Regulatory Impact Analysis for the Final New Source Performance Standards for the Synthetic Organic Chemical Manufacturing Industry and National Emission Standards for Hazardous Air Pollutants for the Synthetic Organic Chemical Manufacturing Industry and Group I and II Polymers and Resins Industry


*	*

KwJ

PRO'&

Regulatory Impact Analysis for the Final New
Source Performance Standards for the Synthetic
Organic Chemical Manufacturing Industry and
National Emission Standards for Hazardous Air
Pollutants for the Synthetic Organic Chemical
Manufacturing Industry and Group I & II
Polymers and Resins Industry

-------

-------
EP A-452/R-24-001
March 2024

Regulatory Impact Analysis for the Final New Source Performance Standards for the Synthetic
Organic Chemical Manufacturing Industry and National Emission Standards for Hazardous Air
Pollutants for the Synthetic Organic Chemical Manufacturing Industry and Group I & II

Polymers and Resins Industry

U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Research Triangle Park, NC

-------
CONTACT INFORMATION

This document has been prepared by staff from the Office of Air and Radiation, U.S.
Environmental Protection Agency. Questions related to this document should be addressed to the
Air Economics Group in the Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Office of Air and Radiation, Research Triangle Park, North Carolina 27711
(email: OAQPSeconomics@epa.gov).

ACKNOWLEDGEMENTS

In addition to U.S. EPA staff from the Office of Air and Radiation, personnel from RTI
International contributed data and analysis to this document.

-------
TABLE OF CONTENTS

Table of Contents	i

List of Tables	iv

List of Figures	vii

1	Executive Summary	1

1.1	Introduction	1

1.1.1	NESHAP for subparts F, G, H, I, U, & W	2

1.1.2	NSPS subparts III, NNN, RRR. & YYb	4

1.2	Market Failure	5

1.3	Results for the Final Action	6

1.3.1	Baseline for the Regulation	6

1.3.2	Overview of Costs and Benefits for the Final Options	7

1.4	Organization of the Report	12

2	Industry Profile	14

2.1	Introduction	14

2.2	SOCMI Industry Profile	14

2.2.1	Oil and Gas Sectors and SOCMI	19

2.2.2	SOCMI Supply Chain Disruptions	21

2.2.3	Ethylene	22

2.3	P&R Groups I and II	28

2.4	Group I Industry Profile	29

2.4.1	Industry Organization of Group I Industries	29

2.4.2	Prices for Group I Industries	31

2.4.3	General Production Description of Group I Industries	34

2.4.4	Product Description of Group I Industries	34

2.5	Group II Industry Profile	39

2.5.1	Industry Organization of Group II Industries	40

2.5.2	Prices for Group II Industries	40

2.5.3	Product Description and Markets of Group II Industries	43

3	Emissions and Engineering Cost Analysis	44

3.1	Introduction	44

3.2	HON	44

3.3	P&R I (Subpart U)	45

3.4	P&R II (Subpart W)	46

3.5	Emission Points and Controls	47

3.5.1 Flares	47

l

-------
3.5.2	Fenceline Monitoring	49

3.5.3	Pressure Relief Devices	50

3.5.4	Storage Vessels	51

3.5.5	Heat Exchange Systems	52

3.5.6	Process Vents	53

3.5.7	Wastewater	58

3.5.8	Equipment Leaks	59

3.6	Engineering Cost Analysis Summary Results	60

3.7	Secondary Air Emission Impacts	66

4 Benefits of emission reductions	69

4.1	Introduction	69

4.2	Health Effects from Exposure to Hazardous Air Pollutants (HAP)	70

4.2.1	Ethylene oxide	71

4.2.2	Chloroprene	71

4.2.3	Benzene	72

4.2.4	1,3-Butadiene	72

4.2.5	Ethylene dichloride (1,2-dichloroethane)	72

4.2.6	Vinyl chloride	73

4.2.7	Chlorine	73

4.2.8	Maleic anhydride	74

4.2.9	Acrolein	74

4.2.10	Other Hazardous Air Pollutants (HAP)	74

4.3	Ozone-related Human Health Benefits	75

4.3.1	Estimating Ozone Related Health Impacts	75

4.3.2	Selecting Air Pollution Health Endpoints to Quantify	76

4.3.3	Quantifying Cases of Ozone-Attributable Premature Mortality	77

4.4	Economic Valuation	78

4.4.1	Benefit-per-Ton Estimates	80

4.4.2	Ozone Vegetation Effects	82

4.4.3	Ozone Climate Effects	82

4.5	Ozone- andNOx- Related Benefits Results	83

4.6	Characterization of Uncertainty in the Monetized Benefits	86

4.7	Climate Impacts	87

4.8	Total Monetized Benefits	107

Chapter 4 Appendix	Ill

ii

-------
5	Economic Impact Analysis	115

5.1	Introduction	115

5.2	Economic Impact Analysis	115

5.3	Description of Approach, Model, and Framework	116

5.3.1	Data Limitations	116

5.3.2	Benchmark Data	116

5.3.3	Control Data	117

5.4	Small Business Impacts Analysis	127

5.5	Screening Analysis	129

5.6	Employment Impact Analysis	131

6	Comparison of Costs and Benefits	135

6.1	Results	135

6.2	Uncertainties and Limitations	140

7	References	143

in

-------
LIST OF TABLES

Table 1-1 Monetized Benefits, Compliance Costs, Emission Reductions and Net	8

Table 1-2 Monetized Benefits, Compliance Costs, and Net Benefits for Amendments to P&R I (dollars in
million 2021$)a	9

Table 1-3	Monetized Benefits, Compliance Costs, and Net Benefits for Amendments	9

Table 1-4	Monetized Benefits, Compliance Costs, and Net Benefits for NSPS	10

Table 1-5	Monetized Benefits, Compliance Costs, and Net Benefits for	10

Table 1-6	Total Monetized Benefits, Compliance Costs, Emission Reductions, and Net	12

Table 2-1	Select SOCMI Chemicals by Feedstock*	15

Table 2-2	Top 10 Globally Produced SOCs by Total Market Value (2021$)	18

Table 2-3	Polymers and Resin Group I Industries	29

Table 2-4	Concentration Findings of Affected Group I Industries	30

Table 2-5	Chemical Manufacturing (NAICS 325) Product Price Index, 2012-2021	31

Table 2-6	Producer Price Index of Synthetic Rubber, 2012-2021 (Index for 2012 is normalized to 100)	33

Table 2-7	Polymers and Resin Group II Industries	39

Table 2-8	Concentration Findings of Affected Group II Industries	40

Table2-9	Producer Price Index of Epoxy and Resins, 2012-2021 (2012= 100)	42

Table 3-1 Nationwide Cost Impacts (2021$) for Flares in the SOCMI Source Category that Control Emissions
from HON Processes including P&R I Flares	48

Table 3-2 Nationwide VOC and HAP Emissions Reductions and Cost-Effectiveness for Flares in the SOCMI
Source Category that Control Emissions from	48

Table 3-3 Nationwide Cost Impacts (2021$) for Flares that Control Emissions from P&R I Processes	49

Table 3-4 Nationwide VOC and HAP Emissions Reductions and Cost-Effectiveness for Flares that Control
Emissions from P&R I Processes	49

Table 3 -5 Nationwide Cost Impacts of Fenceline Monitoring for HON	50

Table 3 -6 Nationwide Cost Impacts of Fenceline Monitoring for P&R 1	50

Table 3-7 Nationwide Cost Impacts of Control Options Considered	51

Table 3 -8 Summary of Storage Vessel Control Options Evaluated for the HON	51

Table 3 -9 Nationwide Emissions Reductions and Cost Impacts of Control	52

Table 3-10 Nationwide Emissions Reductions and Cost Impacts of Control Options Considered for Storage
Vessels at P&R I Facilities (not collocated with HON facilities)	52

Table 3-11 VOC and HAP Cost Effectiveness for the Control Option Evaluated	53

iv

-------
Table 3-12 Summary of Continuous Process Vent Control Options Evaluated for the HON and P&R I NESHAP
54

Table 3-13 Nationwide Emissions Reductions and Cost Impacts of Control Options	55

Table 3-14 Nationwide Emissions Reductions and Cost Impacts of Control Options	55

Table 3-15 Average Cost and Emission Reductions for Process Vents Subject to the HON Used for the Suite of
Process Vent Requirements Evaluated	56

Table 3-16 Nationwide Emissions Reductions and Cost Impacts of Control Options Considered for Non-HON
Vent Streams Triggering NSPS Subparts	58

Table 3-17 Nationwide Emissions Reductions and Cost Impacts of Control Options	59

Table 3-18 Nationwide Emissions Reductions and Cost Impacts of Control Options	59

Table 3-19 Nationwide Emissions Reductions and Cost Impacts of Control Options	60

Table 3-20 Detailed Costs for the HON Source Category by Emission Point for	61

Table 3-21	Detailed Costs for the P&R I Source Category by Emission Point	64

Table 3-22 Detailed Costs for the P&R II Source Category by Emission Point	64

Table 3-23	Summary of the Total Costs by Final Rule ($2021)	64

Table 3-24 Discounted Costs, for the Final Amendments to the HON, P&R I, and P&R II NESHAP, and
Subparts Wb, Ilia, NNNa, and RRRa NSPS, 2024-2038 (million 2021$, discounted to 2023)	65

Table 3 -25 Summary of the HAP and VOC Emission Reductions per Year by Rule	66

Table 3-26 Summary of Emission Changes (Increases or Reductions) Other	67

Table 3-27 Summary of Monetized Greenhouse Gas Emission Changes in Tons Per Year	68

Table 4-1 Human Health Effects of Ambient Ozone and Whether they were Quantified And/or Monetized in
this RIA	77

Table 4-2 Synthetic Organic Chemicals: Benefit per Ton Estimates of Avoided Ozone-Attributable
Premature Mortality and Illness for the Rule, 2024-2038 (2021$)	83

Table 4-3 Synthetic Organic Chemicals: Benefit per Ton Estimates of Avoided NOx- Attributable
Premature Mortality and Illness for the Rule, 2024-2038 (2021$)	84

Table 4-4 Total Benefits Estimates of Ozone- and NOx- -Attributable Avoided Premature Mortality and Illness
(million 2021S) d	3-85

Table 4-5 Undiscounted Total Benefits Estimates of Ozone- and NOx-Attributable	86

Table 4-6 Annual Rounded SC-CO . SC-CII:. and SC-N () Values, 2024-2038	99

Table 4-7 Monetized Impacts of Estimated CO2, CH4, N20 Changes for the HON Amendments, P&R I and
P&R II	101

Table 4-8 Summary of Monetized Benefits PV/EAV for the HON Amendments, 2024-2038 (million 2021$),
Discounted to 2023	108

Table 4-9 Summary of Monetized Benefits PV/EAV for the P&R I Amendments, 2024-2038 (million 2021$),
Discounted to 2023	109

v

-------
Table 4-10 Summary of Monetized Benefits PV/EAV for the Cumulative Impact of the HON Amendments,
P&R I and P&R IINESHAP and Subpart Wb, Ilia, NNNa, and RRRa NSPS Amendments, 2024-2038 (million

2021$), Discounted to 2023	110

Table 4A-1 Interim Social Cost of Carbon Values, 2024-2038 (2021$/Metric Ton CO2)	111

Table 4A-2: Interim Social Cost of Methane Values, 2024-2038 (2021$ /Metric Ton CH4)	112

Table 4A-3: Interim Social Cost of Nitrous Oxide Values, 2024-2038 (2021$ /Metric TonN20)	112

Table 4A-4 Monetized Benefits of Estimated CO2, CH4, N20 Changes of the Proposed HON Amendments, P&R
I and P&R II NESHAP and Subpart Wb, Ilia, NNNa, and RRRa NSPS Amendments, 2024-2038, (million 2021$)
113

Table 5-1 Prices, Production, and Trade Quantities for the Seven Synthetic Organic Chemical Commodities

Selected (in Metric Tons)	117

Table 5 -2 Control Costs Attributed to Each Chemical Modeled (2021$)	119

Table 5-3 Elasticity Parameter Values and Sources	120

Table 5-4 Butadiene Results	124

Table 5-5 Styrene Simulation Results	124

Table 5-6 Acrylonitrile Simulation Results	125

Table 5-7 Acetone Simulation Results	125

Table 5-8 Ethylene Dichloride Simulation Results	126

Table 5-9 Ethylene Glycol Simulation Results	126

Table 5-10 Ethylene Oxide Simulation Results	127

Table 5-11 SBA Size Standards by NAICS Code	128

Table 5-12 Summary Statistics of Potentially Affected Entities	129

Table 5-13 Distribution of Estimated Compliance Costs by Rule and Size for Final Action ($2021)a	130

Table 5-14 Compliance Cost-to-Sales Ratio Distributions for Small Entities, Final Action	130

Table 5-15 Number and Extent of Impacts for Small Entities - Final Action3	131

Table 5-16 Chemical Sector Employment Information	133

Table 6-1 Summary of Monetized Benefits, Compliance Costs, and Net Benefits PV/EAV for HON, 2024-2038
(million 2021$, discounted to 2023)	137

Table 6-2 Summary of Monetized Benefits, Compliance Costs, and Net Benefits PV/EAV for P&R I, 2024-
2038 (million 2021$, discounted to 2023)	138

Table 6-3 Summary of Monetized Benefits, Compliance Costs, and Net Benefits PV/EAV for All Rules, 2024-
2038 (million 2021$, discounted to 2023)	139

vi

-------
LIST OF FIGURES

Figure 2-1 Global Price of Ethylene (USD$/metric ton)	24

Figure 2-2 Global Price of Butadiene from 2017 to 2019 with Estimated Figures for 2020 to 2022	26

Figure 2-3 P&R Group I and II Facilities Map	28

vii

-------
1 Executive Summary

1.1 Introduction

The U.S Environmental Protection Agency (EPA) is finalizing amendments to the
National Emissions Standards for Hazardous Air Pollutants (NESHAP) for subparts (40 CFR
part 63, subparts F, G, H, and I) that apply to the synthetic organic chemical manufacturing
industry (SOCMI) and to equipment leaks from certain non-SOCMI processes1 located at
chemical plants. These four NESHAP are more commonly referred to together as the Hazardous
Organic NESHAP (HON). The HON contains maximum achievable control technology (MACT)
standards for hazardous air pollutants (HAP) from heat exchange systems, process vents, storage
vessels, transfer racks, wastewater, and equipment leaks at chemical plants that are major sources
of HAP-producing SOCMI chemicals (e.g., bulk commodity chemicals) and for equipment leaks
for certain non-SOCMI chemical processes. The EPA is also revising NESHAP requirements for
storage tanks, loading operations, and equipment leaks to reflect cost-effective developments in
practices, processes, or controls of HAP.

The EPA is also finalizing amendments to the NESHAP for subparts (40 CFR part 63,
subparts U and W) that apply to the polymers and resins (P&R) Group I and II industries. P&R
Group I refers to major source facilities that produce certain elastomers and regulates HAP
emissions from nine different source categories. P&R Group I contains MACT standards for
HAP from storage tanks, process vents, equipment leaks, wastewater systems, and heat exchange
systems. P&R Group II applies to existing and new epoxy resins and non-nylon polyamides
production operations that are located at major sources. Similarly, P&R Group II contains
MACT standards for HAP from storage tanks, process vents, equipment leaks, and wastewater
systems.

The EPA is also finalizing New Source Performance Standards (NSPS) to reflect best
system of emissions reduction for four SOCMI NSPS subparts (40 CFR part 60, subparts III,
NNN, RRR, and VV) for emissions of volatile organic compounds (VOC) from SOCMI air
oxidation unit processes, SOCMI distillation operations, SOCMI reactor processes, and

1 NESHAP Subpart I provides the applicability criteria for the non-SOCMI processes subject to the negotiated
regulation for equipment leaks and requires owners and operators to comply with subpart H.

1

-------
equipment leaks located at SOCMI sources. The rule also includes revisions related to emissions
during periods of startup, shutdown, and malfunction (SSM); additional requirements for
electronic reporting of performance test results, performance evaluation reports, and compliance
reports; revisions to monitoring and operating requirements for control devices; and other minor
technical improvements.

1.1.1 NESHAPfor subparts F, G, H, I, U,&W

The statutory authority for the NESHAP amendments is provided by sections 112 and
301 of the Clean Air Act (CAA), as amended (42 U.S.C. 7401 et seq.). Section 112 of the CAA
establishes a two-stage regulatory process to develop standards for emissions of HAP from
stationary sources. Generally, the first stage involves establishing technology-based standards
and the second stage involves evaluating those standards that are based on maximum achievable
control technology (MACT) to determine whether additional standards are needed to address any
remaining risk associated with HAP emissions. In the second stage of the regulatory process, the
CAA requires the EPA to undertake two different analyses, which we refer to as the technology
review and the residual risk review. Under the technology review, we must review the
technology-based standards and revise them "as necessary (taking into account developments in
practices, processes, and control technologies)" no less frequently than every 8 years, pursuant to
CAA section 112(d)(6). Under the residual risk review, we must evaluate the risk to public
health remaining after application of the technology-based standards and revise the standards, if
necessary, to provide an ample margin of safety to protect public health or to prevent, taking into
consideration costs, energy, safety, and other relevant factors, an adverse environmental effect.
The residual risk review is required within 8 years after promulgation of the MACT standards,
pursuant to CAA section 112(f).

In the first stage of the CAA section 112 standard setting process, the EPA promulgates
technology-based standards under CAA section 112(d) for categories of sources identified as
emitting one or more of the HAP listed in CAA section 112(b). Sources of HAP emissions are
either major sources or area sources, and CAA section 112 establishes different requirements for
major source standards and area source standards. "Major sources" are those that emit or have
the potential to emit 10 tons per year (tpy) or more of a single HAP or 25 tpy or more of any
combination of HAP. All other sources are "area sources." For major sources, CAA section

2

-------
112(d)(2) provides that the technology-based NESHAP must reflect the maximum degree of
emission reductions of HAP achievable (after considering cost, energy requirements, and non-air
quality health and environmental impacts). These standards are commonly referred to as MACT
standards. CAA section 112(d)(3) also establishes a minimum control level for MACT standards,
known as the MACT "floor." In certain instances, as provided in CAA section 112(h), the EPA
may set work practice standards in lieu of numerical emission standards. The EPA must also
consider control options that are more stringent than the floor. Standards more stringent than the
floor are commonly referred to as beyond-the-floor standards. For area sources, CAA section
112(d)(5) allows the EPA to set standards based on generally available control technologies or
management practices (GACT standards) in lieu of MACT standards. For categories of major
sources and any area source categories subject to MACT standards, the second stage in standard-
setting focuses on identifying and addressing any remaining {i.e., "residual") risk pursuant to
CAA section 112(f) and concurrently conducting a technology review pursuant to CAA section
112(d)(6). MACT standards were finalized for the HON source category in 1994. The residual
risk and technology review (RTR) was finalized in 2006.

The MACT standards for P&R Group I (40 CFR part 63, subpart U) were initially
promulgated in 1996. Most recently, the agency conducted its RTR of the Group I NESHAP in
2008, for four source categories, and in 2011, for the remaining source categories. The MACT
standards for P&R Group II (40 CFR part 63, subpart W) were initially promulgated in 1995,
with the agency most recently conducting its RTR of the Group II NESHAP in 2008.

The source categories that are the subject of this rulemaking include the HON source
category (the facilities, sources and processes of which we often refer to as "HON facilities,"
"HON sources," and "HON processes") and several Polymers and Resins Production source
categories covered in P&R Group I and II (see section II.B of the preamble for detailed
information about the source categories). The North American Industry Classification System
(NAICS) code for SOCMI facilities begins with 325, for P&R I is 325212, and for P&R II is
325211. The list of NAICS codes is not intended to be exhaustive, but rather provides a guide for
readers regarding the entities that this final action is likely to affect.

As defined in the Initial List of Categories of Sources Under Section 112(c)(1) of the
CAA Amendments of 1990 (see 57 FR 31576, July 16, 1992) and Documentation for

3

-------
Developing the Initial Source Category List, Final Report (see EPA-450/3-91-030, July 1992),
the SOCMI source category is any facility engaged in "manufacturing processes that produce
one or more of the chemicals [listed] that either (1) use an organic HAP as a reactant or (2)
produce an organic HAP as a product, co-product, by-product, or isolated intermediate." Related
chemicals for the HON and P&R Group I and II source categories are listed in the Industry
Profile section of this report.

This action constitutes another CAA section 112(d)(6) technology review for the HON,
P&R I, and P&R II. This action also constitutes an updated CAA section 112(f) risk review
based on new information for the HON and for affected sources producing neoprene subject to
P&R I. We note that although there is no statutory CAA obligation under CAA section 112(f) for
the EPA to conduct a second residual risk review of the HON or standards for affected sources
producing neoprene subject to P&R I, the EPA retains discretion to revisit its residual risk
reviews where the Agency deems that is warranted; for additional information on statutory
authority please see section II. A of the preamble.

1.1.2 NSPS subparts III, NNN, RRR, & VVb

The EPA's authority for the NSPS is CAA section 111, which governs the establishment
of standards of performance for stationary sources. CAA section 111(b)(1)(A) requires the EPA
Administrator to list categories of stationary sources that in the Administrator's judgment cause
or contribute significantly to air pollution that may reasonably be anticipated to endanger public
health or welfare. The EPA must then issue performance standards for new (and modified or
reconstructed) sources in each source category pursuant to CAA section 111(b)(1)(B). These
standards are referred to as new source performance standards, or NSPS. The EPA has the
authority under CAA section 111(b) to define the scope of the source categories, determine the
pollutants for which standards should be developed, set the emission level of the standards, and
distinguish among classes, type, and sizes within categories in establishing the standards.

Section 111(b)(1)(B) of the CAA requires the EPA to "at least every 8 years review and,
if appropriate, revise" new source performance standards. Section 111(a)(1) of the CAA provides
that performance standards are to "reflect the degree of emission limitation achievable through
the application of the best system of emission reduction which (taking into account the cost of
achieving such reduction and any non-air quality health and environmental impact and energy

4

-------
requirements) the Administrator determines has been adequately demonstrated." We refer to this
level of control as the best system of emission reduction or "BSER." The term "standard of
performance" in CAA 111(a)(1) makes clear that the EPA is to determine both the BSER for the
regulated sources in the source category and the degree of emission limitation achievable
through application of the BSER. The EPA must then, under CAA section 111(b)(1)(B),
promulgate standards of performance for new sources that reflect that level of stringency. These
subparts were originally promulgated pursuant to CAA section 111(b) on June 29, 1990
(subparts III and NNN); August 31, 1993 (subpart RRR); and November 16, 2007 (subpart VV).

The emission sources covered by these NSPS subparts are comparable (and in many
instances are the same) as HON sources subject to the standards for process vents (NSPS III,
NNN, and RRR) and equipment leaks (NSPS VV), though regulated pollutants and definitions of
what constitutes an affected source/affected facility are different between the NESHAP and
NSPS. The emissions limits for these NSPS regulate VOC that are in many instances also HAP.

1.2 Market Failure

Many regulations are promulgated to correct market failures, which otherwise lead to a
suboptimal allocation of resources within a market. Air quality and pollution control regulations
address "negative externalities" whereby the market does not internalize the full opportunity cost
of production borne by society as public goods such as air quality are unpriced.

While recognizing that the optimal social level of pollution may not be zero, HAP and
VOC emissions impose costs on society, such as negative health and welfare impacts (that are
discussed in the benefits chapter of this RIA), that are not reflected in the market price of the
goods produced through the polluting process. For this regulatory action, the goods produced are
chemical products (e.g., butadiene, ethylene oxide). If processes of producing chemicals, such as
butadiene or ethylene oxide, yield pollution emitted into the atmosphere, the social costs
imposed by the pollution will not be borne by the polluting firms but rather by society as a
whole. In addition, in the case of this regulation, it's not just the pollution from the process that
poses a negative externality, but the fugitive emissions of the goods themselves (e.g., ethylene
oxide itself is harmful to society, not just the pollution from the process of making ethylene
oxide). Thus, the producers are imposing a negative externality, or a social cost from these
emissions, on society. The equilibrium market price of chemical products such as butadiene or

5

-------
ethylene oxide may fail to incorporate the full opportunity cost to society of consuming the
chemical product. Consequently, absent a regulation or some other action to limit such
emissions, producers will not internalize the negative externality of pollution due to emissions
and social costs will be higher as a result. This regulation will serve to address this market failure
by causing affected producers to begin internalizing the existing negative externality associated
with HAP and other emissions also affected by this rule such as VOC.

1.3 Results for the Final Action

We present benefits and costs for each final rule and their more and less stringent
alternatives. We group the impacts of the Ilia, NNNa, and RRRa NSPS together for
presentational clarity and consistency with the presentation of impacts for these three NSPS in
the preamble and the remainder of the materials for this final action. All benefits and costs
outlined in this RIA are estimated as the change from the baseline, which reflects the current
business practice for the affected sources. There is no monetized estimate of the benefits for the
HAP emission reductions expected to occur as a result of this final action. We do present
monetized estimates for other impacts of this action, such as benefits from both short- and long-
term reduced exposure to ozone caused by VOC emissions reductions, benefits from decreases in
CH4 emissions, and disbenefits from increases in CO2 and N2O emissions.

1.3.1 Baseline for the Regulation

The impacts of regulatory actions are evaluated relative to a baseline that represents, to
the extent possible, the world without the regulatory action. Throughout this document, the EPA
focuses the analysis on the final regulatory requirements that result in quantifiable compliance
cost or emissions changes compared to the baseline as identified above. For each rule and most
emissions sources, the EPA assumed each facility achieved emissions control meeting current
standards and estimated emissions reductions and cost relative to this baseline. The baseline for
the analysis includes "excess emissions" from SOCMI sources, which are emissions from the
source category that should be controlled to current standards but in practice are not.

We calculate cost and emissions reductions relative to the baseline (inclusive of excess
emissions) for the period 2024-2038. This time frame spans the time period from when the
NSPSs take effect (under the presumption that these rules are finalized in 2024 as per the consent

-------
decree under which this rulemaking is being prepared) through the lifetime of the typical capital
equipment (15 years) expected to be installed as a result of the NESHAP and NSPS amendments.

The summaries of impact results below are for the final options. In accordance with the
OMB Circular A-4 (US OMB, 2003),2 we also present impact results for a more stringent and
less stringent set of options as defined by that circular, which is the guidance for regulatory
analysis to be followed by Federal agencies preparing an RIA such as this one. These alternatives
are defined in Chapter 6, where results are presented for these options along with those for the
final policy option.

1.3.2 Overview of Costs and Benefits for the Final Options

The amendments to this chemical sector package constitute a significant regulatory
action. This action is significant, under Executive Order 14094, because it likely to have an
annual effect on the economy of $200 million or more or adversely affect in a material way the
economy, a sector of the economy, productivity, competition, jobs, the environment, public
health or safety, or state, local, or tribal governments or communities. The EPA monetized the
projected benefits of reducing VOC emissions in terms of the value of avoided ozone and NOx-
attributable deaths and illnesses, both short- (ST) and long-term (LT). The EPA also monetized
the benefits and disbenefits from changes in emissions of greenhouse gases (GHG) such as
carbon dioxide (CO2), nitrous oxide (N2O), and methane (CH4). In addition, we include a
discussion of the unmonetized benefits that will result from this final rulemaking.

Table 1-1 through Table 1-5 present the projected ozone health benefits, climate benefits
from CH4 reductions, climate disbenefits from C02and N2O, compliance costs, net benefits, and
unmonetized HAP emission reductions from the amendments to each regulatory action included
in this rulemaking. The projected climate disbenefits are caused by combusting emissions in
flares, thermal oxidizers, and increased electricity usage for the emission controls included in the
cost analysis. Projected climate benefits are caused by reduction of CH4 emissions from control

2 U.S. Office of Management and Budget. Circular A-4, "Regulatory Analysis." September 17, 2003. Available at
https://www.whitehouse.gov/wp-content/uploads/legacY drupal files/omb/circulars/A4Za-4.pdf. Draft revisions
to Circular A-4 became final in November 2023 but will not become effective until after this rule is promulgated.
Hence, the 2003 Circular A-4 will be a basis for the analyses in this RIA.

-------
of flares. Certain control options analyzed in this RIA lead to chemical product recovery, which
has been monetized as product recovery credits that are incorporated into the annual cost
calculations. Net compliance costs are calculated as total compliance costs minus product
recovery credits. For a discussion of product recovery, see Chapter 3. Monetized net benefits are
projected to be negative under each discount rate used in the analysis. The benefits from HAP
reductions (annually) and VOC reductions outside of the ozone season (May-September) have
not been monetized for this action. If we were able to monetize these beneficial impacts, it would
increase the net benefits of this analysis. In addition, the emissions control of HAP reductions is
tied to bringing the cancer risks down to an acceptable level, so while we are unable to monetize
what that HAP reduction is, that the reduction quantity is tied to a health-based threshold.

As mentioned earlier, we calculate cost and emissions reductions relative to the baseline
for the period 2024-2038, with costs discounted to 2023.

Table 1-1 Monetized Benefits, Compliance Costs, Emission Reductions and Net
Benefits for Amendments to the HON (dollars in million 2021$3)a

3 Percent Discount Rate

7 Percent Discount Rate

EAV

Monetized Health
Benefits'3

70 and 630

5.9 and 53

48 and 420

5.2 and 46

Climate Disbenefits0

140

Net Compliance
Costs'1

1,550

130

1,200

130

Compliance Costs

1,560

130

1,200

130

Value of Product

Recovery

Net Benefits

(1,600) and (1,100)

(140) and (88)

(1,300) and (920)

(140) and (95)

1,107 tons of HAP emission reductions. Health effects from reduced exposure to
ethylene oxide, chloroprene, benzene, 1,3-butadiene, vinyl chloride, ethylene
dichloride, chlorine, maleicanhydride, and acrolein

Nonmonetized
Benefits

a Totals may not sum due to independent rounding. Numbers rounded to two significant digits unless otherwise
noted. A number in parentheses denotes a negative value.

b Monetized health benefits include ozone related health benefits associated with reductions in VOC emissions. The
health benefits are associated with several point estimates and are presented at real discount rates of 3 and 7 percent.
The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions (annually) and VOC
reductions (outside of the ozone season) remain unmonetized and are thus not reflected in the table. The
unmonetized effects also include disbenefits resulting from the secondary impact of an increase in CO emissions.
0 Monetized climate benefits and disbenefits are based on increases in CO2 and N20 emissions and decreases in CH4
emissions and are calculated using three different estimates of the social cost of each greenhouse gas (SC-GHG)

3 When necessary, dollar figures in this RIA have been converted to 2021$ using the annual GDP Implicit Price
Deflator values in the U. S. Bureau of Economic Analysis' (BEA) NIPA Table 1.1.9 found at
.

-------
(under 1.5 percent, 2.0 percent, and 2.5 percent near-term Ramsey discount rates). For the presentational purposes of
this table, we show the net climate disbenefits associated with the SC-GHG at a 2 percent near-term Ramsey
discount rate.

dNet compliance costs are the engineering control costs minus the value of recovered product.

Table 1-2 Monetized Benefits, Compliance Costs, and Net Benefits for Amendments to
	P&R I (dollars in million 2021$)a	



3 Percent Discount Rate

7 Percent Discount Rate



PV

EAV

PV

EAV

Health Benefits'3

(0.2) and (1.7)

(0.02) and (0.1)

(0.2) and (1.5)

(0.02) and (0.2)

Climate Disbenefits0

22

2

22

2

Net Compliance Costs'1

140

12

110

12

Compliance Costs

140

12

110

12

Value of Product Recovery

1

0.2

1

0.1

Net Benefits

(160) and (160)

(14) and (14)

(130) and (130)

(14) and (14)

Nonmonetized Benefits

264 tons per year (tpy) of HAP reductions, including an approximate 14
tpy reduction in chloroprene emissions

a Totals may not sum due to independent rounding. Numbers rounded to two significant digits unless otherwise
noted. A number in parentheses denotes a negative value.

b Monetized health benefits include ozone related health benefits associated with reductions in VOC emissions. The
health benefits are associated with several point estimates and are presented at real discount rates of 3 and 7 percent.
The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions and VOC reductions
outside of the ozone season remain unmonetized and are thus not reflected in the table. The unmonetized effects also
include disbenefits resulting from the secondary impact of an increase in CO emissions.

0 Monetized climate benefits and disbenefits are based on increases in CO2 and N20 emissions and decreases in CH4
emissions and are calculated using three different estimates of the social cost of each greenhouse gas (SC-GHG)
(under 1.5 percent, 2.0 percent, and 2.5 percent near-term Ramsey discount rates). For the presentational purposes of
this table, we show the net climate disbenefits associated with the SC-GHG at a 2 percent near-term Ramsey
discount rate.

dNet compliance costs are the engineering control costs minus the value of recovered product. A negative net
compliance costs occurs when the value of the recovered product exceeds the compliance costs.

Table 1-3 Monetized Benefits, Compliance Costs, and Net Benefits for Amendments

to P&R II (dollars in million 2021$)a



3 Percent Discount Rate

7 Percent Discount Rate



PV

EAV

PV

EAV

Monetized Health Benefits'3

0

0

0

0

Climate Disbenefits

0

0

0

0

Net Compliance Costs0

15

1.3

11

1.3

Compliance Costs

15

1.3

11

1.3

Value of Product Recovery

0

0

0

0.0

Net Benefits

(15)

(1.3)

(11)

(1.3)

Nonmonetized Benefits

1 ton/year of HAP emission reduction. Reduced health exposure to
	epichlorohydrin	

Totals may not sum due to independent rounding. Numbers rounded to two significant digits unless otherwise
noted.

b There are no monetized health benefits for this final rule. There are also no climate benefits or disbenefits for the
final amendments to P&R II. The unmonetized effects also include disbenefits resulting from the secondary impact
of an increase in CO emissions.

9

-------
0 Net compliance costs are the engineering control costs minus the value of recovered product. A negative net
compliance costs occurs when the value of the recovered product exceeds the compliance costs. For the final
amendments to P&R II, there is no product recovery.

Table 1-4 Monetized Benefits, Compliance Costs, and Net Benefits for NSPS
subpart Wb (dollars in million 2021$)a

3 Percent Discount Rate 7 Percent Discount Rate

PV EAV PV EAV

Monetized Health Benefits'3 1.3 and 12 0.1 and 1.0 0.9 and 7.9 0.1 and 0.9

Net Compliance Costs0 11 0.9 8.0 0.9

Compliance Costs 15 1.2 11 1.2

Value of Product Recovery 3.7 0.3 2.8 0.3

Net Benefits (9.7) and 1 (0.8) and 0.1 (7.1) and (0.1) (0.8) and (0.0)

a Totals may not sum due to independent rounding. Numbers rounded to two significant digits unless otherwise
noted.

b Monetized benefits include ozone related health benefits associated with reductions in VOC emissions. The health
benefits are associated with several point estimates and are presented at real discount rates of 3 and 7 percent. The
two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The estimates
do not represent lower- and upper-bound estimates. Benefits from HAP reductions and VOC reductions outside of
the ozone season remain unmonetized and are thus not reflected in the table. There are no climate benefits or
disbenefits associated with this NSPS.

0 Net compliance costs are the engineering control costs minus the value of recovered product. A negative net
compliance costs occurs when the value of the recovered product exceeds the compliance costs.

Table 1-5 Monetized Benefits, Compliance Costs, and Net Benefits for

Amendments to Subparts Ilia, NNNa, and RRRa (dollars in million 2021$)a

3 Percent Discount Rate 7 Percent Discount Rate

EAV

Monetized Health Benefits'3

6 and 54

0.5 and 4.6

5.3 and 47

0.6 and 5.2

Climate Disbenefits0

0.3

Net Compliance Costs'1

4.9

5.2

Compliance Costs

4.9

5.2

Value of Product Recovery

Net Benefits

(56) and (8)

(4.7) and (0.6)

(46) and (4)

(4.9) and (0.3)

a Totals may not sum due to independent rounding. Numbers rounded to two significant digits unless otherwise
noted. A number in parentheses denotes a negative value.

b Monetized health benefits include ozone related health benefits associated with reductions in VOC emissions. The
health benefits are associated with several point estimates and are presented at real discount rates of 3 and 7 percent.
The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions and VOC reductions
outside of the ozone season remain unmonetized and are thus not reflected in the table. The unmonetized effects also
include disbenefits resulting from the secondary impact of an increase in CO emissions.

0 Monetized climate benefits and disbenefits are based on increases in CO2 and N20 emissions and decreases in CH4
emissions and are calculated using three different estimates of the social cost of each greenhouse gas (SC-GHG)
(under 1.5 percent, 2.0 percent, and 2.5 percent near-term Ramsey discount rates). For the presentational purposes of
this table, we show the net disbenefits associated with the SC-GHG at a 2 percent near-term Ramsey discount rate.
dNet compliance costs are the engineering control costs minus the value of recovered product.

-------
1.3.2.1 Overview of Costs and Benefits for All Rules (Cumulative Impacts)

Table 1-6 presents the cumulative projected monetized health benefits, climate
disbenefits (inclusive of climate benefits as with the HON summary impacts table above), and
compliance costs (net of product recovery). This final action will reduce HAP and VOC
emissions from HON, P&R I, and P&R II emission sources as well as the NSPS SOCMI air
oxidation unit processes, distillation operations, reactor processes, and equipment leaks sources.
Considering reported emissions inventories for ethylene oxide and chloroprene, we estimate that
the final amendments to the NESHAP will reduce overall HAP emissions from the SOCMI
source category by approximately 1,107 tpy, reduce overall HAP emissions from the P&R I
source categories by approximately 264 tpy, and reduce overall HAP emissions from the P&R II
source categories by approximately 1 tpy. We note that these emissions reductions do not
consider the potential excess emissions reductions from flares that could result from the final
monitoring requirements; we estimate flare excess emissions reductions of 4,858 tpy for HAP
and 19,889 tpy for VOC. Based on our analysis of the finalized actions described in sections
III.B.2, III.D.2, and III.E of this preamble for the NSPS, we estimate that the final amendments
to the NSPS would reduce VOC emissions from the SOCMI source category by approximately
1,622 tpy. Emission reductions and secondary impacts (e.g., emission increases associated with
supplemental fuel or additional electricity) by rule are listed below. The only change in air
impacts since proposal stems from our reevaluation related to the TRE removal for HON and
P&R I (based on comments received as discussed in sections IV.B.3.a.i and IV.B.3.b.i of the
preamble).

There are also emission increases per year in criteria pollutants of 17.4 tons of particulate
matter, 349 tons of nitrogen oxides (NOx), and 1.37 tons of sulfur dioxide (SO2) due to
additional energy usage from the controls applied in the cost analysis. Finally, there are emission
increases per year of 741,102 tons of carbon dioxide (CO2) and 6.86 tons of nitrous oxide (N2O),
and emission decreases per year of 22,951 tons of methane (CH4). Table 3-26 contains the
changes in emissions other than for HAP and VOC. Thus, there are projected climate benefits
caused by CH4 emission reductions, and projected climate disbenefits caused by CO2 and N2O
emissions increases. Table 3-27 contains the GHG emission estimates that were monetized for
this rule. While benefits from HAP reductions and VOC reductions outside of the ozone season

11

-------
have not been monetized for this action, the EPA expects these benefits are positive. As
mentioned earlier, we calculate cost and emissions reductions relative to the baseline for the
period 2024-2038, discounted to 2023.

Table 1-6 Total Monetized Benefits, Compliance Costs, Emission Reductions, and Net
Benefits for the Final Rule (dollars in million 2021$)a

3 Percent Discount Rate 7 Percent Discount Rate

PV EAV PV EAV

Monetized Health Benefits'3 77 and 690 6.5 and 58 53 and 475 5.9 and 52

Climate Disbenefits0 160 13 160 13

Net Compliance Costsd 1,770 150 1,370 150

Compliance Costs 1,790 150 1,380 150

Value of Product Recovery 16 1.3 12 1.3

Net Benefits (1,900) and (1,200) (160) and (110) (1,500) and (1,100) (160) and (110)

6,053 tons/year of HAP

.. . , „ r. Health effects of reduced exposure to ethylene oxide, chloroprene, benzene, 1,3-

Nnnmnnpti7Pf1 Rptiptits

butadiene, vinyl chloride, ethylene dichloride, chlorine, maleic anhydride and
acrolein

a Totals may not sum due to independent rounding. Numbers rounded to two significant digits unless otherwise
noted. A number in parentheses denotes a negative value.

b Monetized health benefits include ozone related health benefits associated with reductions in VOC emissions. The
health benefits are associated with several point estimates and are presented at real discount rates of 3 and 7 percent.
The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions and VOC reductions
outside of the ozone season remain unmonetized and are thus not reflected in the table. The unmonetized effects also
include disbenefits resulting from the secondary impact of an increase in CO emissions.

1.4 Organization of the Report

The remainder of this report details the methodology and the results of the RIA. Chapter
2 presents a profile of the SOCMI and P&R Group I and II industries, which also cover the
industries with sources affected by the NSPS amendments included in this rulemaking. Chapter 3
describes emissions, emissions control options, and engineering costs. Chapter 4 presents the
benefits analysis, including the monetized health benefits from VOC and other emission
reductions, a qualitative discussion of the unmonetized benefits associated with HAP emissions
reductions, and the monetized benefits associated with climate emissions decreases (CH4) and
disbenefits associated with climate (CO2 and N2O) emissions increases. Chapter 5 presents
analyses of economic impacts, impacts on small businesses, and a narrow analysis of

-------
employment impacts. The economic impacts include estimates of price and output changes in
response to the costs of different final rules in this rulemaking. The small business impact
analysis includes estimates of annual cost to sales calculations for affected small businesses and
concludes that no rule in this rulemaking will have a significant impact on a substantial number
of small entities (or SISNOSE). Chapter 6 presents a comparison of the benefits and costs.
Chapter 7 contains the references for this RIA.

13

-------
2 INDUSTRY PROFILE

2.1 Introduction

This chapter will provide a profile of SOCMI and P&R Group I and II industries affected
under this combined rulemaking. While there is overlap between these rules, affected facilities
and area sources are distinct enough that this chapter will provide separate sub-chapters for
SOCMI and P&R Group I and II below. The EPA constructed facility lists for these rules are
based on data from the January 2021 version of the 2017 National Emissions Inventory (NEI).
However, instances where facility-specific data was not available in the 2017 NEI, more recent
data was collected from the 2018 inventory or recent state submittals to the Emissions Inventory
System (EIS).4 The construction of the facility list is described in the preamble for the final
action.

2.2 SOCMI Industry Profile

The SOCMI source category includes chemical manufacturing processes producing
commodity chemicals, see sections I.B and II.B of the preamble for detailed information about
these source categories. An EPA regulatory impact analysis from 1994 identified approximately
30 key chemicals that represent a large portion of output from the industry.5 This profile revisits
these chemicals and their feedstocks, listed in Table 2-1, to provide an updated industry profile.

4 Revenue and employment information was collected through manual search of D&B Hoover's database in 2022.

5 U.S. EPA. 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 Other Process Subject to the Negotiated Regulation for Equipment Leaks. EPA-453/R-94-019.
March 1994.

-------
Table 2-1 Select SOCMI Chemicals by Feedstock*

Benzene

Methane

Styrene-Butadiene Rubber

Formaldehyde

Cyclohexylamine

Chloroform

Hydroquinone

Methyl Tertiary Butyl Ether

Styrene

Methyl Chloride

Acetone

Ethylene

Bisphenol-A

Butadiene

Propylene Glycol

Polybutadiene

Toluene

Ethylene Dichloride

Benzoic Acid

Ethylene Oxide

Xylene

Ethylene Glycol

Terephthalic Acid

Triethylene Glycol

Phthalic Anhydride

Propylene

Naphthalene

Acrylonitrile

Ether

Butylene

*This list of chemicals is from the HON Regulatory Impact Analysis (EPA, 1994)

Synthetic organic chemicals (SOCs) are derived from chemical reactions using
feedstocks containing carbon, such as fossil fuels like oil and natural gas. Supply relies on the
market prices of these feedstocks, but advancements in technology and energy efficiency have
resulted in large production economies of scale. The main source of demand for SOCs is from
plastics manufacturers. In addition, there is demand from a multitude of other industries that
produce, but not limited to, rubber, paints, adhesives, food, and pesticides (Barnicki, 2017).

Existing overall market and industry research for SOCs is scarce. The SOC market in the
United States was valued at $168 billion in 2022 (IBISWorld). SOCMI is a subsector of the
much larger organic chemicals market, which includes natural organic chemicals. Seven of the
eight major feedstocks (excluding naphthalene) belong to a subset of SOCs called
petrochemicals, named for their derivation from crude oil and natural gas, in addition to other
possible sources like coal or vegetable oils. Petrochemicals can be used to make a variety of
products, including rubber, fuel, cleaning agents, and plastics (ScienceDirect, 2022).

-------
The global petrochemicals market size value was $556.1 billion in 2021. An industry
report from Grand View Research prepared in 2021 forecasts the petrochemical market to grow
at a compound annual growth rate (CAGR) of 6.2 percent from 2022 to 2030. The growth in
demand is expected to result from an increase in demand for downstream products from various
end-use industries such as construction, pharmaceuticals, and automotive. Because crude oil is
the basic raw material in production, crude oil price volatility significantly affects production and
the final costs of petrochemical products (Grand View Research, 2021).

Ethylene had the largest revenue share for the petrochemical industry, over 40.0 percent
in 2021. The large revenue share from the ethylene market is due to the wide variety of everyday
products that use this chemical. Ethylene is developed into four different compounds that
produce many products, including:

• Polyethylene (Plastics) - used to make food packaging, bottles, bags, and other
plastics-based goods.

• Ethylene Oxide / Ethylene Glycol - used in polyester for textiles, as well as
antifreeze for airplane engines and wings.

• Ethylene Dichloride - used in a vinyl product used in PVC pipes, siding, medical
devices, and clothing.

• Styrene - synthetic rubber found in tires and foam insulation.

Ongoing industrialization and growing automotive and packaging sectors in emerging
economies such as India, Brazil, Vietnam, and Thailand are forecasted to drive up demand for
ethylene products. Butadiene was the second-largest product segment in 2021. Methanol is
predicted to have the fastest revenue growth, a CAGR of 7.8 percent, over the forecast period.
Methanol is a chemical building block for hundreds of everyday products, including plastics,
paints, car parts and construction materials. Methanol also is a clean energy resource used to fuel
cars, trucks, buses, ships, fuel cells, boilers and cook stoves. There is increasing demand for
methanol from industries such as construction, paints and adhesives, pharmaceuticals, plastics,
and automotive (OEC, 2022a).

The Asia Pacific region has a volume share of over 50.0 percent of the petrochemicals
industry. Increasing natural gas exploration activities in the United States and Canada will grow
the petrochemicals market in North America over the coming years; additionally, this provides

-------
an opportunity for substituting some raw materials with natural gas in the production of several
petrochemicals (Grand View Research, 2021).

In 2018, total global trade of petrochemicals was valued at $123 billion. From 2017 to
2018, exports of petrochemicals worldwide grew by 42.5 percent, from $86.5 billion. The top
exporters in 2018 were Saudi Arabia ($17.0 billion), the United States ($12.8 billion), Germany
($9.8 billion), Belgium ($7.5 billion), and Thailand ($7.0 billion). Of United States exports, 26.4
percent went to Mexico, 21.2% to Canada, and 9.0 percent to China.

The top importers of petrochemicals were China ($19.4 billion), Germany ($7.2 billion),
the United States ($6.4 billion), Italy ($4.7 billion), and Turkey ($4.2 billion). The United States
imported 42 percent of its petrochemicals from Canada, 19.9 percent from Mexico, 6.2 percent
from Germany, and the remainder largely from Asia (OEC, 2022a). Hence, the U.S. was a net
exporter in 2018 with exports at nearly twice the size of imports in monetary terms.

As more natural gas is tapped in the United States and prices decrease (as of 2021), the
United States has become increasingly cost competitive in this market worldwide. U.S.
production of ethylene, produced primarily in natural gas processing plants, has grown rapidly
since 2013. Production has nearly doubled from 0.95 million barrels per day (b/d) in the first
quarter of 2013 to 1.85 million b/d in the first quarter of 2021 (EIA, 2021).

More U.S. firms are keeping production in the United States instead of outsourcing to
foreign countries or using imported oil. For example, firms like Dow Chemical, Exxon Mobile,
Chevron Phillips Chemical, and Royal Dutch Shell have all invested in new ethylene plants and
projects in the United States over the last several years, causing a significant increase in ethylene
production (Pearce, 2014). As of 2021, the United States held approximately 40 percent of the
world's ethane petrochemical production capacity (Novicio, 2021). In 2022, the world gas
supply chain was disrupted by the war in Ukraine. However, because of limited domestic LNG
shipping capacity, U.S. gas prices are likely to remain below global market prices, continuing to
give the U.S. petrochemical manufactures a slight competitive advantage.

Table 2-2 summarizes the top ten globally produced SOCs by total market value, with
U.S. trade statistics and their common use cases. U.S. exports and imports include their global
rank (in parentheses) if they fall within the top five global importers and exporters of that
commodity.

-------
Table 2-2 Top 10 Globally Produced SOCs by Total Market Value (2021$)

Chemical

Total Production

Total

U.S.

U.S.

Uses



(year)

Global
Trade

Exports
(global
rank)

Imports
(global
rank)



Xylene

$178.45B (2021)

$2.18B

$49.9M

$26.1M

Production of drugs and
plastics; solvents;
intermediate for dyes and
organic synthesis,
especially isophthalic acid;
insecticides; aviation fuel;
manufacturing of polyester
and alkyl resins; fibers,
films, and resins;
herbicide; production of
polyester polyurethanes
used in paints and sealants

Propylene

$96.47B (2021)

$5.59B

$559M (3)

$142M

Plastics and carpet fibers;
chemical intermediate for
the manufacture of
acetone, isopropylbenzene,
isopropanol, isopropyl
halides, propylene oxide,
acrylonitrile, and cumene;
production of gasoline or
used as a fuel in oil
refineries

Ethylene

$81.34B (2020)

$4.95B

$401M (5)

$191K

Oxyethylene welding;
chemical manufacturing;
fruit ripening; general
anesthetic; common
ingredient in household
products, such as plastics,
certain foods, and some
detergents; manufacturing
ethylene oxide;
polyethylene for plastics,
alcohol, mustard gas, and
other organics

Benzene

$68.3B (2021)

$4.75B

$38.7M

$632M (2)

Solvent for chemical
synthesis, constituent in
motor fuels, detergents,
explosives,

pharmaceuticals, dyestuffs

Terephthalic Acid

$49.2B (2020)

$4.12B

$23.3M

$414M (2)

Feedstock for the
production of polyesters,
such as PET; wool
processing; production of
plastic films and sheets;
added to certain poultry
feeds and antibiotics to
increase effectiveness

Styrene

$34.23B (2022)

$7.22B

$1.65B (1)

$387M (5)

Polystyrene production
(low cost, low friction

18

-------


plastic used in packaging,
textiles, and construction)

Toluene

$21.15B (2021)

$1.62B

$40.2M

$133M (4)

Solvent in aviation and
automotive fuels; chemical
production; production of
paints, paint thinners,
fingernail polish, lacquers,
adhesives, and rubber;

Bisphenol-A

$16.23B (2020)

$1.46B

$2.88M

$37.2M

printing and leather tanning
processes; production of
benzene, TNT, nylon,
plastics, and polyurethanes
Production of
polycarbonate plastics and
epoxy resins

Acrylonitrile

$12.9B (2020)

$2.03B

$584M (1)

$9.26M

Manufacture of acrylic and
modacrylic fibers,
production of plastics

Styrene-Butadiene
Rubber

$10.24B (2020)

$4.49B

$259M

$472M (2)

Rubber products such as
gloves, tires, and adhesives

2.2.1 Oil and Gas Sectors and SOCMI

Olefins6 (ethylene, propylene, butadiene and butenes) are derived from both natural gas
and petroleum. The aromatics (benzene, toluene, and xylenes) are derived from petroleum and,
minorly, coal. Whether natural gas fractions or petroleum are used for olefins varies throughout
the world depending on the availability of natural gas and demand for gasoline. Both light and
heavy naphthas7 are petroleum fractions that can be used to make olefins. But they can also be
used to make gasoline (Wittcoff, 2012).

In the United States, approximately 95 percent of all organic chemicals by weight are
derivatives of petroleum and natural gas. There has historically been ample natural gas supply in
the United States, resulting in cheaper processing of ethane and propane, as opposed to more
expensive petroleum cracking processes8 for liquids, and naphtha.

In addition, the United States has had an ample supply of propylene, because it is
produced in steam cracking for other products and because catalytic cracking is a required

6	Olefins are a class of chemicals made up of hydrogen and carbon with one or more pairs of carbon atoms linked by

a double bond. They are used as building block materials for products such as plastics, detergents, and adhesives.

7	Naphthas are any of various volatile, highly flammable liquid hydrocarbon mixtures used chiefly as solvents and

diluents and as raw materials for conversion to gasoline.

8	Cracking is the process by which heavy hydrocarbon molecules are broken up into lighter molecules by means of

heat and usually pressure and sometimes catalysts. Cracking is the most important process for the commercial
production of gasoline and diesel fuel.

19

-------
process in the gasoline industry. The propylene industry is based on this reaction that occurs in
the catalytic cracking process, yielding billions of pounds of product generated (Wittcoff, 2012).

Because of low-cost and high-domestic availability in the United States, there is an
incentive for U.S. manufacturers to use natural gas as a feedstock, replacing heavier liquid gases
such as naphtha. Changes in incentives for raw material use also affects byproduct production
prices, because byproducts, such as butadiene resulting from ethylene cracking, could be affected
by new technologies or production processes. Most prices for raw materials will respond in the
same direction as the changes in price for natural gas. Material costs respond in the opposite
direction of natural gas prices, while costs for byproducts respond in tandem with natural gas
prices (DeRosa, 2015).

The chemical industry is considered an upstream industry, because it purchases raw
materials such as petroleum, natural gas, coal, and metallic or nonmetallic minerals and does not
usually sell these products to final consumers. About one-fifth of materials are sold to other firms
in the chemical industry for additional processing, and then the remainder is sold to other
industries to assist in product manufacturing or services (Wittcoff, 2012).

It is often the case that oil refineries become integrated with nearby petrochemicals
plants. This integration allows both plants to exchange supply chain streams. The petrochemical
facility receives streams of raw materials from the oil refinery, and the refinery receives back
streams from the petrochemical plant that can be used again for petroleum products (e.g.,
gasoline blending). The petrochemical plants produce high-value products like ethylene,
propylene, styrene, butadiene, and benzene. Furthermore, these base petrochemicals can be
transformed further into other products like plastics, polyvinyl chloride (PVC), polystyrene,
polyethylene, polypropylene, elastomers, and aromatics-based products.

-------
2.2.2 SOCMI Supply Chain Disruptions

Supply chain disruptions can happen either upstream or downstream,9 but it is worth
noting that within the chemical industry upstream suppliers tend to be of greater concern to
continuity of operations (Kotze, 2017).

In addition to geopolitical risk factors, natural hazard disruptions affect many facets of
the petrochemical supply chain, resulting in longer recovery periods before production continues
(Stamber, 2011). As an example, Hurricane Ike in 200810 damaged readily available electric and
water infrastructure, raw materials, logistics, and production sites that negatively affected efforts
to begin operations post-disaster. These disruptions can have impacts in upstream and
downstream markets, affecting recovery.

Increases in the use of different feedstocks, such as natural gas, can also provide insights
into production and market cost effects that can occur in chemical supply chains. In the United
States, prices for natural gas use affect price patterns for byproducts of petrochemicals, such as
benzene, butadiene, and propylene. The cost of benzene, a byproduct of naphtha, stays relatively
constant during changes in natural gas price. The cost of butadiene, in contrast, increases as
natural gas prices decrease.

On a global scale, COVID-19, and the Russia-Ukraine war both affected oil and chemical
market prices. Continuing trends remain to affect the industry through changing societal
concerns for environmental issues, preferences for sustainable products, accelerated energy
transition, capacity demand and growth, and the continuous adoption of digitization. These
trends disrupted pertinent supply chains. The first quarter of 2020 saw an unanticipated downturn
for the oil, gas, and chemical industries as oversupply issues were exacerbated, and global oil
price collapses narrowed domestic feedstock cost advantages that petrochemical companies in
the United States benefitted from (Deloitte Insights, 2022).

While the COVID-19 crisis has abated worldwide since 2020, the Russia-Ukraine war
has also been a key factor in oil price changes in 2022. Consumer demand reduced as oil prices

9 The upstream market stage in the petrochemical industry refers to the exploration and production of natural gas

and crude oil. The downstream market focuses on taking natural gas and crude oil in finished products for
consumers.

10 https://cen.acs.org/articles/86/web/2008/09/Texas-Weathers-Ike.html

-------
increased, thus eroding profitability in the chemical industry. China has surpassed the United
States as the world's largest chemical market; it now accounts for more than 45 percent of
worldwide chemical sales. Some European chemical companies are also feeling this pressure, as
they expect a drop in 2022 profit (Stokes, 2022).

2.2.3 Ethylene

Ethylene is a valuable chemical product in both the U.S. and the world. It is the third
most valuable synthetic organic chemical product as of 2020 with $81.34 billion in revenue
worldwide. U.S. exports of ethylene were $401 million as of 2020. U.S. demand for ethylene
currently exceeds that of other top petrochemicals and is forecast to grow by 45 percent between
2020 and 2028 and by 51 percent between 2019 and 2035. Ethylene demand is driven by the
demand for its derivatives among which the demand for polyethylene is the highest, followed
(U.S. DOE, 2022 by ethylene dichloride (EDC) and ethylene oxide. U.S. market revenue from
ethylene is projected to rise by approximately $47 billion from 2022 to 2028.

Ethylene by-products are valuable due to their many important uses in common products.
One of those by-products is ethylene oxide. Ethylene oxide is used in the synthesis of ethylene
glycol, as a sterilizing agent for medical supplies and foods, as a fumigant, and as an
insecticide.11

Ethylene is a hydrocarbon gas that is produced by some fruits and vegetables through
natural processes. Ethylene is a by-product during the decomposition of organic material. It is a
common ingredient in various household products, including plastic, certain foods, and some
detergents. In 2020, ethylene was the world's 596th most traded product, with a total trade of
$4.95 billion. Between 2019 and 2020 the exports of ethylene decreased by 27.8 percent, from
$6.85 billion to $4.95 billion, in part due to the Covid-19 pandemic. Trade in ethylene represents
0.03 percent of total world trade (OEC, 2022b).

11 Observatory of Economic Complexity (2022). "Oxirane (ethylene oxide)." https://oec.world/en/profile/hs/oxirane-
ethylene-oxide

-------
Ethylene is used to produce fabricated plastics, antifreeze, and fibers. It is also used in the
process to produce ethylene oxide and to produce polyethylene for plastics, alcohol, mustard gas,
and other organics (National Center for Biotechnology Information, 2022a). Ethylene is a
product of steam cracking of petroleum hydrocarbons. Multiple feedstocks produce ethylene,
including ethane, propane, butanes, naphthas, and gas oils. Naphthas are the primary raw
material used in Western Europe and Japan, accounting for more than three-fourths of ethylene
produced. Ethane is the primary feedstock in the United States, followed by propane, naphthas,
gas oils, and butane. Small amounts of ethylene are recovered from other feedstocks, such as
retrograde-field condensates and refinery waste gases. Dehydration of ethanol is the third
commercial process for producing ethylene (National Center for Biotechnology Information,
2022a).

In 2020, the top exporters of ethylene were the Netherlands ($682 million), South Korea
($608 million), the United Kingdom ($587 million), and the United States ($401 million). Of
U.S.'s ethylene exports, 38.5 percent were exported to Taiwan, 34.2 percent to China, 9.78
percent to Indonesia, 9.03 percent to Belgium. In the United States from 2019 to 2020, the export
value was $401 million, an increase of 82.6 percent from a 2018 to 2019 value of $219 million
(Fernandez, Ethylene Prices Globally 2022, 2022).

In 2020, the top importers of ethylene were China ($1.35 billion), Belgium ($921
million), Indonesia ($552 million), Germany ($432 million), and Sweden ($360 million). In the
United States from 2019 to 2020, the import value was $190,000, an increase of 135.4 percent
from a 2018 to 2019 value of $81,000 (Fernandez, Ethylene Prices Globally 2022, 2022).

The average price of ethylene worldwide was approximately $697 per metric ton in 2020.
By July 2021, the average price for the year had risen to $1,014 per metric ton (see Figure 2-1),
45 percent higher than the previous year. "The global production capacity of ethylene is expected
to grow from approximately 200 million tons in 2020 to some 300 million tons by 2025"
(Fernandez, Ethylene Prices Globally 2022, 2022).

-------
1,500

1,250

1,155

2017	2018	2019	2020	2021*

Figure 2-1 Global Price of Ethylene (USD$/metric ton)

The global ethylene market is expected to grow from $81.34 billion in 2020 to $161.61
billion by 2028 at a CAGR of 8.3 percent during the forecast period of 2021 to 2028. The
increased use of coal as a feedstock for producing ethylene and the rising demand for ethylene
products in the construction industry are some of the factors fueling the ethylene market (Global
Newswire, 2022).

Key companies in the global ethylene market are Saudi Basic Industries Corp., Exxon
Mobil Corporation, Dow DuPont Inc., Royal Dutch Shell pic, China Petroleum & Chemical
Corporation, Chevron Phillips Chemical Company LLC, LyondellBasell Industries N.V., The
National Petrochemical Company, BASF SE, and Lonza Group, among others (Polaris Market
Research, 2021).

U.S. demand for ethane (the primary feedstock for ethylene) has been growing steadily
because of capacity expansions of ethylene crackers in the petrochemical industry, which use
ethane as a feedstock. Ethylene is a basic chemical used to produce plastics and resins. It is
estimated that the U.S. petrochemical industry expanded its capacity to produce ethylene from
almost 27 million metric tons per year (mt/y) in the first quarter of 2013 (when the first capacity
additions to ethylene crackers in over a decade came online) to almost 40 million mt/y in 2020.

24

-------
This growth in ethylene capacity caused domestic demand for ethane as a feedstock to grow
from 960,000 b/d in the first quarter of 2013 to 1.83 million b/d in the fourth quarter of 2020
(U.S. EIA, 2021).

2.2.3.1 Butadiene

Butadiene (1,3-Butadiene) is a synthetic, colorless gas that is basically insoluble in water
but soluble in ethanol, ether, acetone, and benzene (National Center for Biotechnology
Information, 2022b). Butadiene emits acrid fumes and is flammable when it is heated. When
butadiene is oxidized, it can form explosive peroxides. Butadiene rubber is a primary material
used in the production of car tires, gaskets, hoses, synthetic brushes, and synthetic carpets
(National Center for Biotechnology Information, 2022b).

Butadiene is used as a monomer in the manufacturing process of many different types of
polymers and copolymers. It is also used as a chemical intermediate in the production of
industrial chemicals. Butadiene is manufactured primarily as a co-product of ethylene production
from steam cracking in the United States, Western Europe, and Japan (National Center for
Biotechnology Information, 2022b).

The global 1,3-Butadiene market is expected to reach $33.01 billion by 2020. Growing
demand for tires of all types "on account of an upturn in the automotive industry (particularly in
China, India, and Brazil) is expected to remain a key driving factor for the global market" (Grand
View Research, 2015).

As shown in Figure 2-2, the price of Butadiene has decreased since 2017. At the start of
the first quarter of 2022, prices remained low initially. "During January (2022), the prices
dropped significantly by 10 percent as compared to last quarter of 2021. The initial decline in the
prices was attributed to the abundant supplies and weak trading activities. Demand from
downstream Styrene Butadiene Rubber (SBR) and Acrylonitrile Butadiene Styrene (ABS) plastic
has remained bearish in the region. As the upstream Crude and Natural gas prices rallied
upwards by the mid quarter, the Butadiene sentiments shifted marginally towards the upward
side in U.S. Korea, a major exporter of Butadiene, exported the product to the U.S. at sky high
values due to soaring freight charges. The price of Butadiene was last assessed at $l,445/[metric
ton] during March of 2022 in the U.S. Moreover, robust demand from downstream derivatives

-------
[styrene-butadiene rubber] (SBR) and [Polybutadiene rubber] (PBR) kept the Butadiene prices
on the higher side (Fernandez, 2021).

1 750

1 505.5

2017	2018	2019	2020*	2021"	2022*

Figure 2-2 Global Price of Butadiene from 2017 to 2019 with Estimated Figures for 2020

to 2022

2.2.3.2 Ethylene Oxide

Ethylene oxide is a colorless, flammable, toxic gaseous cyclic ether with a sweet ether-
like smell (National Center for Biotechnology Information, 2022c). "Ethylene oxide is used
especially in the synthesis of ethylene glycol and as a sterilizing agent for medical supplies and
foods, as a fumigant and as an insecticide" (OEC, 2022c).

Exposure to ethylene oxide can be highly irritating to the eyes, skin, and respiratory tract,
and can cause nausea, vomiting, and central nervous system depression (National Center for
Biotechnology Information, 2022c). The EPA's Integrated Risk Information System (IRIS)
characterized ethylene oxide as "carcinogenic to humans" by the inhalation route of exposure
based on the total weight of evidence (U.S. EPA, 2016), in accordance with EPA's Guidelines
for Carcinogen Risk Assessment (USEPA, 2005). EPA concluded that there was strong, but less

26

-------
than conclusive on its own, epidemiological evidence of lymphohematopoietic cancers and
breast cancer in ethylene oxide-exposed workers (U.S. EPA, 2016).

Nearly all production of ethylene oxide in the United States uses the direct vapor phase
oxidation process. "This process oxidizes ethylene with air or oxygen in the presence of a silver
catalyst to produce ethylene oxide" (OEC, 2022c).

In 2020, the top exporters of ethylene oxide were Germany ($161 million), the
Netherlands ($123 million), Belgium ($40 million), France ($28.9 million), and Russia ($15.8
million) (OEC, 2022c). In the United States from 2019 to 2020, the export value was $8.82
million, a decrease of 14.7 percent from a 2018 to 2019 value of $10.3 million.

In 2020, the top importers of ethylene oxide were Belgium ($88.9 million), Italy ($80.4
million), Germany ($73.6 million), France ($40.5 million), and the United Kingdom ($19.2
million). In the United States from 2019 to 2020, the import value was $68,900, an increase of
975 percent from a 2018 to 2019 value of $6,410.

Prices of nonyl phenol ethoxylates (upstream product of ethylene oxide) in the United
States grew over 2021 and the first quarter of 2022 "in response to the higher Ethylene Oxide
feedstock prices and outstretching demand" (ChemAnalyst, 2022).

Supply shocks can have a significant impact on the relatively concentrated market. For
example, "the curtailed operations in ExxonMobil's Baytown refinery following an explosion in
late December 2021 have continued to create a supply deficiency of upstream olefins and
consequently caused its prices to gain significant numbers. The high upstream pricing, which got
transferred to its downstream Ethylene Oxide, weighed on the input cost of Nonylphenol
Ethoxylates. Prompting the manufacturers for a price increase, thus, the Nonylphenol
Ethoxylates US discussions reached $1,923/MT FOB Gulf Coast in the quarter ending March
2022" (U.S. EPA, 2016)12.

12 "MT FOB" stands for "metric ton free on board." This refers to the price of one metric ton of a chemical product,
which includes the cost of the product and the cost of loading it onto a vessel for transportation. "FOB" means
that the cost of transportation from the point of origin to the port of shipment is included in the price, but the cost
of shipping the product to its final destination is not included.

-------
2.3 P&R Groups I and II

This sub-chapter focuses on the industries of the Polymers and Resin Group I and II
NESHAP. The economic and financial information in this chapter characterizes the conditions in
these industries which are likely to determine the nature of economic impacts associated with the
implementation of the NESHAP.

Section 2.4 provides an overview of the Group I synthetic rubber industries. Section 2.4.1
details the production processes, properties, and unique market characteristics for each
elastomer. Section 2.5 provides an overview of the industries covered by Polymers and Resin
Group II. Sections 2.5.1 describes epoxy resins and non-nylon polyamides production and their
markets.

Figure 2-3 provides a geographic overview of where Group I and Group II facilities
affected by this rule are located across the U.S. Group I facilities are clustered in the South with
most based in Louisiana and Texas and others spread across the Midwest. There are fewer Group
II facilities affected under this rule; four facilities are distributed in the South, while one is in
Oregon.

A P&R II Facility
• P&R I Facility

O A

A

Figure 2-3 P&R Group I and II Facilities Map

28

-------
2.4 Group I Industry Profile

This section reviews the organization, processes, and products of the affected synthetic
rubber industries. The affected firms are further identified by size and economic impacts in a
later section. Each facility considered in the production of Group I elastomers is categorized by a
North American Industry Classification System (NAICS) code. This code is a "standard used by
Federal statistical agencies in classifying business establishments for the purpose of collecting,
analyzing, and publishing statistical data related to the U.S. business economy" and used for
defining industries. Across the identified facilities in Group I, there are four unique NAICS
industries with varying representation in the associated NESHAP and the U.S. economy.13 Table
2-3 provides 2017 data for these industries in the U.S. economy, not only considering facilities
directly impacted by this rulemaking.14 Data on industries is sourced from the quinquennial
Economic Census which last occurred in 2017.

Table 2-3 Polymers and Resin Group I Industries

Number of Facilities

Total Industry

Total

NAICS

Name of Industry

Impacted (% of
Total Facilities
Impacted)

Revenue in
2017 (in
Billions)

Industry
Employment
in 2017

325110

Petrochemical Manufacturing

1 (5.6%)

$52.97

9,369

325211

Plastics Material and Resin
Manufacturing

3 (16.7%)

$89.52

75,998

325212

Synthetic Rubber Manufacturing

13 (72.2%)

$8.39

9,661

325998

All Other Miscellaneous Chemical
Product and Preparation Manufacturing

1 (5.6%)

$21.85

36,900

2.4.1 Industry Organization of Group I Industries

This section provides information on the structure of the covered synthetic rubber
industries and the characteristics of the market organization of the affected Group I industries.
This is an attempt to characterize the impacts regulation can have in more detailed terms.

Table 2-4 shows how the firms in each product category can be characterized by market
concentration: the market share percentage for the 50 largest firms of the affected industries by
NAICS code. The standard economic framework is that the higher the market concentration, the

13 U.S. Census Bureau (2022). North American Industry Classification System. Retrieved from

https://www.census.gov/naics/.

14 Data available at https://www2.census.gov/programs-surveys/susb/tables/2017/us_6digitnaics_rcptsize_2017.xlsx.

Accessed 11/7/2022. We note the publication of data from the 2022 Economic Census will not occur until late
2023 or early 2024.

-------
more that changes in input price caused by regulation will lead to rises in output price. An
example of this is presented in a 2018 report by Abdela and Steinbaum, which concludes that
there is a market concentration problem in U.S. production, generally (Abdela, 2018). This
assumption has been criticized by Newmark, who argues that "[P]rice-concentration studies are
severely flawed. In industries in which sellers compete on quality and amenities, a positive price-
concentration relation could result, not from coordinated effects, but from competitive
superiority" (Newmark, 2004).

The table provides additional evidence to examine these issues. In addition to the percent
values claimed by the largest companies in the product categories, it includes the Herfindahl-
Hirschman Index (HHI) for industries based on value added. This index provides a signal of how
concentrated market power is across a particular industry.

The U.S. Department of Justice (DO J) and the Federal Trade Commission (FTC) use the
HHI to identify markets where there are potential anti-trust concerns. They consider markets
with an HHI below 1,000 to be unconcentrated; markets with HHI between 1,000 and 1,800 are
considered moderately concentrated, and markets with HHI above 1,800 to be "highly
concentrated" (U.S. EPA, 2013). For a given market, the HHI is calculated by squaring the
market share of each firm competing in the market, then summing the squared shares, as shown
in the following equation:

HHI = sf, where Si is the market share of the ith firm.

Table 2-4 Concentration Findings of Affected Group I Industries

NAICS

Name of Industry

HHI Value*

Finding

325110

Petrochemical Manufacturing

2,868.4

Concentrated

325211

Plastics Material and Resin Manufacturing

409.9

Unconcentrated

325212

Synthetic Rubber Manufacturing

652.6

Unconcentrated

325998

All Other Miscellaneous Chemical Product and

164.8

Unconcentrated

Preparation Manufacturing

Notes: *HHI is based on the 50 largest companies for each NAICS code. HHI for years after 2012 are not available
since comprehensive concentration data from the 2017 Economic Census is not yet available. This value is found
using "value added" which is a measure of manufacturing activity. Value added is derived by subtracting the cost of
purchased inputs from the value of shipments.

Source: US Census. Economic Census (August 2015).

https://www.census.gov/data/tables/2012/econ/census/manufacturing-reports.html

By the DOJ definition above, the only product category for which markets could be
considered concentrated by the HHI was petrochemical manufacturing in 2012. Market
concentration can develop due to various barriers to entry, pushing out potential entrants into a

-------
market. Barriers such as supply chain connection, capital requirements, and access to human
capital can push certain industries to concentrate. Factors like these may be relevant in the case
of the concentration of petrochemical manufacturing. However, there is evidence of a
competitive market for the other synthetic rubber industries impacted by this regulation.

Competition for the synthetic rubber industries can arise due to a number of factors. The
products of these industries can in some cases be substituted for one another. Other natural
rubbers and imported products can also act as substitutes for these products. The presence of
these alternatives can create excess capacity and can lead to falling prices for these industries.

2.4.2 Prices for Group I Industries

From 2012 to 2021, product prices for the wider chemical manufacturing sector (NAICS
325) have increased overall, but the trend has been marked with some volatility over the years.
Prices began falling in 2015, stayed steady and then increased considerably in 2017 and 2018,
and increased sharply in 2021 by almost 12 percent year-over-year. Table 2-5 shows this erratic
pattern in detail.

Table 2-5 Chemical Manufacturing (NAICS 325) Product Price Index, 2012-2021

(2012 = 100)

2012-

NAICS 325

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

Chemical
Manufacturing

100.0

101.4

103.4

101.7

102.6

106.6

111.5

112.2

111.0

124.0

Change from
Previous Year

1.4

2.0

-1.7

0.9

4.0

4.8

0.7

-1.2

13.0

24.0

% Change
from Previous

1.4%

2.0%

-1.6%

0.9%

3.9%

4.5%

0.7%

-1.1%

11.7%

24.0%

Year

Source: U.S. Department of Labor, Bureau of Labor Statistics. Industries at a Glance: Chemical Manufacturing - NAICS 325.
https://www.bls.gov/iag/tgs/iag325.htm

Table 2-6 shows a closer look at prices for synthetic rubber and related products from
2012 to 2021. Note that changes in product prices do not directly relate to the changes in
synthetic rubber prices. Over this period, synthetic rubber prices fell considerably with a clear
surge around 2017. This increase was a result of various supply-side conditions including
growing costs in inputs to Styrene Butadiene Rubber (SBR), technical issues related to several
manufacturing locations equipment failures affecting around 40 percent of U.S. styrene
production, shortage of butadiene as well as its feedstock materials due to production issues, and
other production issues. Tires and pneumatic tire prices remained relatively steady with only

-------
modest year to year changes, whereas rubber and plastics hose prices increased overall with
some volatility in years 2013, 2015, and 2021.

Additionally, while not shown in the RIA, prices for natural rubber did not directly match
the price changes in synthetic rubber as might be assumed from their substitutability. This is due
to natural rubber following the supply-side constraint of agricultural inputs, whereas synthetic
rubbers face the constraint of the availability and price of hydrocarbons. Moreover, the demand
for rubbers is application dependent and may be based on the physical properties such as heat
resistance and tear strength necessary for a particular use. This means substitutability of synthetic
rubbers cannot be widely determined without knowing the needs of certain rubber applications
(Wagner, 2020).

32

-------
* E.O. 12866 Review - Draft - Do Not Cite, Quote, or Release During Review *

Table 2-6 Producer Price Index of Synthetic Rubber, 2012-2021 (Index for 2012 is normalized to 100)



2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2012-2021

Synthetic Rubber

100.0

88.3

85.7

75.3

73.8

80.2

82.3

78.9

73.3

85.8



YoY Change



-11.7

-2.6

-10.4

-1.5

6.4

2.1

-3.4

-5.6

12.5

-14.2

% YoY Change



-11.7%

-2.9%

-12.1%

-2.0%

8.7%

2.6%

-4.1%

-7.1%

17.1%

-14.2%

Tires, Tubes,























Tread, and Repair

100.0

98.2

95.8

93.8

92.4

93.4

94.6

95.6

95.9

100.1



Materials























YoY Change



-1.8

-2.4

-2.0

-1.4

1.0

1.2

1.0

0.3

4.2

0.1

% YoY Change



-1.8%

-2.4%

-2.1%

-1.5%

1.0%

1.3%

1.1%

0.3%

4.4%

0.1%

Pneumatic Tires























(on-road, off-road,

100.0

98.2

95.7

93.6

92.1

93.1

94.3

95.3

95.6

99.7



and other)























YoY Change



-1.8

-2.5

-2.1

-1.5

1.0

1.2

1.1

0.3

4.1

-0.3

% YoY Change



-1.8%

-2.5%

-2.2%

-1.6%

1.1%

1.3%

1.1%

0.3%

4.3%

-0.3%

Rubber and
Plastics Hose

100

104.1

105.2

104.0

105.3

106.0

109.2

112.7

114.0

122.1



YoY Change



4.1

1.1

-1.2

1.3

0.7

3.2

3.5

1.3

8.1

22.1

% YoY Change



4.1%

1.0%

-1.1%

1.2%

0.6%

3.1%

3.2%

1.1%

7.1%

22.1%

Rubber and























Plastics Hose (for
on- and off-road

100

106.7

106.5

102.1

105.2

105.3

109.3

111.8

112.7

114.6



vehicles)























YoY Change



6.7

-0.18

-4.4

3.0

0.1

4.0

2.5

0.8

1.9

14.6

% YoY Change



6.7%

-0.2%

-4.1%

3.0%

0.1%

3.8%

2.3%

0.7%

1.7%

14.6%

Note: "YoY" is an acronym for "Year over Year."

Source: Federal Reserve Economic Data, Economic Research Division, Federal Reserve Bank of St. Louis, https://fred.stlouisfed.org/seriesAVPU071102,
https://fred.stlouisfed.org/series/WPU0712, https://fred.stlouisfed.org/series/WPU0712010, https://fred.stlouisfed.org/series/WPU07130411,
https://fred.stlouisfed.org/series/WPU071304118

33

-------
* E.O. 12866 Review - Draft - Do Not Cite, Quote, or Release During Review *

2.4.3 General Production Description of Group I Industries

Synthetic rubber production requires the synthesis of monomers (derived from
petrochemicals), followed by their polymerization. This process results in an aqueous suspension
of rubber particles, or the latex, which may then be processed into marketable, dry, raw rubber.
Synthetic rubbers are usually compounded with various additives and then molded, extruded, or
calendared into the desired solid form. A percentage of elastomer production is also supplied in
the form of water dispersions, called latexes (primarily used in foam rubber). HAP emission
sources in synthetic rubber facilities include equipment leaks, process vents, wastewater, and
storage tanks. It is important to note that elastomer production sites subject to this standard may
be collocated with other production facilities that are, or will be, subject to NESHAP standards
other than the Group I NESHAP. For example, a refining facility, chlorine plant, SOCMI
facility, or non-elastomer polymer facility could be located on the same site as Group I
production units.

2.4.4 Product Description of Group I Industries

The affected Group I elastomers are classified as synthetic rubbers which have specific
elasticity and yield properties. Synthetic rubbers are either used as stand-alone products, or are
compounded with natural rubber, other thermoplastic materials, or additives, depending on the
desired end-use characteristics. This section describes the properties of each elastomer
individually and identifies its primary end uses. Portions of this section are adapted from the
Economic Impact Analysis for the Polymers and Resin Group I NESHAP (U.S. EPA, 1995).

2.4.4.1 Butyl Rubber (Including Halobutyl Rubber)

Butyl rubbers are copolymers of isobutylene (also known as isobutene) and other
monomers. Other typical monomers include isoprene and methylstyrene. Most butyl rubber is
produced by precipitation polymerization, although other methods may be used. Halobutyl
rubber is a type of butyl rubber elastomer produced using halogenated copolymers and is
typically used as a sealant. Characteristics of butyl rubber include low permeability to gases and
high resistance to tear and aging. Butyl rubber is used as an input to the production of tires,

-------
* E.O. 12866 Review - Draft - Do Not Cite, Quote, or Release During Review *

tubes, and tire products. It is also used into the production of inner tubes because of its low air
permeability. Butyl rubber is used in the production of automotive and mechanical goods,
adhesives, and caulks like halobutyl rubber (Thomas, 2022a).

Butyl rubber had a total trade of $592 million in 2020. Between 2019 and 2020, the
exports of butyl rubber decreased by 0.71 percent, from $596 million to $592 million. Russia had
the highest total exports of butyl rubber, valued at $193 million, while the U.S. was third at $65.5
million. For the same year, China was the largest importer at $197 million (OEC, 2022d).

Halobutyl rubber had a total trade of $1.59 billion in 2020. Between 2019 and 2020, the
exports of halobutyl rubber decreased by 19.3 percent, from $1.97 billion to $1.59 billion.
Singapore had the highest total exports of halobutyl rubber, valued at $266 million, while the
U.S. was fourth at $239 million. For the same year, China was the largest importer at $291
million, whereas the U.S. was fifth highest at $108 million (OEC, 2022e).

2.4.4.2 Epichlorohydrin Elastomers (EPI)

The production of EPI uses epichlorohydrin, ethylene oxide, and allyl glycetal ether,
which are combined in a polymerization process. Due to its low gas permeability and resistance
to heat, fuel, and abrasion, EPI is primarily used in automotive applications including gaskets,
hoses, and seals.15 Market information on epichlorohydrin elastomers markets was limited.

2.4.4.3 Ethylene-Propylene Rubber (EPDM)

The ethylene-propylene category includes both ethylene-propylene copolymers (EPD)
and ethylene-propylene terpolymers (EPDM). EPDM is produced from the polymerization of
ethylene and propylene, which may occur in either a solution process or a suspension process.
EPDM is characterized by lower cost relative to other elastomers, and resistance to cracking, low
temperature flexibility, and weather. EPDM compounds have been developed for many different
applications including automotive, industrial, construction, and HVAC. End uses include roofing
membranes, impact modifiers, oil additives, automobile parts, gaskets and seals, and hoses and

15 BRP Manufacturing (2000). Epichlorohydrin Rubber, https://brpmfg.com/epichlorihydrin-rubber/epichlorohydrin-
rubber/

-------
* E.O. 12866 Review - Draft - Do Not Cite, Quote, or Release During Review *

belts. The wide range of uses of this elastomer is attributable to its multifunctional nature
(Thomas, 2022b).

EPDM had a total trade value of $1.57 billion in 2020. Between 2019 and 2020, the
exports of EPDM rubber decreased by 18.2 percent, from $1.92 billion to $1.57 billion. The U.S.
had the highest total exports of EPDM rubber, valued at $444 million. For the same year, China
was the largest importer at $282 million, whereas the U.S. was fourth highest, at $90.8 million
(OEC, 2022f).

2.4.4.4 Chlorosulfonated Polyethylene (Hypalon™)

Chlorosulfonated polyethylene, also known by the discontinued trade name "Hypalon™"
(Anixter, 2022), is formed from polyethylene through a chlorination and chlorosulfonation
process. Chlorosulfonated polyethylene is characterized by excellent resistance to ozone,
oxidation, UV rays, and weather. Uses of chlorosulfonated polyethylene include coatings for
roofs and tarpaulins, hose construction, wire coverings, industrial rolls, and sporting goods
(Industrial Rubber Goods, 2022). Market information on Chlorosulfonated polyethylene markets
was limited.

2.4.4.5 Nitrile Butadiene Latex (NBL)

Nitrile butadiene latex (NBL) is a polymer consisting primarily of unsaturated nitriles and
dienes, usually acrylonitrile and 1,3-butadiene, that is sold as a latex. NBL is polymerized by
free radical emulsion through advanced techniques. It is used in several applications because of
its low cost, ease of processing, and low flammability. NBL is also used in several applications
like gloves because of its oil resistance (Senlos Chem, 2022).

NBL had a total trade value of $1.6 billion in 2020. Between 2019 and 2020, the exports
of NBL rubber increased by 33.8 percent, from $1.19 billion to $1.6 billion. South Korea had the
highest total exports of NBL, valued at $886 million. For the same year, Malaysia was the largest
importer at $1.08 billion. The U.S. was not among the five largest exporters or importers of NBL
(OEC, 2022g).

-------
* E.O. 12866 Review - Draft - Do Not Cite, Quote, or Release During Review *

2.4.4.6 Nitrile Butadiene Rubber (NBR)

Nitrile butadiene rubber (NBR) is a copolymer of acrylonitrile and butadiene. Its most
significant characteristic is its resistance to mineral oils, vegetable oils, and hydraulic fluids.
NBR is the preferred product for gasoline hoses, gaskets, and printing rolls. Many of the
properties of nitrile rubber are directly related to the proportion of acrylonitrile in the rubber.
NBR is used in many hose applications where oil, fuel, chemicals, and solutions are transported.
In powder form, NBR is used in cements, adhesives, and brake linings, and in plastics
modification. NBR is also used in belting and cable, in addition to its uses in O-rings and seals,
adhesives and sealants, sponges, and footwear (Polymerdatabase, 2022a).

NBR had a total trade value of $890 million in 2020. Between 2019 and 2020, the exports
of NBR decreased by 16.8 percent, from $1.07 billion to $890 million. South Korea had the
highest total exports of NBR, valued at $197 million, while the U.S. was fourth at $101 million.
For the same year, China was the largest importer at $187 million, whereas the U.S. was third
highest at $81.3 million (OEC, 2022h).

2.4.4.7 Neoprene

Polychloroprene, also known as Neoprene, is produced from chloroprene through an
emulsion process. It is characterized by its high flexibility, resistance to oils, strength, and
resistance to abrasion, making it suitable for many diverse uses. Neoprene is similar to NBR in
end uses, given that the primary use is for hoses and belts, with the remaining uses allocated
among mechanical, adhesive, and wire and cable end uses. In latex form, Neoprene is used to
manufacture household and industrial gloves (Polymerdatabase , 2022b).

In 2021, the total neoprene market was valued at $2.23 billion, with the Asia-Pacific
market holding 40 percent of the market share. Acumen Research and Consulting projects a
compound annual growth rate of 2.6 percent from 2022 to 2030. Two identified key market
drivers are an increased adoption of neoprene rubber in the automotive sector and continual
growth in construction and electronics industries in emerging economies. Denka Performance
Elastomer LLC is also a prominent manufacturer in this market, producing over 23 percent of
neoprene globally, and currently owns and operates the only neoprene production facility in the
United States (Acumen Research and Consulting, 2022).

-------
* E.O. 12866 Review - Draft - Do Not Cite, Quote, or Release During Review *

2.4.4.8 Styrene-Butadiene Latex (SBL)

Styrene-butadiene latex (SBL) is a polymer consisting primarily of styrene and butadiene
monomer units produced using an emulsion process and sold as a latex. Most commercial
styrene-butadiene polymers are heavily crosslinked, so they have a high gel content. This
provides a greater degree of toughness, strength, and elasticity compared to other materials,
allowing for its usefulness as a latex. SBL is commonly used as a "coating in paper products,
such as magazines, flyers and catalogs, to achieve high gloss, good printability, and resistance to
oil and water." It's also a popular choice as an adhesive in construction applications (Mallard
Creek Polymers, 2020).

SBL had a total trade value of $939 million in 2020. Between 2019 and 2020, the exports
of SBL decreased by 13.4 percent, from $1.8 billion to $939 million. Germany had the highest
total exports of SBL, valued at $255 million, while the U.S. was second at $122 million. For the
same year, China was the largest importer at $165 million (OEC, 2022i).

2.4.4.9 Styrene-Butadiene Rubber (SBR)

Styrene-butadiene rubber (SBR) is produced in the largest volume of all the synthetic
rubbers. Its chemical properties include favorable performance in extreme temperatures,
resistance to cracking and abrasion, and stability over time. The dominance of SBR among
synthetic rubber types is attributable to its availability and processability. The availability of
styrene and butadiene in fossil hydrocarbons make these two inputs an abundant source of
synthetic rubber, and styrene and butadiene can be combined into rubber compounds which are
easily processed into tire molds. Types of SBR differ in the ratios of styrene to butadiene, their
content of additives, or the type of polymerization process used during the manufacturing
process. The substitutability of SBR with natural rubber is primarily determined by the
fluctuating prices of each, and by the properties required in the end product (Polymerdatabase ,
2022c).

SBR had a total trade value of $4.49 billion in 2020. Between 2019 and 2020, the exports
of SBR rubber decreased by 19.8 percent, from $5.6 billion to $4.49 billion. South Korea had the
highest total exports of SBR, valued at $713 million. For the same year, China was the largest
importer at $627 million, whereas the U.S. was second highest at $472 million (OEC, 2022j).

-------
* E.O. 12866 Review - Draft - Do Not Cite, Quote, or Release During Review *
2.4.4.10 Polybutadiene Rubber (PER)

Polybutadiene rubber (PBR) is formed from butadiene which undergoes emulsion
polymerization. After SBR, polybutadiene rubber is the synthetic rubber produced in the second
highest volume. It is also a relatively low-cost elastomer. The major use of PBR is in tires for the
side walls and treads. To augment properties such as traction, rolling and abrasion resistance, it is
typically compounded with other elastomers such as natural rubber and SBR. Other applications
are golf ball cores, inner tubes of hoses for sandblasting, and covers for pneumatic and water
hoses. Polybutadiene is also used as a toughening agent in the production of certain plastics
(Polymerdatabase, 2022d).

PBR had a total trade of $2.35 billion in 2020. Between 2019 and 2020, the exports of
PBR decreased by 25.1 percent, from $3.14 billion to $2.35 billion. South Korea had the highest
total exports of PBR, valued at $427 million, while the U.S. was third highest, at $241 million.
For the same year, China was the largest importer at $320 million (OEC, 2022k).

2.5 Group II Industry Profile

This section reviews Group II industries, which are characterized by the following source
categories: epoxy and resins production and non-nylon polyamides production. The affected
firms are more explicitly identified by size and economic impacts in a later section. Like the
Group I facilities, each facility is identified using a NAICS code. Among the identified facilities
in Group II, there are three unique NAICS codes represented in the associated NESHAP and the
U.S. economy. Table 2-7 provides 2017 data for these industries in the U.S. economy, not only
considering facilities directly impacted by this rulemaking.

Table 2-7 Polymers and Resin Group II Industries

Number of Facilities

Total Industry

Total

NAICS

Name of Industry

Impacted (% of
Total Facilities
Impacted)

Revenue in
2017 (in
Billions)

Industry
Employment
in 2017

324110

Petroleum Refineries

1 (20%)

$478.60

63,594

325110

Petrochemical Manufacturing

1 (20%)

$52.67

9,369

325211

Plastics Material and Resin
Manufacturing

3 (60%)

$89.52

75,998

-------
* E.O. 12866 Review - Draft - Do Not Cite, Quote, or Release During Review *

2.5.1 Industry Organization of Group II Industries

This section provides information on the structure of the covered epoxy and resins
industries, and the characteristics of the market organization of the affected Group II industries.
This is an attempt to characterize the impacts this final regulation can have in more detailed
terms.

Table 2-8 discusses how the firms in each product category can be characterized by
market concentration: the market share percentage for the 50 largest firms of the affected
industries by NAICS code.

Table 2-8 Concentration Findings of Affected Group II Industries

NAICS

Name of Industry

HHI Value*

Finding

324110

Petroleum Refineries

786

Unconcentrated

325110

Petrochemical Manufacturing

2,868.4

Concentrated

325211

Plastics Material and Resin Manufacturing

409.9

Unconcentrated

Source: US Census. Economic Census (August 2015).

https://www.census.gov/data/tables/2012/econ/census/manufacturing-reports.html

As presented in Section 2.2.1 and based on the given U.S. Department of Justice
definition, only the petrochemical manufacturing markets could be considered concentrated in
2012. Whereas the HHI for both petroleum refineries and plastics materials and resin
manufacturing markets suggest more competitive, unconcentrated industry markets. Like
synthetic rubbers, most affected facilities under this rule inhabit more competitive markets,
which generally suppresses the profit margins for certain firms and increases the price elasticity
of demand. A market with output having higher price elasticity of demand shows a larger change
in quantity demanded relative to a particular change in price, all else equal.

2.5.2 Prices for Group II Industries

Like Group I industries, Group II industries are also a part of the chemical manufacturing
sector (NAICS 325). More information on prices for this sector can be found in Section 2.2.2. In
this section, Table 2-9 shows a closer look at prices for epoxy and resins, and related products,
from 2012 to 2021. Note that changes in the product prices of adhesives, coatings, and paper do

-------
* E.O. 12866 Review - Draft - Do Not Cite, Quote, or Release During Review *

directionally relate to price changes in the broader "Plastics Material and Resins Manufacturing"
(Plastics and Resins) category, but the magnitude of these changes are varied.

Over this period, each product category saw overall prices increase with certain years
with volatility in 2015 and 2019. In 2021, each product category saw its highest single-year spike
in prices with Plastics and Resins increasing by over a third of its 2020 value. This one-year
increase was largely a result of demand-side and supply-side conditions exacerbated by the
COVID-19 pandemic. On the demand-side, COVID-19-related lockdowns increased the demand
for delivered goods and therefore greater plastic packaging, while the healthcare industry
augmented plastic demands for personal protective equipment (PPE). Supply-side issues also
constrained the availability of plastics and resins due to related labor shortages or production
slowdowns, as well as more widespread international supply chain difficulties (Pederson, 2021).

41

-------
Table 2-9 Producer Price Index of Epoxy and Resins, 2012-2021 (2012 = 100)



2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2012-2021

Plastics Material























and Resins

100.0

103.5

107.8

97.3

92.7

97.6

102.3

98.0

94.6

126.1



Manufacturing























YoY Change



3.5

4.3

-10.5

-4.6

4.9

4.7

-4.3

-3.5

31.6

26.1

% YoY Change



3.5%

4.2%

-9.7%

-4.7%

5.3%

4.8%

-4.2%

-3.6%

33.4%

26.1%

Adhesive























Manufacturing:
Synthetic Resin and

100.0

100.7

102.1

101.3

100.2

101.7

104.2

106.5

107.0

113.0



Rubber Adhesives























YoY Change



0.7

1.4

-0.8

-1.1

1.5

2.4

2.4

0.4

6.0

13.0

% YoY Change



0.7%

1.4%

-0.8%

-1.1%

1.5%

2.4%

2.3%

0.4%

5.6%

13.0%

Paint and Coating
Manufacturing

100.0

101.3

102.1

101.0

100.5

101.6

105.3

110.0

112.1

121.3



YoY Change



1.3

0.8

-1.1

-0.6

1.1

3.7

4.7

2.1

9.1

21.3

% YoY Change



1.3%

0.8%

-1.1%

-0.5%

1.1%

3.7%

4.4%

1.9%

8.2%

21.3%

Paper

Manufacturing

100.0

102.2

103.6

103.4

102.6

105.2

108.9

110.2

108.6

117.6



YoY Change



2.2

1.4

-0.2

-0.8

2.6

3.7

1.3

-1.6

9.0

17.6

% YoY Change



2.2%

1.4%

-0.2%

-0.8%

2.5%

3.5%

1.2%

-1.5%

8.3%

17.6%

Note: "YoY" is an acronym for "Year over Year".

Source: Federal Reserve Economic Data, Economic Research Division, Federal Reserve Bank of St. Louis, https://fred.stlouisfed.org/series/PCU325211325211,
https://fred.stlouisfed.org/series/PCU3255203255204, https://fred.stlouisfed.org/series/PCU325510325510, https://fred.stlouisfed.org/series/PCU322322

42

-------
2.5.3 Product Description and Markets of Group II Industries

The epoxy resins production source category involves the manufacture of basic liquid
epoxy resins used in the production of glues, adhesives, plastic parts, and surface coatings. This
source category does not include specialty or modified epoxy resins. The non-nylon polyamides
production source category involves the manufacture of epichlorohydrin crosslinked non-nylon
polyamides used primarily by the paper industry as an additive to paper products. Natural
polymers, such as those contained in paper products, have little cross-linking, which allows their
fibers to change position or separate completely when in contact with water. The addition of
epichlorohydrin cross-linked non-nylon polyamides to these polymers causes the formation of a
stable polymeric web among the natural fibers. Because the polymeric web holds the fibers in
place even in the presence of water, epichlorohydrin cross-linked non-nylon polyamides are also
referred to as wet-strength resins. HAP emission sources in Group II facilities include equipment
leaks, process vents, wastewater, and storage tanks.

The epoxy resin market had a size of $12.8 billion in 2022. Grand View Research
projects a compound annual growth rate of 7.3 percent from 2022 to 2030, forecasting a 2030
revenue of $22.4 billion. A key market driver is an increased demand for paints and coatings due
to increased construction spending in North America and Western Europe. In 2021, "paints and
coating" accounted for more than 37 percent of revenue share of epoxy resin applications and
Asia-Pacific accounted for more than 61 percent of revenue share by region of the market (Grand
View Research, 2022).

The non-nylon polyamide resin market had a size of $3.27 billion in 2020. Between 2019
and 2020, the exports decreased by 10.5 percent, from $3.65 billion to $3.27 billion. Germany
had the highest total exports, valued at $197 million, while the U.S. was second at $529 million.
For the same year, China was the largest importer at $502 million, whereas the U.S. was fourth
highest at $155 million (OEC, 20221).

-------
3 EMISSIONS AND ENGINEERING COST ANALYSIS

3.1 Introduction

In this chapter, we present estimates of the projected emissions reductions and
engineering compliance costs associated with the proposed action. The primary cost and
emissions impacts are calculated at the emission points in the various HON and polymers and
resins processes. This chapter has compliance cost estimates for heat exchange systems, storage
vessels, process vents, wastewater, equipment leaks, flares, and fenceline monitoring. In addition
to emissions reductions, some control options result in product recovery, which can then be sold.
Estimates of annualized cost include the value of the product recovery where applicable.

3.2 HON

In general, the HON applies to chemical manufacturing process units (CMPUs) that: (1)
produce one of the listed SOCMI chemicals, and (2) either use as a reactant or produce a listed
organic HAP in the process. A CMPU refers to the collection of equipment assembled and
connected by pipes or ducts to process raw materials and to manufacture an intended product. A
CMPU consists of more than one unit operation. A CMPU includes air oxidation reactors and
their associated product separators and recovery devices; reactors and their associated product
separators and recovery devices; distillation units and their associated distillate receivers and
recovery devices; associated unit operations; associated recovery devices; and any feed,
intermediate and product storage vessels, product transfer racks, and connected ducts and piping.
A CMPU includes pumps, compressors, agitators, pressure relief devices, sampling connection
systems, open-ended valves or lines, valves, connectors, instrumentation systems, and control
devices or systems. A CMPU is identified by its primary product.

The emissions sources affected by the current HON include heat exchange systems and
maintenance wastewater regulated under NESHAP subpart F; process vents, storage vessels,
transfer racks, and wastewater streams regulated under NESHAP subpart G; equipment leaks
associated with SOCMI processes regulated under NESHAP subpart H; and equipment leaks
from certain non-SOCMI processes at chemical plants regulated under NESHAP subpart I.

-------
As of July 1, 2023, there were 207 facilities that are major sources of HAP emissions in
operation that are subject to the HON. The list of facilities located in the United States that are
major sources of HAP and part of the SOCMI source category with processes subject to HON is
available in the memorandum titled: "Lists of Facilities Subject to the HON, Group I and Group
II Polymers and Resins NESHAPs, and NSPS subparts VV, VVb, III, NNN, and RRR" (ERG,
2023a).

3.3 P&R I (Subpart U)

The P&R INESHAP generally follows and refers to the requirements of the HON, with
an addition of requirements for batch process vents. Generally, the P&R I NESHAP applies to
elastomer product process units (EPPUs) and associated equipment. Similar to a CMPU in the
HON, an EPPU means a collection of equipment assembled and connected by hard-piping or
duct work used to process raw materials and manufacture elastomer product. The EPPU includes
unit operations, recovery operations, process vents, storage vessels, and equipment that are
covered by equipment leak standards and produce one of the elastomer types listed as an
elastomer product {i.e., the list found in Section 1.0 above). An EPPU consists of more than one
unit operation. An EPPU includes, as equipment, pumps, compressors, agitators, pressure relief
devices, sampling connection systems, open-ended valves or lines, valves, connectors, surge
control vessels, bottoms receivers, and instrumentation systems, and control devices or systems.

The emissions sources affected by the current P&R I NESHAP include heat exchange
systems and maintenance wastewater regulated under NESHAP subpart F; storage vessels,
transfer racks, and wastewater streams regulated under NESHAP subpart G; and equipment leaks
regulated under NESHAP subpart H. Process vents are also regulated emission sources but,
unlike the HON, these emissions sources are subdivided into front and back-end process vent
process vents in P&R I. The front-end are unit operations prior to and including the stripping
operations. These are further subdivided into continuous front-end process vents regulated under
NESHAP subpart G and batch front-end process vents that are regulated according to the
requirements within the P&R I NESHAP. Back-end unit operations include filtering,
coagulation, blending, concentration, drying, separating, and other finishing operations, as well
as latex and crumb storage. The requirements for back-end process vents are not subcategorized
into batch or continuous and are also found within the P&R I NESHAP.

45

-------
As of July 1, 2023, there were 19 facilities that are major sources of HAP emissions in
operation that are subject to the P&R INESHAP. The list of facilities located in the United
States that are major sources of HAP is available in the memorandum titled: "Lists of Facilities
Subject to the HON, Group I and Group II Polymers and Resins NESHAPs, and NSPS subparts
VV, VVa, III, NNN, and RRR" (ERG, 2023a).

3.4 P&R II (Subpart W)

The P&R II NESHAP takes a different regulatory and format approach from the P&R I
NESHAP but still refers to HON provisions for a portion of the standards. There are two basic
subcategories of manufacturing sources in the P&R II NESHAP - basic liquid epoxy resins
(BLR) and wet strength resins (WSR). A BLR means resins made by reacting epichlorohydrin
and bisphenol A to from diglycidyl ether of bisphenol-A (DGEBPA). A WSR means
polyamide/epichlorohydrin condensates which are used to increase the tensile strength of paper
products.

The emission sources affected by the P&R II NESHAP are all HAP emission points
within a facility related to the production of BLR or WSR. These emission points include process
vents, storage tanks, wastewater systems, and equipment leaks. Equipment includes connectors,
pumps, compressors, agitators, pressure relief devices, sampling connection systems, open-ended
valves and lines, and instrumentation system in organic HAP service. Equipment leaks are
regulated under the HON {i.e., NESHAP subpart G).

Process vents, storage tanks, and wastewater systems combined are regulated according
to a production-based emission rate {e.g., pounds HAP per million pounds BLR or WSR
produced) standard for existing sources in both BLR (130 pounds) and WSR (10 pounds). For
new sources, BLR requires 98 percent reduction or an overall limit of 5,000 pounds of HAP per
year. New WSR sources are limited to 7 pounds of HAP per million pounds WSR produced.

As of July 1, 2023, there were 5 facilities that are major sources of HAP emissions in
operation that are subject to the P&R II NESHAP. The list of facilities located in the United
States that are major sources of HAP is available in the memorandum titled: "Lists of Facilities
Subject to the HON, Group I and Group II Polymers and Resins NESHAPs, and NSPS subparts
VV, VVa, III, NNN, and RRR" (ERG, 2023a).

-------
3.5 Emission Points and Controls

The EPA evaluated developments in practices, processes, and control technologies for
heat exchange systems, storage vessels, process vents, wastewater, and equipment leaks for
processes subject to the HON, P&R I, and P&R II. Moreover, for the NSPS subpart VVa, we
evaluated BSER for equipment leaks; and for the NSPS subparts III, NNN, and RRR we
evaluated BSER for process vents associated with air oxidation units, distillation operations, and
reactor processes, respectively. We analyzed costs and impacts for each emission source (e.g.,
process vents) by each rule. For the different NSPS, we determined cost-effectiveness, cost per
ton of emissions reduced, on a VOC basis. For each NESHAP, we determined cost-effectiveness
on a HAP basis from the VOC emissions. The estimation of total capital cost (synonymous with
total capital investment) and total annual cost follows the methodology in the EPA Air Pollution
Control Cost Manual (U.S. EPA, 2017). Estimates of total annual cost presented in this chapter
include both operating and maintenance and annualized capital costs (from capital recovery) and
are estimated for each year in the analytic timeframe for this action of 2024-2038. The capital
costs are annualized at 5.5% over a time period of 15 years (2024-2038), a time period that
reflects the life of control equipment and measures applied to reduce HAP emissions. All costs
are in 2021 dollars.

3.5.1 Flares

Flares that control emissions from processes subject to HON or the P&R I NESHAP are
required to meet certain design and operating requirements as specified in 40 CFR 63.11. The
available data at the time these NESHAP were promulgated suggested that flares meeting these
design and operating requirements would achieve a minimum destruction efficiency of no less
than 98 percent emissions control. Relatively recent evidence through Passive Fourier Transform
Infrared spectroscopy (PFTIR) testing suggests that steam- and air-assisted flares can have a
significant degradation in destruction efficiency when operated at high turndown or at other
times when steam- and air-assist rates are too high (EPA, 2012). As many of the flares operated
at HON and P&R I facilities are steam- or air-assisted, concerns of poor flare destruction
efficiency are particularly significant. We note that a substantial portion of the costs, both capital
and annual, for the final HON and P&R I amendments and for this entire final action, are from
the control requirements for flares. Tables 3-1 and 3-2 present cost and emission reductions for

-------
flare control options included as part of the HON amendments under this final action. More
information on these systems and control options can be found in the preamble for this
rulemaking.

Table 3-1 Nationwide Cost Impacts (2021$) for Flares in the SOCMI Source Category
that Control Emissions from HON Processes including P&R I Flares
Collocated with HON Processesa'b



Total Capital Investment

Total Annualized Cost

Control Option

(MM$)

(MM$/yr)

Flare Operational and Monitoring
Requirements

Work Practice Standards for Flares

323.1

67.8

Operating Above Their Smokeless

3.34

0.79

Capacity





Nationwide Total

326.4

68.6

(a)	We were unable to quantify emissions reductions for this option but anticipate some excess emissions reductions.

(b)	VOC and HAP emission reductions are anticipated excess emissions impacts based on ensuring flares achieve
the MACT level of control.

Table 3-2 Nationwide VOC and HAP Emissions Reductions and Cost-Effectiveness for

Flares in the SOCMI Source Category that Control Emissions from
	HON Processes including P&R I Flares Collocated with HON Processes	

Control Option

VOC
Emission
Reductions
(ton/yr)

VOC
Cost-
Effectiveness
($/ton)

HAP
Emissions
Reductions
(ton/yr)

HAP Cost-
Effectiveness
($/ton)

Flare Operational and
Monitoring Requirements(a)
Work Practice Standards for
Flares Operating Above
Their Smokeless Capacity)
Jb)	

19,325

3,512

4,717

14,387

(a)	VOC and HAP emission reductions are anticipated excess emissions impacts based on ensuring flares achieve
the MACT level of control.

(b)	There are no emission reductions for the work practice standards which is why this row is empty.

The nationwide impacts for applying the new flare operating and monitoring
requirements to flares that control emissions from P&R I processes and applying alternative
work practice standards for visible emissions events above the smokeless capacity of each of
these flares are summarized in Tables 3-3 and 3-4.

48

-------
Table 3-3 Nationwide Cost Impacts (2021$) for Flares that Control Emissions from

P&R I Processes

Control Option

Total Capital Investment

(MM$)

Total Annualized Cost
(MM$/yr)

Flare Operational and Monitoring
Requirements

Work Practice Standards for Flares
Operating Above Their Smokeless
Capacity



6.93
0.08



1.46
0.02

Nationwide Total



7.1



1.48

Table 3-4 Nationwide VOC and HAP Emissions Reductions and Cost-Effectiveness for
Flares that Control Emissions from P&R I Processes

Control Option

VOC
Emission
Reductions
(ton/yr)

VOC
Cost-
Effectiveness
($/ton)

HAP
Emissions
Reductions
(ton/yr)

HAP Cost-
Effectiveness
($/ton)

Flare Operational and
Monitoring Requirements(a)
Work Practice Standards for
Flares Operating Above
Their Smokeless Capacity(b)

564

2,594

141

10,378

(a)	VOC and HAP emission reductions are anticipated excess emissions impacts based on ensuring flares achieve
the MACT level of control.

(b)	There are no emission reductions for the work practice standards which is why this row is empty.

With respect to nationwide impacts, we anticipate that there would be no additional cost
impacts for additional monitoring equipment needed to operate these flares at existing HON
operations, as well as no additional emissions reductions realized. Multi-point ground flares are
primarily secondary flares. Facilities that currently use these controls will already have the
necessary monitoring equipment in place.

3.5.2 Fenceline Monitoring

Fenceline monitoring refers to the placement of monitors along the perimeter of a facility
to measure pollutant concentrations. Coupled with requirements for root cause analysis and
corrective action upon triggering an actionable level, this work practice standard is a
development in practices considered under CAA section 112(d)(6) for the purposes of managing
fugitive emissions. Below in Tables 3-5 and 3-6 are costs and emission reductions associated
with several fenceline monitoring scenarios or options in the final HON and P&R I amendments.
More information on fenceline monitoring and implementation options can be found in the
preamble for this rulemaking.

49

-------
Table 3-5 Nationwide Cost Impacts of Fenceline Monitoring for HON

Monitoring

Number of

Monitoring option

Total capital

Total annualized

scenario

Facilities
Impacted

description

investment ($)

costs ($/yr)

1

35

Passives only (1
analyte)*

4,016,000

2,141,000

2

46

Passives only (2
analytes)

2,295,000

1,282,000

3

9

Cannisters only

115,500

5,366,000

4

16

Cannisters and
passives (1 analyte)

1,606,000

10,397,000

5

20

Cannisters and
passives (2 analytes)

1,721,000

12,869,000

* An analyte is the chemical substance being measured. For these fenceline monitoring options, the analyte is a HAP
monitored under HON and P&R I requirements.

	Table 3-6 Nationwide Cost Impacts of Fenceline Monitoring for P&R I	

Monitoring
scenario

Number of

Facilities

Impacted

Monitoring option
description

Total capital
investment ($)

Total annualized
costs ($/yr)

1

2

Cannisters and
passives (2 analytes)
Cannisters only

114,700
12.800

659,000
596,000

3.5.3 Pressure Relief Devices

Pressure relief devices (PRD) are designed to open when the system pressure exceeds a
set pressure, allowing the release of vapors or liquids until the system pressure is reduced to its
normal operating level. When the normal pressure is re-attained, the valve re-seats and a seal is
again formed. Leaks can occur as a result of improper reseating, or if the process is operating too
close to the set pressure of the PRD and the PRD does not maintain its seal. PRD emissions can
be controlled by venting emissions to a closed-vent system or through the use of a rupture disk.

For PRD releases, we are revising the definition of "pressure relief device" for the HON
and P&R I, the definition of "relief valve" for the HON and P&R I, and the definition for
"pressure relief device" in P&R II. Under CAA section 112(h), we are finalizing a work practice
standard for PRDs at 40 CFR 63.165(e) (for HON) and 40 CFR 63.502(a)(1) and (a)(2) (which
references 40 CFR 63.165, for P&R I) that consists of using at least three prevention measures
and performing root cause analysis and corrective action in the event that a PRD does release
emissions directly to the atmosphere. Examples of prevention measures include flow indicators,
level indicators, temperature indicators, pressure indicators, routine inspection and maintenance
programs or operator training, inherently safer designs or safety instrumentation systems, deluge

50

-------
systems, and staged relief systems where the initial PRD discharges to a control system. Table
3-7 provides cost estimates for PRDs in the HON and P&R I and II.

Table 3-7 Nationwide Cost Impacts of Control Options Considered
for Pressure Relief Devices at HON and P&R I and II Facilities

Control option 1

Total capital investment ($)

Total annualized costs ($/yr)

HON

16,829,351

7,481,607

P&R I

504,350

128,475

P&R II

132,724

33,809

3.5.4 Storage Vessels

Storage vessels are used to store liquid and gaseous feedstocks for use in a process, as
well as to store liquid and gaseous products from a process. Most storage vessels are designed
for operation at atmospheric or near atmospheric pressures; high pressure vessels are used to
store compressed gases and liquefied gases. Atmospheric storage vessels are typically cylindrical
with a vertical orientation, and they are constructed with either a fixed roof or a floating roof.
Some, generally small, atmospheric storage vessels are oriented horizontally. High pressure
vessels are either spherical or horizontal cylinders.

Below in Tables 3-8 through 3-10 is a presentation of different control options considered
for storage vessels, and then costs and emissions reductions for the different regulatory options.
More information on these systems and control options can be found in the preamble for this
rulemaking.

Table 3-8 Summary of Storage Vessel Control Options Evaluated for the HON

Storage Vessel

Control Option	Control Option Description	

Revise the HON and P&R INESHAP applicability threshold to require existing storage
vessels between 38 m3 (10,000 gal) and 151m3 (40,000 gal) with a vapor pressure >6.9
SV1	kPa to add control.

	Control is assumed to be a retrofitted IFR	

„ SV1 plus require upgraded deck fittings and controls for guide poles for all IFR storage
io V I* ,
	vessels	

Convert each EFR to an IFR through installation of a geodesic dome plus require
	upgraded deck fittings and controls for guide poles.	

51

-------
Table 3-9 Nationwide Emissions Reductions and Cost Impacts of Control
	Options Considered for Storage Vessels at HON Facilities

Control

Total capital

Total

voc

HAP

HAP cost

option

investment

annualized

emission

emission

effectiveness



($)

costs ($/yr)

reductions

reductions

($/ton)







(tpy)

(tpy)



1

1,727,000

327,400

58.0

40.6

8,070

2

2,191,500

415,500

68.2

47.7

8,710

3

28,916,200

4,065,700

84.3

59.0

68,880

Table 3-10 Nationwide Emissions Reductions and Cost Impacts of Control Options
Considered for Storage Vessels at P&R I Facilities (not collocated with HON facilities)

Control
option

Total capital
investment ($)

Total annualized costs

($/yr)

VOC
emission
reductions
(tpy)

HAP
emission
reductions
(tpy)

HAP cost
effectiveness
($/ton)

1

109,000

20,700

3.7

2.6

7,960

2

131,000

24,800

4.1

2.9

8,550

3

912,200

128,300

2.7

1.9

67,500

3.5.5 Heat Exchange Systems

Heat exchangers are devices or collections of devices used to transfer heat from process
fluids to another process fluid (typically water) without intentional direct contact of the process
fluid with the cooling fluid {i.e., non-contact heat exchanger). There are two types of heat
exchange systems: Closed-loop recirculation systems and once-through systems. Closed-loop
recirculation systems use a cooling tower to cool the heated water leaving the heat exchanger and
then return the newly cooled water to the heat exchanger for reuse. Once-through systems
typically use surface freshwater {e.g., from rivers) as the influent cooling fluid to the heat
exchangers, and the heated water leaving the heat exchangers is then discharged from the
facility.

Based on the HON heat exchange system technology review at proposal, we identified
the following control option (a development in practice as defined in section 112(c)(6)) for heat
exchange systems: quarterly monitoring with the Modified El Paso Method.16 This method uses a

16 ERG, 2023. Clean Air Act Section 112(d)(6) Technology Review for Heat Exchange Systems Located in the
SOCMI Source Category that are Associated with Processes Subject to HON and for Heat Exchange Systems
that are Associated with Processes Subject to Group I Polymers and Resins NESHAP. March 2023. EPA Docket
No. EPA-HQ-OAR-2022-0730.

52

-------
leak action level defined as a total strippable hydrocarbon concentration (as methane) in the
stripping gas of 6.2 parts per million by volume (ppmv). This method does not allow the delay of
repair of leaks for more than 30 days where a total strippable hydrocarbon concentration (as
methane) in the stripping gas of 62 ppmv or higher is found. This option would also require re-
monitoring at the monitoring location where the leak was identified to ensure that any leaks
found are fixed. More information on these systems and control options can be found in the
preamble for this rulemaking.

Table 3-11 VOC and HAP Cost Effectiveness for the Control Option Evaluated

for Heat Exchange Systems at HON Facilities (2021$)

Control
option

Total capital
investment ($)

Total annualized
costs ($/yr)

VOC emission
reductions (tpy)

HAP emission
reductions (tpy)

HAP cost
effectiveness
($/ton)

1

784,000

238,000

934

93

2,441

3.5.6 Process Vents

A process vent is a gas stream that is discharged during the operation of a particular unit
operation (e.g., separation processes, purification processes, mixing processes, reaction
processes). The gas stream(s) may be routed to other unit operations for additional processing
(e.g., a gas stream from a reactor that is routed to a distillation column for separation of
products), sent to one or more recovery devices, sent to a process vent header collection system
(e.g., blowdown system) and air pollution control device (APCD) (e.g., flare, thermal oxidizer,
carbon adsorber), and/or vented to the atmosphere. Process vents may be generated from
continuous and/or batch operations,17 as well as from other intermittent types of operations (e.g.,
maintenance operations). If process vents are required to be controlled prior to discharge to the
atmosphere to meet an applicable emissions standard, then they are typically collected and routed
to an APCD through a closed vent system.

During the public comment period for the proposed amendments, the EPA received
additional information from commenters on costs necessary for control of process vents that emit
greater than or equal to 1.0 lb/hr of total organic HAP.

17 P&R I and P&R II regulate process vents from both continuous and batch operations. The HON and NSPS
subparts III, NNN, and RRR only regulate process vents if some, or all, of the gas stream originates as a
continuous flow.

53

-------
The proposed cost estimate underestimated flow rates needed to route Group 2
continuous process vents with greater than or equal to 1.0 lb/hr of total organic HAP to APCDs.
The revised estimates reflect the limitations of the correlations associated with the EPA's control
cost template which starts with a flow rate of 500 standard cubic feet per minute (scfm). With
these corrections, we estimate the average total capital investment (TCI) to install a new
recuperative thermal oxidizer (for both HON and P&R I) is about $167,000 (as opposed to the
$66,000 proposed).

In light of the fact that commenters were generally concerned about the cost estimate, we
performed additional analyses to evaluate the cost effectiveness to remove the TRE concept from
HON and P&R I. Using a TCI of $1,000,000 as provided by the commenter, and the EPA's
control cost template (for installing a new recuperative thermal oxidizer with 70 percent energy
recovery), we estimated an annual cost of approximately $330,000 (for HON) and $318,000 (for
P&R I).

Tables 3-12 through 3-14 include a summary of control options for process vents, and
costs and emission reductions for each option and type of process vents considered under each of
the rules in this final rulemaking. More information on these systems and control options can be
found in the preamble for this rulemaking.

Table 3-12 Summary of Continuous Process Vent Control Options Evaluated for the

HON and P&R I NESHAP

Process Vent

Control Option Description

Control Option

PV1

• Remove TRE concept in its entirety from HON and P&R I NESHAP.

• Remove 50 ppmv and 0.005 scmm Group 1 process vent thresholds from HON and P&R

I NESHAP.

• Redefine a HON and P&R I NESHAP Group 1 process vent (require control) as any

process vent that emits >1.0 lb/hr of total organic HAP.

PV2

• Same as PV1 but redefine a HON and P&R I NESHAP Group 1 process vent (require

control) as any process vent that emits >0.10 lb/hr of total organic HAP.

PV3

• Keep TRE concept in HON and P&R I NESHAP but change index value threshold from

1.0 to 5.0 Keep 50 ppmv and 0.005 scmm Group 1 process vent thresholds in HON and

P&R I NESHAP.

-------
Table 3-13 Nationwide Emissions Reductions and Cost Impacts of Control Options
Considered for Continuous Process Vents at HON and P&R I Facilities

Control
option 1

Total capital
investment ($)

Total annualized
costs ($/yr)

VOC emission
reductions (tpy)

HAP emission
reductions (tpy)

HAP cost

effectiveness

($/ton)

HON-
PV1

$16,171,549

$5,295,079

538

538

$9,835

PV2
PV3
P&R I -
PV1

$21,903,378
$17,664,900

$3,002,400

$15,624,555
$25,954,031

$966,000

809
441

60.4

533
441

60.4

$29,314
$58,853

$15,992

PV2
PV3

$3,559,327
$3,217,562

$2,208,379
$1,555,812

80.1
54.8

72.4
54.8

$30,502
$28,391

Table 3-14 Nationwide Emissions Reductions and Cost Impacts of Control Options
Considered for Batch Front-end Process Vents at P&R I Facilities

Control
option

Total capital
investment ($)

Total annualized
costs ($/yr)

VOC emission
reductions (tpy)

HAP emission
reductions (tpy)

HAP cost

effectiveness

($/ton)

1

811,000

650,700

105

105

6,200

3.5.6.1 Process Vents - Subpart Ilia, NNNa, and RRRa NSPS

The EPA has used a TRE index value of 1.0 as a basis for distinguishing process vents in
the SOCMI NSPS (40 CFR 60 subparts III, NNN, and RRR) rulemakings. In general, for these
rules, process vents with a TRE index value equal to or less than 1.0 are required to be controlled
(given that process vents with low TRE index values tend to have both higher emission stream
flow rates and higher emission rates than process vents with higher TREs). Increased flow from a
vent generally corresponds with increased size of the unit operation and increased production
rate.

The TRE index is derived from the cost effectiveness associated with VOC control by
thermal oxidation, and is a function of vent stream flowrate, vent stream net heating value,
hourly emissions, and a set of coefficients. The coefficients are based on both the net heating
value of the vent stream and whether or not the vent stream is chlorinated.

We evaluated cost and VOC emissions reductions for the HON for the same process vent
control options for the same unit operations (i.e., air oxidation unit processes, distillation
operations, and reactor processes). Therefore, we used the average cost and emissions reductions
that we determined for process vents subject to the HON to evaluate the costs, VOC emission

55

-------
reductions, and cost-effectiveness of the process vent control options for NSPS subparts III,
NNN, and RRR. Table 3-15 summarizes these average HON cost and VOC emissions
reductions.

Table 3-15 Average Cost and Emission Reductions for Process Vents Subject to the

HON Used for the Suite of Process Vent Requirements Evaluated
	for the NSPS subparts Ilia, NNNa, and RRRa	



Total



Total Annual



Description

Capital

Total Annual

Cost w /

VOC Emission

Investment

($)

Cost ($/yr)

Recovery
Credits ($/yr)

Reductions (tpy)

Flare monitoring requirements1

3,752,200

789,200

789,200

93

Maintenance vent requirements2

-

460

460

-

Revising the standard from a TRE









calculation to control of all vent

39,300

98,400

98,400

9.1

streams3









Adsorber monitoring (carbon

26,500

2,500

2,500

0.21

cannisters)4

1	For additional details, see the document titled Control Option Impacts for Flares Located in the SOCMI Source
Category that Control Emissions from Processes Subject to HON and for Flares that Control Emissions from
Processes Subject to Group I and Group II Polymers and Resins NESHAPs, which is available in the docket for this
rulemaking.

2	For additional details, see the document titled Review of Regulatory Alternatives for Certain Vent Streams in the
SOCMI Source Category that are Associated with Processes Subject to HON and Processes Subject to Group I and
Group II Polymers and Resins NESHAPs, which is available in the docket for this rulemaking.

3	For additional details, see the document titled Clean Air Act Section 112(d)(6) Technology Review for Continuous
Process Vents Located in the SOCMI Source Category that are Associated with Processes Subject to HON,
Continuous Front-end and Batch Front-end Process Vents Associated with Processes Subject to Group I Polymers
and Resins NESHAP, and Process Vents Associated with Processes Subject to Group II Polymers and Resins
NESHAP, which is available in the docket for this rulemaking.

4	For additional details, see the document titled Analysis of Monitoring Costs and Dual Bed Costs for Non-
Regenerative Carbon Adsorbers Used in the SOCMI Source Category that are Associated with Processes Subject to
HON and for Non-Regenerative Carbon Adsorbers that are Associated with Processes Subject to Group I Polymers
and Resins NESHAP, which is available in the docket for this rulemaking.

We analyzed the application of these requirements for process vents, varying the
application based on whether new or existing facilities are affected and if the site is new,
modified or reconstructed. For scenario 1 (i.e., affected facility is at a new greenfield facility),
we assumed only one non-HON greenfield facility will trigger NSPS subpart III, NNN, or RRR
over the next five years. We also assumed this greenfield facility would not be subject to the
EMACT standards, MON, and petroleum refineries NESHAP, and the facility will use one flare
and one non-flare APCD to control all their SOCMI NSPS unit operations.

We used facility responses to the EPA's CAA section 114 request to help us determine
the number of facilities that could potentially trigger scenarios 2, 3, and 4.

56

-------
For scenario 2 (new affected facilities constructed at existing plant sites), we estimate six
new affected facilities will be built and be subject to new requirements in a new NSPS subpart
Ilia, NNNa, or RRRa over the next five years. Facilities responding to the EPA's CAA section
114 request had 500 unit operations subject to either NSPS subpart III, NNN, or RRR; and only
one of these unit operations was new construction in the last five years and not subject to the
HON. We determined that there are currently 284 SOCMI facilities subject to either NSPS
subpart III, NNN, or RRR; and 196 of these are non-HON facilities. Based on our CAA section
114 data, HON facilities have on average 45 unit operations per facility. Assuming non-HON
facilities are smaller, we estimate that non-HON facilities subject to either NSPS subpart III,
NNN, or RRR have 15 unit operations per facility. Assuming the same distribution of new
construction for non-HON facilities, we estimate that six new affected facilities (one new unit
operation per non-HON facility subject to either NSPS subpart III, NNN, or RRR), would have

been constructed in the last five years (— * 15 * 196). This analysis assumes that the same

number of unit operations that were constructed in the last five years would be constructed in the
next five years.

For Scenarios 3 and 4 (existing facility is modified or reconstructed), we estimate 12
existing affected facilities will trigger new requirements in a new NSPS subpart Ilia, NNNa, or
RRRa over the next five years due to modification or reconstruction. Facilities responding to the
EPA's CAA section 114 request had 500 unit operations subject to either III, NNN, or RRR;
however, only two of these unit operations were modified or reconstructed in the last 5 years and
not subject to the HON. Using similar procedures as described above for scenario 2, we estimate
that 12 modified or reconstructed affected facilities (1 modified or reconstructed unit operation
per non-HON facility subject to the NSPS), would have been modified or reconstructed in the

2

last five years (— * 15 * 196). This analysis assumes that the same number of unit operations

that were modified or reconstructed in the last five years would be modified or reconstructed in
the next five years.

Table 3-16 provides nationwide capital and annual costs to comply with the process vent
control options.

57

-------
Table 3-16 Nationwide Emissions Reductions and Cost Impacts of Control Options

Considered for Non-HON Vent Streams Triggering NSPS Subparts
	Ilia, NNNa, and/or RRRa	

Scenario

Total Capital
Investment

($)

Total Annual
Cost ($/yr)

Total Annual
Cost w/
Recovery
Credits ($/yr)

voc

Emission
Reductions

(tpy)

Cost-
effectiveness
w/ Recovery
Credits ($/ton
VOC)

Scenario 1

(i.e., one affected facility
at a new greenfield
facility)

Scenario 2 (i.e., new
affected facility at six
existing facilities)
Scenarios 3 and 4 (i.e.,
12 existing affected
facilities modified or
triggers the
reconstruction
requirements)	

1,665,300

7,609,500

15,192,500

461,000

461,000

93

1,780,000	1,780,000	392

3,558,000 3,558,000	783

4,960

4,540

4,540

Total

24,467,300

5,799,800 5,799,800 1,269

4,570

3.5.7 Wastewater

HAP are emitted into the air from wastewater collection, storage, and treatment systems
that are uncovered or open to the atmosphere through volatilization of organic compounds at the
liquid surface. Emissions occur by diffusive or convective means, or both. Diffusion occurs
when organic concentrations at the water surface are much higher than ambient concentrations.
The organics volatilize, or diffuse into the air, to reach equilibrium between aqueous and vapor
phases. Convection occurs when air flows over the water surface, sweeping organic vapors from
the water surface into the air. The rate of volatilization is related directly to the speed of the air
flow over the water surface.

The HON defines wastewater to mean water that: (1) contains either: (i) an annual
average concentration of compounds of at least 5 part per million by weight (ppmw) and has an
annual average flow rate of 0.02 liter per minute (1pm) or greater or (ii) an annual average
concentration of compounds of at least 10,000 ppmw at any flow rate, and that (2) is discarded
from a CMPU that meets all of the criteria specified in 40 CFR 63.100 (b)(1) through (b)(3).
Wastewater is process wastewater or maintenance wastewater.

P&R II defines wastewater as aqueous liquid waste streams exiting equipment at an
affected source. No further stratification into groups for applicability is specified.

58

-------
Below in Tables 3-17 and 3-18 are costs and emission reductions for control options
considered for wastewater under the final HON amendments, and P&R I. More information on
these systems and control options can be found in the preamble for this final rulemaking.

Table 3-17 Nationwide Emissions Reductions and Cost Impacts of Control Options
Considered for Wastewater Streams at HON Facilities

Control
option

Total capital
investment ($)

Total annualized VOC emission HAP emission
costs ($/yr) reductions (tpy) reductions (tpy)

HAP cost

effectiveness

($/ton)

1

504,766,000

210,739,500 2,755 2,755

76,500

Table 3-18 Nationwide Emissions Reductions and Cost Impacts of Control Options
Considered for Wastewater Streams at P&R I Facilities

Control
option

Total capital
investment ($)

Total annualized VOC emission HAP emission
costs ($/yr) reductions (tpy) reductions (tpy)

HAP cost

effectiveness

($/ton)

1

46,847,800

22,548,200 220 220

102,500

3.5.8 Equipment Leaks

Emissions of VOC and HAP from equipment leaks occur in the form of gases or liquids
that escape to the atmosphere through many types of connection points (e.g., threaded fittings) or
through the moving parts of certain types of process equipment during normal operation.
Equipment regulated by the HON, P&R I, and P&R II includes agitators, compressors,
connectors, instrumentation systems, OEL, PRDs, pumps, sampling collection systems, and
valves18 that contain or contact material that is five percent by weight or more of organic HAP,
operate 300 hours per year or more, and are not in vacuum service.

Based on the costs and emission reductions for the options, we determined that none of
them are cost effective. Therefore, we are not revising the HON, P&R I, and P&R II to reflect
the requirements of these options pursuant to CAA section 112(d)(6). However, tighter
requirements on equipment leaks will be included under Subpart VVb NSPS. Table 3-19
provides costs and emission reductions for these tighter requirements by type of affected

18 We believe P&R II contains a typographical error in that valves are currently excluded from the definition of
equipment leaks at 40 CFR 63.522; see section III.D. 10 of the preamble for this rulemaking for our rationale for
this conclusion.

59

-------
facilities as well as the total costs and emission reductions. More information on these systems
and control options can be found in the preamble for this final rulemaking.

Table 3-19 Nationwide Emissions Reductions and Cost Impacts of Control Options

Considered for Affected Facilities Triggering NSPS Subpart Wb







Total Annual



Cost-
effectiveness w/

Recovery
Credits ($/ton
VOC)

Scenario

Total Capital

Total Annual

Cost w/

VOC Emission

Investment ($)

Cost ($/yr)

Recovery
Credits ($/yr)

Reductions (tpy)

Scenario 1

416,600

77,500

60,900

18

3,380

(i.e., two affected











facilities at a new











greenfield facility)
Scenario 2 (i.e., 34

7,081,700

1,317,900

1,035,800

313

3,310

new affected











facilities)
Scenarios 3 and

208,300

38,800

30,500

9

3,390

(i.e., one modified











existing affected
facility)











Total

7,706,600

1,434,200

1,127,200

340

3,320

3.6 Engineering Cost Analysis Summary Results

Table 3-20 below presents a summary of the costs for the HON amendments by emission
point and in total. Capital and total annual costs are shown, and total annual costs are shown with
and without product recovery. Engineering cost estimates in this chapter include projections of
revenue from product recovery. This is because some control options analyzed in this RIA lead
to the recovery of chemical products (e.g., leak detection and repair (LDAR) for equipment leak
emissions). Recovered chemical product affected by this rulemaking is monetized as recovery
credits by multiplying VOC emissions reductions by a VOC credit of $900/ton (2021 dollars). A
recovery credit of $900 per ton VOC was applied to the VOC emission reductions in the analyses
to calculate the savings in chemicals not being emitted (i.e., lost). The $900 per ton recovery
credit has historically been used by EPA to represent the variety of chemicals that are used as
reactants and produced at synthetic organic chemical manufacturing (SOCMI) facilities (EPA,
2007).

The effect of product recovery on the total annual costs is quite small, as is shown in the
table. The total capital cost of the final HON amendments is about $455 million, and the total
annual cost is about $168 million (with product recovery) and $169 million (without product

60

-------
recovery) in 2021 dollars. The estimation of total capital cost (synonymous with total capital
investment) and total annual cost follows the methodology in the EPA Air Pollution Control Cost
Manual (U.S. EPA, 2017). Estimates of total annual cost includes both operating and
maintenance and annualized capital costs (from capital recovery) and are estimated for each year
in the analytic timeframe for this action of 2024-2038. The inclusion of product recovery reduces
the total annual cost by only 0.5 percent (about $980,000, as shown in the table), but its inclusion
leads to annual cost savings from controls for heat exchange systems.

In the table below the ethylene oxide (EtO) Risk line item refers to costs considered for
EtO specific changes to HON facilities for equipment leaks, wastewater, process vents, flare
maintenance/equip openings, storage vessels, and heat exchange systems. While the three
separate storage vessel line items refer to various measures the rule is addressing for that
emission point. For example, the pressure vessel line is the cost estimate for monitoring, repair,
recordkeeping, and reporting for annual Method 21 monitoring of a pressure vessel. For me
details on the different rule provision please see the preamble for this final rulemaking.

Table 3-20 Detailed Costs for the HON Source Category by Emission Point for
	the Final Rule (2021$)	





Total Annual Cost

Total Annual

Annual

Emission Point

Total Capital Cost

($/yr)

Cost ($/yr)

Recovery

($)

Without Recovery

With Recovery

Credits





Credits

Credits

($/yr)

Flares

326,443,400

68,658,000

68,658,000

0

Fenceline Monitoring

9,754,300

32,055,300

32,055,300

0

Pressure Relief Devices

16,829,400

7,481,600

7,481,600

0

Storage Vessels

2,191,500

415,500

415,500

0

Storage Vessels - Degassing

0

751,500

751,500

0

Storage Vessels - Pressure

77,700

72,900

72,900

0

Vessels

Storage Vessels - 240hr

2,637,400

456,500

456,500

0

Maintenance

Maintenance

0

94,200

94,200

0

Heat Exchange Systems

783,800

237,700

-603,000

840,800

Process Vents

16,171,500

5,294,600

5,294,600

0

EtO Risk

76,695,000

51,453,100

51,315,400

137,651

Dioxins/Furans

3,920,000

2,275,000

2,275,000

0

Carbon Cannisters

53,000

5,000

5,000

0

Total

455,557,700

169,250,100

168,273,200

983,900

Note: The estimation of total capital cost (synonymous with total capital investment) and total annual cost follows
the methodology in the EPA Air Pollution Control Cost Manual (U.S. EPA, 2017). Estimates of total annual cost
includes both operating and maintenance and annualized capital costs (from capital recovery) and are estimated for

each year in the analytic timeframe for this action of 2024-2038. The assumed interest rate for the annualized cost is
5.5 percent.

61

-------
Tables 3-21 and 3-22 below present a summary of the costs for the final P&R I (Table 3-
21) and P&R II (Table 3-22) amendments by emission point and in total. Capital and total annual
costs are shown, and total annual costs are shown with and without product recovery. The effect
of product recovery on the total annual costs is quite small, as is shown in Table 3-21 for P&R I.
The total capital cost of the final P&R I amendments is about $28 million, and the total annual
cost is about $15 million (with product recovery) and $15 million (without product recovery) in
2021 dollars. The estimation of total capital cost (synonymous with total capital investment) and
total annual cost follows the methodology in the EPA Air Pollution Control Cost Manual (EPA,
2017). Estimates of total annual cost includes both operating and maintenance and annualized
capital costs (from capital recovery). The inclusion of product recovery reduces the total annual
cost by only 0.2 percent (about $29,000, as shown in the table), but its inclusion leads to annual
cost savings from controls for heat exchange systems. For the P&R II final amendments, Table
3-22 shows that the total capital cost is about $3 million, with about $2 million in total annual
costs and no product recovery.

For the various NSPS rules, as shown in Table 3-23, the total capital cost for the subpart
VVb is about $8 million, with a total annual cost of just over $1 million with product recovery
included. With product recovery included, the total annual cost is about $300,000 lower. For the
other three NSPS (subpart Ilia, NNNa, RRRa) considered together, the total capital cost is about
$28 million, with a total annual cost of about $6 million as shown in Table 3-23. There is no
product recovery associated with the controls to meet the requirements for these three NSPS.

Finally, the cumulative total capital cost for the entire final action, as shown in Table 3-
23, is about $522 million, with a total annual cost for the entire action of $193 million with
product recovery. Given that the product recovery is just over $1.3 million, the total annual cost
without product recovery is $194 million. The cumulative product recovery is only about 0.6
percent of the total annual costs.

Because the controls considered lead to product recovery, it is possible for the cost of a
control option to be negative once the value of product recovery is considered (the potential
annualized costs may be outweighed by the revenue from product recovery). This observation
may typically support an assumption that owners of facilities would continue to perform the
emissions abatement activity regardless of whether a requirement is in place, because it is in their

-------
private self-interest. However, there may be an opportunity cost associated with the installation
of environmental controls or implementation of compliance activities (for purposes of mitigating
the emission of pollutants) that is not reflected in the control costs. If environmental investment
displaces investment in productive capital, the difference between the rate of return on the
marginal investment displaced by the mandatory environmental investment is a measure of the
opportunity cost of the environmental requirement to the regulated entity. To the extent that any
opportunity costs are not added to the control costs, the compliance costs presented above may
be underestimated. In addition, the hurdle rate is defined as the minimum rate of return on an
investment that a firm would deem acceptable under typical business practices. Thus, if the
hurdle rate is higher on average for firms in this industry than the interest rate used in estimating
the compliance costs (in this final action, 5.5 percent at the time of this analysis), then these
investments in environmental controls may not necessarily be undertaken on average.

From a social perspective, however, the increased financial returns from product recovery
accrue to entities somewhere along the chemical product supply chain and should be accounted
for in a national-level analysis. An economic argument can be made that, in the long run, no
single entity bears the entire burden of compliance costs or fully appropriates the financial gain
of the additional revenues associated with chemical product recovery. The change in economic
surplus resulting from product recovery may be likely to be spread across different market
participants. The simplest and most transparent option for allocating these revenues would be to
assign the compliance costs and revenues to a model plant and not make assumptions regarding
the allocation of costs and revenues across economic agents. In the RIA, we treat these revenues
as an offset to projected compliance costs, while the revenues may also be considered as a
benefit of the regulatory action. However, regardless of whether the revenue from capture of
natural gas is considered a compliance cost offset or a benefit, the net benefits are equivalent.

63

-------
Table 3-21 Detailed Costs for the P&R I Source Category by Emission Point
	for the Final Rule (2021$)	

Emission Point

Total Capital
Cost ($)

Total Annual Cost

($/yr)
Without Recovery
Credits

Total Annual

Cost ($/yr)
With Recovery
Credits

Annual
Recovery
Credits
($/yr)

Flares

6,996,100

1,481,000

1,481,000

0

Fenceline Monitoring

127,600

1,255,000

1,255,000

0

Pressure Relief Devices

504,400

128,500

128,500

0

Storage Vessels

130,900

24,800

24,800

0

Storage Vessels - Degassing

0

12,300

12,300

0

Storage Vessels - Pressure Vessels

2,200

2,100

2,100

0

Storage Vessels - 240hr Maintenance

39,500

6,800

6,800

0

Maintenance

0

8,700

8,700

0

Heat Exchange Systems

48,300

9,900

(19,300)

29,300

Process Vents

3,813,400

1,616,700

1,616,700

0

CP Risk

15,948,900

10,354,500

10,354,500

0

Dioxins/Furans

560,000

325,000

325,000

0

Carbon Cannisters

27,000

2,000

2,000

0

Total

28,198,300

15,227,100

15,198,000

29,200

Table 3-22 Detailed Costs for the P&R II Source Category by Emission Point

for the Final Rule (2021$)





Total Annual Cost

Total Annual Cost

Annual

Emission Point

Total Capital
Cost ($)

($/yr)

Without Recovery
Credits

($/yr)

With Recovery
Credits

Recovery
Credits

($/yr)

Pressure Relief Devices

132,700

33,800

33,800

0

Storage Vessels - Degassing

0

6,200

6,200

0

Maintenance

0

2,300

2,300

0

Dioxins/Furans

2,800,000

1,625,000

1,625,000

0

Total

2,932,500

1,667,200

1,667,200

0

Table 3-23

Summary of the Total Costs by Final Rule ($2021)







Total Annual Cost

Total Annual Cost

Annual

Rule

Total Capital
Cost ($)

($/yr)
Without Recovery
Credits

($/yr)
With Recovery
Credits

Recovery
Credits

($/yr)

HON

455,557,700

169,250,100

168,273,200

983,900

P&R I

28,198,300

15,227,100

15,198,000

29,200

P&R II

2,932,500

1,667,200

1,667,200

0

NSPS Wb

7,706,600

1,434,200

1,127,200

307,000

NSPS Ilia, NNNa, & RRRa

27,844,000

6,269,000

6,269,000

0

Total

522,239,100

193,847,600

192,534,600

1,320,100

We also show the costs in another way - the current day estimate of the costs of these
rules over an analysis time period, and an equivalent annualized value (EAV) of those costs over
the same analysis time period. To facilitate the presentation of these costs, Table 3-24 presents

64

-------
the present value (PV) and EAV of costs over the analysis time period of 2024-2038 for the
cumulative impacts in this rulemaking, discounted to 2023. The present value is a current day
estimate of the costs over the analysis time period for this final rulemaking, and the EAV is the
average annual value of these costs whose sum is the PV. These costs include the value of
product recovery, which is a very small percentage of the costs for the HON (less than 1 percent
of the total annual costs) and cumulatively (also less than 1 percent of the total annual costs).
Showing the costs in this way is consistent with OMB Circular A-4.

Table 3-24 Discounted Costs, for the Final Amendments to the HON, P&R I, and P&R
II NESHAP, and Subparts Wb, Ilia, NNNa, and RRRa NSPS, 2024-2038 (million 2021$,
discounted to 2023)

Year 3 percent 7 percent

Total Annual Cost with Revenue Total Annual Cost with Revenue
from Product Recovery from Product Recovery

2024

169

163

2025

164

152

2026

159

142

2027

125

107

2028

121

100

2029

118

2030

114

2031

111

2032

108

2033

105

2034

102

2035

2036

2037

2038

1,770

1,370

EAV

148

150

Note: Discounted to 2023. Totals may not sum due to independent rounding. Numbers rounded to two significant
digits unless otherwise noted. The EAV is an annualized cost for it is an estimate calculated from annual costs
across the 15-year analytical timeframe.

Table 3-25 contains a summary of the HAP and VOC emission reductions per year for
each final rule within this regulatory action, and cumulative (total) estimates. These emission
reductions are calculated based on a baseline that include the excess emissions from flares as
explained in Chapter 1 of this RIA.

We note that these emissions reductions do not consider the potential excess emissions
reductions from flares that could result from the final monitoring requirements; we estimate flare
excess emissions reductions of 4,858 tpy HAP and 19,889 tpy VOC.

-------
Table 3-25 Summary of the HAP and VOC Emission Reductions per Year by Rule

Rule

HAP Emission Reductions (tons

VOC Emission Reductions (tons



per year)

per year)

HON

1,107

1,919

P&RI

264

278

P&RII

1

1

NSPS Wb

N/A*

340

NSPS Ilia, NNNa, & RRRa

N/A

1,281

Flare excess reductions

4,858

19,889

Total

6,230

23,708

*N/A - not available. No HAP reductions are estimated for the final NSPS included in this rulemaking.

3.7 Secondary Air Emission Impacts

Table 3-26 contains a summary of other pollutants emissions changes (increases and
decreases), both for criteria pollutants and GHGs, cumulatively for this final action. Calculations
for the emissions changes in this table can be found in the technical memo titled Secondary
Impacts for Flares, Thermal Oxidizers, and Wastewater Controls for the SOCMI Source
Category and Source Categories Subject to the Group I Polymers and Resins NESHAP located
in the docket for this rule. As a result of these changes to the flare combustion efficiency
requirements, there will be a change in the GHG emissions that are primarily due to improved
combustion efficiency of steam-assisted flares in these source categories, decreased steam
demand, and increased need for supplemental natural gas.

Additionally, the EPA anticipates that some facilities will need to install and operate
thermal oxidizer controls to better control process vents and streams containing ethylene oxide
and chloroprene. With the installation of thermal oxidizer controls, additional emissions of GHG
and criteria air pollutants (CAP) will occur as a result of controlling HAP and VOC emissions
(i.e., from the combustion of these pollutants plus from the combustion of any auxiliary natural
gas needed in the thermal oxidizer controls).

The increase in CO2 emissions is primarily (63 percent) driven by the requirements for
flares outlined in the preamble of this rulemaking; the additional CO2 emission increases are
from thermal oxidizers.

66

-------
Table 3-26 Summary of Emission Changes (Increases or Reductions) Other

Then HAP and VOC in Tons per Year, Cumulative and
	by Final Rule	

Pollutant

Total

HON

P&RI

IIIa/NNN a/RRRa

CO

845

714

110

21.51

C02

741,102

609,761

115,975

15,366

ch4

-22,951

-20,177

-2,017

-756

n20

6.86

5.27

1.54

0.06

NOx

349

272

73

3.96

PM

17.4

12.7

4.75

0

S02

1.37

0

1.37

0

The CO2 impacts in

Table 3-26 are the emission changes incremental to the baseline that are expected to
occur from combustion flares and thermal oxidizers as well as the additional natural gas added to
those devices. However, even in the absence of combustion-related emissions controls, most of
the carbon in the VOC and CH4 emissions would have eventually oxidized forming CO2 in the
atmosphere. Therefore, the combustion of these carbonaceous substances would be expected to
lead to approximately the same long-run CO2 concentrations as would have occurred without any
controls.19 Therefore, most of the impact of these CO2 contributions to atmospheric
concentrations from the flaring of CH4 and VOC emissions versus future oxidization is not
additional to the impacts that otherwise would have occurred through the atmospheric oxidation
process.

In the case of VOCs, the oxidization time in the atmosphere is relatively short, on the
order of hours to months, so from a climate perspective the difference between emitting the
carbon immediately as CO2 during combustion or as a VOC is expected to be negligible. In the
NSPS OOOOa rulemaking, the EPA solicited comment on the appropriateness of monetizing:
(1) the impact of CO2 emissions associated with combusting methane and VOC emissions from
oil and natural gas sites; and (2) a new potential approach for approximating this value using the
SC-CO2. The illustrative analysis in the NSPS OOOOa RIA provided a method for evaluating

19 The social cost of methane (SC-CH4) used in chapter 4 to monetize the benefits of the CH4 emissions reductions
does not include the impact of the carbon in CH4 emissions after it oxidizes to C02.

67

-------
the estimated emissions outcomes associated with destroying one metric ton of methane by
combusting fossil-based emissions at oil and natural gas sites (flaring) and releasing the CO2
emissions immediately versus releasing them in the future via the methane oxidation process.
The analysis demonstrated that the potential disbenefits of flaring (i.e., an earlier contribution of
CO2 emissions to atmospheric concentrations) are minor compared to the benefits of flaring (i.e.,
avoiding the release of and associated climate impacts from CH4 emissions).

Given the timing of VOCs oxidizing the in the atmosphere and forming CO2 we have
chosen to only monetize the CO2 emissions from the additional natural gas used in the flares and
thermal oxidizers. Table 3-27 provides a summary of the final GHG emission changes that are
monetized in Chapter 4.

Table 3-27 Summary of Monetized Greenhouse Gas Emission Changes in Tons Per Year

Pollutant

Total

HON

P&RI

IIIa/NNN a/RRRa

C02

296,595

258,655

28,615

9,325

ch4

-22,956

-20,181

-2,018

-757

N20

2.07

1.78

0.23

4

68

-------
4 BENEFITS OF EMISSION REDUCTIONS

4.1 Introduction

The emission controls installed to comply with this final action are expected to reduce
emissions of hazardous air pollutants (HAP); volatile organic compounds (VOC) which, in
conjunction with NOx and in the presence of sunlight, form ground-level ozone (O3). This
chapter reports the estimated ozone-related benefits of reducing VOC emissions in terms of the
number and value of avoided ozone-attributable deaths and illnesses. The potential benefits from
reduced ecosystem effects from the reduction in O3 concentrations are not quantified or
monetized here. Time and data limitations for quantifying the effects of this action on biomass
loss and foliar injury and the ensuing loss of ecosystem services prevent an assessment of the
benefits to ecosystems. The EPA provides a qualitative discussion of the benefits of reducing
HAP emissions later in this chapter in Section 4.2. Finally, we include an analysis of the climate
benefits and disbenefits for this final action. We include a presentation of benefits estimates for
each of the final rules in this rulemaking, and also a cumulative estimate with total benefits for
the entire rulemaking.

The lower estimate of the present value (PV) of the cumulative health benefits for the
final rules range from $77 million at a 3 percent discount rate to $54 million at a 7 percent
discount rate with an equivalent annual value (EAV) of $6.5 million to $5.9 million respectively.
The higher estimate of the PV of the benefits for the final rules range from $690 million at a 3
percent discount rate to $480 million at a 7 percent discount rate with an EAV of $58 to $52
million respectively. All estimates are reported in 2021 dollars. Benefits are estimated using two
alternative concentration-response parameters from three epidemiologic studies when
quantifying ozone-related mortality (Di et al., 2017; Turner et al., 2016; and Katsouyanni et al.,
2009)

The monetized climate benefits of reductions of pollutants such as CH4 and disbenefits
resulting from increasing emissions of CO2 and N2O as presented in Chapter 3 are included in
this chapter in Section 4.7. The monetized climate benefits and disbenefits are calculated using
benefit-per-ton estimates of the social cost of greenhouse gases (SC-GHG) estimates as

69

-------
explained later in this RIA chapter and are estimated at negative $1.6 billion PV at a 2 percent
discount rate ($120 million EAV).20

4.2 Health Effects from Exposure to Hazardous Air Pollutants (HAP)

In the subsequent sections, we describe the health effects associated with the main HAP
of concern from SOCMI (found within the HON), P&R I, and P&R II source categories:
ethylene oxide (Section 4.1.1), chloroprene (Section 4.1.2), benzene (Section 4.1.3), 1,3-
butadiene (Section 4.1.4), vinyl chloride (Section 4.1.5), ethylene dichloride (Section 4.1.6),
chlorine (Section 4.1.7), maleic anhydride (Section 4.1.8) and acrolein (Section 4.1.9). This rule
is projected to reduce ethylene oxide emissions from HON processes by approximately 54 tons
per year (tpy) and reduce chloroprene emissions from Neoprene Production processes in P&R I
by approximately 14 tpy. We also estimate that the final amendments to the NESHAP would
reduce other HAP emissions (excluding ethylene oxide and chloroprene) from the HON and
P&R I source categories by approximately 6,230 tpy. The Agency was unable to estimate HAP
emission reductions for the final amendments to the NSPS in this rulemaking.

Quantifying and monetizing the economic value of reducing the risk of cancer and non-
cancer effects is made difficult by the lack of a central estimate of cancer and non-cancer risk
and estimates of the value of an avoided case of cancer (fatal and non-fatal) and morbidity
effects. Due to methodology and data limitations, we did not attempt to monetize the health
benefits of reductions in HAP in this analysis. Instead, we are providing a qualitative discussion
of the health effects associated with HAP emitted from sources subject to control under the final
action.

20 Monetized climate effects are presented under a 2 percent near-term Ramsey discount rate, consistent with EPA's
updated estimates of the SC-GHG. The 2003 version of OMB's Circular A-4 had generally recommended 3
percent and 7 percent as default discount rates for costs and benefits, though as part of the Interagency Working
Group on the Social Cost of Greenhouse Gases, OMB had also long recognized that climate effects should be
discounted only at appropriate consumption-based discount rates. While this RIA was being drafted and
reviewed, OMB finalized an update to Circular A-4, in which it recommended the general application of a 2.0
percent discount rate to costs and benefits (subject to regular updates), as well as the consideration of the shadow
price of capital when costs or benefits are likely to accrue to capital (OMB 2023). Because the SC-GHG
estimates reflect net climate change damages in terms of reduced consumption (or monetary consumption
equivalents), the use of the social rate of return on capital (7 percent under OMB Circular A-4 (2003)) to
discount damages estimated in terms of reduced consumption would inappropriately underestimate the impacts
of climate change for the purposes of estimating the SC-GHG.

70

-------
4.2.1 Ethylene oxide

Ethylene oxide is used as a chemical intermediate in the manufacture of ethylene glycol
(antifreeze), textiles, detergents, polyurethane foam, solvents, medicine, adhesives, and other
products. Health effects from acute (short-term) exposure to ethylene oxide in humans consist
mainly of central nervous system (CNS) depression and irritation of the eyes and mucous
membranes. Chronic (long-term) exposure to ethylene oxide in humans can cause irritation of the
eyes, skin, nose, throat, and lungs, and damage to the brain and nervous system. There is also
some evidence linking ethylene oxide exposure to reproductive effects (ATSDR, 2022). The
EPA's Integrated Risk Information System (IRIS) characterized ethylene oxide as "carcinogenic
to humans" by the inhalation route of exposure based on the total weight of evidence (U.S. EPA,
2016), in accordance with EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005).
EPA concluded that there was strong, but less than conclusive on its own, epidemiological
evidence of lymphohematopoietic cancers and breast cancer in ethylene oxide-exposed workers
(U.S. EPA, 2016).

4.2.2 Chloroprene

Chloroprene is used primarily in the manufacture of poly chloroprene (e.g., Neoprene),
which is used to make diverse products requiring chemical, oil, and/or weather resistance (e.g.,
adhesives, automotive and industrial parts (e.g., belts and hoses), caulks, flame-resistant
cushioning). Health effects from acute (short-term) inhalation exposure to high concentrations of
chloroprene include headache, irritability, dizziness, insomnia, fatigue, respiratory irritation,
cardiac palpitations, chest pains, nausea, dermatitis, and corneal necrosis. Health effects of
chronic (long-term) exposure may include fatigue, chest pains, irritability, dermatitis, and hair
loss. Other effects reported include changes to the nervous system, changes to the cardiovascular
system, and depression of the immune system. There is evidence of an association between
occupational exposure to chloroprene and liver cancer. There is also suggestive evidence of an
association between occupational exposure to chloroprene and lung cancer (U.S. EPA, 2010).
Studies in animals have observed tumors in multiple organs/organ systems (including
reproductive, hepatic, respiratory, gastrointestinal, dermal, and ocular). The EPA's Integrated
Risk Information System (IRIS) database lists chloroprene as "likely to be carcinogenic to

-------
humans" by all routes of exposure (U.S. EPA, 2010), in accordance with EPA's Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 2005.

4.2.3 Benzene

Benzene is used as a constituent in motor fuels and is found in gasoline service station
and motor vehicle exhaust emissions into air. Acute effects of benzene inhalation exposure in
humans include neurological symptoms such as drowsiness, dizziness, headaches, and
unconsciousness. Exposure to benzene vapor can cause eye, skin, and upper respiratory tract
irritation. Chronic exposure to benzene is associated with blood disorders, such as preleukemia
and aplastic anemia (ATSDR, 2007a). The EPA's Integrated Risk Information System (IRIS)
database lists benzene as a known human carcinogen by all routes of exposure. IRIS found a
causal relationship between benzene exposure and acute lymphocytic leukemia and a suggestive
relationship between benzene exposure and chronic non-lymphocytic leukemia and chronic
lymphocytic leukemia (U.S. EPA, 2000). The International Agency for Research on Cancer
(IARC) has also determined that benzene is a human carcinogen (IARC, 2018).

4.2.4 1,3-Butadiene

1,3-Butadiene is used in the production of styrene-butadiene rubber, plastics, and
thermoplastic resins. A variety of reproductive and developmental effects have been observed in
mice exposed to 1,3-butadiene by inhalation (ATSDR, 2012). There are no human data on
reproductive or developmental effects (ATSDR, 2012). Epidemiological studies of workers in
rubber plants have shown an association between 1,3-butadiene exposure and increased
incidence of leukemia (U.S. EPA, 2002). Animal studies have reported tumors at various sites
from 1,3-butadiene exposure. EPA has classified 1,3-butadiene as carcinogenic to humans by the
inhalation route of exposure. (U.S. EPA, 2002).

4.2.5 Ethylene dichloride (1,2-dichloroethane)

Ethylene dichloride is primarily used in the production of vinyl chloride as well as other
chemicals. Inhalation of concentrated ethylene dichloride vapor can induce effects on the human
nervous system, liver, and kidneys, as well as respiratory distress and cardiac arrhythmia. No
information is available on the chronic effects of ethylene dichloride in humans. In animal
studies, chronic (long-term) inhalation exposure to ethylene dichloride has been observed to

-------
cause effects on the liver and kidneys. Decreased fertility and increased embryo mortality have
been observed in inhalation studies of rats (ATSDR, 1992). Epidemiological studies are not
conclusive regarding the carcinogenic effects of ethylene dichloride, due to concomitant
exposure to other chemicals. Following treatment by gavage (experimentally placing the
chemical in the stomach), several tumor types have been induced in rats and mice. An increased
incidence of lung papilloma has been reported in mice after topical application. EPA has
classified ethylene dichloride as a Group B2, probable human carcinogen (U.S. EPA, 1999).

4.2.6 Vinyl chloride

Most of the vinyl chloride produced is used to make polyvinyl chloride (PVC) plastic and
vinyl products. Acute (short-term) exposure to high levels of vinyl chloride in air has resulted in
CNS effects, such as dizziness, drowsiness, and headaches in humans. Chronic (long-term)
exposure to vinyl chloride through inhalation and oral exposure in humans has resulted in CNS
effects and liver damage. Animal studies have reported effects on the liver, kidney, and CNS
from chronic exposure to vinyl chloride (ATSDR, 2006). Vinyl chloride exposure via inhalation
has been shown to increase the risk of a rare form of liver cancer, angiosarcoma of the liver, in
humans. EPA has concluded that vinyl chloride is carcinogenic to humans by the inhalation and
oral routes of exposure, and highly likely to be carcinogenic by the dermal route of exposure
(U.S. EPA, 2000).

4.2.7 Chlorine

Chlorine (Cb) is a gas that is a potent irritant to the eyes and respiratory tract. Exposure
to low levels of chlorine can result in nose, throat, and eye irritation. At higher levels, breathing
chlorine gas may result in changes in breathing rate and coughing, and damage to the lungs.
Studies in volunteers exposed to controlled concentrations of chlorine indicate that exposures to
1-3 ppm produce mild irritation of the nose that can be tolerated for about 1 hour; 5 ppm may
produce eye irritation; headache and throat irritation may occur at concentrations of 5-15 ppm;
30 ppm produces immediate chest pain, nausea and vomiting, dyspnea, and cough; and 40-60
ppm produces toxic pneumonitis and pulmonary edema. Concentrations in typical human
exposure environments are much lower than these levels unless an accident involving chlorine
takes place nearby (e.g., a leak from a chlorine tank or a leak from a facility that produces or uses

-------
chlorine). Chronic (long-term) exposure to chlorine gas in workers has resulted in respiratory
effects, including eye and throat irritation and airflow obstruction (ATSDR, 2010). EPA has not
assessed chlorine for carcinogenicity under the IRIS program (U.S. EPA, 1994).

4.2.8 Maleic anhydride

Maleic anhydride is used in the formulation of resins. Exposure to maleic anhydride may
occur from accidental releases to the environment or in workplaces where it is produced or
used. Acute (short-term) inhalation exposure of humans to maleic anhydride has been observed
to cause irritation of the respiratory tract and eye irritation. Chronic (long-term) exposure to
maleic anhydride has been observed to cause chronic bronchitis, asthma-like attacks, and upper
respiratory tract and eye irritation in workers. In some people, allergies have developed so that
lower concentrations can no longer be tolerated. Kidney effects were observed in rats chronically
exposed to maleic anhydride via gavage (CalEPA, 2001). EPA has not assessed maleic
anhydride for carcinogenicity under the IRIS program (U.S. EPA, 1988).

4.2.9 Acrolein

Acrolein is primarily used as an intermediate in the synthesis of acrylic acid and as a
biocide. It is toxic to humans following inhalation, oral or dermal exposures. Acute (short-term)
inhalation exposure may result in upper respiratory tract irritation and congestion. The major
effects from chronic (long-term) inhalation exposure to acrolein in humans and animals consist
of general respiratory congestion and eye, nose, and throat irritation (ATSDR, 2007b). The EPA
IRIS program noted, in 2003, that the potential carcinogenicity of acrolein cannot be determined
because the existing data are inadequate for an assessment of human carcinogenic potential for
either the oral or inhalation route of exposure (U.S. EPA, 2003).

4.2.10 Other Hazardous Air Pollutants (HAP)

In addition to the compounds described above, other toxic compounds might be affected
by this action. Information regarding the health effects of those compounds can be found in
Health Effects Notebook for Hazardous Air Pollutants (at https://www.epa.gov/haps/health-
effects-notebook-hazardous-air-pollutants) and in the EPA Integrated Risk Information System
(IRIS) database (at https://iris.epa.gov/AtoZ/71ist type=alpha).

-------
4.3 Ozone-related Human Health Benefits

This section summarizes the EPA's approach to estimating the incidence and economic
value of the ozone-related benefits estimated for this action. The Regulatory Impact Analysis
(RIA) Final Revised Cross-State Air Pollution Rule (U.S. EPA, 2021) and its corresponding
Technical Support Document Estimating PM2.5- and Ozone-Attributable Health Benefits (TSD)
(U.S. EPA, 2021) provide a full discussion of the EPA's approach for quantifying the incidence
and value of estimated air pollution-related health impacts. In these documents, the reader can
find the rationale for selecting the health endpoints quantified; the demographic, health and
economic data applied in the environmental Benefits Mapping and Analysis Program—
Community Edition (BenMAP-CE); modeling assumptions; and the EPA's techniques for
quantifying uncertainty.

Implementing this action will affect the distribution of ozone concentrations throughout
the U.S.; this includes locations both meeting and exceeding the NAAQS for O3. This RIA
estimates avoided 03-related health impacts that are distinct from those reported in the RIAs for
the O3 NAAQS (U.S. EPA, 2015). The O3 NAAQS RIAs hypothesize, but do not predict, the
benefits and costs of strategies that states may choose to enact when implementing a revised
NAAQS; these costs and benefits are illustrative and cannot be added to the costs and benefits of
policies that prescribe specific emission control measures.

4.3.1 Estimating Ozone Related Health Impacts

We estimate the quantity and economic value of air pollution-related effects by
estimating counts of air pollution-attributable cases of adverse health outcomes, assigning dollar
values to these counts, and assuming that each outcome is independent of one another. We
construct these estimates by adapting primary research—specifically, air pollution epidemiology
studies and economic value studies—from similar contexts. This approach is sometimes referred
to as "benefits transfer." Below we describe the procedure we follow for: (1) selecting air
pollution health endpoints to quantify; (2) calculating counts of air pollution effects using a
health impact function; and (3) specifying the health impact function with concentration-
response parameters drawn from the epidemiological literature.

-------
4.3.2 Selecting Air Pollution Health Endpoints to Quantify

As a first step in quantifying Cb-related human health impacts, the EPA consults the
Integrated Science Assessment for Ozone (Ozone ISA) (U.S. EPA, 2020) as summarized in the
TSD for the Final Revised Cross State Air Pollution Rule Update (U.S. EPA, 2021). This
document synthesizes the toxicological, clinical, and epidemiological evidence to determine
whether each pollutant is causally related to an array of adverse human health outcomes
associated with either acute {i.e., hours or days-long) or chronic {i.e., years-long) exposure. For
each outcome, the ISA reports this relationship to be causal, likely to be causal, suggestive of a
causal relationship, inadequate to infer a causal relationship, or not likely to be a causal
relationship.

In brief, the ISA for ozone found short-term (less than one month) exposures to ozone to
be causally related to respiratory effects, a "likely to be causal" relationship with metabolic
effects and a "suggestive of, but not sufficient to infer, a causal relationship" for central nervous
system effects, cardiovascular effects, and total mortality. The ISA reported that long-term
exposures (one month or longer) to ozone are "likely to be causal" for respiratory effects
including respiratory mortality, and a "suggestive of, but not sufficient to infer, a causal
relationship" for cardiovascular effects, reproductive effects, central nervous system effects,
metabolic effects, and total mortality.

The EPA estimates the incidence of air pollution effects for those health endpoints listed
above where the ISA classified the impact as either causal or likely-to-be-causal. Table 4-1
reports the effects we quantified and those we did not quantify in this RIA. The list of benefit
categories not quantified shown in that table is not exhaustive. And, among the effects we
quantified, we might not have been able to completely quantify either all human health impacts
or economic values. The table below omits any welfare effects such as biomass loss and foliar
injury. These effects are described in Chapter 7 of the Ozone NAAQS RIA (2015).

-------
Table 4-1 Human Health Effects of Ambient Ozone and Whether they were
	Quantified and/or Monetized in this RIA	

Category

Effect

Effect
Quantified

Effect
Monetized

More
Information

Mortality from

Premature respiratory mortality from
short-term exposure (0-99)

•/

~

Ozone ISA1

exposure to ozone

Premature respiratory mortality from
long-term exposure (age 30-99)

•/

~

Ozone ISA



Hospital admissions—respiratory (ages
65-99)

•/

~

Ozone ISA



Emergency department visits—
respiratory (ages 0-99)

~

V

Ozone ISA



Asthma onset (0-17)

~

V

Ozone ISA



Asthma symptoms/exacerbation
(asthmatics age 5-17)

~

V

Ozone ISA

Nonfatal morbidity
from exposure to

Allergic rhinitis (hay fever) symptoms
(ages 3-17)

V

~

Ozone ISA

Minor restricted-activity days (age 18-
65)

V

~

Ozone ISA

ozone

School absence days (age 5-17)

V

~

Ozone ISA



Decreased outdoor worker productivity
(age 18-65)

—

—

Ozone ISA2



Metabolic effects (e.g., diabetes)

—

—

Ozone ISA2



Other respiratory effects (e.g.,
premature aging of lungs)

—

—

Ozone ISA2



Cardiovascular and nervous system
effects

—

—

Ozone ISA2



Reproductive and developmental effects

—

—

Ozone ISA2

1	We assess these benefits qualitatively due to data and resource limitations for this analysis. In other analyses we

quantified these effects as a sensitivity analysis.

2	We assess these benefits qualitatively because we do not have sufficient confidence in available data or methods.

4.3.3 Quantifying Cases of Ozone-Attributable Premature Mortality

Mortality risk reductions account for the majority of monetized ozone-related benefits.
For this reason, this subsection and the following provide a brief background of the scientific
assessments that underly the quantification of these mortality risks and identifies the risk studies
used to quantify them in this RIA for ozone. As noted above, the Estimating PM2.5- and Ozone-
Attributable Health Benefits TSD describes fully the Agency's approach for quantifying the
number and value of ozone air pollution-related impacts, including additional discussion of how
the Agency selected the risk studies used to quantify them in this RIA. The TSD also includes
additional discussion of the assessments that support quantification of these mortality risk than
provide here.

77

-------
In 2008, the National Academies of Science (NAS) (NRC, 2008) issued a series of
recommendations to EPA regarding the procedure for quantifying and valuing ozone-related
mortality due to short-term exposures. Chief among these was that "short-term exposure to
ambient ozone is likely to contribute to premature deaths," and the committee recommended that
"ozone-related mortality be included in future estimates of the health benefits of reducing ozone
exposures." The NAS also recommended that "the greatest emphasis be placed on the multicity
and [National Mortality and Morbidity Air Pollution Studies (NMMAPS)] ... studies without
exclusion of the meta-analyses" (NRC, 2008). Prior to the 2015 Ozone NAAQS RIA, the
Agency estimated ozone-attributable premature deaths using an NMMAPS-based analysis of
total mortality (Bell et al., 2004), two multi-city studies of cardiopulmonary and total mortality
(Huang et al., 2004; Schwartz, 2005) and effect estimates from three meta-analyses of non-
accidental mortality (Bell et al., 2005; Ito et al., 2005; Levy et al., 2005). Beginning with the
2015 Ozone NAAQS RIA, the Agency began quantifying ozone-attributable premature deaths
using two newer multi-city studies of non-accidental mortality (Smith et al., 2009; Zanobetti and
Schwartz, 2008) and one long-term cohort study of respiratory mortality (Jerrett et al., 2009).
The 2020 Ozone ISA included changes to the causality relationship determinations between
short-term exposures and total mortality, as well as including more recent epidemiologic
analyses of long-term exposure effects on respiratory mortality (U.S. EPA, 2020). In this RIA, as
described in the corresponding TSD, two estimates of ozone-attributable respiratory deaths from
short-term exposures are estimated using the risk estimate parameters from Zanobetti et al.
(2008) and Katsouyanni et al. (2009). Ozone-attributable respiratory deaths from long-term
exposures are estimated using Turner et al. (2016). Due to time and resource limitations, we were
unable to reflect the warm season defined by Zanobetti et al. (2008) as June-August. Instead, we
apply this risk estimate to our standard warm season of May-September.

4.4 Economic Valuation

After quantifying the change in adverse health impacts, we estimate the economic value of
these avoided impacts. Reductions in ambient concentrations of air pollution generally lower the
risk of future adverse health effects for a large population. Therefore, the appropriate economic
measure is willingness to pay (WTP) for changes in risk of a health effect. For some health
effects, such as hospital admissions, WTP estimates are generally not available, so we use the

-------
cost of treating or mitigating the effect. These cost-of-illness (COI) estimates generally (although
not necessarily in every case) 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. The unit values applied in this analysis are provided in
Section 5.1 of the TSD for the MATS rule (U.S. EPA, 2023b).

Avoided premature deaths account for 95 percent of monetized ozone-related benefits.
The economics literature concerning the appropriate method for valuing reductions in premature
mortality risk is still developing. The value for the projected reduction in the risk of premature
mortality is the subject of continuing discussion within the economics and public policy analysis
community. Following the advice of the Scientific Advisory Board's (SAB) Environmental
Economics Advisory Committee (SAB-EEAC), the EPA currently uses the value of statistical
life (VSL) approach in calculating estimates of mortality benefits, because we believe this
calculation provides the most reasonable single estimate of an individual's WTP for reductions
in mortality risk (U.S. EPA-SAB, 2000). The VSL approach is a summary measure for the value
of small changes in mortality risk experienced by a large number of people.

The EPA continues work to update its guidance on valuing mortality risk reductions and
consulted several times with the SAB-EEAC on the issue. Until updated guidance is available,
the EPA determined that a single, peer-reviewed estimate applied consistently best reflects the
SAB-EEAC advice it has received. Therefore, the EPA applies the VSL that was vetted and
endorsed by the SAB in the Guidelines for Preparing Economic Analyses while the EPA
continues its efforts to update its guidance on this issue (U.S. EPA, 2016). This approach
calculates a mean value across VSL estimates derived from 26 labor market and contingent
valuation studies published between 1974 and 1991. The mean VSL across these studies is $10.7
million ($2016).21

The EPA is committed to using scientifically sound, appropriately reviewed evidence in
valuing changes in the risk of premature death and continues to engage with the SAB to identify
scientifically sound approaches to update its mortality risk valuation estimates. Most recently,
the Agency proposed new meta-analytic approaches for updating its estimates which were

21 In 1990$, this base VSL is $4.8 million. In 2016$, this base VSL is $10.7 million.

-------
subsequently reviewed by the SAB-EEAC. The EPA is taking the SAB's formal
recommendations under advisement (U.S. EPA, 2017b).

Because short-term ozone-related premature mortality occurs within the analysis year, the
estimated ozone-related benefits are identical for all discount rates. When valuing changes in
ozone-attributable deaths using the Turner et al. (2016) study, we follow advice provided by the
Health Effects Subcommittee (HES) of the SAB, which found that "there is no evidence in the
literature to support a different cessation lag between ozone and particulate matter. The HES
therefore recommends using the same cessation lag structure and assumptions as for particulate
matter when utilizing cohort mortality evidence for ozone" (U.S. EPA-SAB, 2010).

These estimated health benefits do not account for the influence of future changes in the
climate on ambient concentrations of pollutants (USGCRP, 2016). For example, recent research
suggests that future changes to climate may create conditions more conducive to forming ozone.
The estimated health benefits also do not consider the potential for climate-induced changes in
temperature to modify the relationship between ozone and the risk of premature mortality (Jhun
et al., 2014; Ren et al., 2008a, 2008b).

4.4.1 Benefit-per- Ton Estimates

The EPA did not conduct air quality modeling for this final rulemaking. Rather, we
quantified the value of changes in ozone, and NOx concentrations using a "benefit-per-ton"
(BPT) approach, due to the relatively small number of facilities and the fact that these facilities
are located in a discrete location. Specifically, the EPA believes that the emissions reductions
due to this rule are small and because we cannot be confident of the location of new facilities
under the NSPS included in this final action, we elected to use the benefit-per-ton approach. The
EPA did not expect full air quality modeling to show a significant difference between the policy
and baseline model runs. BPT estimates provide the total monetized human health benefits (the
sum of premature mortality and premature morbidity) of reducing one ton of the VOC precursor
for ozone from a specified source. Specifically, in this analysis, we multiplied the BPT estimates
from the Synthetic Organic Chemicals sector by the corresponding emission reductions. The
method used to derive these estimates is described in the BPT Technical Support Document
(BPT TSD) on Estimating the Benefit per Ton of Reducing Directly-Emitted PM2.5, PM2.5

-------
Precursors and Ozone Precursors from 21 Sectors (U.S. EPA, 2023b). As noted above, we were
unable to quantify the value of changes in exposure to HAP, CO, and NO2.

As noted below in the characterization of uncertainty, all BPT estimates have inherent
limitations. Specifically, all national-average BPT estimates reflect the geographic distribution of
the modeled emissions, which may not exactly match the emission reductions that would occur
due to the action, and they may not reflect local variability in population density, meteorology,
exposure, baseline health incidence rates, or other local factors for any specific location. In an
effort to address this limitation we have generated regional level BPTs for the synthetic organic
chemicals sector. Given sector specific air quality modeling and the small changes in emissions
considered in this action, the difference in the quantified health benefits that result from the BPT
approach compared with if EPA had used a full-form air quality model should be minimal.

The EPA systematically compared the changes in benefits, and concentrations where
available, from its BPT technique and other reduced-form techniques to the changes in benefits
and concentrations derived from full-form photochemical model representation of a few different
specific emissions scenarios. Reduced-form tools are less complex than the full air quality
modeling, requiring less agency resources and time. That work, in which we also explore other
reduced form models, is referred to as the "Reduced Form Tool Evaluation Project" (Project). It
began in 2017, and the initial results were available at the end of 2018. The Agency's goal was to
create a methodology by which investigators could better understand the suitability of alternative
reduced-form air quality modeling techniques for estimating the health impacts of criteria
pollutant emissions changes in the EPA's benefit-cost analysis, including the extent to which
reduced-form models may over- or under-estimate benefits (compared to full-scale modeling)
under different scenarios and air quality concentrations. The EPA Science Advisory Board
(SAB) convened a panel to review this report.22 In particular, the SAB assessed the techniques
the Agency used to appraise these tools; the Agency's approach for depicting the results of
reduced-form tools; and steps the Agency might take for improving the reliability of reduced-
form techniques for use in future Regulatory Impact Analyses (RIAs).

22 85 FR 23823. April 29, 2020.

-------
The scenario-specific emission inputs developed for this project are currently available
online. The study design, methodology, and results of the project are described in the final
report, Evaluating Reduced-Form Tools for Estimating Air Quality Benefits. Final Report (IEc,
2019). Results of this project found that total PM2.5 BPT values were within approximately 10
percent of the health benefits calculated from full-form air quality modeling when analyzing the
pulp and paper sector, a sector used as an example for evaluating the application of the new
methodology in the final report. The ratios for individual PM species varied, and the report found
that the ratio for the directly emitted PM2.5 for the pulp and paper sector was 0.7 for the BPT
approach compared to 1.0 for full-form air quality modeling combined with BenMAP. This
provides some initial understanding of the uncertainty which is associated with using the BPT
approach instead of full-form air quality modeling.

4.4.2 Ozone Vegetation Effects

Exposure to ozone has been found to be associated with a wide array of vegetation and
ecosystem effects in the published literature (U.S. EPA, 2020). Sensitivity to ozone is highly
variable across species, with over 66 vegetation species identified as "ozone-sensitive," many of
which occur in state and national parks and forests. These effects include those that cause
damage to, or impairment of, the intended use of the plant or ecosystem. Such effects are
considered adverse to public welfare and can include reduced growth and/or biomass production
in sensitive trees, reduced yield and quality of crops, visible foliar injury, changes to species
composition, and changes in ecosystems and associated ecosystem services.

4.4.3 Ozone Climate Effects

Ozone is a well-known short-lived climate forcing GHG (U.S. EPA, 2013). Stratospheric
ozone (the upper ozone layer) is beneficial because it protects life on Earth from the sun's
harmful ultraviolet (UV) radiation. In contrast, tropospheric ozone (ozone in the lower
atmosphere) is a harmful air pollutant that adversely affects human health and the environment
and contributes significantly to regional and global climate change. Due to its short atmospheric
lifetime, tropospheric ozone concentrations exhibit large spatial and temporal variability (U.S.
EPA, 2009b). The Intergovernmental Panel on Climate Change's Fifth Assessment Report,

(IPCC AR5,) estimated that the contribution to current warming levels of increased tropospheric

-------
ozone concentrations resulting from human methane, NOx, and VOC emissions was 0.5 watts
per meter squared (W/m2), or about 30 percent as large a warming influence as elevated CO2
concentrations. This quantifiable influence of ground level ozone on climate leads to increases in
global surface temperature and changes in hydrological cycles.

4.5 Ozone- and NOx- Related Benefits Results

Tables 4-2 list the estimated Ozone- and NOx—related benefits per ton applied in this
national level analysis. Benefits are estimated using two alternative concentration-response
parameters from three epidemiologic studies when quantifying ozone-related mortality (Di et al.,
2017; Turner et al., 2016; and Katsouyanni et al., 2009). These results are discounted at 3 and 7
percent for a 2021 currency year. For all estimates, we summarize the monetized health benefits
using discount rates of 3 percent and 7 percent for the 15-year analysis period of this rule,
discounted back to 2023 and rounded to 2 significant figures, as presented in Table 4-4. The
lower estimate of the Present Value (PV) of the cumulative health benefits for the final rules
range from $77 million at a 3 percent discount rate to $54 million at a 7 percent discount rate
with an equivalent annualized value (EAV) of $6.5 million to $5.9 million respectively. The
higher estimate of the PV of the benefits for the final rules range from $690 million at a 3
percent discount rate to $480 million at a 7 percent discount rate with an EAV of $58 to $52
million respectively. All estimates are reported in 2021 dollars. Discounted benefits are
presented by year for the final and less stringent alternative options in Table 4-4 and Table 4-5.
For the full set of underlying calculations see the "Final HONSOCMI Benefits workbook,"
available in the docket for the final action.

Table 4-2 Synthetic Organic Chemicals: Benefit per Ton Estimates of Avoided Ozone-
Attributable Premature Mortality and Illness for the Rule, 2024-2038 (2021$)

Discount Rate

Year

3 Percent

7 Percent

2025

725

and

6,235

652

and

5,579

2030

773

and

6,859

698

and

6,150

2035

815

and

7,537

738

and

6,764

2040

851

and

8,109

773

and

7,272

Note: The standard reporting convention for EPA benefits is to round all results to two significant figures. Here, we
report all significant figures so that readers may reproduce the results reported below.

-------
Table 4-3 Synthetic Organic Chemicals: Benefit per Ton Estimates of Avoided NOx-
	Attributable Premature Mortality and Illness for the Rule, 2024-2038 (2021$)	

Discount Rate

Year



3 Percent



7 Percent

2025

17,572

and

37,685

15,878

and

33,874

2030

19,160

and

39,696

17,254

and

35,673

2035

21,489

and

43,507

19,372

and

39,061

2040

23,500

and

46,682

21,171

and

42,025

Note: The standard reporting convention for EPA benefits is to round all results to two significant figures. Here, we
report all significant figures so that readers may reproduce the results reported below.

84

-------
Table 4-4 Total Benefits Estimates of Ozone- and NOx-Attributable Avoided Premature Mortality and Illness (million

	2021$)a'b'c'	

All Rules





Less Stringent Regulatory Option



Final Regulatory Option

More Stringent Regulatory Option





Discount Rate



Discount Rate

Discount Rate





3 Percent

7 Percent





3 Percent

7 Percent



3 Percent



7 Percent

PV

76

and 680

51 and

460

77

and 690

54 and 480

77

and 690

53

and 470

EAV

6.0

and 53

5.3 and

47

6.5

and 58

5.9 and 52

6.1

and 55

5.5

and 48

Non-Monetized Benefits

Health benefits associated with emission reductions of 6,230 tpy of HAP including hexane, benzene, methanol, 1,3-butadiene, and vinyl acetate.
Health benefits associated with reduction of 54 tpy of ethylene oxide and 14 tpy of chloroprene.

Ecosystem benefits related to the reductions of ozone and nitrogen and sulfur deposition.

a Discounted to 2023.

b Rounded to 2 significant figures. The small differences between the alternatives are masked by the rounding.

0 Benefits are estimated for Ozone and NOx.

3-85

-------
Table 4-5 Undiscounted Total Benefits Estimates of Ozone- and NOx-Attributable

	Premature Mortality and Illness for the Final Option (million 2021$), 2024-2038a'b'c

3 Percent	7 Percent

	Year

2024

2025

2026

2027

2028

2029

2030

2031

2032

2033

2034

2035

2036

2037

2038

a Rounded to 2 significant figures
b Benefits are estimated for Ozone and NOx

c This table presents benefits that are undiscounted for time value of money over the 15-year period. The 3 and 7 percent rates
refer to the cessation lag of impacts for the value of long-term exposure related ozone deaths.

4.6 Characterization of Uncertainty in the Monetized Benefits

In any complex analysis using estimated parameters and inputs from a variety of models,
there are likely to be many sources of uncertainty. This analysis is no exception. This analysis
includes many data sources as inputs, including emission inventories, air quality data from
models (with their associated parameters and inputs), population data, population estimates,
health effect estimates from epidemiology studies, economic data for monetizing benefits, and
assumptions regarding the future state of the world {i.e., regulations, technology, and human
behavior). Each of these inputs are uncertain and generate uncertainty in the benefits estimate.
When the uncertainties from each stage of the analysis are compounded, even small uncertainties
can have large effects on the total quantified benefits. There also remain unmonetized health
benefits resulting from reductions in HAPs and VOCs, and unmonetized disbenefits resulting
from a secondary increase in CO emissions. Additional uncertainties and limitations are outlined
in Chapter 6.2. Therefore, the estimates of annual benefits should be viewed as representative of
the magnitude of benefits expected, rather than the actual benefits that would occur every year.

8.5

73

7.6

65

8.5

73

7.6

65

8.5

73

7.6

65

8.5

73

7.6

65

9.0

80

8.1

72

9.0

80

8.1

72

9.0

80

8.1

72

9.0

80

8.1

72

9.0

80

8.1

72

9.0

83

8.2

75

9.0

83

8.2

75

9.0

83

8.2

75

9.0

83

8.2

75

9.0

83

8.2

75

10

95

9.0

85

86

-------
4.7 Climate Impacts

The EPA estimates the climate impacts of GHG emissions changes expected from the
final rule using estimates of the social cost of greenhouse gases (SC-GHG) that reflect recent
advances in the scientific literature on climate change and its economic impacts and incorporate
recommendations made by the National Academies of Science, Engineering, and Medicine
(National Academies, 2017). The EPA published and used these estimates in the RIA for the
December 2023 Final Oil and Gas NSPS/EG Rulemaking, "Standards of Performance for New,
Reconstructed, and Modified Sources and Emissions Guidelines for Existing Sources: Oil and
Natural Gas Sector Climate Review." The EPA solicited public comment on the methodology
and use of these estimates in the RIA for the agency's December 2022 Oil and Gas NSPS/EG
Supplemental Proposal and has conducted an external peer review of these estimates, as
described further below.

The SC-GHG is the monetary value of the net harm to society associated with a marginal
increase in GHG emissions in a given year, or the benefit of avoiding that increase. In principle,
SC-GHG includes the value of all climate change impacts (both negative and positive), including
(but not limited to) changes in net agricultural productivity, human health effects, property
damage from increased flood risk and natural disasters, disruption of energy systems, risk of
conflict, environmental migration, and the value of ecosystem services. The SC-GHG, therefore,
reflects the societal value of reducing emissions of the gas in question by one metric ton and is
the theoretically appropriate value to use in conducting benefit-cost analyses of policies that
affect GHG emissions. In practice, data and modeling limitations restrain the ability of SC-GHG
estimates to include all physical, ecological, and economic impacts of climate change, implicitly
assigning a value of zero to the omitted climate damages. The estimates are, therefore, a partial
accounting of climate change impacts and likely underestimate the marginal impacts of
abatement.

Since 2008, the EPA has used estimates of the social cost of various greenhouse gases
(i.e., social cost of carbon (SC-CO2), social cost of methane (SC-CH4), and social cost of nitrous
oxide (SC-N2O), collectively referred to as the "social cost of greenhouse gases" (SC-GHG), in
analyses of actions that affect GHG emissions. The values used by the EPA from 2009 to 2016,
and since 2021 - including in the proposal for this rulemaking - have been consistent with those

87

-------
developed and recommended by the Interagency Working Group on the SC-GHG (IWG); and
the values used from 2017 to 2020 were consistent with those required by E.O. 13783, which
disbanded the IWG. During 2015-2017, the National Academies conducted a comprehensive
review of the SC-CO2 and issued a final report in 2017 recommending specific criteria for future
updates to the SC-CO2 estimates, a modeling framework to satisfy the specified criteria, and both
near-term updates and longer-term research needs pertaining to various components of the
estimation process (National Academies, 2017). The IWG was reconstituted in 2021 and E.O.
13990 directed it to develop a comprehensive update of its SC-GHG estimates, recommendations
regarding areas of decision-making to which SC-GHG should be applied, and a standardized
review and updating process to ensure that the recommended estimates continue to be based on
the best available economics and science going forward.

The EPA is a member of the IWG and is participating in the IWG's work under E.O.
13990. As noted in previous EPA RIAs, while that process continues, the EPA is continuously
reviewing developments in the scientific literature on the SC-GHG, including more robust
methodologies for estimating damages from emissions, and looking for opportunities to further
improve SC-GHG estimation.23 As EPA noted in the proposal RIA for this rulemaking, in the
December 2022 Oil and Gas NSPS/EG Supplemental Proposal RIA, the Agency included a
sensitivity analysis of the climate benefits of the Supplemental Proposal using a new set of SC-
GHG estimates that incorporates recent research addressing recommendations of the National
Academies (National Academies, 2017) in addition to using the interim SC-GHG estimates
presented in the Technical Support Document: Social Cost of Carbon, Methane, and Nitrous
Oxide Interim Estimates under Executive Order 13990 (IWG, 2021) that the IWG recommended
for use until updated estimates that address the National Academies' recommendations are
available.

The EPA solicited public comment on the sensitivity analysis and the accompanying draft
technical report, External Review Draft of Report on the Social Cost of Greenhouse Gases:
Estimates Incorporating Recent Scientific Advances, which explains the methodology underlying

23 EPA strives to base its analyses on the best available science and economics, consistent with its responsibilities,
for example, under the Information Quality Act.

88

-------
the new set of estimates, in the December 2022 Supplemental Oil and Gas Proposal. The
response to comments document can be found in the docket for that action.

To ensure that the methodological updates adopted in the technical report are consistent
with economic theory and reflect the latest science, the EPA also initiated an external peer
review panel to conduct a high-quality review of the technical report, completed in May 2023
(EPA, 2023c). The peer reviewers commended the agency on its development of the draft
update, calling it a much-needed improvement in estimating the SC-GHG and a significant step
towards addressing the National Academies' recommendations with defensible modeling choices
based on current science. The peer reviewers provided numerous recommendations for refining
the presentation and for future modeling improvements, especially with respect to climate
change impacts and associated damages that are not currently included in the analysis.

Additional discussion of omitted impacts and other updates have been incorporated in the
technical report to address peer reviewer recommendations. Complete information about the
external peer review, including the peer reviewer selection process, the final report with
individual recommendations from peer reviewers, and the EPA's response to each
recommendation is available on EPA's website (EPA, 2023c).

The remainder of this section provides an overview of the methodological updates
incorporated into the SC-GHG estimates used in this final RIA. A more detailed explanation of
each input and the modeling process is provided in the final technical report, EPA Report on the
Social Cost of Greenhouse Gases: Estimates Incorporating Recent Scientific Advances (EPA,
2023d). The Appendix to Chapter 4 shows the impacts of the final rule using the interim SC-
GHG (IWG 2021) estimates presented in the proposal.

The steps necessary to estimate the SC-GHG with a climate change integrated assessment
model (IAM) can generally be grouped into four modules: socioeconomics and emissions,
climate, damages, and discounting. The emissions trajectories from the socioeconomic module
are used to project future temperatures in the climate module. The damage module then
translates the temperature and other climate endpoints (along with the projections of
socioeconomic variables) into physical impacts and associated monetized economic damages,
where the damages are calculated as the amount of money the individuals experiencing the
climate change impacts would be willing to pay to avoid them. To calculate the marginal effect

89

-------
of emissions, i.e., the SC-GHG in year t, the entire model is run twice - first as a baseline and
second with an additional pulse of emissions in year t. After recalculating the temperature effects
and damages expected in all years beyond t resulting from the adjusted path of emissions, the
losses are discounted to a present value in the discounting module. Many sources of uncertainty
in the estimation process are incorporated using Monte Carlo techniques by taking draws from
probability distributions that reflect the uncertainty in parameters.

The SC-GHG estimates used by the EPA and many other federal agencies since 2009
have relied on an ensemble of three widely used IAMs: Dynamic Integrated Climate and
Economy (DICE) (Nordhaus W. ,2010) Climate Framework for Uncertainty, Negotiation, and
Distribution (FUND) (Anthoff & Tol, 2013); (Anthoff & Tol, 2013b) and Policy Analysis of the
Greenhouse Gas Effect (PAGE) (Hope C. ,2013) In 2010, the IWG harmonized key inputs
across the IAMs, but all other model features were left unchanged, relying on the model
developers' best estimates and judgments. That is, the representation of climate dynamics and
damage functions included in the default version of each IAM as used in the published literature
was retained.

The SC-GHG estimates in this RIA no longer rely on the three IAMs (i.e., DICE, FUND,
and PAGE) used in previous SC-GHG estimates. As explained in EPA (2023d), EPA uses a
modular approach to estimate the SC-GHG, consistent with the National Academies' (National
Academies, 2017) near-term recommendations. That is, the methodology underlying each
component, or module, of the SC-GHG estimation process is developed by drawing on the latest
research and expertise from the scientific disciplines relevant to that component. Under this
approach, each step in the SC-GHG estimation improves consistency with the current state of
scientific knowledge, enhances transparency, and allows for more explicit representation of
uncertainty.

The socioeconomic and emissions module relies on a new set of probabilistic projections
for population, income, and GHG emissions developed under the Resources for the Future (RFF)
Social Cost of Carbon Initiative (Rennert K. P., 2021) (Rennert, et al., 2022a). These
socioeconomic projections (hereafter collectively referred to as the RFF-SPs) are an internally
consistent set of probabilistic projections of population, GDP, and GHG emissions (CO2, CH4,
and N2O) to 2300. Based on a review of available sources of long-run projections necessary for

-------
damage calculations, the RFF-SPs stand out as being most consistent with the National
Academies' recommendations. Consistent with the National Academies' recommendation, the
RFF-SPs were developed using a mix of statistical and expert elicitation techniques to capture
uncertainty in a single probabilistic approach, taking into account the likelihood of future
emissions mitigation policies and technological developments, and provide the level of
disaggregation necessary for damage calculations. Unlike other sources of projections, they
provide inputs for estimation out to 2300 without further extrapolation assumptions. Conditional
on the modeling conducted for the SC-GHG estimates, this time horizon is far enough in the
future to capture the majority of discounted climate damages. Including damages beyond 2300
would increase the estimates of the SC-GHG. As discussed in U.S. EPA (EPA, 2023c) the use of
the RFF-SPs allows for capturing economic growth uncertainty within the discounting module.

The climate module relies on the Finite Amplitude Impulse Response (FaIR) model,
(Smith, et al., 2018; IPCC, Climate Change 2021 - The Physical Science Basis, 2021; Millar,
Nicholls, Friedlingstein, & Allen, 2017) a widely used Earth system model which captures the
relationships between GHG emissions, atmospheric GHG concentrations, and global mean
surface temperature. The FaIR model was originally developed by Richard Millar, Zeb Nicholls,
and Myles Allen at Oxford University, as a modification of the approach used in IPCC AR5 to
assess the GWP and GTP (Global Temperature Potential) of different gases. It is open source,
widely used (e.g., (IPCC, 2018); (IPCC, 2021a) and was highlighted by the National Academies
(National Academies, 2017) as a model that satisfies their recommendations for a near-term
update of the climate module in SC-GHG estimation. Specifically, it translates GHG emissions
into mean surface temperature response and represents the current understanding of the climate
and GHG cycle systems and associated uncertainties within a probabilistic framework. The SC-
GHG estimates used in this RIA rely on FaIR version 1.6.2 as used by the IPCC (IPCC, 2021). It
provides, with high confidence, an accurate representation of the latest scientific consensus on
the relationship between global emissions and global mean surface temperature and offers a code
base that is fully transparent and available online. The uncertainty capabilities in FaIR 1.6.2 have
been calibrated to the most recent assessment of the IPCC (which importantly narrowed the
range of likely climate sensitivities relative to prior assessments). See U.S. EPA (EPA, 2023c)
for more details.

-------
The socioeconomic projections and outputs of the climate module are inputs into the
damage module to estimate monetized future damages from climate change24. The National
Academies' recommendations for the damage module, scientific literature on climate damages,
updates to models that have been developed since 2010, as well as the public comments received
on individual EPA rulemakings and the IWG's February 2021 TSD, have all helped to identify
available sources of improved damage functions. The IWG (e.g., (IWG, 2010) (IWG, 2016a)
(IWG, 2021)), the National Academies (2017), comprehensive studies (e.g., (Rose, et al., 2014)),
and public comments have all recognized that the damages functions underlying the IWG SC-
GHG estimates used since 2013 (taken from DICE 2010 (Nordhaus W. , 2010); FUND 3.8
(Anthoff & Tol, 2013b); (Anthoff & Tol, 2013); and PAGE 2009 (Hope C. , 2013)) do not
include all the important physical, ecological, and economic impacts of climate change. The
climate change literature and the science underlying the economic damage functions have
evolved, and DICE 2010, FUND 3.8, and PAGE 2009 now lag behind the most recent research.

The challenges involved with updating damage functions have been widely recognized.
Functional forms and calibrations are constrained by the available literature and need to
extrapolate beyond warming levels or locations studied in that literature. Research and public
resources focused on understanding how these physical changes translate into economic impacts
have been significantly less than the resources focused on modeling and improving our
understanding of climate system dynamics and the physical impacts from climate change
(Auffhammer, 2018). Even so, there has been a large increase in research on climate impacts and
damages in the time since DICE 2010, FUND 3.8, and PAGE 2009 were published. Along with
this growth, there continues to be wide variation in methodologies and scope of studies, such that
care is required when synthesizing the current understanding of impacts or damages. Based on a
review of available studies and approaches to damage function estimation, the EPA uses three
separate damage functions to form the damage module. They are: a subnational-scale, sectoral

24 In addition to temperature change, two of the three damage modules used in the SC-GHG estimation require
global mean sea level (GMSL) projections as an input to estimate coastal damages. Those two damage modules
use different models for generating estimates of GMSL. Both are based off reduced complexity models that can
use the FaIR temperature outputs as inputs to the model and generate projections of GMSL accounting for the
contributions of thermal expansion and glacial and ice sheet melting based on recent scientific research. Absent
clear evidence on a preferred model, the SC-GHG estimates presented in this RIA retain both methods used by
the damage module developers. See (EPA, 2023c) for more details.

92

-------
damage function (based on the Data-driven Spatial Climate Impact Model (DSCIM) developed
by the Climate Impact Lab (CIL, 2023) (Carleton, 2022) (Rode, et al., 2021), a country-scale,
sectoral damage function (based on the Greenhouse Gas Impact Value Estimator (GIVE) model
developed under RFF's Social Cost of Carbon Initiative (Rennert, et al., 2022) and a meta-
analysis-based damage function (based on (Howard & Sterner, 2017)). The damage functions in
DSCIM and GIVE represent substantial improvements relative to the damage functions
underlying the SC-GHG estimates used by the EPA to date and reflect the forefront of scientific
understanding about how temperature change and SLR lead to monetized net (market and
nonmarket) damages for several categories of climate impacts. The models' spatially explicit and
impact-specific modeling of relevant processes allow for improved understanding and
transparency about mechanisms through which climate impacts are occurring and how each
damage component contributes to the overall results, consistent with the National Academies'
recommendations. DSCIM addresses common criticisms related to the damage functions
underlying current SC-GHG estimates (e.g., (Pindyck, 2017)) by developing multi-sector,
empirically grounded damage functions. The damage functions in the GIVE model offer a direct
implementation of the National Academies' near-term recommendation to develop updated
sectoral damage functions that are based on recently published work and reflective of the current
state of knowledge about damages in each sector. Specifically, the National Academies noted
that "[t]he literature on agriculture, mortality, coastal damages, and energy demand provide
immediate opportunities to update the [models]" (National Academies, 2017), which are the four
damage categories currently in GIVE. A limitation of both models is that the sectoral coverage is
still limited, and even the categories that are represented are incomplete. Neither DSCIM nor
GIVE yet accommodate estimation of several categories of temperature driven climate impacts
(e.g., morbidity, conflict, migration, biodiversity loss) and only represent a limited subset of
damages from changes in precipitation. For example, while precipitation is considered in the
agriculture sectors in both DSCIM and GIVE, neither model takes into account impacts of
flooding, changes in rainfall from tropical storms, and other precipitation related impacts. As
another example, the coastal damage estimates in both models do not fully reflect the
consequences of SLR-driven salt-water intrusion and erosion, or SLR damages to coastal tourism
and recreation. Other missing elements are damages that result from other physical impacts (e.g.,
ocean acidification, non-temperature-related mortality such as diarrheal disease and malaria) and

93

-------
the many feedbacks and interactions across sectors and regions that can lead to additional
damages25. See U.S. EPA (EPA, 2023c) for more discussion of omitted damage categories and
other modeling limitations. DSCIM and GIVE do account for the most commonly cited benefits
associated with CO2 emissions and climate change - CO2 crop fertilization and declines in cold
related mortality. As such, while the GIVE- and DSCIM-based results provide state-of-the-
science assessments of key climate change impacts, they remain partial estimates of future
climate damages resulting from incremental changes in CO2, CH4, and N2O26.

Finally, given the still relatively narrow sectoral scope of the recently developed DSCIM
and GIVE models, the damage module includes a third damage function that reflects a synthesis
of the state of knowledge in other published climate damages literature. Studies that employ
meta-analytic techniques27 offer a tractable and straightforward way to combine the results of
multiple studies into a single damage function that represents the body of evidence on climate
damages that pre-date CIL and RFF's research initiatives. The first use of meta-analysis to
combine multiple climate damage studies was done by (Tol, 2009) and included 14 studies. The
studies in (Tol, 2009) served as the basis for the global damage function in DICE starting in
version 2013R (Nordhaus W. , 2014). The damage function in the most recent published version
of DICE, DICE 2016, is from an updated meta-analysis based on a rereview of existing damage
studies and included 26 studies published over 1994-2013. Howard and Sterner provide a more
recent published peer-reviewed meta-analysis of existing damage studies (published through
2016) and account for additional features of the underlying studies (Howard & Sterner, 2017).
This study address differences in measurement across studies by adjusting estimates such that the
data are relative to the same base period. They also eliminate double counting by removing
duplicative estimates. Howard and Sterner's final sample is drawn from 20 studies that were
published through 2015. (Howard & Sterner, 2017) present results under several specifications

25	The one exception is that the agricultural damage function in DSCIM and GIVE reflects the ways that trade can

help mitigate damages arising from crop yield impacts.

26	One advantage of the modular approach used by these models is that future research on new or alternative

damage functions can be incorporated in a relatively straightforward way. DSCIM and GIVE developers
have work underway on other impact categories that may be ready for consideration in future updates (e.g.,
morbidity and biodiversity loss).

27	Meta-analysis is a statistical method of pooling data and/or results from a set of comparable studies of a problem.

Pooling in this way provides a larger sample size for evaluation and allows for a stronger conclusion than can
be provided by any single study. Meta-analysis yields a quantitative summary of the combined results and
current state of the literature.

94

-------
and shows that the estimates are somewhat sensitive to defensible alternative modeling choices.
As discussed in detail in U.S. EPA (EPA, 2023c), the damage module underlying the SC-GHG
estimates in this RIA includes the damage function specification (that excludes duplicate studies)
from (Howard & Sterner, 2017) that leads to the lowest SC-GHG estimates, all else equal.

The discounting module discounts the stream of future net climate damages to its present
value in the year when the additional unit of emissions was released. Given the long-time
horizon over which the damages are expected to occur, the discount rate has a large influence on
the present value of future damages. Consistent with the findings of (National Academies, 2017),
the economic literature, OMB Circular A-4's guidance for regulatory analysis, and IWG
recommendations to date (IWG, 2010) (IWG, 2013) (IWG, 2016a) (IWG, 2016b) (IWG, 2021),
the EPA continues to conclude that the consumption rate of interest is the theoretically
appropriate discount rate to discount the future benefits of reducing GHG emissions and that
discount rate uncertainty should be accounted for in selecting future discount rates in this
intergenerational context. OMB's Circular A-4 points out that "the analytically preferred method
of handling temporal differences between benefits and costs is to adjust all the benefits and costs
to reflect their value in equivalent units of consumption and to discount them at the rate
consumers and savers would normally use in discounting future consumption benefits" (OMB,
2003). The damage module described above calculates future net damages in terms of reduced
consumption (or monetary consumption equivalents), and so an application of this guidance is to
use the consumption discount rate to calculate the SC-GHG. Thus, EPA concludes that the use of
the social rate of return on capital (7 percent under current OMB Circular A-4 guidance), which
does not reflect the consumption rate, to discount damages estimated in terms of reduced
consumption would inappropriately underestimate the impacts of climate change for the
purposes of estimating the SC-GHG28.

For the SC-GHG estimates used in this RIA, EPA relies on a dynamic discounting
approach that more fully captures the role of uncertainty in the discount rate in a manner
consistent with the other modules. Based on a review of the literature and data on consumption
discount rates, the public comments received on individual EPA rulemakings, and the February

28 See also the discussion of the inappropriateness of discounting consumption-equivalent measures of benefits and
costs using a rate of return on capital in Circular A-4 (2023) (OMB, 2003).

95

-------
2021 TSD (IWG, 2021), and the (National Academies, 2017)29 recommendations for updating
the discounting module, the SC-GHG estimates rely on discount rates that reflect more recent
data on the consumption interest rate and uncertainty in future rates. Specifically, rather than
using a constant discount rate, the evolution of the discount rate over time is defined following
the latest empirical evidence on interest rate uncertainty and using a framework originally
developed by (Ramsey, 1928) that connects economic growth and interest rates. The Ramsey
approach explicitly reflects (1) preferences for utility in one period relative to utility in a later
period and (2) the value of additional consumption as income changes. The dynamic discount
rates used to develop the SC-GHG estimates applied in this RIA have been calibrated following
the (Newell, Pizer, & Prest, 2022) approach, as applied in (Rennert, et al., 2022) (Rennert, et al.,
2022a). This approach uses the discounting formula (Ramsey, 1928) in which the parameters are
calibrated such that (1) the decline in the certainty-equivalent discount rate matches the latest
empirical evidence on interest rate uncertainty estimated by (Bauer & Rudebusch, Interest rates
under falling stars, 2020) (Bauer & Rudebusch, 2023) and (2) the average of the certainty-
equivalent discount rate over the first decade matches a near-term consumption rate of interest.
Uncertainty in the starting rate is addressed by using three near-term target rates (1.5, 2.0, and
2.5 percent) based on multiple lines of evidence on observed market interest rates.

The resulting dynamic discount rate provides a notable improvement over the constant
discount rate framework used for SC-GHG estimation in previous EPA RIAs. Specifically, it
provides internal consistency within the modeling and a more complete accounting of
uncertainty consistent with economic theory (Arrow K. , et al., 2013) (Cropper, Freeman,
Groom, & Pizer, 2014) and the (National Academies, 2017) recommendation to employ a more
structural, Ramsey-like approach to discounting that explicitly recognizes the relationship
between economic growth and discounting uncertainty. This approach is also consistent with the
(National Academies, 2017) recommendation to use three sets of Ramsey parameters that reflect
a range of near-term certainty-equivalent discount rates and are consistent with theory and
empirical evidence on consumption rate uncertainty. Finally, the value of aversion to risk

29 Similarly, OMB's Circular A-4 (2023) points out that "The analytically preferred method of handling temporal
differences between benefits and costs is to adjust all the benefits and costs to reflect their value in equivalent
units of consumption before discounting them" (OMB, 2003).

96

-------
associated with net damages from GHG emissions is explicitly incorporated into the modeling
framework following the economic literature. See U.S. EPA (EPA, 2023c) for a more detailed
discussion of the entire discounting module and methodology used to value risk aversion in the
SC-GHG estimates.

The methodologies adopted in this SC-GHG estimation process allow for a more holistic
treatment of uncertainty than past estimates used by the EPA. The updates incorporate a
quantitative consideration of uncertainty into all modules and use a Monte Carlo approach that
captures the compounding uncertainties across modules. The estimation process generates nine
separate distributions of discounted marginal damages per metric ton - the product of using three
damage modules and three near-term target discount rates - for each gas in each emissions year.
These distributions have long right tails reflecting the extensive evidence in the scientific and
economic literature that shows the potential for lower-probability but higher-impact outcomes
from climate change, which would be particularly harmful to society. The uncertainty grows
over the modeled time horizon. Therefore, under cases with a lower near-term target discount
rate - that give relatively more weight to impacts in the future - the distribution of results is
wider. To produce a range of estimates that reflects the uncertainty in the estimation exercise
while also providing a manageable number of estimates for policy analysis, the EPA combines
the multiple lines of evidence on damage modules by averaging the results across the three
damage module specifications. The full results generated from the updated methodology for
methane and other greenhouse gases (SC-CO2, SC-CH4, and SC-N2O) for emissions years 2020
through 2080 are provided in U.S. EPA (EPA, 2023c).

Table 4-6 summarizes the resulting averaged certainty-equivalent SC-GHG estimates
under each near-term discount rate that are used to estimate the climate impacts of the GHG
emission changes expected from the final rule. These estimates are reported in 2021 dollars but
are otherwise identical to those presented in U.S. EPA (EPA, 2023c). The SC-GHGs increases
over time within the models — i.e., the societal harm from one metric ton emitted in 2030 is
higher than the harm caused by one metric ton emitted in 2027 — because future emissions
produce larger incremental damages as physical and economic systems become more stressed in
response to greater climatic change, and because GDP is growing over time and many damage
categories are modeled as proportional to GDP.

-------
The methodological updates described above represent a major step forward in bringing
SC-GHG estimation closer to the frontier of climate science and economics and address many of
the (National Academies, 2017) near-term recommendations. Nevertheless, the resulting SC-
GHG estimates presented in Table 4-6, still have several limitations, as would be expected for
any modeling exercise that covers such a broad scope of scientific and economic issues across a
complex global landscape. There are still many categories of climate impacts and associated
damages that are only partially or not reflected yet in these estimates and sources of uncertainty
that have not been fully characterized due to data and modeling limitations. For example, the
modeling omits most of the consequences of changes in precipitation, damages from extreme
weather events, the potential for nongradual damages from passing critical thresholds (e.g.,
tipping elements) in natural or socioeconomic systems, and non-climate mediated effects of
GHG emissions. More specifically for methane, the SC-CH4 estimates do not account for the
direct health and welfare impacts associated with tropospheric ozone produced by methane. As
discussed further in U.S. EPA (EPA, 2023c), recent studies have found the global ozone-related
respiratory mortality benefits of CH4 emissions reductions, which are not included in the SC-CH4
values presented in Table 4-6, to be, in 2021 dollars, approximately $2,500 per metric ton of
methane emissions in 2030. (McDuffie, et al., 2023). In addition, the SC-CH4 estimates do not
reflect that methane emissions lead to a reduction in atmospheric oxidants, like hydroxyl
radicals, nor do they account for impacts associated with CO2 produced from methane oxidizing
in the atmosphere. Importantly, the updated SC-GHG methodology does not yet reflect
interactions and feedback effects within, and across, Earth and human systems. For example, it
does not explicitly reflect potential interactions among damage categories, such as those
stemming from the interdependencies of energy, water, and land use. These, and other,
interactions and feedbacks were highlighted by the National Academies as an important area of
future research for longer-term enhancements in the SC-GHG estimation framework.

Table 4-6 summarizes the resulting averaged certainty-equivalent SC-GHG estimates
under each near-term discount rate that are used to monetize the climate impacts of the GHG
emission changes expected from the final rule. These estimates are reported in 2021 dollars but
are otherwise identical to those presented in U.S. EPA () The SC-GHG increases over time
within the models — i.e., the societal harm from one metric ton emitted in 2030 is higher than
the harm caused by one metric ton emitted in 2025 — because future emissions produce larger

98

-------
incremental damages as physical and economic systems become more stressed in response to
greater climatic change, and because GDP is growing over time and many damage categories are
modeled as proportional to GDP.

Table 4-6 Annual Rounded SC-CO2, SC-CH4, and SC-N2O Values, 2024-2038

SC-GHG and Near-term Ramsey Discount Rate

Emission
Year

SC-CO2

(2021 dollars per metric ton of
C02)

SC-CH4

(2021 dollars per metric ton of
CH4)

SC-N2O

(2021 dollars per metric ton of

N20)



Near-term Ramsey discount rate

Near-term Ramsey discount rate

Near-term Ramsey discount rate



2.5%

2%

1.5%

2.5%

2%

1.5%

2.5%

2%

1.5%

2024

130

220

370

1,600

2,000

2,800

41,000

62,000

98,000

2025

140

220

390

1,700

2,100

2,900

42,000

63,000

100,000

2026

140

220

380

1,700

2,200

3,000

43,000

64,000

100,000

2027

140

230

390

1,800

2,300

3,000

44,000

66,000

100,000

2028

150

230

390

1,900

2,400

3,100

45,000

67,000

100,000

2029

150

240

400

1,900

2,400

3,200

46,000

68,000

110,000

2030

150

240

400

2,000

2,500

3,300

47,000

69,000

110,000

2031

150

240

410

2,100

2,600

3,400

48,000

71,000

110,000

2032

160

250

410

2,200

2,700

3,500

49,000

72,000

110,000

2033

160

250

420

2,300

2,800

3,600

50,000

73,000

110,000

2034

160

260

420

2,300

2,900

3,700

51,000

75,000

110,000

2035

170

260

430

2,400

3,000

3,800

52,000

76,000

120,000

2036

170

260

430

2,500

3,100

3,900

53,000

77,000

120,000

2037

170

270

440

2,600

3,200

4,100

54,000

79,000

120,000

2038

170

270

440

2,700

3,200

4,200

55,000

80,000

120,000

Source: U.S. EPA (2023f).

Note: These SC-GHG values are identical to those reported in the technical report U.S. EPA (2023d) adjusted for
inflation to 2021 dollars using the annual GDP Implicit Price Deflator values in the U. S. Bureau of Economic
Analysis' (BEA) NIPA Table 1.1.9 (U.S. BEA, 2021). The values are stated in $/metric ton GHG and vary
depending on the year of GHG emissions. This table displays the values rounded to two significant figures. The
annual unrounded values used in the calculations in this RIA are available in Appendix A.5 of U.S. EPA (2023f) and
at: www.epa.gov/environmental-economics/scghg.

Table 4-7 shows the estimated monetary value of the estimated changes in CO2, CH4, and
N2O emissions expected to occur over 2024 through 2038 for this rule. The EPA estimated the
dollar value of the GHG-related effects for each analysis year between 2024 and 2038 by
applying the SC-GHG estimates presented in Table 4-6 to the estimated changes in GHG
emissions in the corresponding year as shown in Chapter 3. The EPA then calculated the present
value (PV) and equivalent annualized value (EAV) of benefits from the perspective of 2023 by

99

-------
discounting each year-specific value to the year 2023 using the same discount rate used to
calculate the SC-GHG.30

100

-------
Table 4-7 Monetized Impacts of Estimated CO2, CH4, N2O Changes for the HON Amendments, P&R I and P&R II

NESHAP and Subpart Wb, Ilia, NNNa, and RRRa NSPS Amendments, 2024-2038, (million 2021$)





CO2 (Millions of 2021$)





CH4 (Millions of 2021$)





N2O (Millions of 2021$)



Near-term Ramsey Discount Rate

Year

2.5%

2%

1.5%

2.5%

2%

1.5%

2.5%

2%

1.5%

2024

(40)

(65)

(110)

37

47

64

(0.1)

(0.1)

(0.2)

2025

(40)

(66)

(114)

38

49

66

(0.1)

(0.1)

(0.2)

2026

(41)

(67)

(113)

40

50

68

(0.1)

(0.1)

(0.2)

2027

(42)

(68)

(115)

41

52

70

(0.1)

(0.1)

(0.2)

2028

(43)

(69)

(116)

43

54

72

(0.1)

(0.1)

(0.2)

2029

(44)

(70)

(118)

45

56

74

(0.1)

(0.1)

(0.2)

2030

(45)

(71)

(119)

46

58

76

(0.1)

(0.1)

(0.2)

2031

(46)

(73)

(121)

48

60

78

(0.1)

(0.1)

(0.2)

2032

(47)

(74)

(122)

50

62

81

(0.1)

(0.1)

(0.2)

2033

(47)

(75)

(123)

52

64

83

(0.1)

(0.2)

(0.2)

2034

(48)

(76)

(125)

54

66

86

(0.1)

(0.2)

(0.2)

2035

(49)

(77)

(127)

56

68

88

(0.1)

(0.2)

(0.2)

2036

(50)

(78)

(128)

57

70

91

(0.1)

(0.2)

(0.2)

2037

(51)

(79)

(129)

59

72

93

(0.1)

(0.2)

(0.2)

2038

(52)

(80)

(131)

61

75

95

(0.1)

(0.2)

(0.3)

NPV

(560)

(926)

(1,607)

590

764

1,045

(1.2)

(1.9)

(3.0)

EAV

(45)

(72)

(120)

48

59

78

(0.1)

(0.1)

(0.2)

Note: Monetized climate impacts are based on changes (increases) in CO2 and N20 emissions and decreases in CH4 emissions and are calculated using three
different estimates of the social cost of each greenhouse gas (SC-GHG) (2.5 percent, 2 percent, and 1.5 percent discount rates) from U.S. EPA (EPA, 2023c). A
number in parenthesis represents a negative value.

101

-------
Unlike many environmental problems where the causes and impacts are distributed more
locally, GHG emissions are a global externality making climate change a true global challenge.
GHG emissions contribute to damages around the world regardless of where they are emitted.
Because of the distinctive global nature of climate change, in the RIA for this final rule the EPA
centers attention on a global measure of climate impacts from GHG emissions. Consistent with
all IWG recommended SC-GHG estimates to date, the SC-GHG values presented in Table 4-6
provide a global measure of monetized damages from CO2, CH4 and N2O and Table 4-7 present
the monetized global climate impacts of the CO2, CH4 and N2O emission changes expected from
the final rule. This approach is the same as that taken in EPA regulatory analyses from 2009
through 2016 and since 2021. It is also consistent with guidance in (OMB, 2003) that states when
a regulation is likely to have international effects, "these effects should be reported"31. EPA also
notes that EPA's cost estimates in RIAs, including the cost estimates contained in this RIA,
regularly do not differentiate between the share of compliance costs expected to accrue to U.S.
firms versus foreign interests, such as to foreign investors in regulated entities32. A global
perspective on climate effects is therefore consistent with the approach EPA takes on costs.

There are many reasons, as summarized in this section - and as articulated by OMB and in IWG
assessments (IWG, 2010) (IWG, 2013) (IWG, 2016a) (IWG, 2016b) (IWG, 2021), the 2015
Response to Comments (IWG, 2015) and in detail in U.S. EPA (EPA, 2023c) and in Appendix A
of the Response to Comments document for the December 2023 Final Oil and Gas NSPS/EG

31 While OMB Circular A-4 recommends that international effects we reported separately, the guidance also

explains that "[d]ifferent regulations may call for different emphases in the analysis, depending on the nature
and complexity of the regulatory issues" (OMB 2003). Circular A-4 (2023) states that "In certain contexts, it
may be particularly appropriate to include effects experienced by noncitizens residing abroad in your primary
analysis. Such contexts include, for example, when:

• assessing effects on noncitizens residing abroad provides a useful proxy for effects on U.S. citizens
and residents that are difficult to otherwise estimate;

• assessing effects on noncitizens residing abroad provides a useful proxy for effects on U.S. national
interests that are not otherwise fully captured by effects experienced by particular U.S. citizens and
residents (e.g., national security interests, diplomatic interests, etc.);

• regulating an externality on the basis of its global effects supports a cooperative international approach
to the regulation of the externality by potentially inducing other countries to follow suit or maintain
existing efforts; or

• international or domestic legal obligations require or support a global calculation of regulatory effects"
(OMB, 2003).

32 For example, in the RIA for the 2018 Proposed Reconsideration of the Oil and Natural Gas Sector Emission

Standards for New, Reconstructed, and Modified Sources, the EPA acknowledged that some portion of
regulatory costs will likely "accru[e] to entities outside U.S. borders" through foreign ownership, employment,
or consumption. In general, a significant share of U.S. corporate debt and equities are foreign-
owned, including in the oil and gas industry.

102

-------
Rulemaking - why the EPA focuses on the global value of climate change impacts when
analyzing policies that affect GHG emissions.

International cooperation and reciprocity are essential to successfully addressing climate
change, as the global nature of greenhouse gases means that a ton of GHGs emitted in any other
country harms those in the U.S. just as much as a ton emitted within the territorial U.S.
Assessing the benefits of U.S. GHG mitigation activities requires consideration of how those
actions may affect mitigation activities by other countries, as those international mitigation
actions will provide a benefit to U.S. citizens and residents by mitigating climate impacts that
affect U.S. citizens and residents. This is a classic public goods problem because each country's
reductions benefit everyone else, and no country can be excluded from enjoying the benefits of
other countries' reductions. The only way to achieve an efficient allocation of resources for
emissions reduction on a global basis — and so benefit the U.S. and its citizens and residents —
is for all countries to base their policies on global estimates of damages. A wide range of
scientific and economic experts have emphasized the issue of international cooperation and
reciprocity as support for assessing global damages of GHG emission in domestic policy
analysis. Using a global estimate of damages in U.S. analyses of regulatory actions allows the
U.S. to continue to actively encourage other nations, including emerging major economies, to
also assess global climate damages of their policies and to take steps to reduce emissions. For
example, many countries and international institutions have already explicitly adapted the global
SC-GHG estimates used by EPA in their domestic analyses (e.g., Canada, Israel) or developed
their own estimates of global damages (e.g., Germany), and recently, there has been renewed
interest by other countries to update their estimates since the draft release of the updated SC-
GHG estimates presented in the December 2022 Oil and Gas NSPS/EG Supplemental Proposal
RIA33. Several recent studies have empirically examined the evidence on international GHG
mitigation reciprocity, through both policy diffusion and technology diffusion effects. See U.S.
EPA (EPA, 2023c) for more discussion.

33 In April 2023, the government of Canada announced the publication of an interim update to their SC-GHG
guidance, recommending SC-GHG estimates identical to the EPA's updated estimates presented in the
December 2022 Supplemental Proposal RIA. The Canadian interim guidance will be used across all Canadian
federal departments and agencies, with the values expected to be finalized by the end of the year. See more at
https://www.canada.ca/en/environment-climate-change/services/climate-change/science-research-data/social-
cost-ghg.html.

103

-------
For all of these reasons, the EPA believes that a global metric is appropriate for assessing
the climate impacts of GHG emissions in this final RIA. In addition, as emphasized in the
(National Academies, 2017) recommendations, "[i]t is important to consider what constitutes a
domestic impact in the case of a global pollutant that could have international implications that
impact the United States." The global nature of GHG pollution and its impacts means that U.S.
interests are affected by climate change impacts through a multitude of pathways and these need
to be considered when evaluating the benefits of GHG mitigation to U.S. citizens and residents.
The increasing interconnectedness of global economy and populations means that impacts
occuring outside of U.S. borders can have significant impacts on U.S. interests. Examples of
affected interests include direct effects on U.S. citizens and assets located abroad, international
trade, and tourism, and spillover pathways such as economic and political destabilization and
global migration that can lead to adverse impacts on U.S. national security, public health, and
humanitarian concerns. Those impacts point to the global nature of the climate change problem
and are better captured within global measures of the social cost of greenhouse gases.

In the case of these global pollutants, for the reasons articulated in this section, the
assessment of global net damages of GHG emissions allows EPA to fully disclose and
contextualize the net climate impacts of GHG emission changes expected from this final rule.
The EPA disagrees with public comments received on the December 2022 Oil and Gas
NSPS/EG Supplemental Proposal that suggested that the EPA can or should use a metric focused
on benefits resulting solely from changes in climate impacts occuring within U.S. borders. The
global models used in the SC-GHG modeling described above do not lend themselves to be
disaggregated in a way that could provide comprehensive information about the distribution of
the rule's climate impacts to citizens and residents of particular countries, or population groups
across the globe and within the U.S. Two of the models used to inform the damage module, the
GIVE and DSCIM models, have spatial resolution that allows for some geographic
disaggregation of future climate impacts across the world. This permits the calculation of a
partial GIVE and DSCIM-based SC-GHG measuring the damages from four or five climate
impact categories projected to physically occur within the U.S., respectively, subject to caveats.
As discussed at length in U.S. EPA (EPA, 2023c), these damage modules are only a partial
accounting and do not capture all of the pathways through which climate change affects public
health and welfare. As discussed at length in U.S. EPA (EPA, 2023c), these damage modules are

104

-------
only a partial accounting and do not capture all of the pathways through which climate change
affects public health and welfare. Thus, they only cover a subset of potential climate change
impacts. Furthermore, the damage modules do not capture spillover or indirect effects whereby
climate impacts in one country or region can affect the welfare of residents in other countries or
regions—for example through the movement of refugees.

Additional modeling efforts can and have shed further light on some omitted damage
categories. For example, the Framework for Evaluating Damages and Impacts (FrEDI) is an
open-source modeling framework developed by the EPA34 to facilitate the characterization of net
annual climate change impacts in numerous impact categories within the contiguous U.S. and
monetize the associated distribution of modeled damages (Sarofim, et al., 2021) (EPA, 2021).
The additional impact categories included in FrEDI reflect the availability of U.S.-specific data
and research on climate change effects. As discussed in U.S. EPA (EPA, 2023c) results from
FrEDI show that annual damages resulting from climate change impacts within the contiguous
U.S. (CONUS) (i.e., excluding Hawaii, Alaska, and U.S. territories) and for impact categories
not represented in GIVE and DSCIM are expected to be substantial. As discussed in U.S. EPA
(EPA, 2021), results from FrEDI show that annual damages resulting from climate change
impacts within the contiguous U.S. (CONUS) (i.e., excluding Hawaii, Alaska, and U.S.
territories) and for impact categories not represented in GIVE and DSCIM are expected to be
substantial. For example, FrEDI estimates a partial SC-CO2 of $38/mtC02 for damages
physically occurring within CONUS for 2030 emissions (under a 2 percent near-term Ramsey
discount rate) (Hartin, et al., 2023), compared to a GIVE and DSCIM-based U.S.-specific SC-
CO2 of$17/mtC02 and $15/mtC02, respectively, for 2030 emissions (2021 USD). While the
FrEDI results help to illustrate how monetized damages physically occurring within CONUS
increase as more impacts are reflected in the modeling framework, they are still subject to many
of the same limitations associated with the DSCIM and GIVE damage modules, including the

34 The FrEDI framework and Technical Documentation have been subject to a public review comment period and
an independent external peer review, following guidance in the EPA Peer-Review Handbook for Influential
Scientific Information (ISI). Information on the FrEDI peer-review is available at the EPA Science Inventory

105

-------
omission or partial modeling of important damage categories35'36. Finally, none of these modeling
efforts - GIVE, DSCIM, and FrEDI - reflect non-climate mediated effects of GHG emissions
experienced by U.S. populations (other than CO2 fertilization effects on agriculture). As one
example of new research on non-climate mediated effects of methane emissions, (McDuffie, et
al., 2023) estimate the monetized increase in respiratory-related human mortality risk from the
ozone produced from a marginal pulse of methane emissions. Using the socioeconomics from the
RFF-SPs and the 2 percent near-term Ramsey discounting approach, this additional risk to U.S.
populations is on the order of approximately $330/mtCH4 (2021 USD) for 2030 emissions.

Applying the U.S.-specific partial SC-GHG estimates derived from the multiple lines of
evidence described above to the GHG emission changes expected under the final rule would also
yield impacts. For example, the present value of the climate benefits of the final rule as measured
by FrEDI from climate change impacts in CONUS are estimated to be $35 million (under a 2
percent near-term Ramsey discount rate)37. However, the numerous explicitly omitted damage
categories and other modeling limitations discussed above and throughout U.S. EPA ((EPA,
2023c) make it likely that these estimates underestimate the impacts to U.S. citizens and
residents of the GHG emission changes from the final rule; the limitations in developing a U.S.-
specific estimate that accurately captures direct and spillover effects on U.S. citizens and
residents further demonstrates that it is more appropriate to use a global measure of climate

35 Another method that has produced estimates of the effect of climate change on U.S.-specific outcomes uses a

top-down approach to estimate aggregate damage functions. Published research using this approach include
total-economy empirical studies that econometrically estimate the relationship between GDP and a climate
variable, usually temperature. As discussed in U.S. EPA (EPA, 2023c) the modeling framework used in the
existing published studies using this approach differ in important ways from the inputs underlying the SC-
GHG estimates described above (e.g., discounting, risk aversion, and scenario uncertainty) and focus solely on
SC-C02. Hence, we do not consider this line of evidence in the analysis for this RIA. Updating the framework
of total-economy empirical damage functions to be consistent with the methods described in this RIA and ibid,
would require new analysis. Finally, because total-economy empirical studies estimate market impacts, they do
not include any non-market impacts of climate change (e.g., heat related mortality) and therefore are also only a
partial estimate. The EPA will continue to review developments in the literature and explore ways to better
inform the public of the full range of GHG impacts.

36 FrEDI estimates a partial SC-CH4 (N20) of $610/mtCH4 ($ 11,000/mtN2O) for damages physically occurring

within CONUS for 2030 emissions (under a 2 percent near-term Ramsey discount rate) (Hartin, et al., 2023)
compared to a GIVE and DSCIM-based U.S.-specific SC-CH4 of $290/mtCH4 ($5,200/mtN20) and
$78/mtCH4 ($4,000/mtN20), respectively, for 2030 emissions (USD 2021).

37 DCIM and GIVE use global damage functions. Damage functions based on only U.S.-data and research, but not
for other parts of the world, were not included in those models. FrEDI does make use of some of this U.S. -
specific data and research and as a result has a broader coverage of climate impact categories.

106

-------
impacts from GHG emissions. The EPA will continue to review developments in the literature,
including more robust methodologies for estimating the magnitude of the various damages to
U.S. populations from climate impacts and reciprocal international mitigation activities, and
explore ways to better inform the public of the full range of GHG impacts.

4.8 Total Monetized Benefits

Table 4-8 through Table 4-10 present a summary of monetized benefits for the final
amendments to rules included in this rulemaking, both individually and cumulatively. Net
benefits in each table are calculated as the sum of monetized health benefits and climate benefits
(including climate disbenefits). Non-monetized benefits are included qualitatively in each table.
Benefits related to both short- and long-term exposure of ozone are estimated. Tables presenting
benefits list both estimates, with short-term exposure benefits listed first. A complete
presentation of benefits relative to costs appears in Chapter 6 of this RIA. We note, as we
mentioned in Chapter 1, that there are minimal monetized benefits for the P&R II amendments,
and hence there is no table of benefits for this final rule below. In addition, the benefits for the
Subpart VVb and Ilia, NNNa, and RRRa NSPS are the same for the less and more stringent
options, and thus those estimates are already presented earlier in this chapter. Hence, there is no
table of benefits for each of these final rules below.

107

-------
Table 4-8 Summary of Monetized Benefits PV/EAV for the HON Amendments, 2024-
	2038 (million 2021$), Discounted to 2023	



Final Rule

Less Stringent Alternative

More Stringent Alternative

3%

PV

EAV

PV

EAV

PV

EAV

Monetized
Health Benefits

70 and 630

5.9 and 53

70 and 630

5.9 and 52

71 and 640

6 and 54

Climate













Disbenefits

140

11

140

11

140

11

(2%)













Total Benefits

(70) and 490

(5.1) and 42

(70) and 490

(5.1) and 41

(69) and 500

(5) and 43

7%

PV

EAV

PV

EAV

PV

EAV

Monetized
Health Benefits

48 and 420

5.2 and 46

48 and 420

5.2 and 46

49 and 430

5.3 and 47

Climate













Disbenefits

140

11

140

11

140

11

(2%)













Total Benefits

(92) and 280

(5.8) and 35

(92) and 280

(5.8) and 35

(91) and 290

(5.7) and 36

Nonmonetized 1,107 tons of HAP emission reductions. Health effects from reduced exposure to ethylene
Benefits oxide, chloroprene, benzene, 1,3-butadiene, vinyl chloride, ethylene dichloride, chlorine,
maleicanhydride, and acrolein

Note: Monetized air-quality related health benefits include ozone related health benefits associated with reductions
in VOC emissions. The health benefits are associated with several point estimates and are presented at real discount
rates of 3 and 7 percent. The two benefits estimates are separated by the word "and" to signify that they are two
separate estimates. The estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions
and VOC reductions outside of the ozone season remain unmonetized and are thus not reflected in the table. The
unmonetized effects also include disbenefits resulting from a secondary increase in CO emissions. Monetized
climate benefits and disbenefits are based on changes (increases) in CO2 and N20 emissions and decreases in CH4
emissions and are calculated using three different estimates of the social cost of each greenhouse gas (SC-GHG) (2.5
percent, 2 percent, and 1.5 percent discount rates). For the presentational purposes of this table, we show the benefits
and disbenefits associated with the SC-GHG at a 2 percent discount rate. Parentheses around a number denotes a
negative value. Negative climate disbenefits is a positive value. Rows may not appear to add correctly due to
rounding.

108

-------
Table 4-9 Summary of Monetized Benefits PV/EAV for the P&R I Amendments, 2024-

	2038 (million 2021$), Discounted to 2023	

	Final Rule	Less Stringent Alternative More Stringent Alternative

3%	PV	EAV	PV	EAV	PV	EAV

(0.03) and (0.3)
2

(1.7) and (2)
EAV

(0.03) and (0.3)
2

(1.7) and (2)

Nonmonetized 264 tons per year (tpy) of HAP emission reductions, including an approximate 14 tpy

	Benefits	reduction in chloroprene emissions	

Note: Monetized benefits include ozone related health benefits associated with reductions in VOC emissions. The
health benefits are associated with several point estimates and are presented at real discount rates of 3 and 7 percent.
The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions and VOC reductions
outside of the ozone season remain unmonetized and are thus not reflected in the table. The unmonetized effects also
include disbenefits resulting from a secondary increase in CO emissions. Monetized climate benefits and disbenefits
are based on changes (increases) in CO2 and N20 emissions and decreases in CH4 emissions and are calculated using
three different estimates of the social cost of each greenhouse gas (SC-GHG) (2.5 percent, 2 percent, and 1.5 percent
discount rates). For the presentational purposes of this table, we show the benefits and disbenefits associated with
the SC-GHG at a 2 percent discount rate.

Monetized Health	(0.2)and(1.7)	(0.02) and (0.1)	(0.5) and (4)	-0.04 and-0.3	(0.4) and (3)

ClmrlalC	22	2	22	2	22

Disbenefits (2%)

Total Benefits	(22) and (24)	(1.7) and (1.8)	(23) and (26)	(1.7) and (2)	(22) and (25)

7%	PV	EAV	PV EAV	PV

Monetized Heaith	(0.2) and (1.5)	(0.02) and (0.2) (0.02) and (0.2)	(0.3) and (2.4)

Climate	22	2	22 2	22

Disbenefits (2%)

Total Benefits (22) and (24) (1.7) and (1.9) (22) and (24) (1.7) and (1.9) (22) and (24)

109

-------
Table 4-10 Summary of Monetized Benefits PV/EAV for the Cumulative Impact of the

HON Amendments, P&R I and P&R IINESHAP and Subpart Wb, Ilia, NNNa, and
	RRRa NSPS Amendments, 2024-2038 (million 2021$), Discounted to 2023	

Final Rule	Less Stringent Alternative More Stringent Alternative

3%

PV

EAV

PV

EAV

PV

EAV

Health
Benefits

71 and 640

6 and 53

71 and 640

5.9 and 53

72 and 650

6.1 and 53

Climate
Disbenefits

160

13

160

13

160

13

Net Benefits

(89) and 480

(7) and 40

(89) and 480

(7.1) and 40

(88) and 490

(6.9) and 40

7%

PV

EAV

PV

EAV

PV

EAV

Health
Benefits

48 and 430

5.3 and 47

48 and 430

5.3 and 47

49 and 440

5.3 and 47

Climate
Disbenefits

160

13

160

13

160

13

Net Benefits

(110) and 270

(7.7) and 34

(110) and 270

(7.7) and 34

(110) and 280

(7.7) and 34

Non-monetized Benefits: Health benefits associated with emission reductions of 6,230 tpy of HAP including
hexane, benzene, methanol, 1,3-butadiene, and vinyl acetate. Health benefits associated with reduction of 54 tpy
of ethylene oxide and 14 tpy of chloroprene. Ecosystem benefits related to the reductions of ozone and nitrogen

and sulfur deposition.	

Note: Monetized benefits include ozone related health benefits associated with reductions in VOC emissions. The
health benefits are associated with several point estimates and are presented at real discount rates of 3 and 7 percent.
The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions and VOC reductions
outside of the ozone season remain unmonetized and are thus not reflected in the table. The unmonetized effects also
include disbenefits resulting from a secondary increase in CO emissions. Monetized climate benefits and disbenefits
are based on changes (increases) in CO2 and N20 emissions and decreases in CH4 emissions and are calculated using
three different estimates of the social cost of each greenhouse gas (SC-GHG) (2.5 percent, 2 percent, and 1.5 percent
discount rates). For the presentational purposes of this table, we show the benefits and disbenefits associated with
the SC-GHG at a 2 percent discount rate.

110

-------
CHAPTER 4 APPENDIX

Table 4A-1 through 4A-3 summarize the interim SC-CO2, SC-CH4, and SC-N2O
estimates that were used in the RIA for the years 2024-2038. These estimates are reported in
2021 dollars but are otherwise identical to those presented in the IWG's 2021 TSD and 2016
TSD (IWG, 2021). The SC-GHG increases over time within the models {i.e., the societal harm
from one metric ton emitted in 2030 is higher than the harm caused by one metric ton emitted in
2025) because future emissions produce larger incremental damages as physical and economic
systems become more stressed in response to greater climatic change, and because GDP is
growing over time and many damage categories are modeled as proportional to GDP.

Table 4A-4 shows the estimated monetary value of the estimated changes in CO2, CH4,
N2O, and total GHG emissions expected to occur over 2024 through 2038 from the final rule
using the IWG's interim SC-GHG estimates.

Table 4A-1 Interim Social Cost of Carbon Values, 2024-2038 (2021$/Metric Ton CO2)
Emissions Year	Discount Rate and Statistic



5% Average

3% Average

2.5% Average

3% 95th
Percentile

2024

$17

$58

$85

$170

2025

$18

$59

$86

$180

2026

$18

$60

$88

$180

2027

$19

$61

$89

$180

2028

$19

$62

$90

$190

2029

$20

$63

$92

$190

2030

$20

$64

$93

$190

2031

$21

$66

$95

$200

2032

$21

$67

$96

$200

2033

$22

$68

$97

$210

2034

$23

$69

$99

$210

2035

$23

$70

$100

$210

2036

$24

$71

$100

$220

2037

$24

$73

$100

$220

2038

$25

$74

$110

$230

Note: These SC-CO2 values are identical to those reported in the IWG's 2021 TSD (IWG, 2021) and 2016 TSD
(IWG, 2016a) adjusted to 2021 dollars using the annual GDP Implicit Price Deflator values in the U.S. Bureau of
Economic Analysis' (BEA) NIPA Table 1.1.9 (U.S. BEA 2022). This table displays the values rounded to the
nearest dollar; the annual unrounded values used in the calculations in this analysis are available on OMB's website:
https://www.whitehouse.gOv/omb/information-regulatorv-affairs/regulatorv-matters/#scghgs.

Source: Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under
Executive Order 13990 (IWG 2021)

111

-------
Table 4A-2: Interim Social Cost of Methane Values, 2024-2038 (2021$ /Metric Ton CH4)

Emissions Year	Discount Rate and Statistic



5% Average

3% Average

2.5% Average

3% 95th
Percentile

2024

$810

$1,700

$2,300

$4,600

2025

$840

$1,800

$2,300

$4,700

2026

$860

$1,800

$2,400

$4,900

2027

$890

$1,900

$2,400

$5,000

2028

$920

$1,900

$2,500

$5,100

2029

$950

$2,000

$2,600

$5,300

2030

$980

$2,000

$2,600

$5,400

2031

$1,000

$2,100

$2,700

$5,600

2032

$1,000

$2,200

$2,700

$5,700

2033

$1,100

$2,200

$2,800

$5,900

2034

$1,100

$2,300

$2,900

$6,000

2035

$1,200

$2,300

$2,900

$6,200

2036

$1,200

$2,400

$3,000

$6,400

2037

$1,200

$2,400

$3,100

$6,500

2038

$1,300

$2,500

$3,100

$6,700

Note: These SC-CH4 values are identical to those reported in the IWG's 2021 TSD (IWG, 2021) and 2016 TSD
(IWG, 2016a) adjusted to 2021 dollars using the annual GDP Implicit Price Deflator values in the U.S. Bureau of
Economic Analysis' (BEA) NIPA Table 1.1.9 (U.S. BEA 2022). This table displays the values rounded to two
significant figures; the annual unrounded values used in the calculations in this analysis are available on OMB's
website: https://www.whiteh0use.g0v/0mb/inf0rmati0n-regulat0rv-affairs/regulat0rv-matters/#scghgs.

Source: Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under
Executive Order 13990 (IWG 2021)

Table 4A-3: Interim Social Cost of Nitrous Oxide Values, 2024-2038 (2021$ /Metric Ton

	N2O)	

Emissions Year	Discount Rate and Statistic



5% Average

3% Average

2.5% Average

3% 95th
Percentile

2024

$6,900

$21,000

$31,000

$55,000

2025

$7,100

$21,000

$31,000

$57,000

2026

$7,300

$22,000

$32,000

$58,000

2027

$7,500

$22,000

$32,000

$59,000

2028

$7,700

$23,000

$33,000

$60,000

2029

$7,900

$23,000

$33,000

$62,000

2030

$8,100

$24,000

$34,000

$63,000

2031

$8,400

$24,000

$35,000

$64,000

2032

$8,600

$25,000

$35,000

$66,000

2033

$8,900

$25,000

$36,000

$67,000

2034

$9,200

$26,000

$37,000

$69,000

2035

$9,400

$26,000

$37,000

$70,000

2036

$9,700

$27,000

$38,000

$71,000

2037

$9,900

$27,000

$39,000

$73,000

2038

$10,000

$28,000

$39,000

$74,000

Note: These SC-N20 values are identical to those reported in the IWG's 2021 TSD (IWG, 2021) and 2016 TSD
(IWG, 2016a) adjusted to 2021 dollars using the annual GDP Implicit Price Deflator values in the U.S. Bureau of
Economic Analysis' (BEA) NIPA Table 1.1.9 (U.S. BEA 2022). This table displays the values rounded to two
significant figures; the annual unrounded values used in the calculations in this analysis are available on OMB's
website: https://www.whiteh0use.g0v/0mb/inf0rmati0n-regulat0rv-affairs/regulat0rv-matters/#scghgs.

Source: Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under
Executive Order 13990 (IWG 2021)

112

-------
Table 4A-4 Monetized Benefits of Estimated CO2, CH4, N2O Changes of the HON Amendments, P&R I and P&R II
	NESHAP and Subpart Wb, Ilia, NNNa, and RRRa NSPS Amendments, 2024-2038, (million 2021$)	

SC-CP2 (Millions of 2021S)

SC-CH4 (Millions of 2021S)

SC-N2Q (Millions of 2021S)

Discount rate and statistic

Year

2024

2025

2026

2027

2028

2029

2030

2031

2032

2033

2034

2035

2036

2037

2038

5%
Average

$(13)
$(13)
$(13)
$(14)
$(14)
$(15)
$(15)
$(15)
$(16)
$(16)
$(17)
$(17)
$(18)
$(18)
$(19)

3%
Average

$(43)
$(44)
$(44)
$(45)
$(46)
$(47)
$(48)
$(49)
$(49)
$(50)
$(51)
$(52)
$(53)
$(54)
$(55)

2.5%
Average

$(63)
$(64)
$(65)
$(66)
$(67)
$(68)
$(69)
$(70)
$(71)
$(72)
$(73)
$(74)
$(75)
$(76)
$(77)

3% 95'"
Percentile

$(128)
$(131)
$(133)
$(136)
$(139)
$(141)
$(144)
$(147)
$(150)
$(153)
$(156)
$(159)
$(162)
$(165)
$(168)

5%
Average

$19
$19
$20
$20
$21
$22
$22
$23
$24
$25
$26
$27
$27
$28
$29

Discount rate and statistic

3%	2.5%

Average

$40
$41
$42
$43
$44
$46
$47
$48
$49
$51
$52
$53
$55
$56
$57

Average

$52
$53
$55
$56
$57
$59
$60
$61
$63
$65
$66
$68
$69
$71
$72

3% 95th
Percentile

$106
$109
$112
$115
$118
$121
$124
$128
$131
$135
$139
$142
$146
$150
$153

5%
Average

$(0.05)
$(0.05)
$(0.05)
$(0.05)
$(0.05)
$(0.05)
$(0.06)
$(0.06)
$(0.06)
$(0.06)
$(0.06)
$(0.06)
$(0.07)
$(0.07)
$(0.07)

Discount rate and statistic

3%	2.5%

Average

$(0.14)
$(0.15)
$(0.15)
$(0.15)
$(0.16)
$(0.16)
$(0.16)
$(0.17)
$(0.17)
$(0.17)
$(0.18)
$(0.18)
$(0.18)
$(0.19)
$(0.19)

Average

$(0.21)
$(0.21)
$(0.22)
$(0.22)
$(0.23)
$(0.23)
$(0.23)
$(0.24)
$(0.24)
$(0.25)
$(0.25)
$(0.26)
$(0.26)
$(0.26)
$(0.27)

3% 95'"
Percentile

$(0.38)
$(0.39)
$(0.40)
$(0.41)
$(0.41)
$(0.42)
$(0.43)
$(0.44)
$(0.45)
$(0.46)
$(0.47)
$(0.48)
$(0.49)
$(0.50)
$(0.51)

NPV
EAV

($149)
$(15)

($558)
$(48)

($842)
$(70)

($1,690)
$(146)

$225
$23

$552
$48

$738
$61

$1,469
$127

($1)

($0.1)

($2)
($0.2)

($3)
($0.2)

($5)
($0.4)

113

-------
114

-------
5 ECONOMIC IMPACT ANALYSIS

5.1 Introduction

The final amendments to the NESHAP for the HON constitute a significant action
according to Executive Order 14094. As discussed in the previous section, the emissions
reductions estimated under the action are projected to produce substantial VOC health benefits.
At the same time, these final HON amendments are projected to result in environmental control
expenditures by the SOCMI to comply with the rule. The final amendments to the NESHAPs for
P&R Group I and II, and their respective NSPS subparts (III, NNN, RRR, & VV) are not
projected to be significant individually, but they are expected to result in VOC health benefits
and increased environmental control expenditures.

Economic impact analyses focus on changes in market prices and output levels. If
changes in market prices and output levels in the primary markets are significant enough,
impacts on other markets may also be examined. Both the magnitude of costs needed to comply
with a rule and the distribution of these costs among affected facilities can have a role in
determining how the market will change in response to a rule. This chapter analyzes three sets of
economic impact, small entity, and distributional analyses for each individual rule included in
this final action. These analyses are directed toward complementing the benefit-cost analysis and
include a partial equilibrium analysis of market impacts of three sets of NESHAP amendments
from this rule package, analysis of impacts to potentially affected small entities, and employment
impacts.

5.2 Economic Impact Analysis

This section summarizes the economic analysis of environmental control costs for the
SOCMI. This analysis models the impact of two sets of control costs for three different final
NESHAP amendments for the HON and P&R Group I and II, specifically. The analysis does not
include economic impacts calculated for four final New Source Performance Standards (NSPS) -
subparts III, NNN, RRR, and VV - that are part of the same rulemaking in which the three
different NESHAP are also included.

This section outlines the data and sources used to calibrate and parameterize a simplified
partial equilibrium model representing elastic domestic and foreign sources of supply and a

115

-------
single domestic consumer with elastic preferences. The model uses different cost shocks,
including domestic compliance costs, foreign inflation, and combinations.

Economic analysis was conducted for seven synthetic organic chemicals in the SOCMI
list. These chemicals were selected for their relative market size and the availability of data to
conduct the economic analysis. The seven chemicals for which market analysis was conducted
are butadiene, styrene, acrylonitrile, acetone, ethylene dichloride, ethylene glycol, and ethylene
oxide.

5.3 Description of Approach, Model, and Framework

5.3.1 Data Limitations

There were several limitations to data inputs for the economic modeling, including the
availability of production data, the allocation of the control costs from facilities to individual
chemical markets, and the ability to find specific market dynamics (elasticity) data. For
production data, the primary source for most of the chemicals in this analysis included the
American Fuel & Petrochemical Manufacturers (AFPM) Petrochemical Statistics dataset
published in the first quarter of 2022.

For cost data, control costs were calculated by facility, but little detail was initially
available on facility-level production, such as chemical or quantities produced. To allocate the
control costs to specific chemicals, data was sourced from Securities and Exchange Commission
(SEC) filings, company websites, and specific industry reports to identify chemicals produced at
each facility to support the method for cost allocation from facilities to individual chemical
markets for the partial equilibrium analysis.

For market supply and demand elasticities, no sources of previous SOCMI economic
modeling were found, so elasticities were assigned to chemicals based on markets associated
with end use products for the chemicals (such as different types of plastics or PVC piping).

These limitations are discussed more in detail in the following two subsections.

5.3.2 Benchmark Data

There are approximately 400 synthetic organic chemicals, nearly all of which are
contained within a single 6-digit North American Industry Classification System (NAICS) code

116

-------
(325199). A detailed description of the approximately 25 largest SOCMI markets is contained in
the SOCMI Industry Profile prepared for this final action.38 The main limitation of collecting
data for the industry profile was a lack of domestic production for each of the synthetic organic
chemicals.

Trade data do, however, provide more granular information on products via harmonized
system (HS) commodity codes. For the economic analysis, we selected seven of the largest
SOCMI sectors, drawing domestic production quantities from the AFPM Petrochemical
Statistics dataset, prices from Intratec commodities data (Intratec, 2023), and trade quantities
from the United Nations "Comtrade" dataset (United Nations, 2022). Table 5-1 summarizes the
physical quantities and prices for each of the seven chemicals included in the analysis. All
physical quantities here are in tonnes (Metric Tons), which are 10 percent larger by mass than
English tons.

Table 5-1 Prices, Production, and Trade Quantities for the Seven Synthetic Organic

Chemical Commodities Selected (in Metric Tons)

Chemical

U.S. Production
(tonnes)

Exports
(tonnes)

Imports
(tonnes)

Price
($/tonne)

Butadiene

1,218,232

46,261

391,496

1,220

Styrene

3,659,415

1,099,780

211,776

1,841

Acrylonitrile

848,390

529,330

9,003

1,040

Acetone

1,514,800

42,229

84,459

1,839

Ethylene dichloride

9,731,200

727,454

475

723

Ethylene glycol

1,578,142

1,269,166

236,920

1,220

Ethylene oxide

2,400,027

928

1,486

5.3.3 Control Data

Control cost data is available for 225 facilities subject to HON or P&R I, not including
the value of product recovery, which occurs in the HON and P&R I cost analyses due to
detection and repair of equipment leaks. Inclusion of the value of product recovery would lead to
double counting of impacts from a social welfare perspective. More information on how product
recovery is estimated and monetized can be found in Section 3.3 of this RIA. The control cost
data did not include the production processes, or a detailed accounting of chemicals produced at

38 RTI International. SOCMI Industry Profile. Prepared for the U.S. EPA. July 2022.

117

-------
the affected facilities. To develop a method to allocate the control costs to specific chemicals, we
reviewed annual SEC filings (for most companies, the 2021 10-k Report) and facility websites
for all facilities to identify the chemicals produced at the facility. We found general information
about the types of chemicals produced at 116 (51 percent) of those facilities and more detailed
production data by chemical names for 79 (36 percent) of the total population of facilities.

For the 79 facilities with the best {i.e., most complete) chemical production data, we used
this information to identify which of the facilities were producing butadiene, styrene,
acrylonitrile, and acetone. We allocated the control costs to these chemicals at the 79 facilities by
equally distributing the control costs based on the total number of chemicals identified in
production at that facility. The control costs for each chemical were then scaled up to the entire
population based on the percentage of total control costs represented in the population of 79
facilities with specific chemical data.

For ethylene dichloride, ethylene glycol, and ethylene oxide, we reviewed publicly
available reports to identify facilities that produced each specific chemical and then allocated a
portion of the facility's HON costs to their production. The P&R costs apply only to polymers
and resin products, so only HON costs were relevant for these three chemicals. For ethylene
dichloride, a technical report on the conditions of use for ethylene dichloride listed the 15
facilities producing ethylene dichloride in the United States in 2018 (Material Research L3C,
2019). We allocated half of the HON control costs at each of these facilities to ethylene
dichloride because most of these facilities are PVC production facilities, and ethylene dichloride
is one of the two major chemical inputs, but there was not enough data to determine how much
of the HON costs would be applied to the other chemicals produced.

The ethylene oxide facilities were identified from a Bloomberg Law article (Saiyid,
2019) that listed the top 16 sources of ethylene oxide emissions, 10 of which were production
facilities and the other 6 were medical sterilization facilities. Facilities affected by this rule
include ethylene oxide production facilities and not medical sterilization facilities. Sterilization
facilities have also not been found to be co-located with any of the production facilities. The
American Chemistry Society (2023) stated that at the end of 2018, the United States had 15
ethylene oxide facilities. The ethylene glycol facilities were identified from a toxicology profile
from the Department of Health and Human Services (DHHS) (2010). The list of ethylene glycol
facilities was from 2008, so it may not include all current production facilities. The DHHS report

118

-------
noted that nine ethylene oxide production facilities were included in the HON control cost data,
six of which also produced ethylene oxide. These facilities did not generally provide detailed
production data by chemical; for ethylene glycol and ethylene oxide, half of the HON costs for
each facility were allocated to each chemical. Because only 10 of the 15 ethylene glycol
production facilities were identified, the HON costs were multiplied by 1.5 to represent the entire
population of facilities. Table 5-2 details the control costs for each chemical used in the model.
Three of the seven chemicals are found to have total control costs of more than one percent of
total domestic production value, though none reach two percent or higher.

Table 5-2 Control Costs Attributed to Each Chemical Modeled (2021$)

Chemical

HON Control
Cost (USD/yr)

P&R Control
Costs (USD/yr)

Total Control
Costs (USD/yr)

Domestic
Production Value

(USD/yr)

Total Control

Cost % of
Domestic Value

Acrylonitrile

13,370,931

886,272

14,257,202

880,400,521

1.619%

Acetone

4,816,333

1,965,581

6,781,915

2,779,501,547

0.244%

Butadiene

4,921,906

3,895,051

8,816,958

1,483,128,462

0.594%

Ethylene
dichloride

8,684,650

0

8,684,650

7,022,564,654

0.124%

Ethylene
glycol

22,905,950

0

22,905,950

1,633,103,944

1.403%

Ethylene
oxide

6,441,600

0

36,441,600

3,559,565,671

1.024%

Styrene

11,487,402

4,794,858

16,282,260

6,722,358,387

0.242%

All 255

fdpilitip«

163,572,000

16,514,700

180,086,700





5.3.3.1 Synthetic Organic Chemicals Manufacturing Industries (SOCMI) Model

For the analysis, the EPA developed a simplified partial equilibrium model that can be
calibrated to the benchmark data above. The model represents elastic domestic and foreign
production and consumption (four elasticities) and a domestic consumer foreign-domestic
substitution elasticity (one elasticity). Elasticity estimates are scarce for the specific chemicals.
We identified several key elasticity values to populate our five elasticity parameters. For supply
elasticities, we used a value of 0.54 from Chambers and Lichtenberg (1994) econometric
estimation of long-run fertilizer supply for all chemicals except acetone. For acetone, we used a
supply elasticity value of 0 because it is a byproduct of the production of phenol, a much higher
value product.

For demand elasticities, we used values associated with a predominant end use industry
for the product, if possible. For styrene, butadiene, ethylene glycol, ethylene dichloride, and

119

-------
ethylene oxide, we used a value of-0.38 from Trangadisaikul's (2011) econometric estimation
of global tire demand, a proxy market for rubber production.

Ethylene oxide is most commonly used for sterilizing medical equipment, and ethylene
glycol and ethylene dichloride are more commonly used in plastic production, but demand
elasticity data for those markets were not found in our research, so we used the secondary market
of a rubber production input. For acrylonitrile and acetone, we used a value of-1.04 from
Martinez's (2012) estimation of demand elasticity for human-made fabrics in the textile industry
because textile fiber production is a common use for those two chemicals. Last, we took a
foreign-domestic consumer demand substitution elasticity estimate of-2.4 from Ahmad and
Riker (2019) in order to account for substitution in consumer demand between domestic and
foreign production of these chemicals. Our current estimates are summarized in Table 5-3.

Elasticity

Symbol

Value

Source

Supply [domestic (y), imports (m)]

bS

II

b*

0.54

Chambers and Lichtenberg (1994)

Demand [domestic (d). foreign (/)]

adD = afD
adD = afD

-0.38
-1.04

Trangadisaikul's (2011)
Martinez (2012)

Consumer substitution

adf

-2.4

Ahmad and Riker (2019)

The model is a modified version of that specified in Riker and Schreiber (Riker, 2019) in
combination with the calibrated share form of the constant elasticity of substitution (CES)
function detailed in Rutherford (2002). However, the original distinction between different
import sources is removed since our model is only intended to cover domestic compliance cost
shocks. Elastic domestic and foreign production is specified as:

- '®f

- (Pm\am	(2)

M = m(^)	V 7

UV

where Y is the value of domestic production, M is the value of imports, P is the corresponding
price, and bars ( ) denote the benchmark value of a variable. Domestic and foreign demand are
specified similarly to production as:

" = ®(jf	<3)

X= X\^-
Sy,

„ (Pv\GfD	(4)

120

-------
where D is total domestic demand and X is total export demand. We specified total domestic
consumption as an aggregate of domestic- and foreign-produced goods using the calibrated share

dD = 0D I —

PXC	(5)

Py,

Pc \°c	(6)

Pm'

fD = ( 1-0)dM-)

\pm/

form of the CES function as:

where dD is domestic consumer demand for domestic goods, and fD is domestic consumer
demand for foreign goods, and:

_ py dD	(7)

py D+ pmjD

The final conditions for the model require market clearance (i.e., that supply equal demand in the

domestic and foreign markets for the chemical). We specified this requirement as:

0 = Y - X - dD	(9)

0 = M-fD	(10)

These nine equations (excluding the 6 parameter definition in equation 9) form the basis

of our model with the six quantity (Y, M, D, X, dD, fD) and three price (py, pm, pc) variables,

which makes a square system of equations that we can implement in a constrained nonlinear

system (CNS) mathematical program in the GAMS software language using a constrained

optimization solver.

5.3.3.2 SOCMI Model Simulations and Results

For each of the seven chemicals in this analysis, we implemented five counterfactual
shocks to the model to simulate new market outcomes. To implement counterfactual cases
including increased production costs from regulatory compliance or inflation, we included cost
shock parameters in equations (11) and (12) as follows:

K=p(—M*	(11>

\(1 + Cy)Pyj

121

-------
M = M

(1 cf^Pm

(12)

where the cost parameters, cy and cy, are expressed in percentage terms and, when positive,
reduce the effective prices received by suppliers.

Our analysis of the economic impact of costs includes three scenarios:

1) compliance costs due to the HON rule,

2) the compliance costs due to the two P&R rules,

3) the total compliance costs due to the three rules.

We also modeled two foreign market inflation scenarios to investigate the
impact/interactions of an increased price of natural gas in foreign producing countries:

1) inflation caused by rises in foreign natural gas (NG) prices only—primarily energy

inputs. This is referred to as foreign low inflation (INFLO).

2) inflation caused by rises in foreign NG and natural gas liquid (NGL) prices associated

with the product inputs. This is referred to as foreign inflation high (INF HI).

Note: We assumedforeign gas prices do not affect domestic production costs.

For the inflation caused by NG and NGL price increases, we applied the average annual
spot price increase in German NG prices from 2018 to 2021 ($8.57/mmBtu and $15.91 /mmBtu,
respectively39) to the price increases due to NG and NGL price changes from the ACS study on
the impact of NG and NGL prices on the U.S. chemical manufacturing industry (DeRosa, 2015).
The chemicals with benzene as a feedstock—butadiene, styrene, and acetone—do not have any
production cost increase associated with increased NG or NGL prices because benzene is a
production by-product of other higher value products and does not have a cost change due to NG
or NGL prices.

For each of the seven chemicals, we ran a business-as-usual (BAU) benchmark
replication {i.e., essentially a baseline model run) and the following five model runs:

39 https://Ycharts.com/indicators/germanv natural gas border price

122

-------
• BAU: cy = 0% and cy = 0% (no compliance costs and no foreign inflation)

• HON: cyl = HON compliance costs only

• PR: cy2 = P&R compliance costs only

• CC TOT : cyl + cy2 = HON and P&R total compliance.

• INF: cy = Foreign inflation costs only (no compliance costs)

• CC+INF: cyl + cy2 + Cf = HON and P&R compliance costs and foreign inflation costs

Model results for each of the seven chemicals included in the analysis are presented in
Tables 5-4 through Table 5-10. The simulated market impacts from the 35 model runs using
compliance costs as an input are consistent with our expectations in that compliance costs result
in higher market prices and lower output and foreign inflation leads to domestic output
percentage increases and dampens the impact of the regulation's compliance costs.

• Butadiene shows modest impacts of domestic production, decreasing about 0.20 percent.
Because butadiene is a benzene by-product, there is no impact on foreign production
costs due to NG or NGL price changes.

• Styrene and ethylene dichloride see the smallest impacts on domestic production,
decreasing less than 0.05 percent.

• Acrylonitrile sees the largest output drop in response to compliance costs (0.57 percent)
because of its smaller market size and relatively higher P&R compliance costs.
Acrylonitrile also sees a high percentage drop in imports from higher foreign NG prices
because of its high sensitivity to NGL prices.

• Acetone is the only product that does not see (at our significance level) a drop in
production or an increase in price due to the compliance costs because of its very low
relative compliance cost increase (only 0.17 percent of production costs).

• Ethylene glycol has a production decrease of about 0.5 percent with compliance costs.
However, it sees a net increase in production of about 0.25 percent with the foreign NG
prices increase.

123

-------
• Ethylene oxide faces the highest compliance costs for chemical products affected by this
final action due to the expected use of large add-on control technologies such as thermal
oxidizers for compliance with the HON and P&R I requirements. It only sees a modest
decrease in domestic production, however, because of its large domestic market and very
low imports. There have been almost no ethylene oxide imports to the United States in
the past 5 years.

Table 5-4 Butadiene Results



BAU

HON

PR CCTOT

INF

CC+INF

Quantities











Output

1,218

1,217

1,217 1,216

1,218

1,216





-0.11%

-0.09% -0.20%

0.00%

-0.20%

Exports

46

46

46 46

46

46





-0.05%

-0.04% -0.08%

0.00%

-0.08%

Imports

391

392

392 392

391

392





0.12%

0.09% 0.21%

0.00%

0.21%

Demand

1,563

1,563

1,563 1,562

1,563

1,562





-0.06%

-0.04% -0.10%

0.00%

-0.10%

Prices ($000/tonne)











Domestic

1.22

1.22

1.22 1.22

1.22

1.22





0.16%

0.08% 0.25%

0.00%

0.25%

Consumption

1.22

1.22

1.22 1.22

1.22

1.22





0.16%

0.08% 0.25%

0.00%

0.25%

Import

1.22

1.22

1.22 1.23

1.22

1.23





0.25%

0.16% 0.41%

0.00%

0.41%



Table 5-5

Styrene Simulation Results







BAU

HON

PR CCTOT

INF

CC+INF

Quantities











Output

3,659

3,658

3,659 3,657

3,659

3,657





-0.04%

-0.02% -0.06%

0.00%

-0.06%

Exports

1,100

1,099

1,100 1,099

1,100

1,099





-0.04%

-0.01% -0.05%

0.00%

-0.05%

Imports

212

212

212 212

212

212





0.08%

0.03% 0.11%

0.00%

0.11%

Demand

2,771

2,770

2,771 2,770

2,771

2,770





-0.04%

-0.02% -0.05%

0.00%

-0.05%

Prices ($000/tonne)











Domestic

1.84

1.84

1.84 1.84

1.84

1.84





0.11%

0.05% 0.11%

0.00%

0.11%

Consumption

1.84

1.84

1.84 1.84

1.84

1.84





0.11%

0.05% 0.11%

0.00%

0.11%

Import

1.84

1.84

1.84 1.85

1.84

1.85





0.16%

0.05% 0.22%

0.00%

0.22%

124

-------
Table 5-6

Acrylonitrile Simulation Results



BAU

HON

PR

CCTOT

INF

CC+INF

Quantities













Output

848

844

848

844

850

845





-0.54%

-0.04%

-0.57%

0.18%

-0.39%

Exports

529

527

529

526

528

525





-0.52%

-0.03%

-0.55%

-0.34%

-0.90%

Imports

9

9

9

9

6

6





0.52%

0.03%

0.57%

-38.34%

-38.00%

Demand

328

326

328

326

328

326





-0.53%

-0.04%

-0.57%

0.07%

-0.50%

Prices ($000/tonne)













Domestic

1.04

1.05

1.04

1.05

1.04

1.05





0.48%

0.00%

0.58%

0.29%

0.87%

Consumption

1.04

1.05

1.04

1.05

1.04

1.05





0.48%

0.00%

0.58%

-0.10%

0.48%

Import

1.04

1.05

1.04

1.05

0.85

0.86





0.96%

0.10%

1.06%

-18.37%

-17.50%

	Table 5-7 Acetone Simulation Results	

BAU HON	PR CCTOT CC+INF_LO CC+INF_HI

Quantities

Output

1,515

1,515

1,515

1,515

1,515

1,515





0.00%

0.00%

0.00%

0.00%

0.00%

Exports

42

42

42

42

42

42





0.00%

0.00%

0.00%

0.00%

0.00%

Imports

84

84

84

84

84

84





0.00%

0.00%

0.00%

0.00%

0.00%

Demand

1,557

1,557

1,557

1,557

1,557

1,557





0.00%

0.00%

0.00%

0.00%

0.00%

Prices ($000/tonne)













Domestic

1.84

1.84

1.84

1.84

1.84

1.84





0.00%

0.00%

0.00%

0.00%

0.00%

Consumption

1.84

1.84

1.84

1.84

1.84

1.84





0.00%

0.00%

0.00%

0.00%

0.00%

Import

1.84

1.84

1.84

1.84

1.84

1.84





0.00%

0.00%

0.00%

0.00%

0.00%

125

-------
Table 5-8 Ethylene Dichloride Simulation Results



BAU

HON

PR

CC TOT

INF

CC+INF

Quantities













Output

9,731

9,729

9,731

9,729

9,731

9,729





-0.03%

0.00%

-0.03%

0.00%

-0.03%

Exports

727

727

727

727

727

727





-0.03%

0.00%

-0.03%

0.00%

-0.03%

Imports

0

0

0

0

0

0





0.21%

0.00%

0.21%

-8.00%

-8.00%

Demand

9,004

9,002

9,004

9,002

9,004

9,002





-0.03%

0.00%

-0.03%

0.00%

-0.03%

ices ($000/tonne)













Domestic

0.72

0.72

0.72

0.72

0.72

0.72





0.14%

0.00%

0.14%

0.00%

0.14%

Consumption

0.72

0.72

0.72

0.72

0.72

0.72





0.14%

0.00%

0.14%

0.00%

0.14%

Import

0.72

0.72

0.72

0.72

0.70

0.70





0.14%

0.00%

0.14%

-3.46%

-3.32%

	Table 5-9 Ethylene Glycol Simulation Results	

BAU HON	PR CCTOT	INF	CC+INF

Quantities

Output

1,578

1,571

1,578

1,571

1,589

1,582





-0.43%

0.00%

-0.43%

0.67%

0.26%

Exports

1,269

1,266

1,269

1,266

1,263

1,260





-0.23%

0.00%

-0.23%

-0.47%

-0.71%

Imports

237

239

237

239

222

223





0.79%

0.00%

0.79%

-6.44%

-5.73%

Demand

546

544

546

544

548

546





-0.37%

0.00%

-0.37%

0.31%

-0.05%

Prices ($000/tonne)













Domestic

1.04

1.04

1.04

1.04

1.05

1.06





0.58%

0.00%

0.58%

1.25%

1.93%

Consumption

1.04

1.05

1.04

1.05

1.03

1.04





0.96%

0.00%

0.96%

-0.77%

0.10%

Import

1.04

1.05

1.04

1.05

1.00

1.01





1.45%

0.00%

1.45%

-3.66%

-2.31%

126

-------
Table 5-10 Ethylene Oxide Simulation Results



BAU

HON

PR

CCTOT

INF

CC+INF

Quantities













Output

2,400

2,395

2,400

2,395

2,400

2,395





-0.23%

0.00%

-0.23%

0.00%

-0.23%

Exports

1

1

1

1

1

1





-0.22%

0.00%

-0.22%

0.00%

-0.22%

Imports

0

0

0

0

0

0





0.00%

0.00%

0.00%

0.00%

0.00%

Demand

2,399

2,394

2,399

2,394

2,399

2,394





-0.23%

0.00%

-0.23%

0.00%

-0.23%

Prices ($000/tonne)













Domestic

1.49

1.50

1.49

1.50

1.49

1.50





0.61%

0.00%

0.61%

0.00%

0.61%

Consumption

1.49

1.50

1.49

1.50

1.49

1.50





0.61%

0.00%

0.61%

0.00%

0.61%

Import

1.49

1.50

1.49

1.50

1.46

1.47





0.94%

0.00%

0.94%

-1.75%

-0.87%

5.4 Small Business Impacts Analysis

For the final rule, the EPA performed a small entity screening analysis for impacts on all
affected facilities by comparing compliance costs to historic revenues at the ultimate parent
company level. This is known as the cost-to-revenue or cost-to-sales test, or the "sales test." The
sales test is an impact methodology the EPA employs in analyzing entity impacts as opposed to a
"profits test," in which annualized compliance costs are calculated as a share of profits. The sales
test is frequently used because revenues or sales data are commonly available for entities
impacted by the EPA regulations, and profits data normally made available are often not the true
profit earned by firms because of accounting and tax considerations. Also, the use of a sales test
for estimating small business impacts for a rulemaking is consistent with guidance offered by the
EPA on compliance with the Regulatory Flexibility Act (RFA)40 and is consistent with guidance
published by the U.S. Small Business Administration's (SB A) Office of Advocacy that suggests
that cost as a percentage of total revenues is a metric for evaluating cost increases on small
entities in relation to increases on large entities (SB A, 2017).

For purposes of assessing the impacts of this action on small entities, a small entity is
defined as: (1) a small business as defined by the Small Business Administration's (SBA)

40 The RFA compliance guidance to the EPA rule writers can be found at



127

-------
regulations at 13 CFR 121.201; (2) a small governmental jurisdiction that is a government of a
city, county, town, school district or special district with a population of less than 50,000; and (3)
a small organization that is any not-for-profit enterprise that is independently owned and
operated and is not dominant in its field. Businesses in the Gasoline Distribution source category
predominately have NAICS codes 325199 (All Other Basic Organic Chemical Manufacturing).
For the SBA small business size standard definition for each NAICS classification, see below in
Table 5-11.

	Table 5-11 SBA Size Standards by NAICS Code	





Size

Size
Standards
(Number of
employees)

NAICS
Codes



Standards

NAICS U.S. Industry Title

(million$ of
annual

sales/revenues)

325110

Petrochemical Manufacturing



1,300

325120

Industrial Gas Manufacturing



1,200

325130

Synthetic Dye and Pigment Manufacturing



1,050

325180

Other Basic Inorganic Chemical Manufacturing
Cyclic Crude, Intermediate, and Gum and Wood Chemical



1,000

325194

Manufacturing



1,250

325199

All Other Basic Organic Chemical Manufacturing



1,250

325211

Plastics Material and Resin Manufacturing



1,250

325212

Synthetic Rubber Manufacturing



1,000

325220

Artificial and Synthetic Fibers and Filaments Manufacturing



1,050

325311

Nitrogenous Fertilizer Manufacturing



1,050

325320

Pesticide and Other Agricultural Chemical Manufacturing



1,150

325412

Pharmaceutical Preparation Manufacturing



1,300

325620

Toilet Preparation Manufacturing



1,250

325920

Explosives Manufacturing

All Other Miscellaneous Chemical Product and Preparation



750

325998

Manufacturing



650

The EPA constructed a facility list for the HON and P&R Group I and II source
categories. For information on how this list was constructed, see Section 2. The initial facility
lists consisted of 207 HON facilities, 19 P&R I facilities (and ten of the P&R I facilities are
collocated with HON processes), and five P&R II facilities (and three of the P&R II facilities are
collocated with HON processes). However, revised counts of active and unique facilities reduced
the size of these lists. The EPA identified the ultimate parent company along with revenue and
employment information for facilities using D&B Hoover's database. In total, the EPA identified
98 ultimate parent companies as owners of the 214 facilities, of which ten of these ultimate
parent companies were identified as small entities (counts of parent companies do not sum over

128

-------
rules due to some companies owning facilities subject to multiple rules). These companies,
including the small entities, operate in the SOCMI industry, which is marginally competitive as a
whole as mentioned in Chapter 2 of this RIA. Summary statistics for these ultimate parent
companies are in Table 5-12 below.

Table 5-12 Summary Statistics of Potentially Affected Entities

Rule

Size

No. of Ultimate Parent
Companies

Number of
Facilities

Mean Revenue
(million 2021$)

Median
Revenue
(million 2021$)



Small

8

9

299

97.3

HON













Not Small

88

194

22,500

4,900



Small

1

2

290

290

P&R I













Not Small

11

16

40,900

8,940



Small

0

0

_

_

P&R II













Not Small

4

5

78,900

22,900

Rules Combined

Small
Not Small

9

89

10
204

273
22,500

62.2
4,970

Note: Some facilities are affected by more than one rule and therefore, to avoid double counting, "Rules Combined"
will not equal the sum of facilities noted in individual rules.

5.5 Screening Analysis

Using the facility list discussed in the above section, the EPA conducted cost-to-sales
analysis for the final action to screen small entities for potentially significant impacts. We
present results specifically for each of the HON, P&R I and P&R II rules, and a total estimate for
all of these three final rules. We are unable to provide an estimate of small entity impacts for the
NSPS in this final action due to an inability to link impacts to specific known facilities and
ultimate parent owners. While a sales test can provide some insight as to the economic impact of
an action such as this one, it assumes that the impacts of a rule are solely incident on a directly
affected firm (therefore, no impact to consumers of the affected product), or solely incident on
consumers of output directly affected by this action (therefore, no impact to companies that are
producers of the affected product). Thus, an analysis such as this one is best viewed as providing
insight on the polar opposites of economic impacts: maximum impact to either directly affected
companies with no impact on their consumers, or vice versa. A sales test analysis does not
consider shifts in supply and demand curves to reflect intermediate economic outcomes. For a

129

-------
partial equilibrium analysis of the economic impacts of this action that attempts to parse impacts
on consumers relative to producers, see section 5.2.

The results of this analysis for the final options are presented below. Table 5-13 shows
the distribution of average costs for ultimate parent companies by final rule. Tables-5-14 and 5-
15 below show the distribution of cost-to-sales ratios (CSRs) by rule and the percentage of CSRs
clearing 1 percent and 3 percent for each rule. We present the results both with costs including
product recovery and without product recovery. The results are virtually identical regardless of
whether or not product recovery is included.

Table 5-13 Distribution of Estimated Compliance Costs by Rule and Size for Final
Action ($2021)a

Rule

Size

No. of Firms

Average Cost with
Product Recovery

Average Cost without
Product Recovery

HON

Small

309,000

313,000

Not Small

852,000

857,000

P&RI

Small

43,900

Not Small

944,000

946,000

P&RII

Small

Not Small

333,000

Small

Rules Combined

Not Small

291,000
893,000

295,000
897,000

a There are some firms, including one small firm, that are impacted by more than one final rule. This explains why

the totals of combined impacted firms are less than the straight summation across the final rules.

Table 5-14

Compliance Cost-to-Sales Ratio Distributions for Small Entities, Final Action

With Product Recovery

Without Product

Rule

Included

Recovery Included

Mean Maximum
CSR CSR

HON

0.516% 1.26%

0.554% 1.40%

P&RI

No. of Small Entities

0.030% 0.030%

P&RII

All

No. of Small Entities

0.469% 1.26%

0.504% 1.40%

a There is one small firm that is impacted by more than one final rule. This explains why the totals of combined
impacted firms are less than the straight summation across the final rules.

130

-------
Table 5-15 Number and Extent of Impacts for Small Entities - Final Action3

, „ , . „ T i j j Without Product Recovery

With Product Recovery Included T i j j

Kuie

No. of Small

% of Small

No. of Small

% of Small

Entities

No. of Small Entities

100%

HON

Greater than 1%

25%

Greater than 3%

0.0%

No. of Small Entities

100%

P&R I

Greater than 1%

0.0%

Greater than 3%

0.0%

No. of Small Entities

P&R II

Greater than 1%

Greater than 3%

No. of Small Entities

100%

All

Greater than 1%

22%

Greater than 3%

0.0%

a There is one small firm, that is impacted by more than one final rule. This explains why the totals of combined
impacted firms are less than the straight summation across the final rules.

Given the relatively low average CSR for small entities (both with and without product
recovery), as well as there being only two small entities with a CSR of at least one percent and
no small entities with a CSR of at least three percent for the final HON amendments, we
conclude that it is unlikely that the changes to the HON would have a significant impact on a
substantial number of small entities (SISNOSE), and therefore we certify that there is no
SISNOSE for this final rule. Given that there are no small entities with a CSR of at least one
percent for either the P&R I or P&R IINESHAP, we conclude that we can certify no SISNOSE
for either of these final rules.

5.6 Employment Impact Analysis

This section presents a qualitative overview of the various ways that environmental
regulation can affect employment. Employment impacts of environmental regulations are
generally composed of a mix of potential declines and gains in different areas of the economy
over time. Regulatory employment impacts can vary across occupations, regions, and industries;
by labor and product demand and supply elasticities; and in response to other labor market
conditions. Isolating such impacts is a challenge, as they are difficult to disentangle from
employment impacts caused by a wide variety of ongoing, concurrent economic changes. The
EPA continues to explore the relevant theoretical and empirical literature and to seek public

131

-------
comments in order to ensure that the way the EPA characterizes the employment effects of its
regulations is reasonable and informative.41

Employment impacts of environmental regulations are generally composed of a mix of
potential declines and gains in different areas of the economy over time. Regulatory employment
impacts can vary across occupations, regions, and industries; by labor and product demand and
supply elasticities; and in response to other labor market conditions. Isolating such impacts is a
challenge, as they are difficult to disentangle from employment impacts caused by a wide variety
of ongoing, concurrent economic changes including macroeconomic phenomena such as the
Covid 19 pandemic or general inflation.

In the long run, environmental regulation "typically affects the distribution of
employment among industries rather than the general employment level" (Arrow, et al., 1996;
Hafstead and Williams, 2020). The expectation is that there will be a movement of labor towards
jobs that are associated with greater environmental protection, and away from those that are not.
Even if impacts are small after long-run market adjustments to full employment, many regulatory
actions move workers in and out of jobs and industries, which are potentially important
distributional impacts of environmental regulations (Walker 2013). The consequences of
transitional job losses may be particularly important for workers who might not readily find
alternative work. This might include workers with skills that do not transfer easily to other
workplaces, or who operate in declining industries or occupations, have limited capacity to
migrate, or reside in communities or regions with high unemployment rates.

This section discusses the anticipated employment impacts of the final rule, focusing on
the impact of the SOCMI requirements. To the extent possible, it describes the baseline
characteristics of affected labor markets in regulated industries, including trends in employment
numbers and labor intensities. It also describes the rule's potential incremental impacts on
employment in those same industries. As previously mentioned, nearly all synthetic organic
chemicals are contained within the single 6-digit North American Industry Classification System
(NAICS) code "All other basic organic chemical manufacturing" (NAICS 325199). Table 5-16
summarizes baseline employment characteristics in All other basic organic chemical

41 The employment analysis in this RIA is part of EPA's ongoing effort to "conduct continuing evaluations of

potential loss or shifts of employment which may result from the administration or enforcement of [the Act]"
pursuant to CAA section 321(a).

132

-------
manufacturing (NAICS 325199) and in its 4-digit NAICS parent, the Basic chemical
manufacturing industry (NAICS 3251) using data from the Bureau of Labor Statistics (BLS) and
the U.S. Census' Annual Survey of Manufactures (ASM). Table 5-16 shows that across
occupations, average wages in the Basic chemical manufacturing industry and in All other basic
organic chemical manufacturing are high, and employment has been increasing since 2017,
particularly in NAICS 325199. According to the ASM, the ratio of employees to the value of
output in the Basic chemical manufacturing was just 0.576 employees per million dollars of
output in 2021 (the most recent year of data available from the ASM), while the ratio of
employees to value of output in NAICS 325199 was 0.771 per million dollars of output.

Table 5-16 Chemical Sector Employment Information

Industry

NAICS

Employment
in 2022
(1000s)

5-Year
Change in
Employment
(percent)

Avg.
Annual
Wage in
2022
(1000s)

Employment/Output
in 2021

Basic chemical
manufacturing

3251

151.4

1.65

117

0.576

All other basic
organic chemical
manufacturing

325199

41.8

10.33

107.7

0.771

Notes: NAICS is North American Industry Classification System. The data source for all columns except the final
column is the Bureau of Labor Statistics Quarterly Census of Earnings and Wages (BLS QCEW), downloaded
from https://www.bls.gov/cew/downloadable-data-files.htm. The data source for the final column is the U.S.

Census Annual Survey of Manufactures, downloaded from the FTP site at https://www2.census.gov/programs-
surveys/asm/data.

The figures in Table 5-16 can be combined with EPA's economic impact analysis in
Section 5.2 to better understand the potential employment impacts of the rule. Table 5-1 displays
the estimated annual value of output for seven of the largest SOCMI sectors, which totaled over
$24 billion in 2021. To estimate total employment in these sectors, we multiply this total value
by the employment-to-output ratio of 0.771 in Table 5-16, effectively assuming that SOCMI
production has the same average employment intensity as the other basic organic chemical
manufacturing industry more broadly. This results in an estimate of approximately 18,566
individuals employed in the SOCMI sectors analyzed in Section 5.2, or about 44% of
employment in NAICS industry 325199. EPA's economic impact analysis also estimated
percentage output and price changes for each SOCMI sector resulting from the combination of

133

-------
HON and P&R compliance, presented in Table 5-4 through Table 5-10. EPA's modelled changes
in output and prices suggest that the total value of production across all sectors may decline on
the order of 1%, which suggests a reduction of approximately 186 employees across seven of the
largest SOCMI sectors, as modelled.

It is important to caveat that the estimates presented above are unable to capture certain
aspects of firm responses to the final rule. By applying a baseline average employment-to-
production intensity, the analysis implicitly holds production technologies fixed even as output
changes in the regulated industry. It also doesn't account for heterogeneity among regulated
firms, which may lead individual firms to respond in a variety of ways to the control
requirements. Lastly, the control costs presented in Table 5-2 may require increased labor
(suggesting further incremental employment gains) or may result in a shift toward more capital-
intensive production processes (suggesting potential incremental employment losses).

EPA also acknowledges that employment impacts, both positive and negative, are
possible in indirectly affected sectors upstream and downstream from the regulated sector, or in
sectors producing substitute or complementary products. This might include gains at upstream
facilities that manufacture the equipment necessary for pollution control or conversions to
alternative technologies, as well as losses in employment in oil and gas production due to
reduced demand from SOCMI producers. Similarly, output price increases may lead to
downstream adjustments in employment levels by plastics manufacturers or other industries that
demand SOCs. Finally, EPA acknowledges that to the extent domestic SOC production is
displaced by imports, this may result in some net employment loss in the short run (incorporated
in the calculations above), as well as some potential shifts in domestic employment over the
longer run.

134

-------
6 Comparison of Costs and Benefits

In this chapter, we present a comparison of the benefits and costs of this final action. We
present benefits and costs for each final rule and their more and less stringent alternatives. We
group the impacts of the Ilia, NNNa, and RRRa NSPS together for presentational clarity and
consistency with the presentation of impacts for these three NSPS in the preamble and the
remainder of the materials for this final action. As explained in the previous chapters, all costs
and benefits outlined in this RIA are estimated as the change from the baseline, which reflects
the current business practice for the affected sources as mentioned in Chapter 1, particularly with
regard to emissions from flares. As stated earlier in this RIA, there is no monetized estimate of
the benefits for the HAP emission reductions expected to occur as a result of this final action. We
do present monetized estimates for other impacts of this action, such as benefits from both short-
and long-term reduced exposure to ozone caused by VOC emissions reductions and benefits
from decreases in CH4 emissions and disbenefits from increases in CO2, N2O, and NOx
emissions.

6.1 Results

As part of fulfilling analytical guidance with respect to E.O. 12866, EPA presents
estimates of the present value (PV) of the benefits and costs over the period 2024 to 2038. To
calculate the PV of the social net benefits of the final action, annual benefits and costs are in
2021 dollars and are discounted to 2023 at 3 percent and 7 percent discount rates as directed by
OMB's Circular A-4.42'43 The EPA also presents the equivalent annualized value (EAV), which
represents a flow of constant annual values that would yield a sum equivalent to the PV. The
EAV represents the value of a typical cost or benefit for each year of the analysis, consistent with
the estimate of the PV, in contrast to year-specific estimates.

42 We note that this RIA uses the guidance in the previous Circular A-4 version that was issued on September 17,

2003. The effective date for the new, revised Circular A-4 version issued in November 2023 is at a date after
the completion of this final rule RIA; hence, guidance in the previous Circular A-4 version is used in this RIA.

43 The climate benefits and disbenefits from methane emission reductions and CO2 and N20 emission increases are

estimated at a 2 percent discount rate, reflecting the use of the updated SCC, and these estimates are included in
the estimates of total and net benefits for this final action. Explanation of the use of a 2 percent discount rate for
climate benefits and disbenefits is found in Chapter 4 of this RIA.

135

-------
The presentation of impacts in this chapter includes those for more and less stringent
options to those for the final as a whole (that is, across all final rules). The more stringent option
is the same as the final rules except that tighter controls for HON process vents and storage
vessels, and also such controls on P&R I process vents (or PV, when discussing types of affected
sources) and storage vessels (SV), are included. The tighter process vent controls in the more
stringent option are defined as option PV2 in Table 3-12 of this RIA, and the tighter storage
vessels controls are defined as option SV3 in Table 3-8 of this RIA. The less stringent option is
the same as the final except that weaker controls for storage vessels defined as option SV1 in
Table 3-8 of this RIA are included. The less stringent option does not include any other
differences in options and thus no change in costs and impacts between the proposal and the
final. Thus, the differences in stringency for analyses in the RIA reflect different stringencies
primarily in the final HON options as described in Chapter 3. Since the differences in stringency
occur only for options considered under the final HON amendments, we present impacts below
for the final HON, P&R I, and cumulative. More and less stringent options were not available for
the other final rules.

Tables 6-1 through 6-3 presents a summary of the monetized benefits, compliance costs,
and net benefits (including climate disbenefits) of the HON, P&R I, and cumulatively, and the
more and less stringent alternatives in terms of present value (PV) and equivalent annualized
value (EAV). Tables presenting benefits list both figures, with short-term benefits listed first.

136

-------
Table 6-1 Summary of Monetized Benefits, Compliance Costs, and Net Benefits
	PV/EAV for HON, 2024-2038 (million 2021$, discounted to 2023)	

Final	Less Stringent Alternative	More Stringent Alternative

3%

PV

EAV

PV

EAV

PV

EAV

Monetized Health
Benefits

70 and 630

5.9 and 53

70 and 630

5.9 and 52

71 and 640

6 and 54

Climate
Disbenefits (2%)

140

11

140

11

140

11

Net Compliance
Costs

1,550

130

1,500

130

1,600

140

Compliance Costs

1,560

130

1,500

130

1,600

140

Value of Product

12

1

12

1

12

1

Recovery

Net Benefits

(1,600) and
(1,100)

(140) and (88)

(1,600) and
(1,000)

(140) and (89)

(1,700) and
(1,100)

(150) and (97)

7%

PV

EAV

PV

EAV

PV

EAV

Monetized Health
Benefits

48 and 420

5.2 and 46

48 and 420

5.2 and 46

49 and 430

5.3 and 47

Climate
Disbenefits (2%)

140

11

140

11

140

11

Net Compliance
Costs

1,200

130

1,200

130

1,300

140

Compliance Costs

1,200

130

1,200

130

1,300

140

Value of Product



1



1



1

Recovery







Net Benefits

(1,300) and
(920)

(140) and (95)

(1,300) and
(920)

(140) and (95)

(1,400) and
(1,000)

(150) and
(100)

Nonmonetized 1,107 tons of HAP emission reductions. Health effects from reduced exposure to ethylene oxide,

Benefits chloroprene, benzene, 1,3-butadiene, vinyl chloride, ethylene dichloride, chlorine,
	maleicanhydride, and acrolein	

Note: Monetized benefits (incorporating disbenefits) include those related to public health and climate. Monetized
air quality related health benefits include ozone related health benefits associated with reductions in VOC emissions.
The health benefits are associated with several point estimates and are presented at real discount rates of 3 and 7
percent. The two benefits estimates are separated by the word "and" to signify that they are two separate estimates.
The estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions and VOC
reductions outside of the ozone season remain unmonetized and are thus not reflected in the table. The unmonetized
effects also include disbenefits resulting from a secondary increase in CO emissions. Monetized net climate
disbenefits are based on reductions in methane emissions and increases in CO2 and N20 emissions and are
calculated using three different estimates of the social cost of each greenhouse gas (SC-GHG) (under 1.5 percent,
2.0 percent, and 2.5 percent near-term Ramsey discount rates). For the presentational purposes of this table, we
show the net climate disbenefits associated with the SC-GHG at a 2 percent Ramey discount rate, but climate
benefits and disbenefits are presented using the other near-term discount rates in Section 4.7. Net compliance costs
are the compliance costs minus the value of product recovery from compliance with the rule. Parentheses around a
number denotes that is has a negative value. Rows may not appear to add correctly due to rounding.

137

-------
Table 6-2 Summary of Monetized Benefits, Compliance Costs, and Net Benefits
	PV/EAV for P&R I, 2024-2038 (million 2021$, discounted to 2023)	

Final	Less Stringent Alternative More Stringent Alternative

3%

PV

EAV

PV

EAV

PV

EAV

Health Benefits
Climate

(0.2) and (1.7)
22

(0.02) and
(0.1)

(0.5) and (4)
22

(0.04) and
(0.3)

(0.4) and (3)
22

(0.03) and
(0.3)

Disbenefits (2%)







Net Compliance
Costs

140

12

140

12

150

12

Compliance Costs

140

12

140

12

150

12

Value of Product

1

0.2

1

0.2

1

0.2

Recovery

Net Benefits

(160) and
(160)

(14) and (14)

(160) and
(170)

(14) and
(14)

(170) and
(180)

(14) and
(14)

7%

PV

EAV

PV

EAV

PV

EAV

Health Benefits
Climate

(0.2) and (1.5)
22

(0.02) and
(0.2)

(0.2) and (1.5)
22

(0.02) and
(0.2)

(0.3) and (2.4)
22

(0.03) and
(0.3)

Disbenefits (2%)

2

2

2

Net Compliance
Costs

110

12

100

12

110

12

Compliance Costs

110

12

100

12

110

12

Value of Product

1

0.1

1

0.1

1

0.1

Recovery

Net Benefits

(130) and
(130)

(14) and (14)

(120) and
(120)

(14) and
(14)

(130) and
(130)

(14) and
(14)

Nonmonetized 264 tons per year (tpy) of HAP emission reductions, including an approximate 14 tpy
Benefits	reduction in chloroprene emissions	

Note: Monetized benefits (incorporating disbenefits) include those related to public health and climate. Monetized
air quality related health benefits include ozone related health benefits associated with reductions in VOC emissions.
The health benefits are associated with several point estimates and are presented at real discount rates of 3 and 7
percent. The two benefits estimates are separated by the word "and" to signify that they are two separate estimates.
The estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions and VOC
reductions outside of the ozone season remain unmonetized and are thus not reflected in the table. The unmonetized
effects also include disbenefits resulting from a secondary increase in CO emissions. Monetized net climate
disbenefits are based on reductions in methane emissions and increases in CO2 and N20 emissions and are
calculated using three different estimates of the social cost of each greenhouse gas (SC-GHG). For the presentational
purposes of this table, we show the net climate disbenefits associated with the SC-GHG at a 2 percent Ramsey
discount rate, but climate benefits and disbenefits are presented using the other near-term discount rates in Section
4.7. Net compliance costs are the compliance costs minus the value of product recovery from compliance with the
rule. Rows may not appear to add correctly due to rounding. A number in parentheses denotes a negative value.

138

-------
Table 6-3 Summary of Monetized Benefits, Compliance Costs, and Net Benefits
PV/EAV for All Rules, 2024-2038 (million 2021$, discounted to 2023)	

Final	Less Stringent Alternative	More Stringent Alternative

3%

PV

EAV

PV

EAV

PV

EAV

Health Benefits

77 and 690

6.5 and 58

76 and 680

6 and 53

77 and 690

6.1 and 55

Climate
Disbenefits (2%)

160

13

160

13

160

13

Net Compliance
Costs

1,770

150

1,770

150

1,880

160

Compliance Costs

1,790

150

1,790

150

1,900

160

Value of Product

16

1.3

16

1.3

16

1.3

Recovery

Net Benefits

(1,900) and

(160) and

(1,900) and

(160) and

(2,000) and

(170) and



(1,200)

(110)

(1,300)

(110)

(1,400)

(120)

7%

PV

EAV

PV

EAV

PV

EAV

Health Benefits

53 and 475

5.9 and 52

52 and 460

5.4 and 47

53 and 470

5.5 and 48

Climate
Disbenefits (2%)

160

13

160

13

160

13

Net Compliance
Costs

1,370

150

1,370

150

1,500

160

Compliance Costs

1,380

150

1,380

150

1,500

160

Value of Product

12

1.3

12

1.3

12

1.3

Recovery

Net Benefits

(1,500) and

(160) and

(1,500) and

(160) and

(1,600) and

(170) and



(1,100)

(110)

(1,100)

(120)

(1,200)

(130)

Nonmonetized 6,230 tons/year of HAP emission reductions. Health effects of reduced exposure to ethylene

Benefits oxide, chloroprene, benzene, 1,3-butadiene, vinyl chloride, ethylene dichloride, chlorine,
	maleic anhydride and acrolein	

Note: Monetized benefits (incorporating disbenefits) include those related to public health and climate. Monetized
air quality related health benefits include ozone related health benefits associated with reductions in VOC emissions.
The health benefits are associated with several point estimates and are presented at real discount rates of 3 and 7
percent. The two benefits estimates are separated by the word "and" to signify that they are two separate estimates.
The estimates do not represent lower- and upper-bound estimates. Benefits from HAP reductions and VOC
reductions outside of the ozone season remain unmonetized. The unmonetized effects also include disbenefits
resulting from a secondary increase in CO emissions. Monetized net climate disbenefits are based on reductions in
methane emissions and increases in CO2 and N20 emissions and are calculated using three different estimates of the
social cost of each greenhouse gas (SC-GHG) (under 1.5 percent, 2.0 percent, and 2.5 percent near-term Ramsey
discount rates). For the presentational purposes of this table, we show the net climate disbenefits associated with the
SC-GHG at a 2 percent Ramsey discount rate, but climate benefits and disbenefits are presented using the other
near-term discount rates in Section 4.7. Net compliance costs are the compliance costs minus the value of product
recovery from compliance with the rule. A number in parentheses denotes a negative value. Rows may not appear to
add correctly due to rounding.

139

-------
Given these results, the EPA expects that implementation of the final HON, based solely
on an economic efficiency criterion, should provide society with a relatively potential net gain in
welfare, notwithstanding the expansive set of health and environmental benefits and other
impacts we were unable to quantify such as monetization of benefits from VOC emission
reductions occurring outside of the ozone season (the months of October-April). The same holds
true for the final P&R I and IINESHAP and for all final amendments (including the NSPS)
considered cumulatively. Further quantification of directly emitted VOC and HAP would
increase the estimated net benefits of each final action and cumulatively.

6.2 Uncertainties and Limitations

Throughout the RIA, we considered a number of sources of uncertainty, both
quantitatively and qualitatively, regarding the benefits, and costs of the final amendments. We
summarize the key elements of our discussions of uncertainty here:

Projection methods and assumptions: Over time, more facilities are newly established
or modified in each year, and to the extent the facilities remain in operation in future years, the
total number of facilities subject to the action could change. We assume 100 percent compliance
as these final rules and existing rules are implemented, starting from when the source becomes
affected. If sources do not comply with these rules, at all or as written, the cost impacts and
emission reductions may be overestimated. Additionally, new control technology and approaches
may become available in the future at lower cost, and we are unable to predict exactly how
industry will comply with the final rules in the future.

Years of analysis: In addition, the counts of units projected to be affected by this final
action are held constant. Given our analytical timeframe of 2024-2038, it is possible that the
affected unit counts may change. The years of the cost analysis are 2024, to represent the first-
year facilities that the NSPS proposed in this rulemaking will be effective, through 2038, to
represent impacts of the action over the life of installed capital equipment, as discussed in
Chapter 3. Extending the analysis beyond 2038 would introduce substantial and increasing
uncertainties in projected impacts of the final regulations.

Compliance Costs: There may be an opportunity cost associated with the installation of
environmental controls (for purposes of mitigating the emission of pollutants) that is not
reflected in the compliance costs included in Chapter 3. If environmental investment displaces

140

-------
investment in productive capital, the difference between the rate of return on the marginal
investment (which is discretionary in nature) displaced by the mandatory environmental
investment is a measure of the opportunity cost of the environmental requirement to the
regulated entity. To the extent that any opportunity costs are not included in the control costs, the
compliance costs presented above for this final action may be underestimated.

BPT estimates: As discussed earlier in Chapter 4, all national-average BPT estimates
reflect the geographic distribution of the modeled emissions, which may not exactly match the
emission reductions that would occur due to the action, and they may not reflect local variability
in population density, meteorology, exposure, baseline health incidence rates, or other local
factors for any specific location. Recently, the EPA systematically compared the changes in
benefits, and concentrations where available, from its BPT technique and other reduced-form
techniques to the changes in benefits and concentrations derived from full-form photochemical
model representation of a few different specific emissions scenarios. Reduced form tools are less
complex than the full air quality modeling, requiring less agency resources and time. That work,
in which we also explore other reduced form models is referred to as the "Reduced Form Tool
Evaluation Project" (Project), began in 2017, and the initial results were available at the end of
2018. The Agency's goal was to better understand the suitability of alternative reduced-form air
quality modeling techniques for estimating the health impacts of criteria pollutant emissions
changes in the EPA's benefit-cost analysis. The EPA continues to work to develop refined
reduced-form approaches for estimating benefits. The scenario-specific emission inputs
developed for this project are currently available online. The study design and methodology are
described in the final report summarizing the results of the project, available at:

https://www.epa.gov/sites/production/files/2019-
1/documents/rft combined report 10.31.19 final.pdf

Non-monetized benefits: Numerous categories of health and welfare, and climate-
related benefits are not quantified and monetized in this RIA. These unquantified benefits,
including benefits from reductions in emissions of pollutants such as HAP which are to be
reduced by this final action, are described in detail in Chapter 4 of this RIA and various NAAQS
RIAs.

141

-------
VOC health impacts: In this RIA, we quantify an array of adverse health impacts
attributable to emissions of VOC. The Integrated Science Assessment for Particulate Matter
("ISA") (U.S. EPA, 2019) identifies the human health effects associated with ambient particles,
which include premature death and a variety of illnesses associated with acute and chronic
exposures.

Monetized climate benefits and disbenefits: The EPA considered the uncertainty
associated with the social cost of greenhouse gas (SC-GHG) estimates, which were used to
calculate the climate benefits and disbenefits from the increase in CO2 and N2O emissions and
the decrease in CH4 emissions projected under the final amendments in this rulemaking. Some
uncertainties are captured within the analysis, while other areas of uncertainty have not yet been
quantified in a way that can be modeled. A full discussion of uncertainties in the analysis of
monetized climate benefits and disbenefits can be found in Chapter 4 of this RIA.

142

-------
7 References

Abdela, A. a. (2018, September). The United States Has a Market Concentration Problem - Reviewing
Concentration Estimates in Antitrust Markets, 2000-Present. Retrieved from

https://rooseveltinstitute.org/wp-content/uploads/2020/07/RI-US-market-concentration-problem-brief-
201809.pdf

Acumen Research and Consulting. (2022). Neoprene Market Size - Global Industry, Share, Analysis, Trends and
Forecast 2022 - 2030. Retrieved from https://www.acumenresearchandconsulting.com/neoprene-market

Agency for Toxic Substances and Disease Registry (ATSDR). 1992. Toxicological Profile for 1,2-Dichloroethane.
Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.

Agency for Toxic Substances and Disease Registry (ATSDR). 2006. Toxicological Profile for Vinyl Chloride.

Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.
https ://www.atsdr. cdc. gov/toxprofiles/tp20 .pdf

Agency for Toxic Substances and Disease Registry (ATSDR). 2007a. Toxicological Profile for Benzene. U.S.

Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.

Agency for Toxic Substances and Disease Registry (ATSDR). August 2007b. Toxicological Profile for Acrolein.
Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.
https://www.atsdr.cdc.gov/toxprofiles/tpl24.pdf

Agency for Toxic Substances and Disease Registry (ATSDR). November 2010. Toxicological Profile for Chlorine.
Atlanta, GA: U.S. Department of Health and Human Services. Available at
https://www.atsdr.cdc.gov/toxprofiles/tpl72.pdf.

Agency for Toxic Substances and Disease Registry (ATSDR).2012. Toxicological Profile for 1,3-Butadiene. Public
Health Service, U.S. Department of Health and Human Services, Atlanta, GA.
https://www.atsdr.cdc.gov/ToxProfiles/tp28.pdf

Agency for Toxic Substances and Disease Registry (ATSDR). August 2022. Toxicological Profile for Ethylene
Oxide. U.S. Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.
https://www.atsdr.cdc.gov/toxprofiles/tpl37.pdf

Ahmad, S. &. (2019). A Method for Estimating the Elasticity of Substitution and Import Sensitivity by Industry.
Retrieved from US International Trade Commission. Washington, D.C.: Retrieved from
https://usitc.gov/data/pe modeling/a method for estimating the elasticity of substitution and import se
nsitivitv by industrv.pdf

American Chemistry Society. (2023). Ethylene oxide. Retrieved from

https://www.americanchemistrv.com/chemistrv-in-america/chemistries/ethylene-oxide

Anixter. (2022). Hypalon: Hype About Discontinuing. Retrieved from

https://www.anixter.com/en us/resources/literature/wire-wisdom/hvpalon-hype-about-discontinuing.html

Anthoff, D. a. (2013a). The uncertainty about the social cost of carbon: a decomposition analysis using FUND.

Climatic Change(lll), 515-530.

Anthoff, D. a. (2013b). Erratum to: The uncertainty about the social cost of carbon: A decomposition analysis using
FUND. Climatic Change, 121(2), 413.

143

-------
Arrow, K. J., Cropper, M. L., Eads, G. C., Hahn, R. J., Lave, L. B., Noll, R. J.,. . . Stavins, R. N. (1996). Benefit-
Cost Analysis in Environmental, Health, and Safitey Regulation: A Statement of Principles. American
Enterprise Institute Press.

Barnicki, S. D. (2017, August 2). Synthetic Organic Chemicals. Handbook of Industrial Chemistry and

Biotechnology, 423-530. Retrieved from Springer, Cham.: https://doi.org/10.1007/978-3-319-52287-6 7

California Environmental Protection Agency (CalEPA). 2008. Technical Supporting Document for Noncancer
RELs, Appendix D3. Chronic RELs and toxicity summaries using the previous version of the Hot Spots
Risk Assessment guidelines (OEHHA 1999). Office of Environmental Health Hazard Assessment. Maleic
Anhydride, last revised December 2001, pp. 348-353.
https://oehha.ca.gov/media/downioads/crnr/appendixd3finai.pdf

California v. Bernhardt, 472 F Supp. 3d 573 (N.D. Cal 2020).

Chambers, R. G. (1994). Simple Econometrics of Pesticide Productivity. American Journal of Agricultural
Economics, 76(3), 407-417. doi:doi: 10.2307/1243653

ChemAnalyst. (2021). Petrochemicals Market Analysis. Retrieved from https://www.chemanalvst.com/industrv-
report/global-petrochemical-market-308.

ChemAnalyst. (2022). Ethylene Oxide Priced Trend and Forecast. Retrieved from
https://www.chemanalvst.com/Pricing-data/ethvlene-oxide-1102

Deloitte Insights. (2022). The shifting landscape of chemicals and materials. Retrieved from

https://www2.deloitte.com/us/en/insights/industrv/oil-and-gas/covid-19-chemicals-industrv-impact.html.

DeRosa, S. a. (2015). Impact of natural gas and natural gas liquids supplies on the United States chemical

manufacturing industry: production cost effects and identification of bottleneck intermediates. ACS
Sustainable Chemistry & Engineering,, 3(3), pp.451-459. Retrieved from
https://pubs.acs.org/doi/10.1021/sc500649k

Fernandez, L. (2021, July). Global Butadiene Prices 2022. Retrieved from Statista:

https://www.statista.com/statistics/1171063/price-butadiene-forecast-globally/

Fernandez, L. (2022, June 23). Ethylene Prices Globally 2022. Retrieved from Statista:
https://www.statista.com/statistics/1171063/price-butadiene-forecast-globally/

Global Newswire. (2022). Ethylene Market to Reach $161.61 Billion by 2028. Retrieved from

https://www.globenewswire.eom/news-release/2022/02/08/2380916/0/en/Ethylene-Market-to-Reach-161"

61-Billion-by-2028-Superior-Ethylene-Cost-Competitiveness-Emerging-Markets-Pandemic-Recovery-

Top-Manufactures-Application-Regional-Growth-The-Brainy-In.html

Grand View Research. (2015). 1,3-Butadiene Market Size Worth $33.01 Billion By 2020. Retrieved from
https://www.grandviewresearch.com/press-release/global-butadiene-market

Grand View Research. (2021). Petrochemicals Market Size & Share Report, 2022-2030. Retrieved from
https://www.grandviewresearch.com/industry-analysis/petrochemical-market

Grand View Research. (2022). Epoxy Resin Market Size, Share & Trends Analysis Report By Application

(Adhesives, Electrical & Electronics, Paints & Coatings, Wind Turbines, Composites, Construction), By
Region, And Segment Forecasts, 2022 - 2030. Retrieved from
https://www.grandviewresearch.com/industrv-analvsis/epoxv-resins-market

144

-------
Gray, W.B., Shadbegian, R. and Wolverton, A., 2023. Environmental Regulation and Labor Demand: What Does
the Evidence Tell Us?. Annual Review of Resource Economics, 15, pp.177-197.

Hafstead, M.A. and Williams III, R.C., 2020. Jobs and environmental regulation. Environmental and Energy Policy
and the Economy, 1(1), pp. 192-240.

Helbig, C. G. (2016). Extending the geopolitical supply risk indicator: Application of life cycle sustainability

assessment to the petrochemical supply chain of poly aery lonitrile-based carbon fibers. Journal of Cleaner
Production, 137, pp.1170-1178. Retrieved from https://doi.org/10.1016/jjclepro.2016.07.214

Hope, C. (2013). Critical issues for the calculation of the social cost of C02: why the estimates from PAGE09 are
higher than those from PAGE2002. Climatic Change(lll), 531-543.

Industrial Rubber Goods. (2022). ChlorosulfonatedPolyethylene (CSM)/Hypalon. Retrieved from
http://www.industrialrubbergoods.com/chlorosulfonated-polyethylene.html

International Agency for Research on Cancer (IARC). 2018. Monographs on the evaluation of carcinogenic risk of
chemicals to humans, Volume 120, Benzene, World Health Organization, Lyon, France.

Intratec. (2023). Primary Commodity Prices. Retrieved February 2023, from
https://www.intratec.us/products/primary-commodity-prices

IPCC. (2007). Core WritingTeam; Pachauri, R.K; and Reisinger, A. (ed.) Climate Change 2007: Synthesis Report,
Contribution of Working Groups I, II, and II to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. (Vols. ISBN 92-9169-122-4). IPCC.

IPCC. (2014). Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change. (C. W. Team, R. Pachauri, & L.
Meyer, Eds.) Geneva, Switzerland.

IPCC. (2018). Global Warming of 1.5°C. An IPCC Special Report on the impacts of global warming of 1.5°C above
pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening
the global response to the threat of climate change,. (Masson-Delmotte, V., P. Zhai, H.-O. Portner, D.
Roberts, J. Skea,. . . T. Waterfield, Eds.)

IPCC. (2019a). Climate Change and Land: an IPCC special report on climate change, desertification, land
degradation, sustainable land management, food security, and greenhouse gas fluxes in terrestrial
ecosystems. (P. Shukla, J. Skea, E. C. Buendia, V. Masson-Delmot, H.-O. Portner, D. C. Roberts,. . . K.
Kissick, Eds.)

IPCC. (2019b). IPCC Special Report on the Ocean and Cryosphere in a Changing Climate. (H.-O. Portner, D.
Roberts, V. Masson-Delmotte, P. Zhai, M. Tignor, E. Poloczanska,. . . N. Weyer, Eds.)

IWG. (2010). Technical Support Document: Social Cost of Carbon for Regulatory Impact Analysis under Executive
Order 12866. Retrieved 2023, from https://www.epa.gov/sites/default/files/2016-
12/documents/scc_tsd_2010.pdf

IWG. (2013). Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact
Analysis Under Executive Order 12866. Retrieved from https://www.ourenergypolicy.org/wp-
content/uploads/2013/06/social_cost_of_carbon_for_ria_2013_update.pdf

IWG. (2015). Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact
Analysis Under Executive Order 12866.

145

-------
IWG. (2016a). Addendum to Technical Support Document on Social Cost of Carbon for Regulatory Impact Analysis
under Executive Order 12866: Application of the Methodology to Estimate the Social Cost of Methane and
the Social Cost of Nitrous Oxide. Retrieved from https://www.epa.gov/sites/default/files/2016-
12/documents/addendum_to_sc-ghg_tsd_august_2016.pdf

IWG. (2016b). Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact
Analysis Under Executive Order 12866. Retrieved 2023, from https://www.epa.gov/sites/default/files/2016-
12/documents/sc_co2_tsd_august_2016 .pdf

IWG. (2021). Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates
under Executive Order 13990. Retrieved 2023, from https://www.whitehouse.gov/wp-
content/uploads/2021/02/TechnicalSupportDocument_SocialCostofCarbonMethaneNitrousOxide.pdf

Kotze, T. B. (2017). Buyer-supplier collaboration and supply chain resilience: A case study in the petrochemical
industry. South African Journal of Industrial Engineering, 28(4), pp.183-199. Retrieved from
http ://www. scielo .org. za/pdf/saj ie/v28

Mallard Creek Polymers. (2020, October). Retrieved from Styrene-Butadiene Latex.
https://www.mcpolymers.com/library/styrene-butadiene-latex

Marten, A. L., Kopits, E. A., Griffiths, C. W., Newbold, S. C., & Wolverton, A. (2015). Incremental CH4 and N20
mitigation benefits consistent with the US Government's SC-CO 2 estimates. Climate Policy, 15(2), 272-
298.

Martinez, L. (2012). The country-specific nature of apparel elasticities and impacts of the multi-fibre arrangement.
Retrieved from Macalester College:

https://digitalcommons.macalester.edu/cgi/viewcontent.cgi?article=1049&context=economics_honors_proj
ects

Material Research L3C. (2019). Ethylene dichloride (EDC): Technical report on the conditions of use. Prepared for
Earthjustice. Retrieved from https://downloads.regulations.gov/EPA-HQ-OPPT-2018-0427-
0015/attachment_2 .pdf

National Academies. (2016b). Attribution of Extreme Weather Events in the Context of Climate Change.

Washington, D.C.: The National Academies Press. doi:10.17226/21852.

National Academies. (2017). Valuing Climate Damages: Updating Estimation of the Social Cost of Carbon Dioxide.
Washington, D.C.: The National Academies Press.

National Academies. (2019). Climate Change and Ecosystems. Washington, D.C.: The National Academies Press,
doi: 10.17226/25504.

National Center for Biotechnology Information. (2022a). PubChem Compound Summary for CID 6325, Ethylene.
Retrieved from https ://pubchem.ncbi. nlm. nih.gov/compound/ethylene#section=Uses

National Center for Biotechnology Information. (2022b). PubChem Compound Summary for CID 7845, 1,3-
Butadiene. Retrieved from https://pubchem.ncbi.nlm.nih.gov/compound/divinyl

National Center for Biotechnology Information. (2022c). PubChem Compound Summary for CID 6354, Ethylene
oxide. Retrieved from https://pubchem.ncbi.nlm.nih.gov/compound/ethylene-oxide

Newmark, R. C. (2004, February 14). Price-Concentration Studies: There You Go Again. Retrieved from U.S.
Department of Justice: https://www.justice.gov/atr/price-concentration-studies-there-you-go-again

146

-------
Nordhaus, W. (2010). Economic aspects of global warming in a post-Copenhagen environment. Proceedings of the
National Academy of Sciences, 107(26), 11721-11726.

Novicio, T. (2021). 15 Biggest Petrochemical Companies in the World. Retrieved from

https://www.insidermonkey.com/blog/15-biggest-petrochemical-companies-in-the-world-913519/

OEC. (2022a). Petrochemicals. Retrieved from https://oec.world/en/profile/sitc/petrochemicals

OEC. (2022b). Ethylene. Retrieved from https://oec.world/en/profile/hs/ethylene

OEC. (2022c). Oxirane (ethylene oxide). Retrieved from https://oec.world/en/profile/hs/oxirane-ethylene-oxide

OEC. (2022d). Isobutene-isoprene (butyl) rubber (11R). Retrieved from https://oec.world/en/profile/hs/isobutene-
isoprene-butyl-rubber-iir. Accessed 11/7/2022.

OEC. (2022e). Halo-isobutene-isoprene rubber (C11RJB11R). Retrieved from https://oec.world/en/profile/hs/halo-
isobutene-isoprene-rubber-ciirbiir. Accessed 11/7/2022.

OEC. (2022f). Ethylene-propylene diene rubber (EPDM). Retrieved from https://oec.world/en/profile/hs/ethylene-
propylene-non-conj-diene-rubber-epdm. Accessed 11/7/2022.

OEC. (2022g). Acrylonitrile-butadiene rubber (NBR), latex. Retrieved from

https://oec.world/en/profile/hs/acrylonitrile-butadiene-rubber-nbr-latex. Accessed 11/7/2022.

OEC. (2022h). Acrylonitrile-butadiene rubber (NBR) except as latex. Retrieved from

https://oec.world/en/profile/hs/acrylonitrile-butadiene-rubber-nbr-except-as-latex?redirect=true. Accessed
11/7/2022.

OEC. (2022i). Styrene-butadiene rubber (SBR/XSBR) latex. Retrieved from https://oec.world/en/profile/hs/styrene-
butadiene-rubber-sbrxsbr-latex. Accessed 11/7/2022.

OEC. (2022j). Styrene-butadiene rubber (SBR/XSBR) except latex. Retrieved from

https://oec.world/en/profile/hs/styrene-butadiene-rubber-sbrxsbr-except-latex. Accessed 11/7/2022.

OEC. (2022k). Butadiene rubber (BR). Retrieved from https://oec.world/en/profile/hs/butadiene-rubber-
br?redirect=true. Accessed 11/7/2022.

OEC. (20221). Polyamides NES in Primary Forms. Retrieved from https://oec.world/en/profile/hs/polyamides-nes-
in-primary-forms. Accessed 11/7/2022.

Office of Management and Budget. (2015). 2015 Report to Congress on the Benefits and Costs of Federal

Regulations and Agency Compliance with the Unfunded Mandates Reform Act. U.S. Office of Management
and Budget, Office of Information and Regulatory Affairs.

OMB. (2003). Circular A-4. Retrieved from https://www.whitehouse.gov/wp-
content/uploads/legacy_drupal_files/omb/circulars/A4/a-4.pdf

Pearce, J. A. (2014). Why domestic outsourcing is leading America's reemergence in global manufacturing.

Business Horizons, 57, 27-36. Retrieved from http://dx.doi.Org/10.1016/j.bushor.2013.08.007

Pederson, J. (2021, July). Plastic Resin Prices 2021-Resin Shortage & Cost of Plastic Explained. Retrieved from
Summit Packaging: https://summitpackaging.com/plastic-resin-prices-2021-resin-shortage/

Polaris Market Research. (2021). Ethylene Market Size, Share, Trends, Value \ Industry Report 2021 - 2029.
Retrieved from https://www.polarismarketresearch.com/industry-analysis/ethylene-market

147

-------
Polymerdatabase . (2022b). CR — Polychloroprene (Neoprene). Retrieved from
https://polymerdatabase.com/Elastomers/Chloroprene.html

Polymerdatabase . (2022c). Styrene-butadiene rubber (SBR). Retrieved from
https://polymerdatabase.com/Polymer%20Brands/SBRs.html

Polymerdatabase. (2022a). NBR - Butadiene Nitrile Rubber. Retrieved from
https://polymerdatabase.com/Elastomers/NBR. html

Polymerdatabase. (2022d). BR - Polybutadiene. Retrieved from https://polymerdatabase.com/Elastomers/BR.html

Riker, D. &. (2019). An Euler Method Approach to Simulating Changes in Trade Policy. Retrieved from US
International Trade Commission, Office of Economics. Washington, D.C.: Retrieved from
https://usitc.gov/data/pe_modeling/an_euler_method_approach_to_simulating_changes_in_trade_policy_8-
8-2019.pdf

Rutherford, T. F. (2002). Lecture Notes on Constant Elasticity Functions. Lecture Notes. Retrieved from University
of Colorado. Boulder, CO.: Retrieved from https://www.gamsworld.org/mpsge/debreu/ces.pdf

Saiyid, A. (2019). Largest sources of cancer-causing ethylene oxide go unnoticed. Bloomberg Law. Retrieved from
https://news.bloomberglaw.com/environment-and-energy/largest-sources-of-cancer-causing-ethylene-
oxide-go-unnoticed

Santos Manzano, F. (2005). Supply chain practices in the petroleum downstream. Doctoral dissertation,
Massachusetts Institute of Technology. Retrieved from

https://dspace.mit.edu/bitstream/handle/1721.1/33345/62395452-MIT.pdf?sequence=2

SBA. (2017, August). A Guide For Government Agencies: How To Comply With The Regulatory Flexibility Act.
Retrieved from Small Business Administration, Office of Advocacy: https://cdn.advocacy.sba.gov/wp-
content/uploads/2019/07/01092549/How-to-Comply-with-the-RFA-WEB.pdf

ScienceDirect. (2022). Petrochemical. Retrieved from

https://www.sciencedirect.com/topics/engineering/petrochemical

Senlos Chem. (2022). Acrylonitrile Butadiene Copolymer Latex (NBR Latex). Retrieved from

https://www.senloschemical.com/nbr-latex-series/acrylonitrile-butadiene-copolymer-latex-nbr.html

Stamber, K. V. (2011, March). Consequence and Resilience Modeling for Chemical Supply Chains. In Selected
Papers and Presentations Presented atMODSLM World 2010 Conference Expo. Retrieved from
https://ntrs.nasa.gov/api/citations/20110012079/downloads/20110012079.pdf

Stokes, F. (2022, March). Russia-Ukraine War Pushes Up Gas Prices, Chemical Producers' Profits Under

Pressure. Retrieved from https://www.chemanalyst.com/NewsAndDeals/NewsDetails/russia-ukraine-war-
pushes-up-gas-prices-chemical-producers-profits-under-pressure-9269

Thomas. (2022a). All About Butyl Rubber - Properties, Applications, and Uses. Retrieved from Retrieved from
https://www.thomasnet.com/articles/plastics-rubber/all-about-butyl-rubber/

Thomas. (2022b). All About EPDM Rubber - Properties, Applications and Uses. Retrieved from

https://www.thomasnet.com/articles/plastics-rubber/all-about-epdm-rubber-properties-applications-and-
uses/

Trangadisaikul, S. (2011). Oligopsony in the Tire Industry: A study of its Impacts on the Natural Rubber Industry in
Thailand. Thailand and The World Economy, 29(1), 128-169. Retrieved from https://so05.tci-
thaii o. ore/index. php/TER/article/view/13 7396

148

-------
U.S. DOE. 2022. U.S. Ethane: Market Issues and Opportunities: Report to Congress. September 2022. Available at:
https://www.energy.gOv/sites/default/files/2023-06/U.S.Ethane-Market-Issues-and-Opportunities.pdf

U.S. Department of Health and Human Services. (2010). Retrieved from Toxicological profile for ethylene glycol.
https://www.atsdr.cdc.gov/ToxProfiles/tp96.pdf

U.S. EIA. (2021). U.S. ethane production to grow, along with expanding domestic consumption and exports.
Retrieved from https://www.eia.gov/todavinenergy/detail.php?id=48056

U.S. EPA. 1988. Integrated Risk Information System (IRIS) on Maleic Anhydride. National Center for
Environmental Assessment, Office of Research and Development, Washington, DC.

https://iris.epa.gOv/static/pd:fs/0307 siimmarv.pdf and
https://iris.epa.gov/Che mica lLanding/&substance n.mbr=307.

U.S. EPA. June 1994. Integrated Risk Information System (IRIS) on Chlorine. National Center for Environmental
Assessment, Office of Research and Development, Washington, DC.
https://iris.epa.gov/Che mica lLanding/&substance n.mbr=405

U.S. EPA. 1999. Integrated Risk Information System (IRIS) on 1,2-Dichloroethane. National Center for
Environmental Assessment, Office of Research and Development, Washington, DC.

U.S. EPA. 2000. Integrated Risk Information System (IRIS) on Benzene. National Center for Environmental
Assessment, Office of Research and Development, Washington, DC.

U.S. EPA. May 2000. Toxicological Review of Vinyl Chloride (CAS No. 75-01-4) In Support of Summary
Information on the Integrated Risk Information System (IRIS). National Center for Environmental
Assessment, Office of Research and Development, Washington, DC. EPA/635R-00/004.

https://iris.epa.gov/static/pdfs/1001tr.pdf and
https://iris.epa.gov/ChemiealLanding/&substance n.mbr= 1.001.

U.S. EPA. October 2002. Health Assessment of 1,3-Butadiene. Integrated Risk Information System (IRIS) on 1,3-
Butadiene. National Center for Environmental Assessment, Office of Research and Development,
Washington, DC. EPA/600/P-98/001F. https://cfpnb.epa.gov/ncea/iris drafis/reeordisptav.cfm?deid=54499
and https://iris.epa.gov/ChemicalLanding/&substanee n.mbr=.1.39.

U.S. EPA. June 2003. Toxicological Review of Acrolein (CAS No. 107-02-8) In Support of Summary Information
on the Integrated Risk Information System (IRIS). National Center for Environmental Assessment, Office
of Research and Development. Washington, DC. EPA/635/R-03/003.

https://iris.epa.gov/static/pdfs/0364tr.pdf and https://iris.epa.gov/ChemicalLanding/&substance nmbr=364.

U.S. EPA. Guidelines for Carcinogen Risk Assessment (2005). U.S. Environmental Protection Agency, Washington,
DC, EPA/630/P-03/001F, 2005

U.S. EPA. September 2010. Toxicological Review of Chloroprene (CAS No. 126-99-8) In Support of Summary
Information on the Integrated Risk Information System (IRIS). National Center for Environmental
Assessment, Office of Research and Development. Washington, DC. EPA/635/R-09/010F.

https://iris.epa.gov/static/pdfs/1021tr.pdf and
https://iris.epa.gov/ChemiealLanding/&substance n.mbr= 1.021.

U.S. EPA. (2013, August). Memorandum, "Evaluation of Competitive Harm from Disclosure of "Inputs to
Equations" Data Elements Deferred to March 31, 2015". Retrieved from Available at
https://www.epa.gov/sites/default/files/2015-07/documents/ste

149

-------
U.S. EPA. December 2016. Evaluation of the Inhalation Carcinogenicity of Ethylene Oxide (CASRN 75-21-8) In
Support of Summary Information on the Integrated Risk Information System (IRIS). National Center for
Environmental Assessment, Office of Research and Development. Washington, DC. EPA/635/R-16/350Fa.
https://cfpnb.epa.gov/ncea/iris/iris doeuments/documents/toxreviews/1025tr.pdfU.S. EPA. (2012a).
Regulatory Impact Analysis for the Particulate Matter National Ambient Air Quality Standards. Available
at: https://www3.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf.

U.S. EPA. (2013, August). Memorandum, "Evaluation of Competitive Harm from Disclosure of "Inputs to
Equations" Data Elements Deferred to March 31, 2015". Retrieved from Available at
https://www.epa.gov/sites/default/files/2015-07/documents/ste

U.S. EPA. (2015a). Proposed Rulemaking for Greenhouse Gas Emissions and Fuel Efficiency Standards for

Medium - and Heavy-Duty Engines and Vehicles-Phase 2 (Vols. (EPA-420-D-15-900)). Washington, D.C.:
Office of Transportation and Air Quaility, Assessment and Standards Division. Retrieved 2023, from
https://nepis.epa.gov/Exe/ZyPDF.cgi/P100MKYR.PDF?Dockey=P100MKYR.PDF

U.S. EPA. (2015b). Regulatory Impact Analysis for the Proposed Revisions to the Emission Guidelines for Existing
Sources and Supplemental Proposed New Source Performance Standards in the Municipal Solid Waste
Landfills Sector. Retrieved January 2023, from https://www.regulations.gov/document?D=EPA-HQ-OAR-
2014-0451-0086

U.S. EPA. (2015c). Regulatory Impact Analysis of the Proposed Emission Standards for New and Modified Sources
in the Oil and Natural Gas Sector. Retrieved January 2023, from
https://www.regulations.gov/document/EPA-HQ-OAR-2010-0505-5258

U.S. EPA. (2015a). Proposed Rulemaking for Greenhouse Gas Emissions and Fuel Efficiency Standards for

U.S. EPA. (2016). Ethylene Glycol. Retrieved from https://www.epa.gov/sites/production/files/2016-
09/documents/ethylene-glycol.pdf

U.S. EPA. (2021b). Technical Support Document (TSD) for the Final Revised Cross-State Air Pollution Rule

Update for the 2008 Ozone Season NAAQS Estimating PM2.5- and Ozone-Attributable Health Benefits.
Available at https://www.epa.gov/sites/default/files/2021 -03/documents/estimating_pm2.5-_and_ozone-
attributable_health_benefits_tsd.pdf.

U.S. EPA. (2023a). Regulatory Impact Analysis fo the Proposed National Emission Standards for Hazardous Air
Pollutants: Coal- and Oil-Fired Electric Utility Steam Generating Units Review of the Residual Risk and
Technology Review. Office of Air Quality Planning and Standards. EPA-452/R-23-002.

U.S. EPA. (2023b). Estimating the Benefit per Ton of Reducing Directly-Emitted PM2.5, PM2.5 Precursors and
Ozone Precursors from 21 Sectors. Research Triangle Park, NC: Office of Air Quality Planning and
Standards. Available at: https://www.epa.gov/benmap/estimating-benefit-ton-reducing-directlv-emitted-
pm25-pm25-precursors-and-ozone-precursors.

150

-------
U.S. EPA. (2023c). EPA's "Report on the Social Cost of Greenhouse Gases: Estimates Incorporating Recent
Scientific Advances". Available at: https ://www. epa. gov/environmental-economics/sc ghg U.S. EPA.
(2023d). "Supplementary Material for the Regulatory Impact Analysis for the Final Rulemaking: Standards
of Performance for New, Reconstructed, and Modified Sources and Emissions Guidelines for Existing
Sources: Oil and Natural Gas Sector Climate Review. " EPA Report on the Social Cost of Greenhouse
Gases: Estimates Incorporating Recent Scientific Advances, Washington, DC. doi:Docket ID No. EPA-HO-
OAR-2021-0317.

U.S. GAO. (2020). Social Cost of Carbon: Identifying a Federal Entity to Addess the National Academies'

Recommendations Could Strengthen Regulatory Analysis. GAO-20-254. Retrieved January 2023, from
https://www.gao.gov/assets/gao-20-254.pdf

United Nations. (2022). UN Comtrade Database. Retrieved from https://comtrade.un.org/

USGCRP. (2016). The Impacts of Climate Change on Human Health in the United States: A Scientific Assessment.
(A. Crimmins, J. Balbus, J. Gamble, C. Beard, J. Bell, D. Dodgen,. .. L. Ziska, Eds.) U.S. Global Change
Research Program.

USGCRP. (2018). Impacts, Risks, and Adaptation in the United States: Fourth National Climate Assessment,

Volume II. (D. Reidmiller, C. Avery, D. Easterling, K. Kunkel, K. Lewis, T. Maycock, & B. Stewart, Eds.)
Washington, D.C.: U.S. Global Change Research Program. doi:10.7930/NCA4.2018

Wagner, N. (2020, February). Why the prices of natural and synthetic rubber do not always bounce together.

Retrieved from U.S. Bureau of Labor Statistics: https://www.bls.gov/opub/btn/volume-9/whv-the-prices-of-
natural-and-svnthetic-rubber-do-not-alwavs-bounce-together.htm

Walker, W.R., 2013. The transitional costs of sectoral reallocation: Evidence from the clean air act and the
workforce. The Quarterly journal of economics, 128(4), pp. 1787-1835.

Wittcoff, H. A. (2012). Industrial organic chemicals. Retrieved from John Wiley & Sons:

https://www.academia.edu/download/50154210/Industrial_Organic_Chemistry.pdf

151

-------
United States	Office of Air Quality Planning and Standards Publication No. EPA-

Environmental Protection Health and Environmental Impacts Division 452/R-24-001
Agency	Research Triangle Park, NC	March 2024

152

-------