavior, and other activities associated with changes
in the risk of death. By examining these risk changes and the other characteristics of people's
choices, it is possible to infer information about the monetary values associated with changes in
mortality risk (see Section 8.4). For measurement of health benefits, this analysis captures the
WTP for most use and non-use values, with the exception of the value of avoided hospital
admissions, which only captures the avoided cost of illness because no WTP values were
available in the published literature.
8.3.2 Quantifying Individual Health Effect Endpoints
We use the term "endpoints" to refer to specific effects that can be associated with
changes in air quality. To estimate these endpoints, EPA combines changes in ambient air
quality levels with epidemiological evidence about population health response to pollution
exposure. The most significant monetized benefits of reducing ambient concentrations of ozone
and PM are attributable to reductions in human health risks. EPA's Criteria Documents for
ozone and PM list numerous health effects known to be linked to ambient concentrations of the
pollutants (EPA, 1996a; EPA, 1996b). Chapter 7 described some of these effects. This section
describes methods used to quantify and monetize changes in the expected number of incidences
of various health effects.
The specific ozone and PM endpoints that are evaluated in this analysis include:
Premature mortality
•• Bronchitis - chronic and acute
Hospital admissions - respiratory and cardiovascular
•• Emergency room visits for asthma
•• Asthma attacks
Acute respiratory symptoms
•• Lower and upper respiratory illness
•• Decreased worker productivity
Minor restricted activity days
•• Work loss days
8-10
-------
As is discussed previously, this analysis relies on concentration-response (C-R) functions
estimated in published epidemiological studies relating health effects to ambient air quality. The
specific studies from which C-R functions are drawn are included in Table 8-2. Because we rely
on methodologies used in prior benefit analyses, a complete discussion of the C-R functions used
for this analysis and information about each endpoint are contained in the HDD RIA and in the
benefits Technical Support Document for the RIA of the Heavy Duty Engine/Diesel Fuel rule
[hereafter referred to as the HDD TSD] (Abt Associates, 2000).
While a broad range of serious health effects have been associated with exposure to
elevated ozone and PM levels (described more fully in the EPA's ozone and PM Criteria
Documents), we include only a subset of health effects in this quantified benefit analysis. Health
effects are excluded from this analysis for four reasons: (i) the possibility of double counting
(such as hospital admissions for specific respiratory diseases); (ii) uncertainties in applying
effect relationships based on clinical studies to the affected population; (iii) a lack of an
established C-R relationship; or (iv) lack of resources to estimate some endpoints.
Using the C-R functions derived from the studies cited in this table, we apply that same
C-R relationship to all locations in the United States. Although the C-R relationship may in fact
vary somewhat from one location to another (for example, due to differences in population
susceptibilities or differences in the composition of PM), location-specific C-R functions are
generally not available. A single function applied everywhere may result in overestimates of
incidence changes in some locations and underestimates in other locations, but these location-
specific biases will, to some extent, cancel each other out when the total incidence change is
calculated. It is not possible to know the extent or direction of the bias in the total incidence
change based on the general application of a single C-R function everywhere.
8-11
-------
Table 8-2. Health Outcomes and Studies Included in the Analysis
Health Outcome
Premature Mortality
All-cause premature
mortality from long-term
exposure (Base Estimate)
Short-term exposure
(Alternative Estimate)
Short-term exposure
(Alternative Estimate)
Chronic Illness
Chronic Bronchitis
(pooled estimate)
Hospital Admissions
All Respiratory
COPD
Pneumonia
Asthma
Total Cardiovascular
Asthma-Related ER Visits
Other Effects
Any of 19 Acute
Symptoms
Asthma Attacks
Acute Bronchitis
Upper Respiratory
Symptoms
Lower Respiratory
Symptoms
Decreased Worker
Productivity
Work Loss Days
Applied
Pollutant Population
PM25
PM25
PM10
PM25
PM10
Ozone
PM10
PM10
PM25
PM10
PM10
Ozone
PM10
PM25
PM10
PM25
Ozone
PM25
> 29 years
< 65 years, • 65
years
All ages
All ages
> 26 years
> 29 years
Pooled estimate
(8 studies)
> 64 years
> 64 years
< 65 years
> 64 years
All ages
All ages
Asthmatics, all
ages
Children, 8-12
years
Asthmatic
children,
9-11
Children, 7-14
years
Adults, 18-65
years
Source of Baseline
Source of Effect Estimate Incidence
Krewski et al., 2000
Schwartz etal. (1996)
Schwartz et al. (2000)
Samet et al. (2000)
Schwartz et al. (2000)
Abbey etal., 1995
Schwartz et al., 1993
All ages
Samet etal., 2000
Samet etal., 2000
Sheppard et al., 1999
Samet etal., 2000
Schwartz et al., 1993
Thurston et al., 1992
Whittemore and Korn,
1980
Dockery et al., 1996
Pope etal., 1991
Schwartz et al., 1994
Crocker and Horst, 1981;
EPA, 1994
Ostro, 1987
U.S. Centers for Disease
Control, 1999
U.S. Centers for Disease
Control, 1999
U.S. Centers for Disease
Control, 1999
Abbey etal., 1993
Abbey etal., 1993
Adams and Marano,
Graves and Gillum,
Graves and Gillum,
Graves and Gillum,
Graves and Gillum,
Smith etal., 1997
Graves and Gillum,
Krupnick, 1988
Adams and Marano,
Adams and Marano,
Pope etal., 1991
1995
1997
1997
1997
1997
1997
1995
1995
Schwartz etal., 1994
Adams and Marano,
1995
8-12
-------
Applied Source of Baseline
Health Outcome Pollutant Population Source of Effect Estimate Incidence
Minor Restricted Activity PM25 Adults, 18-65 Ostro and Rothschild., Ostro and Rothschild,
Days (minus asthma years 1989 1989
attacks)
Recently, the Health Effects Institute (HEI) reported findings by investigators at Johns
Hopkins University and others that have raised concerns about aspects of the statistical
methodology used in a number of recent time-series studies of short-term exposures to air
pollution and health effects (Greenbaum, 2002a). Some of the concentration-response functions
used in this benefits analysis were derived from such short-term studies. The estimates derived
from the long-term mortality studies, which account for a major share of the benefits in the Base
Estimate, are not affected. As discussed in HEI materials provided to sponsors and to the Clean
Air Scientific Advisory Committee (Greenbaum, 2002a, 2002b), these investigators found
problems in the default "convergence criteria" used in Generalized Additive Models (GAM) and
a separate issue first identified by Canadian investigators about the potential to underestimate
standard errors in the same statistical package.1 These and other investigators have begun to
reanalyze the results of several important time series studies with alternative approaches that
address these issues and have found a downward revision of some results. For example, the
mortality risk estimates for short-term exposure to PM10 from NMMAPS were overestimated
(the C-R function based on the NMMAPS results used in this benefits analysis uses the revised
NMMAPS results).2 However, both the relative magnitude and the direction of bias introduced
by the convergence issue is case-specific. In most cases, the concentration-response relationship
may be overestimated; in other cases, it may be underestimated. The preliminary renalyses of
the mortality and morbidity components of NMMAPS suggest that analyses reporting the lowest
relative risks appear to be affected more greatly by this error than studies reporting higher
relative risks (Dominici et al., 2002; Schwartz and Zanobetti, 2002).
'Most of the studies used a statistical package known as "S-plus." For further details, see
http://www.healtheffects.org/Pubs/NMMAPSletter.pdf.
2HEI sponsored the multi-city the National Morbidity, Mortality, and Air Pollution Study (NMMAPS). See
http://biosunO 1.biostat.jhsph.edu/~fdominic/NMMAPS/nmmaps-revised.pdf for revised mortality results.
8-13
-------
Our examination of the original studies used in this analysis finds that the health
endpoints that are potentially affected by the GAM issues include: reduced hospital admissions
and reduced lower respiratory symptoms in the both the Base and Alternative Estimates; reduced
lower respiratory symptoms in both the Base and Alternative Estimates; and reduced premature
mortality due to short-term PM10 exposures in the Base Estimate3 and reduced premature
mortality due to short-term PM2 5 exposures in the Alternative Estimate. While resolution of
these issues is likely to take some time, the preliminary results from ongoing reanalyses of some
of the studies used in our analyses (Dominici et al., 2002; Schwartz and Zanobetti, 2002;
Schwartz, personal communication, 2002) suggest a more modest effect of the S-plus error than
reported for the NMMAPS PM10 mortality study. While we wait for further clarification from
the scientific community, we have chosen not to remove these results from the RICE NESHAP
benefits estimates, nor have we elected to apply any interim adjustment factor based on the
preliminary reanalyses. EPA will continue to monitor the progress of this concern, and make
appropriate adjustments as further information is made available.
8.3.2.1 Concentration-Response Functions for Premature Mortality
Both long and short-term exposures to ambient levels of air pollution have been
associated with increased risk of premature mortality. The size of the mortality risk estimates
from these epidemiological studies, the serious nature of the effect itself, and the high monetary
value ascribed to prolonging life make mortality risk reduction the most important health
endpoint quantified in this analysis. Because of the importance of this endpoint and the
considerable uncertainty among economists and policymakers as to the appropriate way to value
reductions in mortality risks, this section discusses some of the issues surrounding the estimation
of premature mortality. For additional discussion on mortality and issues related to estimating
risk for other health effects categories, we refer readers to the discussions presented in EPA's
Heavy-Duty Engine/Diesel Fuel RIA (EPA, 2000d).
3Note that in the Base Estimate, reduced premature mortality from long-term PM2 5 accounts for a large majority of
total monetized benefits. Therefore, although benefits from PM10-related short-term mortality are affected by the
GAM issue, total benefits in the Base Estimate are not greatly altered by the affect of this issue on PM10.
8-14
-------
Health researchers have consistently linked air pollution, especially PM, with excess
mortality. Although a number of uncertainties remain to be addressed by continued research
(NRC, 1998), a substantial body of published scientific literature recognizes a correlation
between elevated PM concentrations and increased mortality rates. Two types of community
epidemiological studies (involving measures of short-term and long-term exposures and
response) have been used to estimate PM/mortality relationships. Short-term studies relate
short-term (often day-to-day) changes in PM concentrations and changes in daily mortality rates
up to several days after a period of elevated PM concentrations. Long-term studies examine the
potential relationship between longer-term (e.g., one or more years) exposure to PM and annual
mortality rates. Researchers have found significant associations using both types of studies.
Base Estimate
Over a dozen studies have found significant associations between measures of long-term
exposure to PM and elevated rates of annual mortality (e.g., Lave and Seskin, 1977; Ozkaynak
and Thurston, 1987). While most of the published studies found positive (but not always
significant) associations with available PM indices such as total suspended particles (TSP), fine
particles components (i.e., sulfates), and fine particles, exploration of alternative model
specifications sometimes found inconsistencies (e.g., Lipfert, 1989). These early "cross-
sectional" studies were criticized for a number of methodological limitations, particularly for
inadequate control at the individual level for variables that are potentially important in causing
mortality, such as wealth, smoking, and diet. More recently, several new long-term studies have
been published that use improved approaches and appear to be consistent with the earlier body of
literature. These new "prospective cohort" studies reflect a significant improvement over the
earlier work because they include information on individuals with respect to measures related to
health status and residence. The most extensive study and analyses has been based on data from
two prospective cohort groups, often referred to as the Harvard "Six-City study" (Dockery et al.,
1993) and the "American Cancer Society or ACS study" (Pope et al., 1995); these studies have
found consistent relationships between fine particle indicators and mortality across multiple
locations in the United States. A third major data set comes from the California based 7th day
Adventist study (e.g., Abbey et al., 1999), which reported associations between long-term PM
8-15
-------
exposure and mortality in men. Results from this cohort, however, have been inconsistent and
the air quality results are not geographically representative of most of the U.S. More recently, a
cohort of adult male veterans diagnosed with hypertension has been examined (Lipfert et al.,
2000). Unlike previous long-term analyses, this study found some associations between
mortality and ozone but found inconsistent results for PM indicators.
Given their consistent results and broad applicability to general U.S. populations, the Six-
City and ACS data have been of particular importance in benefits analyses. The credibility of
these two studies is further enhanced by the fact that they were subject to extensive
reexamination and reanalysis by an independent scientific analysis team (Krewski et al., 2000).
The final results of the reanalysis were then independently peer reviewed by a Special Panel of
the HEI Health Review Committee. The results of these analyses confirmed and expanded those
of the original investigators. This intensive independent reanalysis effort was occasioned both
by the importance of the original findings as well as concerns that the underlying individual
health effects information has never been made publicly available. The HEI re-examination
lends credibility to the original studies but also found unexpected sensitivities concerning (a)
which pollutants are most important, (b) the role of education in mediating the association
between pollution and mortality, and (c) the magnitude of the association depending on how
spatial correlation was handled. Further confirmation and extension of the overall findings using
more recent air quality and ACS health information was recently published in the Journal of the
American Medical Association (Pope et al., 2002). In general, the risk estimates based on the
long-term mortality studies are substantially greater than those derived from short-term studies.
In developing and improving the methods for estimating and valuing the potential
reductions in mortality risk over the years, EPA has consulted with a panel of the Science
Advisory Board. That panel recommended use of long-term prospective cohort studies in
estimating mortality risk reduction (EPA-SAB-COUNCIL-ADV-99-005, 1999c). More
specifically, the SAB recommended emphasis on Pope et al. (1995) because it includes a much
larger sample size and longer exposure interval, and covers more locations (50 cities as
compared to 6 cities in the Harvard data) than other studies of its kind. As explained in the
regulatory impact analysis for the Heavy-Duty Engine/Diesel Fuel rule (EPA, 2000d), more
recent EPA benefits analyses have relied on an improved specification from this data set that was
8-16
-------
developed in the HEI reanalysis of this study (Krewski et al., 2000). The particular specification
estimated a C-R function based on changes in mean levels of PM25, as opposed to the function in
the original study, which used median levels. This specification also includes a broader
geographic scope than the original study (63 cities versus 50). The SAB has recently agreed
with EPA's selection of this specification for use in analyzing mortality benefits of PM
reductions (EPA-SAB-COUNCIL-ADV-01-004, 2001). For these reasons, the present analysis
uses the same C-R function in developing the Base Estimate of mortality benefits related to fine
particles.
Our Base estimate also accounts for a lag between reductions in PM 2.5 concentrations
and reductions in mortality incidence. It is currently unknown whether there is a time lag (a
delay between changes in PM exposures and changes in mortality rates) in the long-term
PM2.5/premature mortality relationship. The existence of such a lag is important for the
valuation of premature mortality incidences because economic theory suggests that benefits
occurring in the future should be discounted. Although there is no specific scientific evidence of
the existence or structure of a PM effects lag, current scientific literature on adverse health
effects, such as those associated with PM (e.g., smoking-related disease) and the difference in
the effect size between chronic exposure studies and daily mortality studies suggest that all
incidences of premature mortality reduction associated with a given incremental change in PM
exposure probably would not occur in the same year as the exposure reduction. This same
smoking-related literature implies that lags of up to a few years are plausible. Adopting the lag
structure used in the Tier 2/Gasoline Sulfur RIA, the HDD RIA, and endorsed by the SAB
(EPA-SAB-COUNCIL-ADV-00-001, 1999), we assume a five-year lag structure, with 25
percent of premature deaths occurring in the first year (in 2005), another 25 percent in the
second year, and 16.7 percent in each of the remaining three years. The mortality incidences
across the 5-year period is then discounted back to our year of analysis, 2005.
For reductions in direct emissions of PM10, we use a different C-R function, based on the
studies of mortality and shorter term exposures to PM. Long-term studies of the relationship
between chronic exposure and mortality have not found significant associations with coarse
particles or total PM10. As discussed earlier in this chapter, concerns have recently been raised
about aspects of the statistical methodology used in a number of recent time-series studies of
8-17
-------
short-term exposures to air pollution and health effects. Due to the "S-Plus" issue identified by
the Health Effects Institute, we use as the basis for the Base estimate the revised relative risk
from the NMMAPS study, reported on the website of the Johns Hopkins School of Public Health
(2002)4. Similar to the PM2 5 lag adjustment discussed above, we also include an adjustment for
PM10 to account for recent evidence that daily mortality is associated with particle levels from a
number of previous days. We use the overall pooled NMMAPS estimate of a 0.224 percent
increase in mortality for a 10 • g/m3 increase in PM10 as the starting point in developing our C-R
function. In a recent analysis, Schwartz (2000) found that elevated levels of PM10 on a given day
can elevate mortality on a number of following days. This type of multi-day model is often
referred to as a "distributed lag" model because it assumes that mortality following a PM event
will be distributed over a number of days following or "lagging" the PM event5. Because the
NMMAPS study reflects much broader geographic coverage (90 cities) than the Schwartz study
(10 cities), and the Schwartz study has not been reanalyzed to account for the "S-Plus" issue, we
choose to apply an adjustment based on the Schwartz study to the NMMAPS study to reflect the
effect of a distributed lag model.
The distributed lag adjustment factor is constructed as the ratio of the estimated
coefficient from the unconstrained distributed lag model to the estimated coefficient from the
single-lag model reported in Schwartz (2000). The unconstrained distributed lag model
coefficient estimate is 0.0012818 and the single-lag model coefficient estimate is 0.0006479.
The ratio of these estimates is 1.9784. This adjustment factor is then multiplied by the revised
estimated coefficients from the NMMAPS study. The NMMAPS coefficient corresponding to
the 0.224 percent increase in mortality risk is 0.000224. The adjusted NMMAPS coefficent is
then 0.000224*1.9784 = 0.000444.
Alternative Estimate
4 Available at http://www.biostat.jhsph.edi^iostat/research/update.main.htm.
5 Both the single day and distributed lag models are likely to be affected to the same degree by the S-Plus
convergence issue. As such, the ratio of the coefficients from the models should not be affected as much by any
changes in the coefficient.
8-18
-------
To reflect concerns about the inherent limitations in the number of studies supporting a
causal association between long-term exposure and mortality, an Alternative benefit estimate
was derived from the large number of time-series studies that have established a likely causal
relationship between short-term measures of PM and daily mortality statistics. A particular
strength of such studies is the fact that potential confounding variables such as socio-economic
status, occupation, and smoking do not vary on a day-to-day basis in an individual area. A
number of multi-city and other types of studies strongly suggest that these effects-relationships
cannot be explained by weather, statistical approaches, or other pollutants. The risk estimates
from the vast majority of the short-term studies include the effects of only one or two-day
exposure to air pollution. More recently, several studies have found that the practice of
examining the effects on a single day basis may significantly understate the risk of short-term
exposures (Schwartz, 2000; Zanobetti et al., 2002). These studies suggest that the short-term
risk can double when the single-day effects are combined with the cumulative impact of
exposures over multiple days to weeks prior to a mortality event.
The fact that the PM-mortality coefficients from the cohort studies are far larger than the
coefficients derived from the daily time-series studies provides some evidence for an
independent chronic effect of PM pollution on health. Indeed, the Base Estimate presumes that
the larger coefficients represent a more complete accounting of mortality effects, including both
the cumulative total of short-term mortality as well as an additional chronic effect. This is,
however, not the only possible interpretation of the disparity. Various reviewers have argued
that (1) the long-term estimates may be biased high and/or (2) the short-term estimates may be
biased low. In this view, the two study types could be measuring the same underlying
relationship.
Reviewers have noted some possible sources of upward bias in the long-term studies.
Some have noted that the less robust estimates based on the Six-Cities Study are significantly
higher than those based on the more broadly distributed ACS data sets. Some reviewers have
also noted that the observed mortality associations from the 1980s and 90s may reflect higher
pollution exposures from the 1950s to 1960s. While this would bias estimates based on more
recent pollution levels upwards, it also would imply a truly long-term chronic effect of
pollution. With regard to possible sources of downward bias, it is of note that the recent
8-19
-------
studies suggest that the single day time series studies may understate the short-term effect on the
order of a factor of two. These considerations provide a basis for considering an Alternative
Estimate using the most recent estimates from the wealth of time-series studies, in addition to
one based on the long-term cohort studies.
In essence, the Alternative Estimate addresses the above noted uncertainties about the
relationship between premature mortality and long-term exposures to ambient levels of fine
particles by assuming that there is no mortality effect of chronic exposures to fine particles.
Instead, it assumes that the full impact of fine particles on premature mortality can be captured
using a concentration-response function relating daily mortality to short-term fine particle levels.
Specifically, a concentration-response function based on Schwartz et al. (1996) is employed,
with an adjustment to account for recent evidence that daily mortality is associated with particle
levels from a number of previous days (Schwartz, 2000), similar to the adjustment for the PM10
mortality C-R function described for the Base Estimate.
There are no PM25 daily mortality studies which report numeric estimates of relative
risks from distributed lag models; only PM10 studies are available. Daily mortality C-R functions
for PM10 are consistently lower in magnitude than PM2 5-mortality C-R functions, because fine
particles are believed to be more closely associated with mortality than the coarse fraction of
PM. Given that the NOx emissions reductions under the RICE NESHAP result primarily in
reduced ambient concentrations of PM25, use of a PM10 based C-R function results in a
significant downward bias in the estimated reductions in mortality. To account for the full
potential multi-day mortality impact of acute PM25 events, we use the same adjustment factor
(1.9784) used in developing the PM10 mortality C-R function, applied to the PM25 based C-R
function reported in Schwartz et al. (1996).
If most of the increase in mortality is expected to be associated with the fine fraction of
PM10, then it is reasonable to assume that the same proportional increase in risk would be
observed if a distributed lag model were applied to the PM25 data. There are two relevant
coefficients from the Schwartz et al. (1996) study, one corresponding to all-cause mortality, and
one corresponding to chronic obstructive pulmonary disease (COPD) mortality (separation by
cause is necessary to implement the life years lost approach detailed below). The adjusted
estimates for these two C-R functions are:
8-20
-------
All cause mortality = 0.001489 * 1.9784 = 0.002946
COPD mortality = 0.003246 * 1.9784 = 0.006422
Note that these estimates, while approximating the full impact of daily pollution levels on
daily death counts, do not capture any impacts of long-term exposure to air pollution. As
discussed earlier, EPA's Science Advisory Board, while acknowledging the uncertainties in
estimation of a PM-mortality relationship, has repeatedly recommended the use of a study that
does reflect the impacts of long-term exposure. The omission of long-term impacts accounts for
approximately 40 percent reduction in the estimate of avoided premature mortality in the
Alternative Estimate relative to the Base Estimate.
8.3.3 Valuing Individual Health Effect Endpoints
The appropriate economic value of a change in a health effect depends on whether the
health effect is viewed ex ante (before the effect has occurred) or ex post (after the effect has
occurred). Reductions in ambient concentrations of air pollution generally lower the risk of
future adverse health affects by a fairly small amount for a large population. The appropriate
economic measure is therefore ex ante WTP for changes in risk. However, epidemiological
studies generally provide estimates of the relative risks of a particular health effect avoided due
to a reduction in air pollution. A convenient way to use this data in a consistent framework is to
convert probabilities to units of avoided statistical incidences. This measure is calculated by
dividing individual WTP for a risk reduction by the related observed change in risk. For
example, suppose a measure is able to reduce the risk of premature mortality from 2 in 10,000 to
1 in 10,000 (a reduction of 1 in 10,000). If individual WTP for this risk reduction is $100, then
the WTP for an avoided statistical premature mortality amounts to $1 million ($100/0.0001
change in risk). Using this approach, the size of the affected population is automatically taken
into account by the number of incidences predicted by epidemiological studies applied to the
relevant population. The same type of calculation can produce values for statistical incidences
of other health endpoints.
For some health effects, such as hospital admissions, WTP estimates are generally not
available. In these cases, we use the cost of treating or mitigating the effect as a Base estimate.
8-21
-------
For example, for the valuation of hospital admissions we use the avoided medical costs as an
estimate of the value of avoiding the health effects causing the admission. These costs of illness
(COI) estimates generally understate the true value of reductions in risk of a health effect. They
tend to reflect the direct expenditures related to treatment but not the value of avoided pain and
suffering from the health effect.
In the HDD RIA and TSD, we describe how the changes in health effects should be
valued and indicate the value functions selected to provide monetized estimates of the value of
changes in health effects. Table 8-3 below summarizes the value estimates per health effect that
we used in this analysis. Note that the unit values for hospital admissions are the weighted
averages of the ICD-9 code-specific values for the group of ICD-9 codes included in the hospital
admission categories.
Adjustments for Growth in Real Income
Our analysis also accounts for expected growth in real income over time. Economic
theory argues that WTP for most goods (such as environmental protection) will increase if real
incomes increase. The economics literature suggests that the severity of a health effect is a
primary determinant of the strength of the relationship between changes in real income and WTP
(Alberini, 1997; Miller, 2000; Viscusi, 1993). As such, we use different factors to adjust the
WTP for minor health effects, severe and chronic health effects, and premature mortality.
Adjustment factors used to account for projected growth in real income from 1990 to 2005 are
1.03 for minor health effects, 1.09 for severe and chronic health effects, and 1.08 for premature
mortality.6
6Details of the calculation of the income adjustment factors are provided in the HDD RIA (EPA, 2000d).
8-22
-------
Table 8-3. Unit Values Used for Economic Valuation of Health Endpoints
Health or Welfare
Endpoint
Estimated Value
Per Incidence
(1999$)
Central Estimate
Derivation of Estimates
Premature Mortality (long-
term exposure endpoint, Base
Estimate)
Premature Mortality (short-
term exposure endpoints,
Alternative Estimate)
Chronic Bronchitis (Base
Estimate)
Value is the mean of value-of-statistical-life estimates from
$6 million per 26 studies (5 contingent valuation and 21 labor market
statistical life studies) reviewed for the Section 812 Costs and Benefits of
the Clean Air Act, 1990-2010 (EPA, 1999).
Varies by age and See section on Valuation of Premature Mortality, Alternative
life years lost Estimate, in text
$331,000
Chronic Bronchitis
(Alternative Estimate)
Hospital Admissions
All Ozone-Related Respiratory
Chronic Obstructive
Pulmonary Disease (COPD)
(ICD codes 490-492, 494-496)
Pneumonia
(ICD codes 480-487)
Asthma admissions
All Cardiovascular
(ICD codes 390-429)
Emergency room visits for
asthma
$107,000
$9,823
$12,378
$14,693
$6,634
$18,387
$299
Value is the mean of a generated distribution of WTP to avoid
a case of pollution-related CB. WTP to avoid a case of
pollution-related CB is derived by adjusting WTP (as
described in Viscusi et al., 1991) to avoid a severe case of CB
for the difference in severity and taking into account the
elasticity of WTP with respect to severity of CB.
Cost of Illness (COI) estimate based on Cropper and
Krupnick (1990).
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total COPD category illnesses)
reported in Elixhauser (1993).
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total pneumonia category
illnesses) reported in Elixhauser (1993).
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total asthma category illnesses)
reported in Elixhauser (1993).
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total cardiovascular illnesses)
reported in Elixhauser (1993).
COI estimate based on data reported by Smith, et al. (1997).
-------
Health or Welfare
Endpoint
Estimated Value
Per Incidence
(1999$)
Central Estimate
Derivation of Estimates
Respiratory Ailments Not Requiring Hospitalization
Any of 19 Acute Symptoms
(ozone-related)
Upper Respiratory Symptoms
(URS)
$22
$24
Lower Respiratory Symptoms
(LRS)
$15
Acute Bronchitis
$57
Combinations of the 3 symptoms for which WTP estimates
are available that closely match those listed by Pope, et al.
result in 7 different "symptom clusters," each describing a
"type" of URS. A dollar value was derived for each type of
URS, using mid-range estimates of WTP (lEc, 1994) to avoid
each symptom in the cluster and assuming additivity of
WTPs. The dollar value for URS is the average of the dollar
values for the 7 different types of URS.
Combinations of the 4 symptoms for which WTP estimates
are available that closely match those listed by Schwartz, et
al. result in 11 different "symptom clusters," each describing
a "type" of LRS. A dollar value was derived for each type of
LRS, using mid-range estimates of WTP (lEc, 1994) to avoid
each symptom in the cluster and assuming additivity of
WTPs. The dollar value for LRS is the average of the dollar
values for the 11 different types of LRS.
Average of low and high values recommended for use in
Section 812 analysis (Neumann, et al., 1994)
Restricted Activity and Work Loss Days
Decreased Worker
Productivity
Work Loss Days (WLDs)
Minor Restricted Activity
Days (MRADs)
$1 per worker per
10% change in
ozone
Variable
$48
Regionally adjusted median weekly wage for 1990 divided by
5 (adjusted to 1999$) (U.S. Bureau of the Census, 1992).
Median WTP estimate to avoid one MRAD from Tolley, et
al. (1986).
8-24
-------
8.3.3.1 Valuation of Reductions in Premature Mortality Risk
Below we present the method for valuing premature mortality in our Base and
Alternative Estimates. In both estimates, the values reflect two alternative discount rates, three
percent and seven percent, used to estimate the present value of the effect. The choice of a
discount rate, and its associated conceptual basis, is a topic of ongoing discussion within the
federal government. We adopted a three percent discount rate to reflect reliance on a "social rate
of time preference" discounting concept, which is recommended by EPA's Guidelines for
Preparing Economic Analyses (EPA, 2000b). We also calculate benefits using a seven percent
rate consistent with an "opportunity cost of capital" concept to reflect the time value of resources
directed to meet regulatory requirements, which is recommended by OMB Circular A-94 (OMB,
1992). In this analysis, the benefit estimates were not significantly affected by the choice of
discount rate. Further discussion of this topic appears in EPA's Guidelines for Preparing
Economic Analyses (EPA, 2000b).
Base Estimate
The monetary benefit of reducing premature mortality risk was estimated using the
"value of statistical lives saved" (VSL) approach, although the actual valuation is of small
changes in mortality risk experienced by a large number of people. The VSL approach applies
information from several published value-of-life studies, which themselves examine tradeoffs of
monetary compensation for small additional mortality risks, to determine a reasonable benefit of
preventing premature mortality. The mean value of avoiding one statistical death is estimated to
be $6 million in 1999 dollars. This represents an intermediate value from a range of estimates
that appear in the economics literature, and it is a value the EPA has used in rulemaking support
analyses and in the Section 812 Reports to Congress.
This estimate is the mean of a distribution fitted to the estimates from 26 value-of-life
studies identified in the Section 812 reports as "applicable to policy analysis." The approach and
set of selected studies mirrors that of Viscusi (1992) (with the addition of two studies), and uses
the same criteria as Viscusi in his review of value-of-life studies. The $6 million estimate is
consistent with Viscusi's conclusion (updated to 1999$) that "most of the reasonable estimates
of the value of life are clustered in the $3.7 to $8.6 million range." Five of the 26 studies are
8-25
-------
contingent valuation (CV) studies, which directly solicit WTP information from subjects; the rest
are wage-risk studies, which base WTP estimates on estimates of the additional compensation
demanded in the labor market for riskier jobs, controlling for other job and employee
characteristics such as education and experience. As indicated in the previous section on
quantification of premature mortality benefits, we assume for this analysis that some of the
incidences of premature mortality related to PM exposures occur in a distributed fashion over the
five years following exposure. To take this into account in the valuation of reductions in
premature mortality, we apply an annual three percent discount rate to the value of premature
mortality occurring in future years.
The economics literature concerning the appropriate method for valuing reductions in
premature mortality risk is still developing. The adoption of a value for the projected reduction
in the risk of premature mortality is the subject of continuing discussion within the economic and
public policy analysis community. Regardless of the theoretical economic considerations,
distinctions in the monetary value assigned to the lives saved were not drawn, even if
populations differed in age, health status, socioeconomic status, gender or other characteristics.
Following the advice of the EEAC of the SAB, the VSL approach was used to calculate
the Base Estimate of mortality benefits (EPA-SAB-EEAC-00-013). While there are several
differences between the risk context implicit in labor market studies we use to derive a VSL
estimate and the particulate matter air pollution context addressed here, those differences in the
affected populations and the nature of the risks imply both upward and downward adjustments.
For example, adjusting for age differences between subjects in the economic studies and those
affected by air pollution may imply the need to adjust the $6 million VSL downward, but the
involuntary nature of air pollution-related risks and the lower level of risk-aversion of the
manual laborers in the labor market studies may imply the need for upward adjustments.
Some economists emphasize that the value of a statistical life is not a single number
relevant for all situations. Indeed, the VSL estimate of $6 million (1999 dollars) is itself the
central tendency of a number of estimates of the VSL for some rather narrowly defined
populations. When there are significant differences between the population affected by a
particular health risk and the populations used in the labor market studies, as is the case here,
some economists prefer to adjust the VSL estimate to reflect those differences.
8-26
-------
There is general agreement that the value to an individual of a reduction in mortality risk
can vary based on several factors, including the age of the individual, the type of risk, the level
of control the individual has over the risk, the individual's attitudes towards risk, and the health
status of the individual. While the empirical basis for adjusting the $6 million VSL for many of
these factors does not yet exist, a thorough discussion of these uncertainties is included in EPA's
Guidelines for Preparing Economic Analyses (EPA, 2000b). The EPA recognizes the need for
investigation by the scientific community to develop additional empirical support for
adjustments to VSL for the factors mentioned above.
As further support for the Base benefits estimate, the SAB-EEAC advised in their recent
report that the EPA "continue to use a wage-risk-based VSL as its Base Estimate, including
appropriate sensitivity analyses to reflect the uncertainty of these estimates," and that "the only
risk characteristic for which adjustments to the VSL can be made is the timing of the risk"(EPA-
SAB-EEAC-00-013). In developing the Base Estimate of the benefits of premature mortality
reductions, we have discounted over the lag period between exposure and premature mortality.
However, in accordance with the SAB advice, we use the VSL in the Base Estimate.
Alternative Estimate
The Alternative Estimate reflects the impact of changes to key assumptions associated
with the valuation of mortality. These include: (1) the impact of using wage-risk and contingent
valuation-based value of statistical life estimates in valuing risk reductions from air pollution as
opposed to contingent valuation-based estimates alone, (2) the relationship between age and
willingness-to-pay for fatal risk reductions, and (3) the degree of prematurity in mortalities from
air pollution.
The Alternative Estimate addresses the first issue by using an estimate of the value of
statistical life that is based only on the set of five contingent valuation studies included in the
larger set of 26 studies recommended by Viscusi (1992) as applicable to policy analysis. The
mean of the five contingent valuation based VSL estimates is $3.7 million (1999$), which is
approximately 60 percent of the mean value of the full set of 26 studies.
The second issue is addressed by assuming that the relationship between age and
willingness-to-pay for fatal risk reductions can be approximated using an adjustment factor
8-27
-------
derived from Jones-Lee (1989). The SAB has advised the EPA that the appropriate way to
account for age differences is to obtain the values for risk reductions from the age groups
affected by the risk reduction. Several studies have found a significant effect of age on the value
of mortality risk reductions expressed by citizens in the United Kingdom (Jones-Lee et al., 1985;
Jones-Lee, 1989; Jones-Lee, 1993).
Two of these studies provide the basis to form ratios of the WTP of different age cohorts
to a base age cohort of 40 years. These ratios can be used to provide Alternative age-adjusted
estimates of the value of avoided premature mortalities. One problem with both of the Jones-Lee
studies is that they examine VSL for a limited age range. They then fit VSL as a function of age
and extrapolate outside the range of the data to obtain ratios for the very old. Unfortunately,
because VSL is specified as quadratic in age, extrapolation beyond the range of the data can lead
to a very severe decline in VSL at ages beyond 75.
A simpler and potentially less biased approach is to simply apply a single age adjustment
based on whether the individual was over or under 65 years of age at the time of death. This is
consistent with the range of observed ages in the Jones-Lee studies and also agrees with the
findings of more recent studies by Krupnick et al. (2000) that the only significant difference in
WTP is between the over 70 and under 70 age groups. To correct for the potential extrapolation
error for ages beyond 70, the adjustment factor is selected as the ratio of a 70 year old
individual's WTP to a 40 year old individual's WTP, which is 0.63, based on the Jones-Lee
(1989) results and 0.92 based on the Jones-Lee (1993) results. To show the maximum impact of
the age adjustment, the Alternative Estimate is based on the Jones-Lee (1989) adjustment factor
of 0.63, which yields a VSL of $2.3 million for populations over the age of 70. Deaths of
individuals under the age of 70 are valued using the unadjusted mean VSL value of $3.7 million
(1999$). Since these are acute mortalities, it is assumed that there is no lag between reduced
exposure and reduced risk of mortality.
Jones-Lee and Krupnick may understate the effect of age because they only control for
income and do not control for wealth. While there is no empirical evidence to support or reject
hypotheses regarding wealth and observed WTP, WTP for additional life years by the elderly
may in part reflect their wealth position vis a vis middle age respondents.
8-28
-------
The third issue is addressed by assuming that deaths from chronic obstructive pulmonary
disease (COPD) are advanced by 6 months, and deaths from all other causes are advanced by 5
years. These reductions in life years lost are applied regardless of the age at death. Actuarial
evidence suggests that individuals with serious preexisting cardiovascular conditions have a
remaining life expectancy of around 5 years. While many deaths from daily exposure to PM
may occur in individuals with cardiovascular disease, studies have shown relationships between
all cause mortality and PM, and between PM and mortality from pneumonia (Schwartz, 2000).
In addition, recent studies have shown a relationship between PM and non-fatal heart attacks,
which suggests that some of the deaths due to PM may be due to fatal heart attacks (Peters et al.,
2001). And, a recent meta-analysis has shown little effect of age on the relative risk from PM
exposure (Stieb et al., 2002), which suggests that the number of deaths in non-elderly
populations (and thus the potential for greater loss of life years) may be significant. Indeed, this
analysis estimates that 21 percent of non-COPD premature deaths avoided are in populations
under 65. Thus, while the assumption of 5 years of life lost may be appropriate for a subset of
total avoided premature mortalitites, it may over or underestimate the degree of life shortening
attributable to PM for the remaining deaths.
In order to value the expected life years lost for COPD and non-COPD deaths, we need to
construct estimates of the value of a statistical life year. The value of a life year varies based on
the age at death, due to the differences in the base VSL between the 65 and older population and
the under 65 population. The valuation approach used is a value of statistical life years (VSLY)
approach, based on amortizing the base VSL for each age cohort. Previous applications have
arrived at a single value per life year based on the discounted stream of values that correspond to
the VSL for a 40 year old worker (EPA, 1999a). This assumes 35 years of life lost is the base
value associated with the mean VSL value of $3.7 million (1999$). The VSLY associated with
the $3.7 million VSL is $163,000, annualized assuming EPA's guideline value of a 3 percent
discount rate, or $270,000, annualized assuming OMB's guideline value of a 7 percent discount
rate. For example, using the 3 percent discount rate, the VSL applied in this analysis is then
built up from that VSLY by taking the present value of the stream of life years. Thus, if you
assume that a 40 year-old dying from pneumonia would lose 5 years of life, the VSL applied to
that death would be $0.79 million. For populations over age 65, we then develop a VSLY from
8-29
-------
the age-adjusted base VSL of $2.3 million. Given an assumed remaining life expectancy of 10
years, this gives a VSLY of $258,000, assuming a 3 percent discount rate. A similar calculation
is used to derive the VSLY estimate using a 7% discount rate. Again, the VSL is built based on
the present value of 5 years of lost life, so in this case, we have a 70 year old individual dying
from pneumonia losing 5 years of life, implying an estimated VSL of $1.25 million. As a final
step, these estimated VSL values are multiplied by the appropriate adjustment factors to account
for changes in WTP over time, as outlined above.
Applying the VSLY approach to the four categories of acute mortality results in four
separate sets of values for an avoided premature mortality based on age and cause of death.
Non-COPD deaths for populations aged 65 and older are valued at $1.4 million per incidence in
2010, and $1.6 million in 2020. Non-COPD deaths for populations aged 64 and younger are
valued at $0.88 million per incidence in 2010, and $1.0 million in 2020. COPD deaths for
populations aged 65 and older are valued at $0.15 million per incidence in 2010, and $0.17
million in 2020. Finally, COPD deaths for populations aged 64 and younger are valued at
$0.096 million per incidence in 2010, and $0.11 million in 2020. The implied VSL for younger
populations is less than that for older populations because the value per life year is higher for
older populations. Since we assume that there is a 5 year loss in life years for a PM related
mortality, regardless of the age of person dying, this necessarily leads to a lower VSL for
younger populations.
Note that the NMMAPS study used to derive the C-R function for PM10 did not provide
separate estimates for different causes of death, so we are unable to determine the proportion of
PM10 deaths that are attributable to COPD or other causes. In the Base analysis, such
distinctions are unnecessary, as all reductions in incidence of premature mortality are valued
equally, regardless of age at death or remaining life expectancy. In the alternative estimate, the
value of avoided incidences of premature mortality is determined by age and remaining life
expectancy, so cause of death and age are important. Given the lack of data on cause of death,
we assume all deaths from PM10 are equivalent (within an age category) and result in the same
number of life years lost, assumed to be equal to 5 years.
8.3.3.2 Valuation of Reductions in Chronic Bronchitis
8-30
-------
Base Estimate
The best available estimate of WTP to avoid a case of chronic bronchitis (CB) comes
from Viscusi et al. (1991). The Viscusi et al. study, however, describes a severe case of CB to
the survey respondents. We therefore employ an estimate of WTP to avoid a pollution-related
case of CB, based on adjusting the Viscusi, et al. (1991) estimate of the WTP to avoid a severe
case. This is done to account for the likelihood that an average case of pollution-related CB is
not as severe. The adjustment is made by applying the elasticity of WTP with respect to severity
reported in the Krupnick and Cropper (1992) study. Details of this adjustment procedure can be
found in the Heavy-Duty Engine/Diesel Fuel RIA and its supporting documentation, and in the
most recent Section 812 study (EPA, 1999).
We use the mean of a distribution of WTP estimates as the central tendency estimate of
WTP to avoid a pollution-related case of CB in this analysis. The distribution incorporates
uncertainty from three sources: (1) the WTP to avoid a case of severe CB, as described by
Viscusi et al.; (2) the severity level of an average pollution-related case of CB (relative to that of
the case described by Viscusi et al.); and (3) the elasticity of WTP with respect to severity of the
illness. Based on assumptions about the distributions of each of these three uncertain
components, we derive a distribution of WTP to avoid a pollution-related case of CB by
statistical uncertainty analysis techniques. The expected value (i.e., mean) of this distribution,
which is about $331,000 (1999$), is taken as the central tendency estimate of WTP to avoid a
PM-related case of CB.
Alternative Estimate
For the Alternative Estimate, a cost-of illness value is used in place of willingness-to-pay
to reflect uncertainty about the value of reductions in incidences of chronic bronchitis. In the
Base Estimate, the willingness-to-pay estimate was derived from two contingent valuation
studies (Viscusi et al., 1991; Krupnick and Cropper, 1992). These studies were experimental
studies intended to examine new methodologies for eliciting values for morbidity endpoints.
Although these studies were not specifically designed for policy analysis, the SAB (EPA-SAB-
COlMCIL-ADV-00-002, 1999a) has indicated that the severity-adjusted values from this study
provide reasonable estimates of the WTP for avoidance of chronic bronchitis. As with other
8-31
-------
contingent valuation studies, the reliability of the WTP estimates depends on the methods used
to obtain the WTP values. In order to investigate the impact of using the CV based WTP
estimates, the Alternative Estimate relies on a value for incidence of chronic bronchitis using a
cost-of-illness estimate based on Cropper and Krupnick (1990) which calculates the present
value of the lifetime expected costs associated with the illness. The current cost-of-illness (COI)
estimate for chronic bronchitis is around $107,000 per case, compared with the current WTP
estimate of $330,000.
8.3.4 Methods for Describing Uncertainty
In any complex analysis using estimated parameters and inputs from numerous models,
there are likely to be many sources of uncertainty.7 This analysis is no exception. As outlined
both in this and preceding chapters, there are many inputs used to derive the final estimate of
benefits, including emission inventories, air quality models (with their associated parameters and
inputs), epidemiological estimates of concentration-response (C-R) functions, estimates of
values (both from WTP and cost-of-illness studies), population estimates, income estimates, and
estimates of the future state of the world (i.e., regulations, technology, and human behavior).
Each of these inputs may be uncertain, and depending on their location in the benefits analysis,
may have a disproportionately large impact on final estimates of total benefits. For example,
emissions estimates are used in the first stage of the analysis. As such, any uncertainty in
emissions estimates will be propagated through the entire analysis. When compounded with
uncertainty in later stages, small uncertainties in emission levels can lead to much larger impacts
on total benefits.
Some key sources of uncertainty in each stage of the benefits analysis are:
•• Gaps in scientific data and inquiry;
7It should be recognized that in addition to uncertainty, the annual benefit estimates for the RICE NESHAP
presented in this analysis are also inherently variable, due to the truly random processes that govern pollutant
emissions and ambient air quality in a given year. Factors such as electricity demand and weather display
constant variability regardless of our ability to accurately measure them. As such, the estimates of annual
benefits should be viewed as representative of the types of benefits that will be realized, rather than the actual
benefits that would occur every year.
8-32
-------
•• Variability in estimated relationships, such as C-R functions, introduced through
differences in study design and statistical modeling;
•• Errors in measurement and projection for variables such as population growth
rates;
Errors due to mis-specification of model structures, including the use of surrogate
variables, such as using PM10 when PM2 5 is not available, excluded variables, and
simplification of complex functions; and
Biases due to omissions or other research limitations.
Some of the key uncertainties in the benefits analysis are presented in Table 8-3.
Information on the uncertainty surrounding particular C-R and valuation functions is provided in
HDD TSD.
Our estimate of total benefits should be viewed as an approximate result because of the
sources of uncertainty discussed above (see Table 8-4). The total benefits estimate may
understate or overstate actual benefits of the rule.
In considering the monetized benefits estimates, the reader should remain aware of the
many limitations of conducting these analyses mentioned throughout this RIA. One significant
limitation of both the health and welfare benefits analyses is the inability to quantify many of the
serious effects discussed in Chapter 7. For many health and welfare effects, such as PM-related
materials damage, reliable C-R functions and/or valuation functions are not currently available.
In general, if it were possible to monetize these benefits categories, the benefits estimates
presented in this analysis would increase. Unquantified benefits are qualitatively discussed in
the health and welfare effects sections of this RIA. The net effect of excluding benefit and
disbenefit categories from the estimate of total benefits depends on the relative magnitude of the
effects.
Table 8-4. Primary Sources of Uncertainty in the Source Benefit Analyses
||_7. Uncertainties Associated With Concentration-Response (C-R) Functions II
8-33
-------
The value of the PM-coefficient in each C-R function.
Application of a single C-R function to pollutant changes and populations in all locations.
Similarity of future year C-R relationships to current C-R relationships.
Correct functional form of each C-R relationship.
Extrapolation of C-R relationships beyond the range of PM concentrations observed in the study.
Application of C-R relationships only to those subpopulations matching the original study population.
2. Uncertainties Associated With PM Concentrations
Responsiveness of the models to changes in precursor emissions resulting from the control policy.
Projections of future levels of precursor emissions, especially ammonia and crustal materials.
Model chemistry for the formation of ambient nitrate concentrations.
3. Uncertainties Associated with PM Mortality Risk
No scientific literature supporting a direct biological mechanism for observed epidemiological evidence.
Direct causal agents within the complex mixture of PM have not been identified.
The extent to which adverse health effects are associated with low level exposures that occur many times in the year
versus peak exposures.
The extent to which effects reported in the long-term exposure studies are associated with historically higher levels
of PM rather than the levels occurring during the period of study.
Reliability of the limited ambient PM25 monitoring data in reflecting actual PM25 exposures.
4. Uncertainties Associated With Possible Lagged Effects
The portion of the PM-related long-term exposure mortality effects associated with changes in annual PM levels
would occur in a single year is uncertain as well as the portion that might occur in subsequent years.
5. Uncertainties Associated With Baseline Incidence Rates
Some baseline incidence rates are not location-specific (e.g., those taken from studies) and may therefore not
accurately represent the actual location-specific rates.
Current baseline incidence rates may not approximate well baseline incidence rates in 2005.
Projected population and demographics may not represent well future-year population and demographics.
6. Uncertainties Associated With Economic Valuation
Unit dollar values associated with health and welfare endpoints are only estimates of mean WTP and therefore have
uncertainty surrounding them.
Mean WTP (in constant dollars) for each type of risk reduction may differ from current estimates due to differences
in income or other factors.
7. Uncertainties Associated With Aggregation of Monetized Benefits
Health and welfare benefits estimates are limited to the available C-R functions. Thus, unquantified or
unmonetized benefits are not included.
8.4 DERIVATION OF BENEFIT TRANSFER VALUES FOR THE RICE NESHAP
8.4.1 Ozone Benefit Transfer Values for Application to NOx Emission Reductions
8-34
-------
The ozone benefits analysis conducted for this RIA includes three categories of ozone
related health benefits,2 but not the ozone related welfare benefits (including changes in
agricultural and forest productivity3). These categories are not included in this analysis due to a
lack of suitable sources for benefit transfer. The agricultural and forestry models used to
generate benefits are national sector models. As such, the outputs of these models are not
suitable for disaggregation to dollar per ton values. Benefits from the omitted welfare
categories (primarily commercial agriculture and forestry) accrue in rural areas, and thus may be
important sources of benefits from reductions in emissions from RICE sources. This will lead to
a downward bias in the reported estimates of benefits from NOx reductions.
The first step of the benefit transfer method is to select an existing air quality analysis
from which to obtain changes in ambient ozone concentrations. Two factors guide the selection
of an ozone air quality analysis for use in the RICE NESHAP benefits analysis: (1) while both
NOx and VOC contribute to ozone formation, this regulation will lead to reductions
predominately in NOx,8 and 2) RICE sources are stationary combustion sources (as opposed to
mobile sources such as vehicles). As such, an existing set of ozone air quality results covering
primarily NOx reductions at stationary combustion sources is the most appropriate match.
We selected an air quality scenario developed for the NOx SIP call. This air quality
scenario uses the Urban Airshed Model, version 5 (UAM-V) to predict ambient ozone
concentration changes in 2007 from a 0.15 Ib/mmBTU limit on NOx emissions for electric
utilities and a 60 percent reduction in NOx emissions for non-utility point sources. UAM-V is a
regional scale ozone model accounting for spatial and temporal variations as well as differences
in the reactivity of emissions. Ozone air quality is modeled for the Ozone Transport Assessment
Group (OTAG) region (essentially the 37 easternmost states). The model segments the area in
the OTAG region into grids, each of which has several layers of ambient conditions that are
considered in the analysis. Using this data, the UAM-V generates predictions of hourly ozone
concentrations for every grid. Results of this process are used to generate ozone profiles at
monitor sites by applying derived adjustment factors to the actual 1990 ozone data at each
8Some VOC reductions are expected from the controls applied to RICE sources. However, we are unable to measure
them with a reasonable level of certainty. As the reductions are expected to be small, we do not anticipate a large
impact on ambient ozone levels.
8-35
-------
monitor site.9 For areas without ozone monitoring data, ozone values are interpolated using data
from monitors surrounding the area. For a more detailed discussion of UAM-V and the air
quality interpolation procedure and the NOx SIP call reduction scenario, see the 1998 NOx SIP
Call RIA and associated air quality technical support document (EPA, 1998; Abt Associates,
1998).
In prior EPA analyses (i.e., the 1997 Regulatory Impact Analysis (RIA) for the integrated
pulp and paper rule and McKeever, 1997), we used air quality results with both NOx and VOC
emission reductions. However, use of these results required the assumption of proportionality
between emission reductions of VOCs and NOx and reductions in ambient ozone concentrations
to obtain benefit transfer values for each pollutant. Subsequent to 1997, EPA has conducted air
quality analyses of changes in ozone concentrations from NOx emissions alone. By using air
quality results based solely on NOx emission reductions, all changes in ozone concentrations
will be directly attributable to the NOx reductions, removing the need for assumptions about the
proportion of changes in ambient ozone attributable to VOC reductions relative to NOx
reductions.
To construct the dollar per ton ($/ton) benefit transfer value based on the NOx SIP call
ozone benefits analysis estimate, we perform the following steps:
1) Adjust the ozone benefits estimated for the NOx SIP call to reflect the current set
of endpoints and benefits assumptions and updated the base year to 1998 dollars.
2) Divide the resulting estimate by the total ozone season tons of NOx reduced
under the NOx SIP call to obtain monetary ozone benefits per ton ($/ton) of NOx
reduced.
9Ten decile adjustment factors are derived based on UAM-V modeled daytime hours (8:00 am-7:59 pm). From the
distribution of these modeled hours, each decile is represented by its middle value. In other words, the first
decile is represented by the 5th percentile value, the second decile by the 15th percentile value, and so on. For
both the baseline and control scenarios, ten adjustment factors are then calculated using the ratio within each
decile of the future year to the base year concentration. The ten adjustment factors for the baseline and control
scenario are then used to adjust 1990 hourly ozone concentrations to projected 2007 concentrations. The lowest
10 percent of the distribution of these hours were multiplied by the first decile adjustment factor, the next 10
percent by the second adjustment factor and so on. Only daytime hours (8:00 am to 7:59 pm) were adjusted.
Nighttime hours were assumed to be constant.
8-36
-------
Step 1 is necessary due to the refinements in the benefits methodology that have occurred
since the NOx SIP call analysis. The benefits analysis for the HDD RIA incorporates the latest
guidance from the Science Advisory Board regarding appropriate endpoints for inclusion and
appropriate valuation methods. For a complete description of the benefits methodology used to
develop the HDD benefits estimates, see the HDD RIA and TSD. The key modification to the
ozone benefits associated with NOx is that reductions in ozone-related mortality are no longer
included in the primary estimate of ozone-related benefits.10
Step 2 converts total benefits into an appropriate dollar per ton metric using NOx
emissions during the ozone season. Ozone season NOx reductions are the basis for the benefits
reported for the NOx SIP call, reflecting the greater impact of NOx reductions on ozone
formation during the ozone season (May through September). Note that annual RICE NOx
reductions will also have to be separated into ozone and non-ozone season tons before
application of the ozone $/ton transfer values. The calculations for this benefit transfer exercise
are laid out in Table 8-5, and will be applied to emission reduction estimates for the RICE
NESHAP.
10At least some evidence has been found linking both PM and ozone with premature mortality. The SAB has raised
concerns that mortality-related benefits of air pollution reductions may be overstated if separate pollutant-
specific estimates, some of which may have been obtained from models excluding the other pollutants, are
aggregated. In addition, there may be important interactions between pollutants and their effect on mortality
(EPA-SAB-Council-ADV-99-012, 1999b). The Pope et al. (1995) study used to quantify PM-related mortality
included only PM, so it is unclear to what extent it may include the impacts of ozone or other gaseous pollutants.
Because of concern about overstating of benefits and because the evidence associating mortality with exposure to
paniculate matter is currently stronger than for ozone, only the benefits of PM-related premature mortality
avoided are included in the total benefits estimate.
8-37
-------
Table 8-5. Ozone $/ton Transfer Values for NOx Reductions
Using Estimates from the NOx SIP Call
Step la
Step Ib
Description
Calculate unadjusted ozone-related health benefits
from NOx SIP Call "Best Estimate" (Hubbell, 1998)
Calculate adjusted ozone health benefits (applying
Outcome (1998$)
$1,690 million
$36 million
SAB recommended current assumptions and
endpoint sets3)
Step 2 Divide adjusted ozone benefits by total ozone $36 million/1.3 million tons
season NOx reductions for the NOx SIP call $28/ton
Includes hospital admissions for all respiratory causes, acute respiratory symptoms, and lost worker productivity.
8.4.2 PM2 j Benefit Transfer Values for Application to NOx Emission Reductions
PM2 5 benefit transfer values associated with NOx reductions are developed using the
same basic approach as for ozone. However, the specific air quality models and health endpoints
differ. The PM2 5 benefits analysis conducted for this RIA includes health benefits associated
with reductions in both PM25 and PM10.n While Table 8-1 lists the endpoints included in this
analysis, not all known health and welfare effects associated with PM are quantified and
monetized for this analysis. Potential benefit categories that have not been quantified and
monetized are listed in Table 8-13 later in this chapter. For more details on the sources and
derivation of C-R functions and unit economic values for specific PM related health endpoints,
see the HDD RIA and TSD.
The first step of the benefit transfer approach for PM25 related to NOx reductions is to
generate an emissions control scenario reflecting the types of reductions expected from the RICE
NESHAP rule. Based on the NET96 emissions inventory, one-half of all NOx emissions from
RICE sources totals 370,877 tons. In developing the RICE NESHAP we estimated NOx
reductions to total 420,000 tons if all new sources (including major and area sources) were
1 'PM2 5 is a fraction of PM10. As such, reductions in NOx that lead to reductions in secondarily formed PM2 5 will
also be equivalent to reductions of PM10 in the same amount. Because PM2 5 may be more strongly associated
with health effects, we use PM2 5 based concentration-response functions where available. However, due to
limited availability of PM2 5 data, many concentration-response functions are estimated using only PM10 data.
8-38
-------
controlled. Thus during the development stage of the NESHAP, we concluded that an air quality
analysis of approximately 50 percent reduction in NOx is reasonable to use to transfer results to
the NESHAPs reductions. Thus, we selected and modeled a 50 percent reduction to NOx
emissions from all RICE sources in the continental U.S. contained in the NET96 inventory. We
recognize that many of the RICE sources included in the modeled air quality analysis will not be
controlled under this NEHSAP, but this scenario provides a close approximation of the influence
of NOx emissions reductions at RICE sources on concentrations of PM for the purpose of
developing benefit transfer values.
PM air quality changes resulting from the 50 percent RICE NOx reduction were analyzed
using a national-scale source-receptor matrix (S-R Matrix) based on the Climatological Regional
Dispersion Model (CRDM) (Latimer and Associates, 1994; E.H. Pechan, 1994, 1996). Ambient
concentrations of PM25 are composed of directly emitted particles and of secondary aerosols of
sulfate, nitrate, ammonium, and organics. Relative to more sophisticated and resource-intensive
three-dimensional modeling approaches, the CRDM and its associated S-R Matrix do not fully
account for all the complex chemical interactions that take place in the atmosphere in the
secondary formation of PM. Instead it relies on more simplistic species dispersion-transport
mechanisms supplemented with chemical conversion at the receptor location.
The S-R Matrix consists of fixed-coefficients that reflect the relationship between annual
average PM concentration values at a single receptor in each county (i.e., a hypothetical monitor
sited at the county population centroid) and the contribution by PM species to this concentration
from each emission source (E.H. Pechan, 1997). The modeled receptors include all U.S. county
centroids as well as receptors in 10 Canadian provinces and 29 Mexican cities/states. The
methodology used in this RIA for estimating PM air quality concentrations is detailed in Pechan-
Avanti (2000) and is similar to the method used in the RIA for the recent Tier 2/Gasoline Sulfur
Rule (EPA, 1999e). For a complete description of the S-R Matrix, see chapter 7 of the Final Tier
2/Gasoline Sulfur RIA.
In the air quality modeling of the 50 percent NOx reduction scenario, results are based on
a baseline 1996 emission inventory applied to populations estimated for the year 2005. The
actual emissions in 2005 may be higher or lower than the 1996 baseline used in this analysis.
8-39
-------
Given the changes in ambient PM2 5 concentrations from the S-R matrix, the following are the
key steps in the approach for developing a PM2 5 benefit transfer value:
Step 1) Apply changes in PM25 concentrations to selected health and welfare
concentration-response functions at the population grid cell level12.
Step 2) Apply valuation functions to the change in endpoint incidences and sum
over endpoints to obtain monetary benefits at the population grid cell
level.
Step 3) Sum monetary benefits over population grid cells to obtain aggregate
monetary benefits estimates for the continental U.S.
Step 4) Divide aggregate monetary benefits by annual NOx emission reductions in
the NET96 inventory to obtain a national $/ton estimate.
The calculations for this benefit transfer exercise are provided in Tables 8-6(a) and (b). Total
reductions in NOx emissions for the 50 percent RICE NOx reduction scenario using the NET96
inventory are 370,877 tons. Dividing total benefits by the NOx emission reductions yields a
$/ton estimate of $1,510 using the Base Estimate with a 3 percent discount rate on mortality, and
$1,430 per ton using a 7 percent discount rate. Using results from the Alternative Estimate, the
$/ton estimate using a 3 percent discount rate is $188, while using a 7 percent discount rate
yields a $/ton of $215. Note that this averaging process implies that all reductions in emissions,
wherever they occur, potentially affect air quality across the entire U.S. population. Thus, no
additional scaling for population is appropriate.
12Changes in ambient pollutant concentrations are input to CAPMS, a custom benefits analysis program, to generate
changes in health and welfare endpoints. CAPMS interpolates pollutant concentrations to population grid cells
for input into concentration-response functions. CAPMS uses census block population data along with the
interpolated changes in pollutant concentrations to estimate changes in endpoints at the population grid cell level.
For more details on CAPMS, see the benefits technical support documents for the Final Tier 2/Gasoline Sulfur
RIA. (Abt Associates, 1998b, 1999)
8-40
-------
Table 8-6(a). Base Estimate of Annual Health Benefits
Resulting from 50 Percent RICE NOx Emission Reduction Scenario"
Monetary Benefits,
Adjusted for Growth
Avoided Incidence6 in Income0
Endpoint (cases/year) (millions 1998$)
Premature mortality 4e (long-term exposure, adults, 30 and over):
-Using a 3% discount rate
-Using a 7% discount rate
Chronic bronchitis (adults, 26 and over, WTP valuation)
Hospital Admissions — Pnuemonia (adults, over 64)
Hospital Admissions — COPD (adults, 64 and over)
Hospital Admissions — Asthma (65 and younger)
Hospital Admissions — Cardiovascular (adults, over 64)
Emergency Room Visits for Asthma (65 and younger)
Asthma Attacks (asthmatics, all ages)
Acute Bronchitis (children, 8-12)
Lower Respiratory Symptoms (children, 7-12)
Upper Respiratory Symptoms (asthmatic children, 9-11)
Work loss days (adults, 18-65)
Minor restricted activity days (Adults, 18-65)
Other NOx, PM, and HAP-related health effectsf
Total PM Health-Related Benefits
-Using a 3% discount rate6
-Using a 7% discount rate6
90
90
60
10
10
10
30
20
1,750
130
2,180
2,150
15,010
79,728
u,
—
—
$535
$505
$20
$<1
<$1
<$1
<$5
<$5
B!
<$1
<$1
<$1
<$5
$5
B2
$560 + BH
$530 + BH
The results presented in this table are based on a 50% reduction of all NOx emissions from RICE sources nationwide based
on a 1996 emissions inventory (370,877 tons) evaluated with a 2005 population.
Incidences are rounded to the nearest 10 and may not add due to rounding. Incidences of unquantified endpoints are
indicated with a U.
Dollar values are rounded to the nearest 5 million and may not add due to rounding.
Note that the estimated value for PM-related premature mortality in the Base Estimate assumes the 5 year distributed lag
structure described in detail in the Regulatory Impact Analysis of Heavy Duty Engine/Diesel Fuel rule.
Results of premature mortality benefits reflect the use of two different discount rates; a 3% rate which is recommended by
EPA's Guidelines for Preparing Economic Analyses (EPA, 2000b), and 7% which is recommended by OMB Circular A-94
(OMB, 1992).
For notational purposes, unquantified benefits are indicated with a "U" to represent avoided incidences and a "B" to
represent monetary benefits. A detailed listing of unquantified NOx, PM, and HAP related health effects is provided in
Table 8-13.
8-41
-------
Table 8-6(b). Alternative Estimate of Annual Health Benefits
Resulting from 50 Percent RICE NOx Emission Reduction Scenario"
Endpoint
Premature mortality 4e (short-term exposure):
-Using a 3% discount rate
-Using a 7% discount rate
Chronic bronchitis (adults, 26 and over, COI valuation)
Hospital Admissions — Pnuemonia (adults, over 64)
Hospital Admissions — COPD (adults, 64 and over)
Hospital Admissions — Asthma (65 and younger)
Hospital Admissions — Cardiovascular (adults, over 64)
Emergency Room Visits for Asthma (65 and younger)
Asthma Attacks (asthmatics, all ages)
Acute Bronchitis (children, 8-12)
Lower Respiratory Symptoms (children, 7-12)
Upper Respiratory Symptoms (asthmatic children, 9-11)
Work loss days (adults, 18-65)
Minor restricted activity days (Adults, 18-65)
Other NOx, PM, and HAP-related health effectsf
Total PM Health-Related Benefits'
-Using a 3% discount rate
-Using a 7% discount rate
Avoided
Incidence6
(cases/year)
50
50
60
10
10
10
30
20
1,750
130
2,180
2,150
15,010
79,730
u,
—
—
Monetary Benefits,
Adjusted for
Growth in Income0
(millions 1998$)
$55
$65
$5
$<1
<$1
<$1
<$5
<$5
B!
<$1
<$1
<$1
<$5
$5
B2
$70 + BH
$80 + BH
The results presented in this table are based on a 50% reduction of all NOx emissions from RICE sources nationwide based
on a 1996 emissions inventory (370,877 tons) evaluated with a 2005 population.
Incidences are rounded to the nearest 10 and may not add due to rounding. Incidences of unquantified endpoints are
indicated with a U.
Dollar values are rounded to the nearest 5 million and may not add due to rounding.
Note that the estimated value for PM-related premature mortality in the Base Estimate assumes the 5 year distributed lag
structure described in detail in the Regulatory Impact Analysis of Heavy Duty Engine/Diesel Fuel rule.
Results of premature mortality benefits reflect the use of two different discount rates; a 3% rate which is recommended by
EPA's Guidelines for Preparing Economic Analyses (EPA, 2000b), and 7% which is recommended by OMB Circular A-94
(OMB, 1992).
For notational purposes, unquantified benefits are indicated with a "U" to represent avoided incidneces and "B" to represent
monetary benefits. A detailed listing of unquantified NOx, PM, and HAP related health effects is provided in Table 8-13.
8-42
-------
Table 8-7. Benefit Value Per Ton of NOx—
Based on a 50% NOx Reduction at RICE Units3
Benefit Per Ton of NOx Reduced
Base Estimate-
Using 3% discount rate $1,510
Using 7% discount rate $1,430
Alternative Estimate-
Using 3% discount rate $188
Using 7% discount rate $215
Results reflect the use of two different discount rates; a 3% rate which is recommended by EPA's Guidelines for Preparing
Economic Analyses (EPA, 2000b), and 7% which is recommended by OMB Circular A-94 (OMB, 1992).
8.4.3 PM10 Benefit Transfer Values for Application to PM10 Emissions Reductions
The RICE NESHAP is expected to reduce direct emissions of PM10. Unlike the
secondary formation of PM2 5 that results from NOx reductions, direct PM10 emissions consist of
all particles whose size are PM10 or smaller. In the prior section, PM2 5 transfer values were
developed to estimate benefits from reduced secondary formation of PM from NOx emissions.
In this section, PM10 transfer functions are developed to value benefits of direct PM emission
reductions, due to a lack of information on the fraction of PM10 from RICE that is PM25.
Directly emitted PM10 benefit transfer values are developed using the same basic
approach as for PM2 5. However, the specific air quality scenario and health endpoints differ.
The only difference in the transfer values for PM2 5 and PM10 is the choice of mortality endpoint
and the exclusion of health effects whose C-R functions are based on PM25. While PM25 is a
component of PM10, it is considered to potentially have a much larger impact on mortality due to
long-term exposures. Given our inability to fractionate total PM10 into fine and coarse particles,
we use the C-R function relating PM10 to premature mortality in developing the direct PM10
benefit transfer value in this section to avoid overstating potential impacts of reductions in total
PM10. Note again that not all known health and welfare effects associated with PM are
8-43
-------
quantified and monetized for this analysis. Potential benefit categories that have not been
quantified and monetized are listed in 8-10 later in this chapter.
The first step of the benefit transfer approach for PM10 is to generate an air quality
scenario reflecting the types of direct PM emissions reductions expected from the RICE
NESHAP rule. We selected a scenario which modeled a 100 percent reduction in PM emissions
from all RICE sources in the continental U.S. These emission reductions were then analyzed
using the S-R matrix described above. While a 100 percent reduction in PM emissions at RICE
sources does not reflect an approximation of the NESHAPs PM reductions, the 100 percent
reduction scenario is necessary to observe results in the national scale air quality model.
Because PM air quality impacts are linear in form, however, the results can be scaled to the
NESHAPs level of control and is considered a representative benefit transfer value.
Following the same steps as used in generating the PM2 5 transfer value for NOx
reductions, the results of the benefit transfer development are presented in Table 8-8(a) and (b).
Total reductions in direct PM emissions for the 100 percent RICE direct PM reduction scenario
are 95,178 tons. Dividing total benefits by the PM emission reductions yields a $/ton estimate of
$6,619 using the Base Estimate with a 3 percent discount rate, and $6,303 per ton with a 7
percent discount rate. Using the Alternative Estimate, the $/ton is $1,628 with a 3 percent
discount rate and $1,681 with a 7 percent discount rate.
8-44
-------
Table 8-8(a). Base Estimate: Annual Health Benefits
Resulting from 100 Percent RICE Direct PM Emission Reduction Scenario"
Endpoint
Premature mortality (short term exposure)6
-Using a 3% discount rate
-Using a 7% discount rate
Chronic bronchitis (adults, 26 and over, WTP valuation)
Hospital Admissions — Pnuemonia (adults, over 64)
Hospital Admissions — COPD (adults, 64 and over)
Hospital Admissions — Cardiovascular (adults, over 64)
Emergency Room Visits for Asthma (65 and younger)
Asthma Attacks (asthmatics, all ages)
Lower Respiratory Symptoms (children, 7-12)
Upper Respiratory Symptoms (asthmatic children, 9-11)
Other NOX, PM, and HAP-related health effects'1
Total PM Health-Related Benefits6
-Using a 3% discount rate
-Using a 7% discount rate
Avoided
Incidence6
(cases/year)
75
75
440
100
80
240
200
15,620
9,120
16,730
u,
—
—
Monetary Benefits,
Adjusted for
Growth in Income0
(millions 1998$)
$465
$440
$155
<$5
<$5
$5
<$1
B!
<$1
<$1
B2
$630 + BH
$600 + BH
The results presented in this table are based on a 100% reduction of all direct PM emissions from RICE sources nationwide
based on a 1996 emissions inventory (95,178 tons) evaluated with a 2005 population.
Incidences are rounded to the nearest 10 and may not add due to rounding.
Dollar values are rounded to the nearest 5 million and may not add due to rounding.
For notational purposes, unquantified benefits are indicated with a "U" to represent avoided incidneces and "B" to represent
monetary benefits. A detailed listing of unquantified NOx, PM, and HAP related health effects is provided in Table 8-13.
Results of premature mortality benefits reflect the use of two different discount rates; a 3% rate which is recommended by
EPA's Guidelines for Preparing Economic Analyses (EPA, 2000b), and 7% which is recommended by OMB Circular A-94
(OMB, 1992).
8-45
-------
Table 8-8(b). Alternative Estimate: Annual Health Benefits
Resulting from 100 Percent RICE Direct PM Emission Reduction Scenario"
Endpoint
Premature mortality (short term exposure)6
-Using a 3% discount rate
-Using a 7% discount rate
Chronic bronchitis (adults, 26 and over)
Hospital Admissions — Pnuemonia (adults, over 64)
Hospital Admissions — COPD (adults, 64 and over)
Hospital Admissions — Cardiovascular (adults, over 64)
Emergency Room Visits for Asthma (65 and younger)
Asthma Attacks (asthmatics, all ages)
Lower Respiratory Symptoms (children, 7-12)
Upper Respiratory Symptoms (asthmatic children, 9-11)
Other PM-related health effects'1
Total PM Health-Related Benefits6
-Using a 3% discount rate
-Using a 7% discount rate
Avoided
Incidence6
(cases/year)
70
70
440
100
80
240
200
15,620
9,120
16,730
u,
—
—
Monetary Benefits,
Adjusted for
Growth in Income0
(millions 1998$)
$95
$100
$50
<$5
<$5
$5
<$1
B!
<$1
<$1
B2
$155 +BH
$160 +BH
The results presented in this table are based on a 100% reduction of all direct PM emissions from RICE sources nationwide
based on a 1996 emissions inventory (95,178 tons) evaluated with a 2005 population.
Incidences are rounded to the nearest 10 and may not add due to rounding.
Dollar values are rounded to the nearest 5 million and may not add due to rounding.
For notational purposes, unquantified benefits are indicated with a "U" to represent avoided incidneces and "B" to represent
monetary benefits. A detailed listing of unquantified NOx, PM , and HAP related health effects is provided in Table 8-13.
Results of premature mortality benefits reflect the use of two different discount rates; a 3% rate which is recommended by
EPA's Guidelines for Preparing Economic Analyses (EPA, 2000b), and 7% which is recommended by OMB Circular A-94
(OMB, 1992).
8-46
-------
Table 8-9. Benefit Value Per Ton of PM10—
Based on a 100% Reduction of Direct PM10 at RICE Units
Benefit Per Ton of PM10
Reduced
Base Estimate
-Using a 3% discount rate $6,619
-Using a 7% discount rate $6,603
Alternative Estimate
-Using a 3% discount rate $ 1,628
-Using a 7% discount rate $1,681
8.5 APPLICATION OF BENEFIT TRANSFER VALUES TO THE RICE NESHAP RULE
Using the ozone and PM benefit transfer values calculated above, we can develop an
estimate of potential benefits associated with reductions in direct PM and NOx emissions at
RICE sources. NOx emission reductions from the RICE NESHAP regulation are expected to be
167,900 tons per year at major sources once the regulation is fully implemented in 2005. Since
no information is available about the distribution of these emission reductions across the year,
we assume that emission reductions are equally distributed over all months. Thus, ozone season
emissions (from May to September) will be approximately equal to 5/12 of annual emissions, or
70,000 tons. Because the NOx SIP call only estimated benefits for the reductions in NOx
emissions in the easternmost 37 states, we must also apportion the emission reductions from the
RICE NESHAP into eastern and western regions. Based on the 1996 NET emissions inventory,
approximately 74 percent of NOx emissions from RICE facilities occurred in the eastern 37
states. Thus, we multiply NOx emission reductions by 0.74 to arrive at the 51,800 NOx tons to
which the ozone benefit transfer value will be applied. For PM benefits, since we use a national
model, total national emission reductions for the full year will be applied to the PM benefit
transfer values (i.e., 167,900 tons NOx and 3,700 tons direct PM).
Using the equation for total benefits, the estimated monetary benefits of the NOx and PM
reductions from the RICE NESHAP for the Base and Alternative Estimates are presented in
Table 8-10.
8-47
-------
Table 8-10. Benefits of the RICE NESHAP
Benefit Transfer Values
(1998$)
Reductions in Emissions
(tons)
NOx
Direct
PM
Ozone- Annual
season Annual Direct
NOxa NOx PM Ozone PM25
Total
Monetized
Benefits"
PM10 (million 1998$)
Base Estimate0
-Using 3% discount rate
-Using 7% discount rate
Alternative Estimate0
-Using 3% discount rate
-Using 7% discount rate
51,800 167,900 3,700 $28 $1,510 $6,619 $280+B
$28 $1,430 $6,603 $265+B
51,800 167,900 3,700
$28 $188 $1,628 $40+ B
$28 $215 $1,681 $45+B
Emission reductions for ozone are for the Eastern United States, and are assumed to equal 5/12 of annual NOx reductions
representing 5 months of the year associated with the ozone season.
For notational purposes, unquantified benefits are indicated with a "B" to represent monetary benefits. A detailed listing of
unquantified NOx, PM, and HAP related health effects is provided in Table 8-13.
Results reflect the use of two different discount rates; a 3% rate which is recommended by EPA's Guidelines for Preparing
Economic Analyses (EPA, 2000b), and 7% which is recommended by OMB Circular A-94 (OMB, 1992).
8.6 LIMITATIONS OF THE ANALYSIS
8.6.1 Uncertainties and Assumptions
Significant uncertainties and potential biases are inherent in any benefits analysis based
on benefit transfer techniques. The degree of uncertainty and bias depends on how divergent the
reality of the policy situation is from the state of the world assumed in the benefit transfer.
For this analysis, several key assumptions may lead to over or underestimation of
benefits. Tables 8-11 and 8-12 list these assumptions, and where possible indicate the expected
direction of the bias. This is by no means an exhaustive list, but captures what we have
identified as key assumptions. In addition to these uncertainties and biases, there are
uncertainties and biases embedded in the original benefits analyses from which the transfer
values were generated. Some of these potential biases and assumptions are discussed in the
preceding sections. For a full discussion of these uncertainties, see the NOx SIP Call RIA and
8-48
-------
the HDD RIA, as well as the Section 812 report to congress on the Benefits and Costs of the
Clean Air Act 1990 to 2010.
Table 8-11. Significant Uncertainties and Biases
in Derivation of the Benefit Transfer Values
Assumption
Direction of Bias
Impact of NOx reductions on PM formation is Unknown
equivalent across all RICE sources
Impact of NOx reductions on ozone formation is Unknown
equivalent across all RICE sources
Population distributions of PM and ozone Unknown
reductions in source analyses are similar to
population distributions of PM and ozone
reductions resulting from the RICE NESHAP
Benefits from source studies do not include all Unknown
benefits and disbenefits
8.6.2 Unqualified Effects
In addition to the monetized benefits presented in the above tables, it is important to
recognize that many benefit categories associated with NOx and PM10 reductions are not
quantified or monetized for this analysis. In addition to agricultural and forestry benefits, other
potentially important unquantified benefit categories are listed in Table 8-13. For a more
complete discussion of unquantified benefits and disbenefits, see the HDD RIA and the NOx SIP
Call RIA.
8-49
-------
Table 8-12. Significant Uncertainties and Biases
in Application of Benefit Transfer Values to RICE NOx and PM Reductions
Assumption
Direction of Bias
Omission of commercial agriculture, forestry, Downward
visibility, and materials damage benefit
categories
Same transfer value applied to all populations Unknown
exposed to NOx and PM emissions from
NESHAP sources
Linear relationship between emission Upward
reductions and benefits
Meteorology in 2005 well-represented by Unknown
modeled meteorology
PM10 reductions are not quantified or monetized for this analysis. In addition to agricultural and
forestry benefits, other potentially important unquantified benefit categories are listed in Table 8-
13. For a more complete discussion of unquantified benefits and disbenefits, see the HDD RIA
and the NOx SIP Call RIA.
8-50
-------
Table 8-13. Unquantified Benefit Categories
Unquantified Benefit Categories
Associated with Ozone
Unquantified Benefit Categories
Associated with PM
Health Airway responsiveness
Categories Pulmonary inflammation
Increased susceptibility to respiratory
infection
Acute inflammation and respiratory cell
damage
Chronic respiratory damage/Premature
aging of lungs
Emergency room visits for asthma
Respiratory hospital admissions for
children
Chronic asthma
Premature mortality (independent of PM
related mortality)
Increased school absence rates
Welfare Ecosystem and vegetation effects in Class
Categories I areas (e.g., national parks)
Damage to urban ornamentals (e.g.,grass,
flowers, shrubs, and trees in urban
areas)
Commercial field crops
Fruit and vegetable crops
Reduced yields of tree seedlings,
commercial and non-commercial
forests
Damage to ecosystems
Materials damage
Changes in pulmonary function
Morphological changes
Altered host defense mechanisms
Cancer
Other chronic respiratory disease
Emergency room visits for asthma
Emergency room visits for non-asthma
respiratory and cardiovascular causes
Lower and upper respiratory symptoms
Acute bronchitis
Shortness of breath
Increased school absence rates
Myocardial infarction (heart attacks)
Materials damage
Damage to ecosystems (e.g., acid
sulfate deposition)
Nitrates in drinking water
Visibility in recreational and residential
areas
8-51
-------
8.7 BENEFIT-COST COMPARISON
Benefit-cost analysis provides a valuable framework for organizing and evaluating
information on the effects of environmental programs. When used properly, benefit-cost
analysis helps illuminate important potential effects of alternative policies and helps set priorities
for closing information gaps and reducing uncertainty. According to economic theory, the
efficient policy alternative maximizes net benefits to society (i.e., social benefits minus social
costs). However, not all relevant costs and benefits can be captured in any analysis. Executive
Order 12866 clearly indicates that unquantifiable or nonmonetizable categories of both costs and
benefits should not be ignored. There are many important unquantified and unmonetized costs
and benefits associated with reductions in PM and NOx emissions, including many health and
welfare effects. Potential PM and NOx benefit categories that have not been quantified and
monetized are listed in Table 8-13 of this chapter. It is also important to recall that this analysis
is only of the monetizable benefits associated with NOx and direct PM reductions. The rule is
designed to reduce HAP emissions to a level mandated by the Clean Air Act - the MACT floor.
It also achieves significant CO reductions. By achieving these emission reductions, the rule
reduces the risks associated with exposures to those pollutants, including the toxic effects and
risk of fatal cancers associated with HAPs, and the effects on the central nervous system and
cardiovascular system associated with CO. The monetized benefit estimates presented in this
chapter are thus expected to underestimate total benefits of the rule.
In addition to categories that cannot be included in the calculated net benefits, there are
also practical limitations for the comparison of benefits to costs in this analysis, which have been
discussed throughout this chapter. Several specific limitations deserve to be mentioned again
here:
The state of atmospheric modeling is not sufficiently advanced to provide a
workable "one atmosphere" model capable of characterizing ground-level
pollutant exposure for all pollutants of interest (e.g., ozone, particulate matter,
carbon monoxide, nitrogen deposition, etc). Therefore, the EPA must employ
several different pollutant models to characterize the effects of alternative policies
on relevant pollutants. Also, not all atmospheric models have been widely
8-52
-------
validated against actual ambient data. Additionally, significant shortcomings
exist in the data that are available to perform these analyses. While containing
identifiable shortcomings and uncertainties, EPA believes the models and
assumptions used in the analysis are reasonable based on the available evidence
and resources.
Qualitative and more detailed discussions of the above and other uncertainties and
limitations are included in detail in earlier sections. Data limitations prevent an
overall quantitative estimate of the uncertainty associated with final estimates.
Nevertheless, the reader should keep all of these uncertainties and limitations in
mind when reviewing and interpreting the results.
The Base benefit estimate does not include the monetary value of several known
ozone and PM-related welfare effects, including commercial forest growth,
recreational and residential visibility, household soiling and materials damage,
and deposition of nitrogen to sensitive estuaries.
The benefit estimates presented in this document do not capture any additional
short-term mortality impacts related to changes in exposure to ambient ozone. A
recent analysis by Thurston and Ito (2001) reviewed previously published time
series studies of the effect of daily ozone levels on daily mortality and found that
previous EPA estimates of the short-term mortality benefits of the ozone NAAQS
(EPA, 1997b) may have been underestimated by up to a factor of two. The
authors hypothesized that much of the variability in published estimates of the
ozone/mortality effect could be explained by how well each model controlled for
the influence of weather, an important confounder of the ozone/mortality effect,
and that earlier studies using less sophisticated approaches to controlling for
weather consistently under-predicted the ozone/mortality effect. They found that
models incorporating a non-linear temperature specification appropriate for the
"U-shaped" nature of the temperature/mortality relationship (i.e., increased deaths
8-53
-------
at both very low and very high temperatures) produced ozone/mortality effect
estimates that were both more strongly positive (a two percent increase in relative
risk over the pooled estimate for all studies evaluated) and consistently
statistically significant. Further accounting for the interaction effects between
temperature and relative humidity produced even more strongly positive results.
Inclusion of a PM index to control for PM/mortality effects had little effect on
these results, suggesting an ozone/mortality relationship independent of that for
PM. However, most of the studies examined by Ito and Thurston only controlled
for PM10 or broader measures of particles and did not directly control for PM2 5.
As such, there may still be potential for confounding of PM25 and ozone mortality
effects, as ozone and PM2 5 are highly correlated during summer months in some
areas13. In its September 2001 advisory on the draft analytical blueprint for the
second Section 812 prospective analysis, the SAB cited the Thurston and Ito
study as a significant advance in understanding the effects of ozone on daily
mortality and recommended re-evaluation of the ozone mortality endpoint for
inclusion in the next prospective study (EPA-SAB-COUNCIL-ADV-01-004,
2001). Thus, recent evidence suggests that by not including an estimate of
reductions in short-term mortality due to changes in ambient ozone, both the Base
and Alternative Estimates may underestimate the benefits of implementation of
the RICE NESHAP.
Nonetheless, if one is mindful of these limitations, the relative magnitude of the benefit-
cost comparison presented here can be useful information. Thus, this section summarizes the
benefit and cost estimates that are potentially useful for evaluating the efficiency of the RICE
NESHAP rulemaking.
The estimated social cost of implementing the RICE program is approximately $255
million (1998$) in the fifth year, while the estimate of NOx and PM-related monetized benefits
are $280 + B million (3 percent discount rate), or 265 + B million (7 percent discount rate) under
13 Short-term ozone mortality risk estimates may also be affected by the statistical issue discovered by the Health
Effects Institute (Greenbaum, 2002a). See page 24 for a more detailed discussion of this issue.
8-54
-------
the Base Estimate. Under the Alternative Estimate, total benefits are $40 + B million (3 percent
discount rate), or $45 + B million (7 percent discount rate). Comparison with costs indicates that
the monetized benefits of NOx and PM reductions exceed costs by approximately $25 + B
million (3 percent discount rate), or $15 million + B (7 percent discount rate) under the Base
Estimate. Under the Alternative Estimate, net benefits are -$215 + B million (3 percent discount
rate), or -$210 + B million (7 percent discount rate). Note that while monetized benefits of PM
and NOx reductions exceed monetized costs only under our Base Estimate, PM and NOx
benefits account for only a portion of the benefits of this rule. Again, with the omission of a
quantified value for any of the benefits of HAPs and CO reductions, total net benefits of the rule
are understated.
With respect to the benefits of reducing exposure to HAPs, EPA has developed a
rudimentary risk analysis focusing only on cancer risks. As discussed above, this analysis
suggests that the proposed rule would reduce cancer incidence by roughly 10 cases per year if it
were implemented at all affected RICE facilities. Placing a value on these impacts would
increase the economic benefits of the rule. This analysis carries significant assumptions,
uncertainties, and limitations. EPA is working with the SAB to develop better methods for
analyzing the cancer and non-cancer benefits of HAP reductions. EPA will include a monetized
estimate of the benefits of reducing HAP emissions with the analysis for the final rule
if it is able to develop better methods before promulgation of this rule.
It is also important to note that not only are entire pollutant categories missing from our
benefit estimate, but also not all benefits of PM and NOx reductions have been monetized.
Categories which have contributed significantly to monetized benefits in past analyses (see the
NOx SIP call and HDD RIAs) include increased productivity of commercial agriculture and
forestry, improved recreational and residential visibility, and reductions in deposition to nitrogen
sensitive estuaries. Table 8-13 lists known unquantified benefits associated with PM and NOx
reductions. Thus, this information should be used in conjunction with information provided in
all other chapters of this report to understand the overall impacts of the rule on society. Table
8-14 and 8-15 summarizes the costs, benefits, and net benefits for the MACT Floor regulatory
option.
8-55
-------
Additionally, we did not attempt to estimate welfare benefits associated with ozone and
PM reductions for this rule because of the difficulty in developing reliable benefit transfer values
for these effects. The SAB has recently reviewed existing studies valuing improvements in
residential visibility and reductions in household soiling and advised that these studies do not
provide an adequate basis for valuing these effects in cost-benefit analyses (EPA-SAB-
COUNCIL-ADV-00-002, 1999; EPA-SAB-COlMCIL-ADV-003, 1999a). Reliable methods do
exist for valuing visibility improvements in Federal Class I areas, however, the benefit transfer
method outlined above does not allow for predictions of changes in visibility at specific Class I
areas. These predictions are necessary to estimate Class I area visibility benefits. As such we
have left this potentially important endpoint unquantified for this analysis. Given the proximity
of some RICE sources to national parks in the west and northwest, these omitted benefits may be
significant.
8-56
-------
Table 8-14. Summary of Costs, Emission Reductions, and Quantifiable Benefits
by Engine Type
Type of
Engine
2SLB-New
4SLB-New
4SRB-
Existing
4SRB-New
CI-New
Total
Total
Annualized
Cost (million
$/yr in 2005)
$3
$66
$38
$48
$99
$255
Emission Reductions3
(tons/yr in 2005)
Quantifiable Annual Monetized
Benefits" c (million $/yr in 2005)
Alternative
HAP
250
4,035
230
215
305
5,035
CO
2,025
36,240
98,040
91,820
6,320
234,445
NOx
0
0
69,900
98,000
0
167,900
PM
0
0
0
0
3,700
3,700
Base Estimate
B!
B3
$105 +B5
$100 + B6
$150 + B9
$140 + B10
$25+B13
$280 + B
$265 + B
B2
B4
$15
$15
$20
$25
Estimate
+ B7
+ B8
+ Bn
+ B12
$5+B14
$40
$45
+ B
+ B
For the calculation of PM-related benefits, total NOx reductions are multiplied by the appropriate benefit per ton value
presented in Table 8-7. For the calculation of ozone-related benefits, NOx reductions are multiplied by 5/12 to account for
ozone season months and 0.74 to account for Eastern States in the ozone analysis. The resulting ozone-related NOx
reductions are multiplied by $28 per ton. Ozone-related benefits are summed together with PM-related benefits to derive
total benefits of NOx reductions. All benefits values are rounded to the nearest $5 million.
Benefits of HAP and CO emission reductions are not quantified in this analysis and, therefore, are not presented in this table.
The quantifiable benefits are from emission reductions of NOx and PM only. For notational purposes, unqualified benefits
are indicated with a "B" to represent monetary benefits. A detailed listing of unquantified NOx, PM, and HAP related health
effects is provided in Table 8-13.
Results reflect the use of two different discount rates; a 3% rate which is recommended by EPA's Guidelines for Preparing
Economic Analyses (EPA, 2000b), and 7% which is recommended by OMB Circular A-94 (OMB, 1992).
8-57
-------
Table 8-15. Annual Net Benefits of the RICE NESHAP in 2005
Million 1998$a
Social Costs" $255
Social Benefits15'c'd:
HAP-related benefits Not monetized
CO-related benefits Not monetized
Ozone- and PM-related Welfare benefits Not monetized
Ozone- and PM-related Health benefits:
Base Estimate $28Q + R
-Using 3% Discount Rate ,.,.,, R
-Using 7% Discount Rate
Alternative Estimate
-Using 3% Discount Rate $40 + B
-Using 7% Discount Rate $45 + B
Net Benefits (Benefits - Costs)0'd:
Base Estimate ^, + R
-Using 3% Discount Rate
-------
REFERENCES
Abbey, D.E., F. Peterson, P.K Mills, and W.L. Beeson. 1993. "Long-term Ambient
Concentrations of Total Suspended Particulates, Ozone and Sulfur Dioxide and
Respiratory Symptoms in a Nonsmoking Population." Archives of Environmental Health.
48(1): 33-46.
Abt Associates Inc. September 1998. Selected Health and Welfare Benefits Methods for the
NOx SIP CallRIA. Report prepared for U.S. EPA, Office of Air Quality Planning and
Standards, under contract no. 68-D-98-001. Research Triangle Park, NC: Abt
Associates Inc.
Abt Associates Inc. 1999. The Final Tier 2/Gasoline Sulfur Rule: Air Quality Estimation,
Selected Health and Welfare Benefits Methods, and Benefit Analysis Results. Report
prepared for U.S. EPA, Office of Air Quality Planning and Standards, under contract no.
68-D-98-001. Research Triangle Park, NC: Abt Associates Inc.
Agency for Toxic Substances and Disease Registry (ATSDR). 2002. Toxicological Profile
Sheet. .
Alpha Gamma Technologies, Inc. 200la. Memorandum to Sims Roy, U.S. Environmental
Protection Agency, "Reciprocating Internal Combustion Engine Distribution at Major
and Area Sources;" November 27, 2001.
Alpha Gamma Technologies, Inc. 2001b. Memorandum to Sims Roy, U.S. Environmental
Protection Agency, "Control Costs for Reciprocating Internal Combustion Engines;"
December 11,2001.
Alpha Gamma Technologies, Inc. 2002a. Memorandum to Sims Roy, U.S. Environmental
Protection Agency, "National Impacts Associated with Reciprocating Internal
Combustion Engines;" January 10, 2002.
R-l
-------
Alpha Gamma Technologies, Inc. 2002b. Memorandum to Sims Roy, U.S. Environmental
Protection Agency, "Estimate of Reciprocating Internal Combustion Engines with
Oxidation Catalysts orNSCR;" June 19, 2002.
Berger, Bill D., and Kenneth E. Anderson. 1978. Modern Petroleum: A Basic Primer of the
Industry. Tulsa, OK: The Petroleum Publishing Co.
Cote, I, L. Cupitt and B. Hassett. 1988. "Toxic Air Pollutants and Non-Cancer Health Risks."
Unpublished paper provided by B. Hassett.
Conces, Elaine. 2000. "Hoover's Online—Energy Industry Snapshot." As obtained on April
18, 2000. .
Crocker, T.D. and R.L. Horst, Jr. 1981. "Hours of Work, Labor Productivity, and
Environmental Conditions: A Case Study." The Review of Economics and Statistics. 63:
361-368.
Dahl, Carol, and Thomas E. Duggan. 1996. "U.S. Energy Product Supply Elasticities: A Survey
and Application to the U.S. Oil Market." Resource and Energy Economics. 18: 243-263.
Dun & Bradstreet. 1997. Industry Norms & Key Business Ratios. Desktop Edition 1996-97.
Murray Hill, NJ: Dun & Bradstreet, Inc.
E.H. Pechan and Associates, Inc. 1997. Integrated Ozone Paniculate Matter and Regional
Haze Cost Analysis: Methodology and Results. Report prepared for the U. S.
Environmental Protection Agency, Office of Air Quality Planning and Standards.
Research Triangle Park, NC: E.H. Pechan and Associates, Inc.
Gas Research Institute (GRI). 2000. Engine Design, Operation, and Control in the Natural Gas
Industry, . As
obtained on May 9, 2000.
Graham, John D., D.R. Holtgrave, and M.J. Sawery. 1989. "The Potential Health Benefits of
Controlling Hazardous Air Pollutants." In: Health Benefits of Air Pollution Control: A
Discussion. Blodgett, J. (ed). Congressional Research Service report to Congress.
CR589-161. Washington, DC.
Heil, Scott F., Ed. 1998. Ward's Business Directory of U.S. Private and Public Companies
1998, Volume 5. Detroit, MI: Gale Research Inc.
Hoover's Incorporated. 2000. Hoover's Company Profiles. Austin, TX: Hoover's
Incorporated. .
R-2
-------
Industrial Combustion Coordinated Rulemaking (ICCR). 1998. Data/Information Submitted to
the Coordinating Committee at the Final Meeting of the Industrial Combustion
Coordinated Rulemaking Federal Advisory Committee. EPA Docket Numbers A-94-63,
II-K-4b2 through -4b5. Research Triangle Park, North Carolina. September 16-17.
Johns Hopkins University School of Public Health. 2002. Updated City-Specific Estimates of
PM10 Relative Risks, .
As obtained in October 2002.
Krupnick, A.J., W. Harrington and B. Ostro. 1990. "Ambient Ozone and Acute Health Effects -
Evidence From Daily Data." Journal of Environmental Economics and Management.
18(1): 1-18.
Lillis, Kevin. 1998. "Crude Petroleum and Natural Gas." The U.S. Industry and Trade Outlook
'98. New York: DRI/McGraw-Hill, Standard & Poor's, and U.S. Department of
Commerce/International Trade Administration.
Mathtech, Inc. May 1992. Benefit Analysis Issues for Section 112 Regulations. Final report
prepared for U.S. Environmental Protection Agency. Office of Air Quality Planning and
Standards. Contract No. 68-D8-0094. Research Triangle Park, NC: Mathtech, Inc.
McKeever, M. 1997. Memorandum to Lisa Conner, Innovative Strategies and Economics
Group, Office of Air and Radiation, U.S. Environmental Protection Agency, November,
1997.
National Academy of Sciences (NAS). 1975. Chlorine and Hydrogen Chloride. NAS/National
Research Council. Chapter 7.
Natural Gas Supply Association (NGS A), Independent Petroleum Association of America,
National Ocean Industries Association, and International Centre for Gas Technology
Information. 2000a. "Transportation of Natural Gas."
. As obtained on March 22, 2000.
Natural Gas Supply Association (NGS A), Independent Petroleum Association of America,
National Ocean Industries Association, and International Centre for Gas Technology
Information. 2000b. "Storage of Natural Gas."
. As obtained on March 22, 2000.
Natural Gas Supply Association (NGS A), Independent Petroleum Association of America,
National Ocean Industries Association, and International Centre for Gas Technology
Information. 2000c. "Natural Gas Production."
. As obtained on March 22, 2000.
-------
Nicholson, Walter. 1998. Microeconomic Theory: Basic Principles and Extensions. Fort
Worth, TX: Dryden Press.
Ostro, B.D. 1987. "Air Pollution and Morbidity Revisited: A Specification Test." Journal of
Environmental Economics and Management. 14: 87-98.
Ostro, B.D. and S. Rothschild. 1989. "Air Pollution and Acute Respiratory Morbidity - an
Observational Study of Multiple Pollutants." Environmental Research. 50(2): 238-247.
Pope, C.A., III, MJ. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer, and
C.W. Heath, Jr. 1995. "Particulate Air Pollution as a Predictor of Mortality in a
Prospective Study of U.S. Adults." Am. J. Respir. Crit. Care Med. 151: 669-674.
Schwartz, J. 1993. "Particulate Air Pollution and Chronic Respiratory Disease." Environmental
Research. 62: 7-13.
Schwartz, J. and R. Morris. 1995. "Air Pollution and Hospital Admissions For Cardiovascular
Disease in Detroit, Michigan." American Journal of Epidemiology. 142(1): 23-35.
Stern, A., H. Wohlers, R. Boubel, and W. Lowry. 1973. Fundamentals of Air Pollution.
Academic Press, New York.
Sorrels, Larry, EPA. August 15, 2000. Personal communication with Mike Gallaher, RTI.
Qualitative Market Assessment—PM NAAQS.
Thurston, G.D., K. Ito, P.L. Kinney and M. Lippmann. 1992. "A Multi-Year Study of Air
Pollution and Respiratory Hospital Admissions in 3 New-York State Metropolitan Areas
—Results For 1988 and 1989 Summers." Journal of Exposure Analysis and
Environmental Epidemiology. 2(4): 429-450.
Thurston, G.D., K. Ito, C.G. Hayes, D.V. Bates and M. Lippmann. 1994. "Respiratory Hospital
Admissions and Summertime Haze Air Pollution in Toronto, Ontario—Consideration of
the Role of Acid Aerosols." Environmental Research. 65(2): 271-290.
Tussing, Arlon R., and Bob Tippee. 1995. The Natural Gas Industry: Evolution, Structure, and
Economics. 2nd Edition. Tulsa, OK: PennWell Publishing Company.
U.S. Department of Commerce, Bureau of the Census. 1995a. 1992 Census of Mineral
Industries, Industry Series: Crude Petroleum and Natural Gas. MIC92-I-13A.
Washington, DC: U.S. Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 1995b. 1992 Census of Mineral
Industries, Industry Series: Natural Gas Liquids. MIC92-I-13B. Washington, DC: U.S.
Department of Commerce.
R-4
-------
U.S. Department of Commerce, Bureau of the Census. 1995c. 1992 Census of Mineral
Industries, Industry Series: Oil and Gas Field Services. MIC92-I-13C. Washington,
DC: U.S. Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 1999a. 1997 Economic Census, Mining,
Industry Series: Crude Petroleum and Natural Gas Extraction. EC97N-2111 A.
Washington, DC: U.S. Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 1999b. 1997 Economic Census, Mining,
Industry Series: Natural Gas Liquid Extraction. EC97N-211 Ib. Washington, DC: U.S.
Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 1999c. 1997 Economic Census, Mining,
Industry Series: Drilling Oil and Gas Wells. EC97N-2131A. Washington, DC: U.S.
Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 1999d. 1997 Economic Census, Mining,
Industry Series: Support Activities for Oil and Gas Operations. EC97N-213 IB.
Washington, DC: U.S. Department of Commerce.
U.S. Department of Commerce, Bureau of the Census. 2000. 1997 Economic Census,
Transportation and Warehousing: Geographic Area Series. EC97T48A-US.
Washington, DC: Government Printing Office.
U.S. Department of Energy, Energy Information Administration (EIA). January 1998. Model
Documentation Report: Industrial Sector Demand Module of the National Energy
Modeling System. DOE/EIA-M064(98). Washington, DC: U.S. Department of Energy.
U.S. Department of Energy, Energy Information Administration (EIA). 1999a. Natural Gas
1998: Issues and Trends. Washington, DC: U.S. Department of Energy.
U.S. Department of Energy, Energy Information Administration (EIA). 1999b. Natural Gas
Annual 1998. Washington, DC: U.S. Department of Energy.
U.S. Department of Energy, Energy Information Administration (EIA). 1999c. Petroleum
Supply Annual 1998, Volume I. Washington DC: U.S. Department of Energy.
U.S. Department of Energy, Energy Information Administration (EIA). 1999d. U.S. Crude Oil,
Natural Gas, and Natural Gas Liquids Reserves 1998 Annual Report. Washington, DC:
U.S. Department of Energy.
U.S. Department of Energy, Energy Information Administration (EIA). 2000a. As obtained on
June 1, 2000. .
R-5
-------
U.S. Department of Energy, Energy Information Administration (EIA). 2000b. Issues in
Midterm Analysis and Forecasting 1999—Table 1. . As obtained on May 8, 2000.
U.S. Department of Energy, Energy Information Administration (EIA). 2000c. Supplement
Tables to the AEO 2000. . As
obtained on June 20, 2000.
U.S. Department of Transportation, Office of Pipeline Safety (OPS). Annual Mileage Totals
Transmission—Distribution, . As obtained on March
22, 2000.
U.S. Environmental Protection Agency (EPA). August 1987. The Risk Assessment Guidelines
of 1986. Office of Health and Environmental Assessment. Washington, DC.
U.S. Environmental Protection Agency (EPA). September 1990. Cancer Risk from Outdoor
Exposure to Air Toxics. Volume I. EPA-450/l-90-004a. Office of Air Quality Planning
and Standards. Research Triangle Park, NC.
U.S. Environmental Protection Agency (EPA). 1991. Ecological Exposure and Effects of
Airborne Toxic Chemicals: An Overview. EPA/6003-91/001. Environmental Research
Laboratory. Corvallis, OR.
U.S. Environmental Protection Agency (EPA). 1992a. Regulatory Impact Analysis for the
National Emissions Standards for Hazardous Air Pollutants for Source Categories:
Organic Hazardous Air Pollutants from the Synthetic Organic Chemical Manufacturing
Industry and Seven Other Processes. Draft Report. Office of Air Quality Planning and
Standards. Research Triangle Park, NC. EPA-450/3-92-009, December.
U.S. Environmental Protection Agency (EPA). 1992b. Draft Regulatory Impact Analysis of
National Emissions Standards for Hazardous Air Pollutants for By Product Coke Oven
Charging, Door Leaks, and Topside Leaks. Office of Air Quality Planning and
Standards, Research Triangle Park, NC.
U.S. Environmental Protection Agency (EPA). 1992c. A Tiered Modeling Approach for
Assessing the Risks due to Sources of Hazardous Air Pollutants. EPA-450/4-92-001.
David E. Guinnup, Office of Air Quality Planning and Standards, USEPA, March.
U.S. Environmental Protection Agency (EPA). October 1994. Methods for Derivation of
Inhalation Reference Concentrations and Applications of Inhalation Dosimetry.
EPA-600/8-90-066F, Office of Research and Development, USEPA.
R-6
-------
U.S. Environmental Protection Agency (EPA). 1995. Regulatory Impact Analysis for the
Petroleum Refinery NESHAP. Revised Draft for Promulgation. Office of Air Quality
Planning and Standards, Research Triangle Park, NC.
U.S. Environmental Protection Agency (EPA). 1997a. Economic Analysis for the National
Emission Standards for Hazardous Air Pollutants for Source Category: Pulp and Paper
Production; Effluent Limitations Guidelines, Pretreatment Standards, and New Source
Performance Standards: Pulp, Paper, and Paperboard Category—Phase 1. Research
Triangle Park, NC: Office of Air Quality Planning and Standards. (1997b in Sec.8)
U.S. Environmental Protection Agency (EPA). 1997b. Health Effects Assessment Summary
Tables. FY-1997 update. Office of Research and Development, Office of Emergency
and Remedial Response. Washington, DC: U.S. Environmental Protection Agency.
EPA/540/R-97/036. NTIS PB 97-921199.
U.S. Environmental Protection Agency (EPA). December 1998. Regulatory Impact Analyses for
the NOx SIP Call, FIP, and Section 126 Petitions. Research Triangle Park, NC: Office of
Air Quality Planning and Standards.
U.S. Environmental Protection Agency (EPA). 1999a. Criteria Document for Carbon
Monoxide. Office of Air Quality Planning and Standards.
U.S. Environmental Protection Agency (EPA). 1999b. Draft Revised Guidelines for
Carcinogen Risk Assessment. NCEA-F-0644. USEPA, Risk Assessment Forum, July
1999. pp 3-9ff. .
U.S. Environmental Protection Agency (EPA). 1999c. Economic Impact Analysis of the Oil and
Natural Gas Production NESHAP and the Natural Gas Transmission and Storage
NESHAP. Research Triangle Park, NC: Office of Air Quality Planning and Standards.
U.S. Environmental Protection Agency (EPA). 1999d. EPA Office of Compliance Sector
Notebook Project: Profile of the Oil and Gas Extraction Industry. Washington, DC:
U.S. Environmental Protection Agency.
U. S. Environmental Protection Agency (EPA). 1999e. Regulatory Impact Analysis—Control of
Air Pollution from New Motor Vehicles: Tier 2 Motor Vehicle Emissions Standards and
Gasoline Sulfur Control Requirements. Ann Arbor, MI: Office of Mobile Sources,
December.
U. S. Environmental Protection Agency (EPA). 1999f OAOPS Economic Analysis Resource
Document. Research Triangle Park, NC: Office of Air Quality Planning and Standards.
U.S. Environmental Protection Agency (EPA). 2000a. Draft Preamble. National Emission
Standards for Hazardous Air Pollutants for Stationary Reciprocating Internal
R-7
-------
Combustion Engines. Research Triangle Park, NC: U.S. Environmental Protection
Agency.
U.S. Environmental Protection Agency (EPA). 2000b. Guidelines for Preparing Economic
Analysis, . As obtained
in September 2000.
U.S. Environmental Protection Agency (EPA). 2000c. Integrated Risk Information System;
updated through April 2000. Website access available at www.epa.gov/ngispgm3/iris.
U.S. Environmental Protection Agency, (EPA). 2000d. Regulatory Impact Analysis: Heavy-
Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control
Requirements. Prepared by Office of Air and Radiation.
. As obtained on April 16, 2001.
U.S. Environmental Protection Agency (EPA). 2000e. Staff Paper for the Carbon Monoxide
National Ambient Air Quality Standard; Office of Air Quality Planning and Standards.
U.S. Environmental Protection Agency, (EPA). 2000f. Supplementary Guidance for
Conducting Health Risk Assessment of Chemical Mixtures. Risk Assessment Forum
Technical Panel. EPA/630/R-00/002. USEPA, August.
http://www.epa.gov/nceawwwl/pdfs/chemmix/chemmix082001.pdf
U.S. Environmental Protection Agency (EPA). 2002a. Integrated Risk Information System;
Substance file—Acetaldehyde; CASRN 75-07-0;
. As obtained in October 2002.
U.S. Environmental Protection Agency (EPA). 2002b. Integrated Risk Information System;
Substance file—Formaldehyde; CASRN 50-00-0;
. As obtained in October 2002.
U.S. Environmental Protection Agency (EPA). 2002c. "The National-scale Air Toxics
Assessment." .
U.S. Environmental Protection Agency, Science Advisory Board (EPA SAB). 1998. Advisory
Council on Clean Air Compliance Analysis Advisory on the Clean Air Act Amendments
(CAAA) of 1990 Section 812 Prospective Study: Overview of Air Quality and Emissions
Estimates: Modeling, Health and Ecological Valuation Issues Initial Studies. EPA-
S AB-COUNCIL-AD V-98-003.
U.S. Environmental Protection Agency, Science Advisory Board (EPA SAB). 1999a. The Clean
Air Act Amendments (CAAA) Section 812 Prospective Study of Costs and Benefits
(1999): Advisory by the Advisory Council on Clean Air Compliance Analysis: Costs and
Benefits of the CAAA. EPA-SAB-COUNCIL-ADV-00-002.
R-8
-------
U.S. Environmental Protection Agency, Science Advisory Board (EPA SAB). 1999b. The Clean
Air Act Ammendments (CAAA) Section 812 Prospective Study of Costs and Benefits
(1999): Advisory by the Health and Ecological Effects Subcommittee on Initial
Assessments of Health and Ecological Effects; Part 1. EPA-SAB-COUNCIL-ADV-99-
012. July.
U.S. Environmental Protection Agency, Science Advisory Board (EPA SAB). 1999c. An SAB
Advisory on the Health and Ecological Effects Initial Studies of the Section 812
Prospective Study: Report to Congres: Advisory by the Health and Ecological Effects
Subcommittee. EPA-SAB-COUNCIL-ADV-99-05. February.
U.S. Environmental Protection Agency, Science Advisory Board (EPA SAB). 2000. An SAB
Report on EPA 's White Paper Valuing the Benefits of Fatal Cancer Risk Reduction.
EPA-SAB- EEAC-00- 013
U.S. Office of Management and Budget (OMB). 1992. Regulatory Impact Guidance. Appendix
V of Regulatory Program of the United States Government. April 1, 1991 - March 31,
1992.
U.S. Small Business Administration (SBA). 1999. "Size Standards by SIC Industry."
.
Voorhees, A., B. Hassett, and I. Cote. 1989. "Analysis of the Potential for Non-Cancer Health
Risks Associated with Exposure to Toxic Air Pollutants." Paper presented at the 82nd
Annual Meeting of the Air and Waste Management Association.
Weinstein, D. and E. Birk. 1989. "The Effects of Chemicals on the Structure of Terrestrial
Ecosystems: Mechanisms and Patterns of Change." In: Levin, S. et al. (eds).
Ecotoxicology: Problems and Approaches. Chapter?, pp. 181-209. Springer-Verlag,
New York.
Whelan, Michael. 1998. "Improving Compressor Station Performance: Getting at the Heart of
the Matter." .
Zimmerman, M.B. 1977. "Modeling Depletion in the Mineral Industry: The Case of Coal."
The Bell Journal of Economics 8(2):41-65.
R-9
-------
APPENDICES
-------
APPENDIX A:
ECONOMIC MODEL OF MARKETS AFFECTED BY THE RICE MACT
Implementation of the proposed MACT standards will affect the costs of production in
U.S. energy markets, thus changing the amount of energy that producers are willing to supply
and leading to a change in price. Because energy is used as an input in the production of most
goods and services, changes in the price of energy will affect almost all of the markets in the
U.S. to some extent. Specifically, the cost of the regulation may cause individual facilities to
decrease their current level of production or even to close. These choices affect, and in turn are
affected by, the market price for each product. As the individual facilities in a market decrease
their current level of production, the market supply will decrease as well.
The Agency developed an economic model of markets affected by the proposed rule to
estimate its economic impact (see Section 5 for details on the conceptual approach). In addition
to the impact on the energy markets, many final product markets where RICE units are used as
part of the production process will also be affected. The economic analysis employs standard
concepts in microeconomics to model the regulation's impacts on production costs, supply,
equilibrium price and quantity, and economic welfare. This appendix presents the structural
equations used in the computer model to estimate these impacts and discusses the method used
for welfare calculations.
A. 1 ENERGY MARKETS MODEL
The operational model includes four energy markets: coal, electricity, natural gas, and
petroleum. The following sections describe supply and demand equations the Agency developed
to characterize these markets. The data source for the price and quantity data used to calibrate
the model is the Department of Energy's Supplemental Tables to the Annual Energy Outlook
2000 (EIA, 2000c).
A-l
-------
A. 1.1 Supply Side Modeling
The Agency modeled the existing market supply of energy markets (Qsi) using a single
representative supplier with an upward-sloping supply curve. The Cobb-Douglas (CD) function
specification is
where
Pi
n
V(Pi-
n
.
(A.1)
the supply of energy product i,
a parameter that calibrates the supply equation to replicate the
estimated 2005 level of production (Btu),
the projected 2005 ($/Btu) market price for product i,
per-unit direct compliance costs generated by dividing the annual
control costs estimated by the engineering analysis by the
production level (Qs),
the domestic supply elasticity for product i, and
indirect effect of changes in energy input prices, where • j*is the
fuel share of energy product] used in producing energy product i.
The fuel share is allowed to vary using a fuel switching rule
relying on cross-price elasticities of demand between energy
sources, as described in Section 5 of the report.
A. 1.2 Demand Side Modeling
Market demand in the energy markets (QDi) is expressed as the sum of the energy,
residential, transportation, industrial, and commercial sectors:
n
QDi= I
(A.2)
A-2
-------
where i indexes the energy market and j indexes the consuming sector. The Agency modeled the
residential, and transportation sectors as single representative demanders using a simple Cobb
Douglas specification:
where p is the market price, • is an assumed demand elasticity (actual values are presented in
Section 5, Table 5-2), and A is a demand parameter used to calibrate the demand equations to
match baseline conditions.
In contrast, energy demand in the energy, industrial and commercial sectors is modeled
as a derived demand resulting from the production/consumption choices in the agricultural,
energy, mining, manufacturing, and service industries. Energy demand for these industries
responds to changes in output as well as fuel switching that occurs in response to changes in
relative energy prices projected in the energy markets. For each sector, energy demand is
expressed as follows:
qDijl = (1 + %AQDj) • (qDij0) • FSW (A.4)
where qD is demand for energy, QD is output in the final product or service market, FSW is a
factor generated by the fuel switching algorithm, i indexes the energy market, j indexes the
market. The subscripts 0 and 1 represent baseline and with regulation conditions, respectively.
A.2 INDUSTRIAL AND COMMERCIAL MARKETS
Given data limitations associated with the scope of potentially affected industrial and
commercial markets, EPA used an alternative approach to estimate the relative changes in price
and quantities in these markets. Rather than using measures of price and quantity as in the
energy markets, data for the industrial and commercial markets was estimated in terms of
percentage changes in prices and quantities relative to baseline values. The estimated percentage
changes in prices and quantities in each market are used to compute changes in economic
welfare as described in Section A.4.
A-3
-------
A. 2.1 Compute Percentage Change in Market Price
First, we computed the change in production costs resulting from changes in the market
price of fuels (determined in the energy markets):
n
%Ac. = £ OjApj , (A.5)
i=l
where • *is the fuel share,1 i indexes the energy market, and j indexes the industrial or
commercial market. We use the results from equation A.5 and the market supply and demand
elasticities to compute the percentage change in market price2:
Sj
i, = %ACj- 8
(A.6)
A.2.2 Compute Percentage Change in Market Quantity
Using the percentage change in the price calculated in Equation A.6 and assumptions
regarding the market demand elasticity, the relative change in quantity was computed. For
example, in a market where the demand elasticity is assumed to be -1 (i.e., unitary), a 1 percent
increase in price results in a 1 percent decrease in quantity. This change was then input into
equation A.4 to determine energy demand.
A. 3 WITH-REGULATION MARKET EQUILIBRIUM DETERMINATION
Market adjustments can be conceptualized as an interactive feedback process. Supply
segments face increased production costs as a result of the rule and are willing to supply smaller
quantities at the baseline price. This reduction in market supply leads to an increase in the
market price that all producers and consumers face, which leads to further responses by
producers and consumers and thus new market prices, and so on. The new with-regulation
'The fuel share is allowed to vary using a fuel switching rule using cross-price elasticities of demand between energy
sources, as described in Section 5.
2The approach is based on a mathematical model of tax incidence analysis decribed in Nicholson (1998) pages 444-
445.
A-4
-------
equilibrium is the result of a series of iterations in which price is adjusted and producers and
consumers respond, until a set of stable market prices arises where total market supply equals
market demand (i.e., Qs = QD) in each market. Market price adjustment takes place based on a
price revision rule that adjusts price upward (downward) by a given percentage in response to
excess demand (excess supply).
The algorithm for determining with-regulation equilibria can be summarized by seven
recursive steps:
1. Impose the control costs on electricity supply segments, thereby affecting their
supply decisions.
2. Recalculate the market supply in the energy markets. Excess demand exists.
3. Determine the new energy prices via a price revision rule.
4. Recalculate energy market supply.
5. Account for fuel switching given new energy prices. Solve for new equilibrium
in final product and service market.
6. Compute energy demand.
7. Compare supply and demand in energy markets. If equilibrium conditions are not
satisfied, go to Step 3, resulting in a new set of energy prices. Repeat until
equilibrium conditions are satisfied (i.e., the ratio of supply to demand is
arbitrarily close to one).
A.4 COMPUTING SOCIAL COSTS
In the energy markets, consumer (residential and transportation sectors) and producer
surplus were calculated using standard methods based on the price and quantity before and after
regulation. In the industrial and commercial markets, however, there is no easily defined price
or quantity due to the wide variety of products that fall under each sector (i.e., NAICS code).
Therefore, methods of calculating consumer and producer surplus are defined based on relative
changes in prices and quantities and total industry sales rather than on the prices and quantities
directly. The following sections describe how we derive welfare estimates for these markets.
A-5
-------
A.4.1 Change in Consumer Surplus
If price and quantities were available, a linear approximation of the change in consumer
surplus can be calculated using the following formula:
•CS = -[(.P)Q1-0.5(.Q)(.P)], (A.7)
where Qx denotes the estimated post-regulation quantity, • P denotes the estimated change in
price resulting from the regulation, and • Q denotes the estimated change in quantity resulting
from the regulation. Given the difficulties associated with defining baseline measures of price
and quantity for broad NAICS codes described above, the model estimates relative changes in
price and quantity for each industrial/commercial market. Thus, changes in consumer surplus
were calculated using these data and total revenue by NAICS code as shown below:
• CS = -[(• P) Q! - 0.5 (• Q) (• P)] (Px Q^/fP, Qt)
• CS = -[%• P - 0.5 (%• P) (%• Q)] (Pi QO. (A.8)
A.4.2 Change in Producer Surplus
If price and quantities were available, a linear approximation could also be used to
compute the change in producer surplus:
• PS —[((CC/QO - • P)(Q! - • Q)]+ 0.5 [(CC/Q! -•?)(• Q)], (A.9)
where CC/Qj equals the per-unit "cost-shifter" of the regulation. Again, we transform this
equation into one that relies only on percentage changes in price and quantity, total revenue,3 and
compliance costs:
• PS = - [((CC/QO - • P)(Q1 - • Q)]+ 0.5 [((CC/QO - • P)(- Q)]^ Q^ Qx)
• PS = - [(% cost shift - %• P)(l - %• Q)+ 0.5 (% cost shift - %• P )(%• Q)][Pi QJ
• PS = - [% cost shift - %• P ][1 - 0.5(%« Q)][TR], (A. 10)
where TR refers to total revenue, which is equal to price multiplied by quantity. This modified
formula no longer requires price and quantity directly4 and can be applied to the final product
markets where this information is not available.
Multiplying price and quantity in an industry yields total industry revenue. The U.S. Census Bureau provides
shipment data for the NAICs codes included in the economic model.
4Only the product of price and quantity is required for this formula. Multiplying price and quantity in an industry
yields total industry revenue. The value used for total industry revenue is derived from industry-level value of
shipments data so that price and quantity do not have to be individually defined.
A-6
-------
APPENDIX B:
ECONOMIC MODEL SENSITIVITY ANALYSIS
Estimates of the economic impacts of the MACT standard are sensitive to the parameters
used in the model. Therefore, a sensitivity analysis was conducted to determine the effects on
the model results of changing several of the key parameters. Sensitivity analyses were
developed for the elasticity of supply in the electricity markets, the demand and supply
elasticities in the manufacturing final product markets, the own- and cross-price elasticities used
to model fuel switching, and the distribution of affected engines in SIC 13 between the natural
gas and petroleum industries. In general, estimates of the change in social welfare are robust.
The distribution of welfare losses across producers and consumers responds moderately to
changes in the selected parameters.
B. 1 ELASTICITY OF SUPPLY FOR ELECTRICITY
The price elasticity of supply in the electricity markets represents the behavioral
responses from existing sources to changes in the price of electricity. However, there is no
consensus on estimates of the price elasticity of supply for electricity, as discussed in Section 4
of the report. Because of deregulation, the market price for electricity has become the
determining factor in decisions to retire older units or to make higher cost units available to the
market, so the price elasticity of supply is becoming more important to utilities' decisionmaking.
To examine how the assumed value of the elasticity of supply for electricity affects the model's
outcomes, welfare impacts were estimated for supply elasticities both higher and lower than the
assumed value of 0.75. Table B-l shows the economic impact estimates as the elasticity of
supply in the electricity markets is varied between 0.5 and 1.0.
B-l
-------
Table B-l. Sensitivity Analysis: Elasticity of Supply in the Electricity Markets ($106)
Change in producer surplus
Change in consumer surplus
Change in social welfare
ES = 0.5
-121.7
-125.8
-247.6
ES = 0.75
-122.1
-125.4
-247.6
ES = 1.0
-122.2
-125.4
-247.6
B.2 FINAL PRODUCT MARKET ELASTICITIES
The final product markets were modeled at the two- or three-digit NAICS code level to
operationalize the economic model. Due to a lack of data on final product supply elasticities, the
elasticity of supply was assumed to equal 0.75 in each of the final product markets. The
elasticity of demand in each final product market was assumed to equal the values in Table 5-4.
The elasticities of supply and demand in the final product markets determine the distribution of
economic impacts between producers and consumers. To examine the change in distribution of
welfare impacts as the elasticities are changed, two alternative assumptions about the elasticities
in the final product markets were used. In the first alternative, supply is assumed to be 25
percent more inelastic than in the model, while the demand elasticity estimate remains the same.
In the second alternative, the supply elasticity is the same as used in the model, but demand is
assumed to be 25 percent more inelastic. Table B-2 shows how the economic impact estimates
vary as the supply and demand elasticities in the final product markets vary. As expected, when
supply becomes more inelastic, producers bear a larger share of the costs relative to the model
results and when demand becomes more inelastic, it is the consumers who bear a larger share of
the cost burden.
B.3 OWN AND CROSS-PRICE ELASTICITIES FOR FUELS
Own- and cross-price elasticities of demand from NEMS were used to capture fuel
switching in the manufacturing sectors in the economic model. However, the NEMS projection
reflects aggregate behavioral responses in the year 2015. Because this is a longer window of
analysis compared to the baseline year 2005, this analysis may overestimate firms' ability to
switch fuels in the short run.
B-2
-------
Table B-2. Sensitivity Analysis: Supply and Demand Elasticities in the Industrial and
Commercial Markets ($106)
Supply Elasticities
Reduced by 25%
Change in producer surplus
Change in consumer surplus
Change in social welfare
-131.3
-116.3
-247.6
Base Case
-122.1
-125.4
-247.6
Demand Elasticities
Reduced by 25%
-111.0
-136.5
-247.6
Table B-3 shows how the economic impact estimates vary as the own- and cross-price
elasticities used in the economic analysis are reduced by 75 percent and 50 percent. Changing
the elasticities used to model fuel switching has only a very small effect on the estimates of
welfare changes.
Table B-3. Sensitivity Analysis: Own- and Cross-Price Elasticities Used to Model Fuel
Switching ($106)
Change in producer surplus
Change in consumer surplus
Change in social welfare
Fuel Price Elasticities
Presented in Table 4-2
-122.1
-125.4
-247.6
Reduced by
75 Percent
-124.3
-123.3
-247.6
Reduced by 50
Percent
-123.9
-123.6
-247.6
B.4 SHARE OF NAICS 211 ASSOCIATED WITH NATURAL GAS AND PETROLEUM
PRODUCTS
Direct costs associated with the regulation are linked to the energy markets in which
engines are operating. Because no information was available on each unit's application, NAICS
codes were used to link engines to specific energy markets. However, for NAICS 211 it was not
possible to distinguish between engines involved in the extraction and production of natural gas
and engines involved in the extraction and processing of petroleum products. In addition,
-------
because petroleum and natural gas are frequently joint products, some engines may be involved
in both markets.
Based on information from industry, it was determined that the majority of the engines
classified under NAICS 211 were involved in natural gas extraction and transportation. The
economic impact estimates presented in Section 5 use an 80/20 percent distribution of control
costs between the natural gas and petroleum markets. Table B-4 shows how the economic
impact estimates vary as the 80/20 percent distribution of control costs between the natural gas
and petroleum markets varies. Once again, there is only a slight difference in the distribution of
costs between producers and consumers under this sensitivity analysis.
Table B-4. Sensitivity Analysis: Distribution of Affected Units in NAICS 211 Between the
Natural Gas and Petroleum Industries ($106)
Natural Gas = 70% Natural Gas = 80% Natural Gas = 90%
Petroleum = 30% Petroleum = 20% Petroleum = 10%
Change in producer
surplus
Change in consumer
surplus
-121.1
-126.4
-122.1
-125.4
-122.6
-124.9
Change in social welfare -247.6 -247.6 -247.6
B-4
-------
APPENDIX C:
ASSUMPTIONS AND LIMITATIONS OF THE ECONOMIC MODEL
In developing the economic model of effects of the RICE NESHAP, several assumptions
were necessary to make the model operational. These assumptions are in addition to those
described in Section 5.2 for the values of supply and demand elasticities. In this section, the
major operational assumptions are listed and explained. Possible impacts and limitations of the
model resulting from each assumption are then described.
Assumption: The domestic markets for energy are perfectly competitive.
Explanation: Assuming that the markets for energy are perfectly competitive implies
that individual producers are not capable of unilaterally affecting the prices they receive for their
products. Under perfect competition, firms that raise their price above the competitive price are
unable to sell at that higher price because they are a small share of the market and consumers can
easily buy from one of a multitude of other firms that are selling at the competitive price level.
Given the relatively homogeneous nature of individual energy products (petroleum, coal, natural
gas, electricity), the assumption of perfect competition at the national level seems to be
appropriate.
Possible Impact: If energy markets were in fact imperfectly competitive, implying that
individual producers can exercise market power and thus affect the prices they receive for their
products, then the economic model would understate possible increases in the price of energy
due to the regulation as well as the social costs of the regulation. Under imperfect competition,
energy producers would be able to pass along more of the costs of the regulation to consumers;
thus, consumer surplus losses would be greater, and producer surplus losses would be smaller in
the energy markets.
Assumption: The domestic markets for industrial products are all perfectly competitive.
Explanation: Assuming that these markets are perfectly competitive implies that the
producers of these products are unable to unilaterally affect the prices they receive for their
products. Because the industries used in this analysis are aggregated across a large number of
individual producers, it is a reasonable assumption that the individual producers have a very
C-l
-------
small share of industry sales and cannot individually influence the price of output from that
industry.
Possible Impact: If these product markets were in fact imperfectly competitive, implying
that individual producers can exercise market power and thus affect the prices they receive for
their products, then the economic model would understate possible increases in the price of final
products due to the regulation as well as the social costs of the regulation. Under imperfect
competition, producers would be able to pass along more of the costs of the regulation to
consumers; thus, consumer surplus losses would be greater, and producer surplus losses would
be smaller in the final product markets.
Assumption: The baseline year of the analysis, 2005, provides representative information
about the impacts on affected industries after new engines subject to the regulation have
been installed.
Explanation: The engineering costs of the regulation are estimated for all engines
projected to exist in 2005 in terms of 1998 dollars. For the economic model to be consistent, all
costs and prices must be denominated in the same year. However, to examine future impacts, the
number of engines projected to exist in 2005 is used in conjunction with costs and prices in 1998
dollars. Because most of the impact of the regulation is borne by new engines, it is more
informative to use a future year that includes some of these new engines rather than the current
year. In the current year, no new engines would be subject to the proposed rule. Choosing a
baseline year 5 years into the future allows an examination of intermediate-run costs resulting
from the regulation.
Possible Impact: If the proj ections for growth in the number of engines of each type
(4SRB, 2SLB, 4SLB, CI) turn out to be incorrect, then the actual costs of the regulation will
differ from the estimated values. Also, it is assumed that the relationships between many
variables stay the same in 2005 as they are in 1998, the year that most of the data are from. For
example, it is assumed that fuel costs remain the same proportion of production costs in 2005 as
in 1998. If these relationships change over time, then the actual cost of the regulation in 2005
will differ from the estimated values. Also, because the number of engines subject to the
regulation is projected to increase over time, the farther into the future the analysis looks, the
higher the costs will be given the current projections. However, extrapolating far into the future
C-2
-------
may not give an accurate picture of the number of engines that will be used because many factors
could change the growth rate of RICE.
Assumption: Fuel costs are a constant proportion of production costs.
Explanation: It is assumed that the percentage of production costs spent on fuels
remains constant as the price of fuel changes. Because the price changes obtained in the model
are so small, it is not unreasonable to assume that producers will not change the mix of inputs
that they use in the production process as a result of the price increase.
Possible Impact: Theoretically, producers could switch their production process to one
that requires less fuel by substituting more labor, capital, etc., for fuel. If producers respond to
the increase in fuel prices by significantly altering their input mix and using less fuel, then the
price in the energy markets will increase less than the estimated value due to the decrease in
demand, and prices in the final product markets will also increase less than expected. In this
case, producers will face higher welfare losses and consumers smaller welfare losses than in the
current model.
Assumption: The amount of fuel required to produce a unit of output in the final product
markets remains constant as output changes.
Explanation: The importance of this assumption is that when output in the final product
markets changes as a result of a change in energy prices, it is assumed that the amount of fuel
used changes in the same proportion as output, although the distribution of fuel usage among
fuel types may change due to fuel switching. This change in the demand for fuels feeds into the
energy markets and affects the equilibrium price and quantity in the energy markets.
Possible Impact: Fuel usage may not actually change in exactly this way. If fuel usage
decreases more than proportionately, then the demand for fuels will decrease more, and there
will be more downward pressure on energy prices than the model results suggest. If fuel usage
decreases less than proportionately, then the demand for fuels will decrease less, and the price
will be higher than the model result.
Assumption: All pipelines are affected by the regulation.
Explanation: It is assumed that new engines will be distributed across all existing
pipelines and any new pipelines so that the cost of distribution rises for all natural gas rather than
only affecting some producers and leaving others unaffected.
C-2
-------
Possible Impact: If only some natural gas producers are affected and others are
unaffected, then the unaffected firms may see their profits rise if the market price increases due
to decreases in output from affected suppliers because the unaffected firms experience no shift in
their cost curves as a result of the regulation. The relative proportion of affected and unaffected
producers would then be important in determining the overall change in equilibrium price and
quantity. If the regulation affected only a very small percentage of the market, then market price
and quantity may not change appreciably.
C-4
-------
APPENDIX D:
SUMMARY OF STUDIES OF THE EFFECTS OF
EMISSIONS OF HAZARDOUS AIR POLLUTANTS
Although we are unable to quantify the effects of the HAPs reduced by this rule, below
we present a qualitative discussion of the toxic effects of the pollutants that are controlled by the
regulation. The information presented is obtained from the EPA's Integrated Risk Information
System (IRIS) (EPA, 2002a; 2002b), which is a resource of health assessment information on
chemical substances that have undergone a comprehensive review by EPA health scientists from
several Program Offices and the Office of Research and Development. The summaries presented
are the result of consensus reached during the review process. While this rule produces
significant reductions in formaldehyde, acetaldehyde, acrolein, methanol, carbon monoxide, and
nitrous oxides, IRIS based risk assessments due to inhalation are only available for formaldehyde
and acetaldehyde.
Formaldehyde:
Based on a review of human epidemiological studies and available animal studies of the
chronic effects from this pollutant, formaldehyde is classified as a "probable human carcinogen"
if inhaled through the air (EPA, 2002b). The human data is "limited,"1 but includes nine studies
that show statistically significant associations between site-specific respiratory neoplasms and
exposure to formaldehyde. Long-term inhalation studies in rats and mice are determined to be
"sufficient"2 and show an increased incidence of cancerous cells in the nasal cavity.
At least 28 epidemiological studies of the effects on humans have been conducted, nine
of which are used for the classification of formaldehyde as a probable human carcinogen.
Among these, two cohort studies (Blair et al., 1986, 1987; Stayner et al., 1988) and one case-
'In general, "limited" means that the studies show a tendency for these effects, but the data used or study findings are
limited to a small set of studies on humans or have a large amount of uncertainty associated with them.
2In general, "sufficient" means that there are a sufficient number of studies with statistically significant findings such
that classification of carcinogenicity is more certain.
D-l
-------
control study (Vaughan et al., 1996a, b) were well conducted according to IRIS and specifically
designed to detect small to moderate increases in formaldehyde-associated human risks. Blair et
al. studied workers at 10 plants who were in some way exposed to formaldehyde and observed
significant excesses in lung and nasopharyngeal cancer deaths. Despite the lack of significant
trends with increasing concentration or cumulative formaldehyde exposure, lung cancer
mortality was significantly elevated in analyses with or without a 20-year latency allowance.
Stayner et al. reported statistically significant excesses in mortality from buccal cavity tumors
among formaldehyde-exposed garment workers. The case-control study conducted by Vaughan
et al. examined occupational and residential exposures, and showed a significant association
between nasopharyngeal cancer and having lived 10 or more years in a mobile home, especially
for mobile homes built in the 1950's to 1970's, a period of increasing formaldehyde-resin usage.
The 25 other reviewed studies had limited ability to detect small to moderate increases in
formaldehyde risks owing to small sample sizes, small numbers of observed site-specific deaths,
and insufficient follow-up. Even with these potential limitations, 6 of the 25 studies reported
significant associations between excess site-specific respiratory (lung, buccal cavity, and
pharyngeal) cancers and exposure to formaldehyde. Although the common exposure in all of
these studies was formaldehyde, the epidemiological evidence is categorized as "limited" in the
IRIS database primarily because of the possible exposures to other agents. Such exposures could
have contributed to the findings of excess cancers.
The data on animal carcinogenicity, however, was found to be sufficient. Consequences
of inhalation exposure to formaldehyde have been studied in rats, mice, hamsters, and monkeys.
Kerns et al. (1983) exposed about 120 mice and rats to 0, 2, 5.6, or 14.3 ppm, 6 hours/day, 5
days/week for 24 months. From the 12th month on, the rats in the highest dose group (14.3 ppm)
showed significantly increased mortality compared to control groups. In the 5.6 ppm group,
male rats showed a significant increase in mortality from 17 months on. Squamous cell
carcinomas were seen in the nasal cavities of 51 out of 117 male rats and 52 out of 117 female
rats at 14.3 ppm by experiment's end.
Tobe et al. (1985) conducted a 28-month study of male rats. Exposure to 15 ppm ended
at 24 months; at that point, mortality was 88 percent. Squamous cell carcinomas were seen at 15
D-2
-------
ppm in 14 out of 27 rats surviving past 12 months, compared with 0 out of 27 rats in the control group.
Based on the results of these studies, IRIS reports quantitative estimates of risk from
inhalation exposure. One form in which risk is presented is an estimate of "unit risk." The unit
risk is the quantitative estimate in terms of risk per ug/cu.m air breathed. Another form in which
risk is presented is an air concentration providing cancer risks of 1 in 10,000, 1 in 100,000, and 1
in 1,000,000. The rationale and methods used to develop the carcinogenicity information in
IRIS are described in The Risk Assessment Guidelines of 1986 (EPA/600/8-87/045) and in the
IRIS Background Document available on the EPA's website. Using these guidelines, IRIS
reports an inhalation unit risk for formaldehyde of 1.3E-5 (ug/cu.m), with a corresponding
chance of cancer of 1 in 10,000 at concentrations of 8E+0 ug/cu.m, a 1 in 100,000 chance of
cancer at concentrations of 8E-1 ug/cu.m, and a 1 in 1,000,000 chance of cancer at
concentrations of 8E-2 ug/cu.m.
Acetaldehyde:
Acetaldehyde is similar in structure to formaldehyde which also produces nasal tumors in
animals exposed to inhalation. When inhaled, acetaldehyde causes cancers in the nose and
trachea of hamsters, and nasal cancers in rats. The epidemiological studies in humans is
determined to be "inadequate,"3 however, based on the evidence in animal studies (which are
determined to be sufficient), acetaldehyde is classified as a probable human carcinogen (EPA,
2002a). Two short-term animal studies conducted by the same research group (Appleman et al.,
1986; Appleman et al., 1982) are the principal studies used in the determination if a Reference
Concentration (RfC) presented in IRIS. The RfC is a benchmark concentration at which risk is
not a public health concern. If the RfC is exceeded, the risk of effects increases to an unsafe
level. While these studies are short-term in duration, together they establish a concentration-
response for lesions after only 4 weeks of exposure. These same types of lesions appear at
longer exposure times and higher exposure levels in chronic studies (Wouterson et al., 1986;
Wouterson and Feron, 1987; Kruysse et al., 1975).
3In general, "inadequate" means that the there are too small a number of human studies to determine the
classification, or the findings of the studies have a large level of uncertainty.
D-3
-------
Appleman et al. (1986) conducted two inhalation studies on male rats. Continuous and
interrupted (define) exposure to 500 ppm did not induce any visible effect on general condition
or behavior, but peak exposures at this level caused irritation. Mean body weights of the group
exposed to 500 ppm with interruption and with peak exposures were statistically significantly
lower than those in the control group. Histopathological changes attributable to exposure were
found only in the nasal cavity. Degeneration of the olfactory epithelium was observed in rats
exposed to 500 ppm. Interruption of the exposure or interruption combined with peak exposure
did not visibly influence this adverse effect.
The Appleman et al. (1982) study found that during the first 30 minutes of each exposure
at the 5000-ppm level, rats exhibited severe dyspnea that gradually became less severe during
the subsequent exposure period. Two animals died at this level and one male died at the 2200-
ppm level. Growth was retarded in males at the three highest exposure concentrations (1000,
2200, and 5000 ppm) and in females at the 5000-ppm level. Compound-related
histopathological changes were observed only in the respiratory system. The nasal cavity was
most severely affected and exhibited a concentration-response function. At the 400-ppm level,
compound-related change included: slight to severe degeneration of the nasal olfactory
epithelium, without hyper- and metaplasia, and disarragement of epithelial cells. At the 1000-
and 2200-ppm levels, more severe degenerative changes occurred, which hyperplastice and
metaplactis changes in the olfactory and respiratory epithelium of the nasal cavity.
Degenerations with hyperpolasia/metaplasia also occurred in the laryngeal and tracheal
epithelium at these levels. At 5000 ppm changes included severe degenerativehyperplastic and
metplastic changes of the nasal, laryngeal, and tracheal epithelium.
Wouterson et al.(1986) exposed rats for 6 hours/day, 5 days/week for up to 28 months to
0, 750, 1500 and 3000 ppm. The highest concentration was gradually decreased because of
severe growth retardation, occasional loss of body weights, and early mortality in this group.
The rats in this high concentration group showed excessive salivation, labored respiration, and
mouth breathing. The respiratory distress was still observed when the concentration was reduced
to 1000 ppm, although fewer were dyspneic. Only a few rats died during the first 6 months of
the study but thereafter a sharp increase in the numbers of deaths occurred in the high-
concentration group. By 25 months, all top concentration rats had died. When the study was
D-4
-------
terminated, only a few animals remained alive in the mid-concentration group. The cause of
early death was nearly always partial or complete occlusion of the nose by excessive amounts of
keratin and inflammatory exudate. Several showed acute bronchophnuemonia occasionally
accompanied by tracheitis. The only exposure-related histopathology occurred in the respiratory
system and showed a concentration-response relationship. The most severe abnormalities were
found in the nasal cavity. Adenocarcinomas occurred at all exposure concentrations and
squamous cell carcinoma at the mid and high concentrations only.
Data on animal carcinogenicity was determined to be sufficient and data from 3 studies is
presented in IRIS. Feron (1979) exposed hamsters to 0 or 1500 ppm acetaldehyde by inhalation
7 hours/day, 5 days/weeks, for 52 weeks. The exposure produced twice the incidence of
squamous cell carcinomas compared to the control group. Feron et al. (1982) found similar
observations that support Feron (1979). In a study of hamsters exposed to acetaldehyde alone or
in combination with benzo(a)pyrene (BaP), the animals showed a slight increase in nasal tumors
and a significantly increased incidence of larygeal tumors. Woutersen and Appelman (1984)
studied albino rats for 27 months and found multiple respiratory tract tumores.
Adenocarcinomas were increased significantly at all exposure levels, and squamous cell
carcinoma incidences showed a clear dose-response relationship.
The critical effect reported in IRIS for acetaldehyde is degenerations of olfactory
epithelium and the inhalation Reference Concentration (RfC) is reported as 9E-3 mg/cu.m. In
general, the RfC is an estimate of a daily inhalation exposure of the human population (including
sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a
lifetime. Thus, if this concentration is exceeded on a daily basis, then degeneration of olfactory
epithelium is likely to occur in humans. Similar to the formaldehyde description above, IRIS
also presents risk of cancer in other terms. IRIS reports the inhalation unit risk for acetaldehyde
as 2.2E-6 per ug/cu.m., with a corresponding risk of cancer of 1 in 10,000 at concentrations of
5E+1 ug/cu.m, a 1 in 100,000 risk of cancer at concentrations of 5E+0 ug/cu.m, and a 1 in
1,000,000 risk of cancer at concentrations of 5E-1 ug/cu.m.
D-5
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before completing)
1. REPORT NO.
EPA-452/R-02-012
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Regulatory Impact Analysis for the Proposed Reciprocating
Internal Combustion Engines NESHAP
5. REPORT DATE
November 2002
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
None
12. SPONSORING AGENCY NAME AND ADDRESS
Steve Page, Director
Office of Air Quality Planning and Standards
Office of Air and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proposed regulation
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report summarizes the benefits, costs, and economic impacts associated with the National Emissions Standard for Hazardous Air
Pollutants (NESHAP) for the Reciprocating Internal Combustion Engines (RICE) source category. This source category includes spark
ignition engines that operate generally with natural gas and compression ignition engines that operate with diesel fuel, and can be classified as
two-stroke, or four-stroke engines. They are also classified by the richness of the fuel mix: rich burn or lean burn. RICE units are typically
used along natural gas pipelines to provide adequate pressure to transmit fuel through the pipeline. Others are also used to provide power in a
remote area of an operation in industries such as health services, energy generation, oil and gas extraction, and quarrying of non-metallic
minerals.
In the 5th year after implementation, the proposed NESHAP for RICE will impact existing and new engine and is expected to reduce
HAP emissions by 5,000 tons per year, 234,400 tons of carbon monoxide (CO) per year, 167,900 tons of nitrogen oxides (NOx) per year, and
3,700 tons of particulate matter (PM10) per year. The total social cost of rule is approximately $255 million (1998$). This cost is spread
across more than 25 different industries, which results in small economic impacts with minimal changes in prices and production levels in most
affected industries. Benefits of the HAP reductions include reduced respiratory illnesses and reduced incidence of cancer, however, we are
unable to quantify these effects. Benefits from NOx and PM reductions include fewer fatalities, and reduced incidence of chronic bronchitis,
asthma, and other respiratory illnesses, which are valued at approximately $280 million per year.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Regulatory Impact Analysis
Health Assessment
Economic Analysis
air pollution control, environmental
regulation, economic impact analysis,
benefits, costs, maximum achievable
control technology, reciprocating internal
combustion engines
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (Report)
Unclassified
21. NO. OF PAGES
115
20. SECURITY CLASS (Page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
------- |