Prioritizing Justice in New York State Climate Policy A
Prioritizing Justice in New
York State Climate Policy:
Cleaner Air for Disadvantaged
Communities?
Alan Krupnick, Molly Robertson, Wesley Look, Eddie Bautista, Victoria Sanders,
Eunice Ko, Dan Shawhan, Joshua Linn, Miguel Jaller, Narasimha Rao, Miguel
Poblete Cazenave, Yang Zhang, Kai Chen, and Pin Wang
Report 23-12
September 2023
Resources for the Future and New York City Environmental Justice Alliance i
About the Authors
Alan Krupnick is a senior fellow at Resources for the Future (RFF).
Molly Robertson is a research associate at RFF.
Wesley Look is a senior research associate at RFF.
Eddie Bautista is the executive director of the New York City Environmental Justice
Alliance (NYC-EJA).
Victoria Sanders is a reseach analyst at NYC-EJA.
Eunice Ko is the deputy director of NYC-EJA.
Dan Shawhan is a fellow at RFF.
Joshua Linn is a senior fellow at RFF.
Miguel Jaller is an associate professor at University of California, Davis.
Narasimha Rao is an associate professor at Yale University.
Miguel Poblete Cazenave is an assistant professor at VU Amsterdam.
Yang Zhang is a professor at Northeastern University.
Kai Chen is an assistant professor at Yale University.
Pin Wang is a postdoctoral associate at Yale University.
Acknowledgements
We would like to acknowledge the research contributions of Steven Witkin (RFF),
Christoph Funke (RFF), Daniel Chu (NYC-EJA), Kevin Garcia (NYC-EJA), and Annel
Hernandez (NYC-EJA).
We would also like to provide special thanks to all the New York community groups
and policy experts that supported this project. The following groups and consultants
provided valuable time and feedback on environmental and climate justice policy
priorities and the policy landscape in New York: NY Renews, UPROSE, El Puente,
Brooklyn Movement Center, The Point CDC, Nos Quedamos, GOLES, PUSH Bualo,
Long Island Progressive Coalition, Citizen Action of NY, New York Lawyers for Public
Interest, Alliance for Green Energy, Kinetic Communities, Environmental Advocates NY,
Prioritizing Justice in New York State Climate Policy ii
Earthjustice, and Lew Daly (independent consultant). Finally, we would like to thank those
who made this report possible through their financial support including Environmental
Defense Fund (EDF) and the many others who support RFF. The work does not represent
the positions of EDF.
About RFF
Resources for the Future (RFF) is an independent, nonprofit research institution in
Washington, DC. Its mission is to improve environmental, energy, and natural resource
decisions through impartial economic research and policy engagement. RFF is committed
to being the most widely trusted source of research insights and policy solutions leading
to a healthy environment and a thriving economy.
The views expressed here are those of the individual authors and may dier from those of
other RFF experts, its oicers, or its directors.
About NYC-EJA
The New York City Environmental Justice Alliance (NYC-EJA) is a non-profit, 501(c)3
citywide membership network linking grassroots organizations from lowincome
neighborhoods and communities of color in their struggle for environmental justice.
NYC-EJA empowers its member organizations to advocate for improved environmental
conditions and against inequitable environmental burdens by the coordination of
campaigns designed to inform City and State policies.
Sharing Our Work
Our work is available for sharing and adaptation under an Attribution-NonCommercial-
NoDerivatives 4.0 International (CC BY-NC-ND 4.0) license. You can copy and
redistribute our material in any medium or format; you must give appropriate credit,
provide a link to the license, and indicate if changes were made, and you may not apply
additional restrictions. You may do so in any reasonable manner, but not in any way
that suggests the licensor endorses you or your use. You may not use the material for
commercial purposes. If you remix, transform, or build upon the material, you may not
distribute the modified material. For more information, visit https://creativecommons.
org/licenses/by-nc-nd/4.0/.
Resources for the Future and New York City Environmental Justice Alliance iii
Contents
1. Introduction 1
2. The New York Climate Policy and Environmental Justice Landscape 3
2.1. The Climate Leadership and Community Protection Act (CLCPA) 3
2.2. Environmental Justice 4
2.2.1. Community Leadership and Solutions 5
3. Our Research 5
3.1. Methodology 6
4. The Cases 7
4.1. Business-as-Usual (BAU) Case 7
4.2. CAC-Inspired Policy Case (CPC) 8
4.3. Stakeholder Policy Case (SPC) 8
5. Results 14
5.1. Economic Modeling Results 14
5.2. Greenhouse Gas, PM
2.5
, and Precursor Emissions Results 16
5.3. Location of Emissions Changes 18
5.4. Air Quality Results 24
5.4.1. Air Quality Changes in Each Policy Case 25
5.4.2. Outcomes for Disadvantaged Communities 28
5.4.3. Contextualizing Air Quality Changes 31
6. Conclusion 34
7. References 38
Appendix A. Building the Policy Cases 41
Appendix B. Background on Economic Models 43
B.1. Economic Models 43
B.1.1. Power Sector 43
B.1.2. Light-Duty Vehicles 43
B.1.3. Medium- and Heavy-Duty Vehicles 45
B.1.4. Residential Buildings 48
Prioritizing Justice in New York State Climate Policy iv
Appendix C. Background on Air Quality Modeling 52
Appendix D. Identifying Disadvantaged Communities 55
Appendix E. Supplementary Methodologies 61
E.1. Model Integration and Coordination 61
E.2. Ancillary Pollutant Valuation 61
E.3. Methane 62
Appendix F. Comparison with New York State’s Analysis 63
Appendix G. Research Limitations and Caveat 65
Appendix H. Economic Modeling Results 68
H.1. Electricity Sector 68
H.1.1. Electricity Demand and Price 68
H.1.2. Generation Mix 69
H.2. Residential Building Sector 71
H.2.1. Electricity Demand 72
H.3. Transportation Sector 73
H.3.1. Light-Duty Vehicles 73
H.3.2. Medium- and Heavy-Duty Vehicles 74
Appendix I. Greenhouse Gas, PM
2.5
, and Precursor Emissions Results 77
I.1. Power Sector Detail 79
I.2. Residential Buildings Sector Detail 79
I.3. Transportation Sector Detail 80
I.3.1. Light-Duty Vehicle Fleet 80
I.3.2. Medium and Heavy-Duty Vehicle Fleet 80
Appendix J. Location of Emissions Changes 81
J.1. Electricity Sector 81
J.2. Residential Buildings Sector 86
J.3. Transportation Sector 87
J.3.1. Light-Duty Vehicle Fleet 87
J.3.2. Medium- and Heavy-Duty Vehicle Fleet 88
Resources for the Future and New York City Environmental Justice Alliance v
Appendix K. Additional Results Context: Nonlinearities and Excluded Emissions 91
K.1. Nonlinearities 91
K.2. Modeling Choices 91
K.3. Traveling Air Pollution 94
Appendix L. Distribution of PM
2.5
Concentration Reductions by
Scenario and DACs vs. Non-DACs 96
Prioritizing Justice in New York State Climate Policy 1
1. Introduction
As a result of historically unjust systems and policies, the neighborhoods where low-income
communities and communities of color live, work, learn, and play are often sites for or
aected by polluting infrastructure, vehicle congestion, and other environmental hazards.
Racist systems and policies along with economic discrimination continue to diminish the
health and quality of life of communities of color and low-income communities and make
them more at risk to other hazards like climate change (Peña-Parr 2020; Donaghy et al.
2023). As fossil fuel consumption and pollution have increased exponentially over the past
century, not only has the climate change outlook worsened, but vulnerable communities
have also disproportionally suered injury, disease, death, displacement, and loss of property
because of these same trends (Resnik 2022).
To address the growing inequities, community leaders have advocated for clean air, water,
and land and fought against trash incinerators, highways, and fossil fuel power plants being
placed in their neighborhoods. National and state climate policy has rapidly evolved in the
past decade—not only are eorts to reduce greenhouse gases (GHGs) increasing, but
recent eorts have also put racial, social, and economic justice at the forefront of the policy
conversation. With leadership from communities aected first and worst by pollution and
other environmental and climate risks (often referred to as “frontline communities”), policies
are now expected to achieve not only climate goals but also improvements in environmental
and economic justice. Another key goal is to ensure a just transition, making sure that
low-income communities and communities of color don’t face disproportionate burdens
or exclusion from the benefits of the transition from a fossil fuel economy to a resilient,
equitable, and regenerative society.
One of the most prominent examples of justice-oriented climate policy is New York State’s
recent climate law, the Climate Leadership and Community Protection Act (CLCPA). As
the state moves to implement this groundbreaking law, rigorous research and analysis are
needed to shed light on policy design options that can achieve the dual goals of cutting GHG
emissions and improving air quality and other public health outcomes for “disadvantaged
communities,” as defined by the state. This requirement is the motivation for this study.
Bringing together leading environmental justice (EJ) advocates, economic researchers,
public health scientists, and air quality modelers (see Appendix A), our study investigates
EJ impacts in the context of the CLCPA. Specifically, we model the impact of policies on
the electric power, on-road transportation, ports, and residential building sectors, and the
resulting fine particulate (PM
2.5
) air pollution experienced by disadvantaged communities
and nondisadvantaged communities alike. We compare two policy scenarios: one inspired
by the Climate Action Council’s (CAC) scoping plan for implementing the CLCPA, and one
inspired by the policy priorities of environmental and climate justice stakeholders in New
York. A key question driving our research is, to what extent do the policies contemplated in
the CLCPA scoping plan indeed (as required by the CLCPA) “not result in a net increase in
copollutant emissions or otherwise disproportionately burden disadvantaged communities”
and/or “prioritize measures to maximize net reductions of … copollutants in disadvantaged
communities”? And furthermore, how do the scoping plan policies compare—in this regard—
with those policies advocated by New York’s EJ stakeholders?
Resources for the Future and New York City Environmental Justice Alliance 2
Our analysis has revealed 10 major findings and insights:
1. GHG reductions in 2030 are substantial under both cases relative to the
business-as-usual (control) case but are greater under the stakeholder case
(58 percent vs. 34 percent reduction).
2. The stakeholder case leads to greater statewide emissions reductions for all
PM
2.5
precursors (nitrogen oxides, NO
x
; sulfur dioxide, SO
2
; and volatile organic
compounds, VOCs) than the CAC-inspired case.
3. The stakeholder case leads to greater statewide PM
2.5
concentration
reductions (air quality improvements) than the CAC-inspired case in 99
percent of census tracts.
4. In the CAC-inspired case, average air quality improvements in disadvantaged
communities are comparable to the improvements made in nondisadvantaged
communities. In the stakeholder case, improvements in disadvantaged
communities are greater than those in nondisadvantaged communities.
5. Although, on average across the state, both cases improve air quality (reduced
PM
2.5
concentrations), some census tracts do experience a worsening of air
quality (increased PM
2.5
concentrations): in the CAC-inspired case, about 6
percent of the state’s roughly 5,000 tracts experience worse air quality, a
fourth of which are disadvantaged communities, whereas in the stakeholder
case, only three census tracts experience worse air quality, none of which are
disadvantaged communities.
6. The most vulnerable communities (the top 10 percent of tracts in the state’s
social vulnerability measure, and the 10 percent with historically worse
air quality) experience particularly pronounced improvements under the
stakeholder case but experience average air quality improvements in the CAC-
inspired case.
7. Both policy cases make air quality improvements in disadvantaged communities,
but the impacts are too evenly shared with nondisadvantaged communities to
reverse the historical disparity in air pollution concentrations.
8. In an illustrative calculation, the stakeholder case oers the greatest public
health benefits to elderly Black New Yorkers relative to their Hispanic, Asian,
and white counterparts. Although 22 percent of New York City’s 65 and older
population is Black, this group accounts for 42 percent of the avoided deaths
from PM
2.5
reductions; white residents make up 41 percent of the city’s 65 and
older population but account for 37 percent of the avoided deaths.
9. Both cases, as modeled, meet the CLCPA EJ goal to “not disproportionately
burden disadvantaged communities,” if we define burden to mean increasing
PM
2.5
concentrations.
10. The greater benefits associated with the stakeholder case relative to the CAC-
inspired case require greater investment, since the stakeholder policies are
more ambitious and oer more generous subsidies to encourage electrification
of buildings and vehicles. The emissions reductions and air quality improvements
for disadvantaged communities certainly favor the stakeholder case, driving
valuable health and equity outcomes that may well outweigh the policy costs, but
estimating benefits and costs was beyond the scope of this project.
Prioritizing Justice in New York State Climate Policy 3
2. The New York Climate Policy and
Environmental Justice Landscape
2.1. The Climate Leadership and Community
Protection Act (CLCPA)
The CLCPA, passed in 2019 after years of debate and advocacy, sets ambitious GHG
emissions goals, including an 85 percent reduction in economywide GHG emissions
by 2050, 70 percent renewable energy by 2030, and a 100 percent zero-emissions
electricity sector by 2040. The law also mandates that the state achieve 9,000 MW
of oshore wind by 2035; 3,000 MW of energy storage by 2030; 6,000 MW of solar
by 2025; and 22 million tons of carbon reduction through energy eiciency and
electrification.
In addition, the law explicitly sets goals for environmental and climate justice—
addressing the disinvestment and disproportionate environmental burdens that
communities of color and low-income communities have experienced. The preamble
states that “actions undertaken by New York State to mitigate GHG emissions should
prioritize the safety and health of disadvantaged communities, control potential
regressive impacts of future climate change mitigation and adaptation policies on
these communities, and prioritize the allocation of public investments in these areas.”
Not only does the CLCPA require reductions in GHGs like methane and carbon dioxide
(CO
2
), it also targets local air pollutants—what the law refers to as copollutants—
such as PM
2.5
and SO
2
. The law specifically directs the New York State Department
of Environmental Conservation (NYSDEC) to “ensure that activities undertaken to
comply with the regulations do not result in a net increase in copollutant emissions
or otherwise disproportionately burden disadvantaged communities” and to
“prioritize measures to maximize net reductions of GHG emissions and co-pollutants
in disadvantaged communities.” Stated simply, the CLCPA requires state climate
regulations to prioritize air quality in disadvantaged communities—including by
requiring that environmental burdens are not shifted from wealthier communities to
lower-income, minority communities. The CLCPA also establishes a Climate Justice
Working Group (CJWG) tasked with establishing criteria for identifying disadvantaged
communities and representing EJ priorities throughout the various stages of CLCPA
implementation. The final criteria were adopted on March 27, 2023, after a public
comment period and public hearings held across New York State. Additionally, the law
stipulates that 35 to 40 percent of the benefits and investments go to disadvantaged
communities.
Resources for the Future and New York City Environmental Justice Alliance 4
2.2. Environmental Justice
The Climate Justice Alliance, a national network of frontline communities and
organizations demanding a just transition, defines environmental justice as the right of
all people, regardless of race or socioeconomic background, to live, work, and play in
communities that are safe, healthy, and free of life-threatening and harmful conditions.
The alliance works to realize a vision for a just transition, which is a “place-based
set of principles that build economic and political power to shift from an extractive
economy to a regenerative economy.” The alliance states that the “Just Transition must
advance ecological resilience, reduce resource consumption, restore biodiversity and
traditional ways of life, and undermine extractive economies, including capitalism, that
erode the ecological basis of our collective well-being. This requires a re-localization
and democratization of primary production and consumption by building up local
food systems, local clean energy, and small-scale production that are sustainable
economically and ecologically. This also means producing to live well without living
better at the expense of others.”
The Climate Justice Alliance and the EJ movement more broadly are led by (and
advocate for) frontline communities who experience climate and environmental hazards
first and worst. EJ communities are frontline communities: low-income communities
and communities of color who face disproportionate exposure to environmental hazards
due to both intentional design and structural racism. The Climate Justice Alliance
describes the origins of the EJ movement
as growing “out of a response to the system
of environmental racism where communities of color and low-income communities have
been (and continue to be) disproportionately exposed to and negatively impacted by
hazardous pollution and industrial practices. Its roots are in the civil rights movement
and are in sharp contrast to the mainstream environmental movement, which has failed
to understand or address this injustice. The EJ movement emphasizes bottom-up
organizing, centering the voices of those most impacted, and shared community.
Historically, low-income communities and communities of color have been systematically
disinvested from, with racist policies and practices such as redlining used to value
certain neighborhoods and residents above others (Homan et al. 2020). These policies
and systems have caused wealth and resource gaps that endure to this day, investing
in quality-of-life improvements in wealthier areas while pushing polluting industries into
lower-income communities (Homan et al. 2020; Nardone et al. 2020; Schell et al. 2020).
We see these disparities reflected in the location of power plants, transportation depots,
and city parks. The impacts of this unequal investment are clear in public health data,
with environmentally driven poor health outcomes like asthma most prevalent in EJ
communities (New York City Department of Health and Mental Hygiene 2020).
Policymakers have paid little attention to these historical racist practices and the
resulting disparities in distributional eects that have maintained and widened resource
gaps. Without intentional consideration and targeted policy implementation, an unjust
distribution of costs and benefits of policies and programs will continue to cause EJ
communities to experience greater burdens than their white and wealthier counterparts.
Prioritizing Justice in New York State Climate Policy 5
2.2.1. Community Leadership and Solutions
As policymakers seek solutions to climate change, EJ experts and community
leaders have emphasized the importance of doing so in a way that centers racial and
economic justice, addressing this history of abuse. EJ advocates have been calling for
climate policies that not only reduce GHG emissions but also ensure that the costs
of an energy transition do not fall unduly on disadvantaged communities, and make
improvements in the environmental conditions, public health, and adaptive capacity of
disadvantaged communities.
EJ and climate justice stakeholders in New York have played a central role in
representing the needs of underserved communities, as reflected in numerous
provisions in the CLCPA. In the creation and execution of this research project, EJ
and climate justice stakeholders were centrally involved to ensure that community
concerns and expertise were woven into the fabric of the research design and process.
It was crucial that the EJ stakeholder policy case reflect what these EJ stakeholders
are fighting for and most want to see enacted to protect their communities. Accepting
and incorporating their knowledge and leadership are important steps in the process
of dismantling historical inequities and ensuring that all parties involved have a seat at
the negotiating table for environmental policies such as the CLCPA.
3. Our Research
As New York moves to meet the CLCPA decarbonization goals, the state will implement
policies that phase out behaviors and technologies that generate GHG emissions.
Our research seeks to inform this process by analyzing the GHG and air pollution
impacts of three policy cases: a business-as-usual case, meant to represent what
would happen to emissions and air quality without the actions contemplated in the two
policy cases; the stakeholder policy case, meant to reflect EJ policy priorities; and the
CAC-inspired policy case, meant to reflect a plausible set of policies coming out of the
state’s scoping plan process, which defines the policy goals and tools that ought to be
used to meet the legal requirements of the CLCPA. Details on the policy cases can be
found in Section IV. We focus on how the policy cases aect PM
2.5
concentrations in
communities at the census tract level across the state.
To compare policy outcomes between disadvantaged communities and
nondisadvantaged communities, we use an EJ screen (referred to in this report as a
climate health and vulnerability index) and an EJ map. The EJ screen and map reflect
the disadvantaged community criteria and methodology developed by the state
through the CJWG. Using this EJ screen and map, we track the eects of changing
PM
2.5
concentrations on disadvantaged and other communities.
Resources for the Future and New York City Environmental Justice Alliance 6
Several characteristics of our research set it apart from other research eorts. Our
contribution to examining the outcomes of decarbonization policies on EJ communities
at a state level is unique. Additionally, we use a combination of behavioral models and
one of the most sophisticated air quality models to assess and trace the consequences
of the two policy cases for disadvantaged communities (DACs) and non-DACs. Further,
mapping these results visually at the 4km
2
scale gives readers an unprecedented
ability to assess and understand the geographic distribution of results.
3.1. Methodology
This project involves several models that work together to estimate the emissions
and air quality impacts of dierent policies. The first step in our research is to build
and compare the three cases—business-as-usual case, CAC-inspired policy case,
and stakeholder policy case—in consultation with New York policy experts (see
Appendix A and Section IV). Next, we use four economic models that estimate the
future emissions impacts of dierent policies in each of the three cases (see Appendix
B). Models of the electric power sector, the light-duty vehicle market and fleet, the
medium- and heavy-duty vehicle market and fleet, as well as nonmarine port activities
and the residential building sector are included in our analysis.
The emissions estimates produced by the economic models are used as inputs in an air
quality model that projects changes in PM
2.5
resulting from meteorological conditions,
chemical reactions, and other factors (see Appendix C). We then incorporate the
CJWG’s criteria for identifying DACs in our EJ screening and mapping tool to analyze
outcomes for the three cases in 2030 (see Appendix D). By comparing emissions
levels and air quality changes in DACs with non-DACs, we can ascertain the impacts of
dierent implementations of the CLCPA for vulnerable communities. Figure 1 depicts
the flow of research for this project.
Figure 1. Research Process
Prioritizing Justice in New York State Climate Policy 7
The research and findings from this project contribute to the current body of work
investigating the impacts of the state’s climate policies. The most expansive work in this
space was conducted by New York State Energy Research and Development Authority
(NYSERDA) and NYSDEC in partnership with E3 in 2021 to inform the scoping plan
process. Their work focused on establishing estimated pathways of decarbonization
across all sectors aected by the CLCPA. Our work is more focused on a few critical
sectors for which we have robust behavioral models. For a full list of dierences with the
state-sponsored analysis, see Appendix F.
We also acknowledge that important boundaries to our research may influence the
interpretation of the results. For example, our air quality modeling is at a 4km
2
grid
resolution, which in some cases is larger than a DAC boundary. We use one of the
most advanced air quality models for our estimates, which incorporates detailed
representations of atmospheric science and chemical processes. We have selected a
spatial resolution that preserves the accuracy of that model. To aid in the interpretation of
our work, we describe the limitations and caveat for our analysis in Appendix G, including
a small error in the transportation emissions used as an input in the air quality model.
4. The Cases
We focus on 2030 as the year for modeling economic activity and related air pollutant
concentrations throughout New York State. Our modeling begins with a business-as-usual
case, which includes policies in place prior to the passage of the CLCPA and continues
through 2030. We also model two policy cases: the CAC-inspired policy case, which
assumes policies are implemented to meet the goals of the CLCPA as stated in the CAC
scoping plan, and the stakeholder policy case, which assumes policies are implemented
in line with the priorities and preferences of prominent EJ advocates in New York. The
details of all three scenarios are described below.
4.1. Business-as-Usual (BAU) Case
The models incorporate a variety of forward-looking economic and demographic
projections to anticipate future conditions. For example, future oil and natural gas prices,
population projections, and income and wage growth are all considered. The energy
models in this project, like many others, use Energy Information Agency (EIA) projections
for these high-level drivers of change. The projections are found in EIA’s Annual Energy
Outlook (AEO), which is based on runs of the National Energy Modeling System.
The latest AEO available at the start of our project was the AEO2021 and embodies the
agency’s most recent economic activity projections. EIA limits its modeling to sets of
policies (mostly federal, and in some cases, state) that are already in place, not policies
that might be implemented in the future. Its projections do not assume that a carbon tax
or any other federal CO
2
reduction plan is implemented.
Resources for the Future and New York City Environmental Justice Alliance 8
Our modeling work took place over the course of 2022. We were required to lock in
assumptions about what federal and state policies to model in the BAU case in January
2022. As a result, the eects of neither the Infrastructure Investment and Jobs Act
nor the Inflation Reduction Act are included in our baseline. We made the following
adjustments from the AEO2021 reference case to adapt it to our research:
• We used the AEO2020 projections for key transportation parameters, since these
assume the Obama-era fuel economy standards are in place, rather than the
Trump administration rollbacks represented in AEO2021.
• We used the reference cases for both AEO2020 and AEO2021 for oil and gas
supply based on alignment with other modeling exercises and to avoid potentially
biasing results by selecting an AEO case with high or low growth.
1
• We used E3’s New York–specific assumptions about the transportation sector,
including the increase in vehicle miles traveled in the state.
4.2. CAC-Inspired Policy Case (CPC)
The CPC was developed based on the Climate Action Council’s (CAC) draft scoping
plan and conversations with New York State policy experts. We identified policies in
the relevant sectors and adjusted where needed to fit the needs of our models. Table 1
provides a list of the CPC policies we model with brief descriptions.
Not all these policies are explicitly mentioned in the scoping plan. Our modeling
work is based on behavioral responses to economic policies, so we had to add detail
and specificity to policies where none existed. The CPC represents one reasonable
interpretation of how the priorities in the scoping plan may be executed. The details
were established using a mix of New York policy proposals, examples from other state
and federal climate policy proposals, and feedback from New York policy experts.
4.3. Stakeholder Policy Case (SPC)
The SPC was developed with various EJ organizations operating throughout New York.
These groups identified local and statewide priorities and gave feedback on the draft
scoping plan’s proposed policies in the sectors we analyzed through several workshops
and written comments. Table 1 lists the SPC policies that we included in our modeling
and reasons why they diverge from the CPC.
There are many policies considered essential by climate and environmental justice
advocates that we were unable to integrate into our research. We excluded policies for
three primary reasons: (1) the policy applies to sectors that we are not modeling (e.g.,
agriculture policies), (2) we were unable to estimate accurate emissions changes from
1 The dierence in 2030 CO
2
emissions is about 5.6 percent between the reference and
high growth cases and about 10 to 11 percent between the high and low cases. Thus, we
favor an unbiased choice—the reference case.
Prioritizing Justice in New York State Climate Policy 9
the policy because of model limitations (e.g., public vehicle fleets are excluded in our
transportation model), and/or (3) the objective of the policy aects an outcome other
than emissions reductions or air quality improvements (e.g., job training programs).
The exclusion of these policies is in no way a reflection on their importance, feasibility,
or impact but rather a result of the reality of our modeling limitations. See Appendix
A for a list of policies that the stakeholder groups identified as critical to their
communities that we were unable to include in our models.
Furthermore, some if not all of the policies we model will aect conditions and
outcomes that we do not estimate. For example, regional employment, water pollution
levels, and health outcomes are all important metrics of a thriving community that are
shaped by climate policies but not included in our modeling. Consequently, although
we can make inferences about some of these outcomes, they are excluded from our
policy analysis. The decision not to model these outcomes is not a judgment on their
significance; they are simply outside our research scope.
Table 1. Details of Policy Cases
Policy CPC SPC Motivation for stakeholders
Economy-wide
Carbon regulation
2
An economy-wide carbon fee is
established to achieve emissions
reductions across sectors we
are analyzing. Fee is $25/ton
in 2030.
3
Fee was determined
iteratively with our models to
meet state’s target after other
policies were in place, similarly
to how carbon cap would be
modeled.
Carbon fee introduced in 2023 at
$55/ton, increases 5% annually
to $77/ton in 2030.
Copollutant prices ($2017):
NO
x
: $9,025/short ton
SO
2
: $36,382/short ton
PM
2.5
: $231,965/short ton
SPC carbon-pricing scheme
reflects ambition of CCIA
polluter fee. It prices
copollutants based on social
marginal cost in addition to CO
2
.
This could also be achieved
with an economy-wide cap on
pollutants.
2 Economy-wide carbon regulation (cap or fee) in both policy cases is accompanied by a border carbon adjustment for
imported and exported electricity. That border carbon adjustment is a fee on electricity imports to New York State and
an equal subsidy on electricity exports from New York State. The level of the fee and subsidy is the price on New York
in-state CO
2
emissions times the average GHG emissions rate of electricity generation in adjacent regions. In this GHG
emissions rate calculation, each pound of methane is counted the same as 30 pounds of CO
2
.
3 This price was set to meet emissions reductions targets only in the sectors we model. A carbon price may play a larger
role in eliciting emissions reductions in other sectors. The relative cost of emissions reductions in those sectors may
require a higher price to achieve the desired result. Observing further emissions reductions with a low price in our policy
case implies that there are relatively low-cost reductions that were not incentivized by the other modeled policies.
Several policies included in the policy case force relatively high-cost emissions reductions (like the ZEV mandate in the
transportation sector) that would not be achieved with the modeled carbon price alone.
Resources for the Future and New York City Environmental Justice Alliance 10
Electricity sector
Clean energy
standard
70% of electricity must come
from clean energy sources, as
defined in CLCPA.
Same as CPC. —
Distributed solar
target
Mandates 10 GW solar installed
by 2025.
Same as CPC. —
Battery storage
target
Mandates 3 GW battery storage
installed by 2030.
Same as CPC. —
Oshore wind
target
Mandates 9 GW oshore wind
installed by 2035.
Same as CPC. —
Transmission
investment
Two new DC lines will be built
in New York: Clean Path and
Champlain Hudson Power
Express.
Same as CPC. —
Nuclear subsidies
4
Extend ZECs for nuclear until
after 2030; extend nuclear
licenses to 80 years.
End ZECs for nuclear in 2029
when they are set to expire; do
not extend nuclear licenses;
no new generating units to be
developed in NYS.
SPC reflects lack of consensus
on how supporting nuclear
aects electricity costs and
trade-os with supporting other
technologies.
Demand
response policy,
flexible demand,
distributed energy
resource subsidy
5
Shift 6% of peak electricity to
o-peak times based on New
York integration analysis flexible
load assumptions in 2030
(developed by E3 Consulting).
Same as CPC. —
Peaker plant
policy
6
Shut down fossil fuel peaker
plants in line with stated policy,
enforcing NYC NO
x
rule and
Pollution Justice Act of 2022
(Brisport’s S4378B).
All NYS fossil fuel peaker plants
close by 2030.
Peaker plants disproportionately
contribute to air pollution in
disadvantaged communities.
New combustion
fuels, CCUS
Allow biofuels, natural
gas, hydrogen, and CCUS
if economical after other
abatement policies are in place.
Ban use of new natural gas and
CCUS in power sector by 2025.
SPC reflects more ambitious
transition away from polluting
generators, does not support
investment in technologies
that may prolong fossil fuel use
(deemed as “false solutions”).
4 Our modeling took place before the passage of the Infrastructure Investment and Jobs Act or the Inflation Reduction
Act, so the Civil Nuclear Credit program and the nuclear tax credit are not included in our modeling. These would likely
reduce the costs of sustaining nuclear in the CPC and prevent retirement of nuclear in the SPC.
5 Our power sector model has limited ability to model demand response and incentives for distributed energy resources
directly, so we model these as an assumed shift in peak demand. We use the assumptions from the state analysis for this
purpose.
6 Peaker plants generally run only when demand for electricity is very high, or “peaking.” They are generally less eicient
than other fossil fuel generators because they are designed to ramp up energy production quickly when needed.
Prioritizing Justice in New York State Climate Policy 11
Residential buildings
Heat pump
subsidy
7
Starting in 2023, provide $4,750
subsidy for households with 80%
or less of state median income
and multifamily households;
provide $3,000 subsidy for all
other single-family homes.
Starting in 2023, provide full
heat pump cost subsidy to
households with 60% percent
or less of state median income;
$4,750 subsidy to households
with 80% percent or less of
median income; $3,000 subsidy
for all other single-family homes.
SPC reflects more ambitious
support for electrification of low-
income households.
Shell eiciency
upgrades
8
By 2030, shell eiciency
upgrades targeted to 25%
of homes based on building
vintage.
By 2030, shell eiciency
upgrades targeted to homes in
top 25% of energy burden.
SPC prioritizes upgrades for
highest-burdened homes, rather
than most ineicient homes.
Fossil fuel
phaseout
Before 2030, no bans on fossil
fuel appliances.
Starting in 2023,
9
households
cannot replace fossil fuel
appliances at end of their
useful life with more fossil fuel
appliances.
SPC reflects more ambitious
timeline for replacing fossil
fuel technology in residential
buildings.
Building standards
Starting in 2027, NYSERDA
Stretch Code is enforced for
new residential construction
standard.
Same as CPC. —
7 Heat pumps are assumed to be air-source heat pumps, and subsidies are assumed to fully cover the costs associated
with building upgrades needed to install heat pumps.
8 Shell eiciency upgrades are defined as an upgrade of the building standard when the home was built, to the latest
NYSERDA stretch building code. The model considers the eiciency associated with each of these standards.
9 At the time of constructing this policy case, the state was considering a fossil fuel hook-up ban for residential buildings. It
was not included in the 2023 budget but it was after we modeled the SPC.
Resources for the Future and New York City Environmental Justice Alliance 12
Light-duty vehicles
California’s
Advanced Clean
Cars 2 Rule
By 2030, 50% of new-vehicle
sales are PEV (expected to
require 100% ZEV sales by
2035).
Same as CPC. —
Feebate for ZEVs
Means test for rebates and
vouchers: Subsidy begins
at $5,000 for lowest income
group and declines to $1,000
for highest income group, in
increments of $1,000.
Tiered fee system for new-
vehicle purchases based on
vehicle miles per gallon. Reduced
fee based on means test or
household income (exemptions:
100% for low income, 50% for
middle income, 0% for high
income).
Same as CPC. —
Scrappage
incentive
No scrappage incentive.
Subsidy per vehicle (means
tested): $3,000 for households
above 200% of federal poverty
line, $5,000 for households
below 200%.
Eligible vehicles:
any ICE vehicle at
least 15 and not
more than 25 years
old
Scrappage incentives can
accelerate retirement of high-
emissions vehicles and may
provide greater subsidies to low-
income households with older
vehicles.
Electricity rates for
EVs*
Assume 25% reduction in
electricity rates for EV charging
for all households.
Free electricity for EV charging
for all low- to middle-income
households.
SPC reflects more targeted aid
to low-income households.
Infrastructure
investments*
Grants up to $2,000 for Level 2
home charger installation.
Same as CPC. —
Bus service
expansion
By 2040, double service
availability and accessibility of
municipally sponsored upstate
and downstate suburban public
transportation services.
Same as CPC. —
Clean fuel
standard
Standard follows Senator Kevin
Parker’s proposed S4003A
(2019-20 legislative session),
likely with less aggressive
decarbonization pathway.
No low-carbon or clean fuel
standard implemented.
SPC reflects EJ concern that
low-carbon fuel incentives will
delay full electrification.
Prioritizing Justice in New York State Climate Policy 13
Medium- and heavy-duty vehicles
California’s
Advanced Clean
Trucks Rule
New-ZEV sales goals:
By 2030, Classes 2b and 3, 35%;
Classes 4-6, 50%; Classes 7–8
(long haul), 35%.
By 2035, 80% for all classes.
New-ZEV sales goals for all
classes:
By 2030, 50%.
By 2035, 80%.
SPC reflects more ambitious
ZEV goals for 2030 in the MHDV
sector.
California’s
Advanced Clean
Fleets Rule
By 2045, for vehicle fleet owners,
100% of new MHDV purchases
must be ZEV.
Same as CPC.
Feebate for ZEVs
Non-ZEV vehicles incur 5%
purchase fee (assumed as
increased purchase cost).
Purchase incentive levels vary by
vehicle class and year.
Same policy design, with
higher incentives (as with CPC,
purchase incentive levels vary by
vehicle class and year).
SPC incentives reflect more
ambitious ZEV targets.
Investment in ZEV
infrastructure
Grants up to ~$25k for each new
MHDV for charging.
Same as CPC. —
Electricity rates for
EVs*
Reduced electricity rates for
MHDV charging.
Same as CPC. —
Public financing
for EV
procurement*
Assume zero cost of capital for
ZEVs (EV, H2).
Same as CPC. —
Clean fuel
standard
Standard follows Senator Kevin
Parker’s proposed S4003A
(2019-20 legislative session),
likely with less aggressive
decarbonization pathway.
No low-carbon or clean fuel
standard implemented.
SPC reflects EJ concern that
low-carbon fuel incentives will
delay full electrification.
Port electrification
By 2030, assume 100%
electrification of equipment
10
purchased new or used.
Same as CPC. —
Notes: CCIA = Climate Community and Investment Act; CCUS = carbon capture, utilization, and storage; EV = electric vehicle;
H2 = hydrogen vehicle; ICE = internal combustion engine; LDV = light-duty vehicle; MHDV = medium- or heavy-duty vehicle;
PEV = plug-in electric vehicle; ZEC = zero-emissions credit; ZEV = zero-emissions vehicle.
10 This only includes equipment and terminal vehicles (any type F vehicle that operates within terminals). Drayage is
included in the overall truck flows, and so is subject to all of the policies (most prominently, the Advanced Clean Trucks
rule and feebate) listed above that apply to MHDVs.
Resources for the Future and New York City Environmental Justice Alliance 14
5. Results
This results section has four subsections: Economic Modeling Results, which describes
estimated changes in energy demand and technology adoption across our modeled
sectors; Greenhouse Gas, PM
2.5
, and Precursor Emissions Results, which describes
estimated emissions changes in our modeled sectors; Location of Emissions Changes,
which describes the location of estimated changes in PM
2.5
emissions; and finally Air
Quality Results, which describes estimated changes in PM
2.5
concentrations across
community types and provides context for understanding the public health benefits of
changes in air quality.
Our results focus on the dierences between our two main policy cases (CPC and
SPC), with reference to the BAU case. Significant changes in technology adoption and
emissions are estimated to already occur under the BAU case, relative to the historical
baseline. For example, even under the BAU, our modeling projects that New York State’s
electricity sector CO
2
emissions will fall to 17 million short tons in 2030, down from 31
million in 2020 (US EIA 2023), primarily as a result of the addition of solar and wind
generation capacity. State policy to close peaker plants in areas with high population
density also contributes to the steep decline in SO
2
and NO
X
by 2030 under the BAU.
Additionally, federal vehicle emissions standards and the continual retirement of older,
more polluting vehicles leads to significantly lower emissions from cars and trucks.
In the residential sector, the transition to natural gas furnaces in lieu of traditional oil
furnaces significantly reduces PM
2.5
and SO
2
by 2030, even without additional policy.
5.1. Economic Modeling Results
The policies we modeled have a wide range of ambition and vary in their timelines
for implementation. These results illustrate how far each policy case goes in pushing
economic behavior that will lead to decarbonization and air quality improvements. The
full sectoral analysis of economic results can be found in Appendix H. Figure 2 shows a
summary of the key technologies and their adoption rates across each policy case.
Both policy cases prompt a dramatic increase in clean energy generation relative to the
BAU. Compared with the CPC, SPC policies boost renewable generation and storage
capacity. Relative to the CPC, the SPC delivers a 30 percent boost for solar, a 40
percent boost for wind, and nearly a 200 percent increase in storage capacity. The SPC
also cuts nuclear, natural gas, and waste and biomass generation relative to the CPC
(roughly 35 percent less for nuclear and 30 percent less for natural gas and waste).
Prioritizing Justice in New York State Climate Policy 15
Figure 2. Clean Technology Adoption Rates, by Policy Case
(Percentage)
Heat pump adoption is also higher in both policy cases relative to the BAU. The SPC
has the highest heat pump subsidies for low- and middle-income households; the
CPC subsidy level is more modest. The SPC also includes a ban on new gas furnace
purchases, rapidly phasing out fossil fuel heating. This leads to an approximately
90 percent heat pump adoption rate in the SPC, compared with a 54 percent
adoption rate in the CPC. In addition to the statewide adoption rates, we find a range
of adoption across counties in each case. For instance, the adoption rate varies across
counties from 77 to 96 percent in the SPC, from 27 to 78 percent in the CPC, and from 2
to 15 percent in the BAU case. The higher adoption rates tend to be in the southeastern
part of the state, such as Staten Island and Long Island.
Both policy cases also encourage adoption of light-duty zero-emissions vehicles
(ZEVs). In the BAU, New York has about 241,000 EVs—about 2 percent of all on-road
vehicles. The CPC and SPC yield roughly four times more EV sales than the BAU,
driven by their more ambitious ZEV standards. Compared with the BAU, the policy
cases also increase the average fuel economy of on-road vehicles by about 15 percent.
The largest dierence between the policy cases is in fuel consumption, which is driven
by the dierent prices on carbon emissions. Fuel consumption is about 6 percent lower
in the CPC and 12 percent lower in the SPC compared with the BAU. The SPC reduces
fuel consumption more than the CPC because of its higher carbon price.
Similar to the light-duty vehicle findings, the mandate for zero-emissions medium- and
heavy-duty vehicles (MHDVs) is a primary driver of the shift to a cleaner fleet. Because
of the MHDV ZEV rule (see Table 1), by 2030, it is expected that about 14 percent
and 13 percent of the fleet will be ZEV (mostly battery electric) in the SPC and CPC,
respectively. Key drivers of the dierence between cases (although small) are the
Resources for the Future and New York City Environmental Justice Alliance 16
more ambitious ZEV sales mandate and carbon price, as well as the copollutant fees, of
the SPC; and the low-carbon fuel standard program (LCFS), which is specific to the CPC.
Our findings suggest that significant financial incentives will be required to achieve the
desired ZEV adoption. The models estimate that the various fees considered (on carbon,
copollutants, and internal combustion engine vehicles) and the existing BAU vehicle
incentives programs (e.g., New York City Clean Truck, New York Truck Voucher Incentive
Program) will not be enough to fulfill the mandate.
Both policy cases increase total New York electricity demand because of the high rates
of electrification in the residential and transportation sectors. Under the CPC, electricity
demand increases by 17 percent, and under the SPC, by 29 percent, compared with the
BAU. This leads to greater generation needs and contributes to higher electricity prices.
Our research did not include a comprehensive cost-benefit analysis, but we can observe
that the more ambitious investments in the SPC are associated with higher costs. Both
policy cases lead to modest increases in wholesale electricity prices, compared with the
BAU (a 10 percent increase for the CPC and an 18 percent increase for the SPC). The
subsidy levels for heat pumps for low- and middle-income households are much higher in
the SPC than in the CPC, leading to higher government spending. The higher carbon price
in the SPC also contributes to higher fuel costs, which reduces vehicle miles traveled in the
transportation sector.
5.2. Greenhouse Gas, PM
2.5
, and Precursor Emissions
Results
Emissions of multiple types are expected to decline as a result of the CLCPA. That said,
the dierent policy cases lead to significantly dierent emissions outcomes. For example,
in 2030, CPC carbon emissions reductions are about 30 percent below the BAU, and
SPC carbon reductions are estimated to be about 54 percent below the BAU. The 2030
percentage reduction below the BAU for methane is even more dramatic in the SPC (91
percent reduction) than in the CPC (31 percent reduction).
PM
2.5
and precursors are also significantly aected by the dierent policy cases. The
CPC creates estimated statewide reductions below the BAU of 25, 18, and 42 percent for
SO
2
, NO
x
, and direct PM
2.5
, respectively. The SPC creates estimated reductions of 52, 32,
and 75 percent for the same pollutants, relative to the BAU. Figure 3 shows statewide 2030
emissions for the three sectors we model under each case.
Reduced natural gas generation in both policy cases, relative to the BAU, leads to
significant electricity sector emissions reductions in 2030. CPC policies reduce New
York power plant NO
x
and SO
2
emissions by smaller proportions than the other emissions
types because waste-fueled generation accounts for large portions of the state’s power
plant NO
x
and SO
2
emissions, and the CPC policies do not appreciably change waste-
fueled generation. Even in the BAU, waste-fueled generation accounts for more than half
of New York power plant NO
x
and SO
2
emissions, despite producing less than 10 percent
as much generation as natural gas does. The reduction of waste-fueled generation in the
SPC is a significant contributor to the emissions reductions in that scenario.
Prioritizing Justice in New York State Climate Policy 17
Figure 3. Emissions across Policy Cases, by Sector, 2030
0
10
20
30
40
50
60
70
Electricity Residential LDV MHDV
GHG (MMT CO
2
e)
BAU2030 CPC2030 SPC2030
0.00
200.00
400.00
600.00
800.00
1,000.00
1,200.00
1,400.00
Electricity Residential LDV MHDV
SO
2
(Metric Tons)
BAU2030 CPC2030 SPC2030
0.00
500.00
1,000.00
1,500.00
2,000.00
2,500.00
Electricity Residential LDV MHDV
Direct PM
2.5
(Metric Tons)
BAU2030 CPC2030 SPC2030
0.00
5,000.00
10,000.00
15,000.00
20,000.00
Electricity Residential LDV MHDV
NO
X
(Metric Tons)
BAU2030 CPC2030 SPC2030
Similarly, in the residential building sector, both GHGs and local air pollutants decline under
both scenarios because of reductions in fossil fuel (natural gas and diesel) use for heating.
Although emissions decline under both scenarios, SPC reductions in both GHGs and local
air pollutants associated with the residential sector are more than double those of the
CPC, with 90 to 100 percent reductions from the BAU across the various emissions types.
These emissions changes are the result of the adoption of electric heat pumps (Figure 2).
For light-duty vehicles, the emissions changes are more modest than for the other
sectors because of minimal vehicle stock change by 2030. The CPC reduces CO
2
emissions by 6 percent and the SPC by 12 percent below the BAU—the same as the fuel
consumption reductions reported above.
11
Compared with the BAU, the CPC and SPC
reduce direct PM
2.5
emissions of NO
x
and SO
2
by small amounts (3 to 8 percent across
the two cases). The SPC does reduce emissions by about double the CPC, although the
reduction is still modest (e.g., 4 percent in the CPC versus 8 percent in the SPC for SO
2
; see
Table I-1 in Appendix I). The main policy driving this dierence is the ZEV standards, since
EVs do not emit these pollutants directly when running on electricity.
11 Methane (from incomplete combustion and upstream fugitive emissions associated with
gasoline production and distribution) accounts for a trivial share of light-duty vehicles’ GHG
emissions.
Resources for the Future and New York City Environmental Justice Alliance 18
The medium- and heavy-duty vehicle sector is also estimated to have
reduced emissions by 2030, but the dierences between the two policy case
implementations are small. The CPC reduces CO
2
emissions by 16 percent and the
SPC by 18 percent below the BAU (reductions in methane are similar; see Appendix
H), primarily resulting from the penetration of ZEVs, with the SPC having a slightly
larger share of ZEVs by 2030. From the BAU, the CPC achieves a further reduction of
15 percent for SO
2
, 11 percent for NO
x
, and 8 percent for direct PM
2.5
; the SPC reduces
these emissions by 18, 17, and 10 percent below the BAU, respectively (see Table I1 in
Appendix I).
5.3. Location of Emissions Changes
As stated above, the focus of this study is to estimate how New York climate policy
aects PM
2.5
pollution exposure in disadvantaged communities. This requires analyzing
emissions—and ultimately air quality—by location, going beyond the statewide
emissions estimates described above. This is because pollution is not spread uniformly
across the state: where you live matters in terms of the air you breathe, which is
a key aspect of environmental justice. To get at this geography of pollution (and
related disparities in pollution exposure), we begin by studying where emissions
occur—emissions from burning fossil fuels (and some waste and biomass) to generate
electricity, heat homes, and power heavy trucks and passenger vehicles on New York
roads. Identifying the location of emissions is a prerequisite for determining where
pollution ultimately settles (after being mixed and morphed in the atmosphere),
which is how we determine the geography of air quality and associated public health
implications, discussed below. It is important for the reader to make a clear distinction
between emissions and air quality—a distinction we will continue to discuss.
This section covers details about where emissions changes take place, to the greatest
level of spatial detail possible. For simplicity of presentation, we restrict our discussion
to direct emissions of PM
2.5
, even though the models predict changes in NO
X
, SO
2
, and
VOCs (and other pollutants). We focus on direct PM
2.5
because it has the greatest
impact on local air quality. The extent to which other pollutants combine to form
secondary PM
2.5
is covered in the following section, Air Quality Results.
A number of our models indicate the largest emissions reductions (by mass) tend to
occur in densely populated areas. This stands to reason, since the more people who live
in an area, the more fuel combustion tends to take place. This is especially the case for
sectors like transportation and residential buildings, where fuels are burned (resulting in
emissions) in the location where the population is concentrated.
This trend is reflected in Figure 4, which shows the estimated direct PM
2.5
emissions
reductions (in 2030) from LDVs under the CPC. As we can see, the deeper emissions
reductions (darker plots) are located around New York State’s metropolitan areas—
especially New York City, Bualo, Rochester, and Syracuse. It is worth noting that in the
CPC, subsidies are not targeting these metro geographies—in fact, the model shows
relatively uniform percentage reductions in fuel use and emissions across the state.
Simply the fact that there are more people in metro areas means that these uniform
Prioritizing Justice in New York State Climate Policy 19
percentage reductions lead to larger quantity reductions. These trends continue under
the SPC (see Appendix J).
It is also evident from Figure 4 that many DACs are located in these metro areas, which
means this trend is estimated to provide above-average benefits to DACs. That said,
there are numerous DACs located outside metro areas, and these DACs are estimated
to experience smaller emissions improvements associated with changes to the
passenger vehicle fleet (however, this is also because nonmetro DACs experience less
LDV pollution in the first place). It is also important to observe that although nonmetro
DACs may see less emissions reductions from LDVs, none experience an increase in
LDV emissions.
Key Findings for Figure 4
• Relative to the BAU, the CPC leads to direct PM
2.5
emissions reductions
from LDVs across the state.
• No regions experience increases in direct PM
2.5
emissions from LDVs.
• Direct PM
2.5
emissions reductions from LDVs are most pronounced in
New York City and other high-density areas.
Figure 4. Light-Duty Vehicle Direct PM
2.5
Emissions, BAU vs. CPC, 2030
Resources for the Future and New York City Environmental Justice Alliance 20
We see some of this same population density-driven trend in emissions changes
(around metro areas) from the residential sector under the CPC. However, as shown in
Figure 5, other factors influence the distribution of emissions. For example, in addition
to population density, the means testing of heat pump subsidies (favoring low-income
households) leads to greater emissions reductions in low-income communities. The
heterogeneity in building conditions, climate, and other household attributes also
influences the adoption of heat pumps and resulting emissions reductions.
As with LDV emissions, we see that a large number of DACs experience sizable
benefits associated with residential emissions reductions. And again, although some
DACs experience more modest benefits, we do not observe any DACs that experience
an increase in residential emissions.
Key Findings for Figure 5
• Relative to the BAU, the CPC leads to direct PM
2.5
emissions reductions
from residential heating across the state.
• No regions experience increases in direct PM
2.5
emissions from residential
heating.
• Direct PM
2.5
emissions reductions from residential heating are most
pronounced in high density areas like New York City.
Figure 5. Residential Home Heating Direct PM
2.5
Emissions, BAU vs. CPC, 2030
Prioritizing Justice in New York State Climate Policy 21
•
Key Findings for Figure 6
• Relative to the BAU, the CPC leads to direct PM
2.5
emissions reductions
from MDHVs across the state.
• No regions experience increases in direct PM
2.5
emissions from MDHVs.
• Direct PM
2.5
emissions reductions from MDHVs are most pronounced
along major highways.
One of the limitations of the LDV and residential models is that they estimate emissions
at a somewhat coarse geographic scale (at the county level for the LDV model, and
at the level of the public use microdata area, or PUMA, for the residential model).
This limits the ability to identify more localized dierences in pollution exposure. This
limitation can really matter for something like transportation emissions, where pollution
exposure is often a function of how close one is to a specific highway or port depot—a
level of geographic granularity that goes below the county or PUMA level.
Figure 6. Medium- and Heavy-Duty Vehicle Direct PM
2.5
Emissions, BAU vs. CPC, 2030
Note: There are some outlying tracts with particularly pronounced emissions reductions in the MHDV sector. They are marked
with a dark blue outside of the legend scale.
Resources for the Future and New York City Environmental Justice Alliance 22
The MHDV model provides this finer geographic granularity. For the MHDV fleet,
emissions were estimated for each major road segment (“network link”) along the
primary and secondary highway system in New York (see Appendix E). In this analysis
of the location of emissions changes, PM
2.5
is displayed by census tract.
12
As shown in Figure 6, there is some clustering of emissions reductions around metro
areas; however, the predominant trend is emissions reductions along highways—the
major intercity corridors. This is mainly because the majority of long-haul heavy-duty
truck traic occurs on these highway network links, and so any vehicle improvements
(and associated emissions reductions) made to the MHDV fleet will be concentrated
on those highways.
A large percentage of DACs clustered in the urban core of cities (e.g., in Brooklyn and
parts of Manhattan) are not located near these major intercity corridors, and so they
are estimated to experience less benefit from MHDV emissions reductions. However,
a number of nonmetro DACs that experience relatively modest emissions reductions
from policies aecting residential buildings and LDVs (above) are estimated to see
some of the largest benefits from policies aecting the MHDV fleet. Indeed, there
is a close correlation between the location of many nonmetro DACs and New York’s
intercity highway system (and therefore the deepest MHDV emissions reductions).
This could partly be due to the fact that one of the criteria for designating DACs (as
determined by the New York CJWG) is high exposure to traic. No census tracts
experience an increase in PM
2.5
emissions, and reductions are greater under the SPC
than in the CPC, following largely the same geographic distribution.
Our estimates of electricity emissions are even more geographically granular than
the MHDV model, with the specific location of each power plant represented in the
E4ST electricity sector model (see Appendix E). Figure 7 shows the location of 2030
emissions reductions from implementing the CPC, with green dots representing a
decrease in power plant emissions at a given location and red dots indicating an
increase in emissions (this is the first sector where we observe emissions increases).
12 Link emissions are attributed to census tracts based on what percent of the link is in
each tract.
Prioritizing Justice in New York State Climate Policy 23
Figure 7. Change in Direct PM
2.5
Emissions, by Electric Power Generator, BAU vs. CPC,
2030
Key Findings for Figure 7
• Most of the power-generating units in New York decrease their direct
PM
2.5
emissions in the CPC relative to the BAU (green dots).
• The largest decreases are at generating units close to or in New York City.
• Emissions change at many generating units outside the state; some of the
largest increases occur at generating units close to and upwind of New
York City (e.g., in New Jersey); the largest out-of-state decreases occur at
coal-fired generating units in Pennsylvania.
For the electricity sector, we show estimates for adjacent states as well because New
York policy has a greater eect on out-of-state emissions in the electricity sector
than in other sectors, and because electricity emissions get dispersed over a broader
geographic area than emissions from the other sectors we model because of the tall
smokestacks at power plants. Given this, New York electricity policy changes would
aect emissions both in New York and in other states and Canadian provinces.
Just comparing the number of green and red dots, we can see that although most
power-generating units in New York State decrease emissions in both policy cases
relative to the BAU (green dots), the CPC results in a fair number of power-generating
units that increase emissions (red dots); however, they tend to be small increases, and
many are located out of state.
Resources for the Future and New York City Environmental Justice Alliance 24
Examining the size of the dots, we see that the largest decreases in emissions are
at power plants close to or in New York City—again reflecting a trend of emissions
reductions near population centers. As with other metro-centered reductions, the
concentration of emissions reductions near the New York City region is beneficial for
public health, since that is the most densely populated part of the state. However, while
the metro region is estimated to experience reductions at in-state plants, just over the
border in New Jersey, emissions increases are expected.
Looking at the dots outside New York State, we find that emissions both increase and
decrease at many power-generating units. This holds for both policy cases. The largest
increases are at those New Jersey generating units that are close to and upwind of
New York City, and the largest decreases occur at Pennsylvania coal-generating units
that are also upwind of the state but farther away.
13
In terms of how these emissions changes may aect New York communities, as stated
above, reductions or increases in the New York City metro area will tend to have
significant public health eects because of its population density. New York City is also
where many DACs are located, and so emissions changes will have a significant impact
on DACs: if emissions decline in that area, many DACs will benefit, and if emissions
increase, many DACs will be harmed. However, because of the stack heights of power
plants and the tendency of their emissions to be carried long distances by prevailing
winds, the pollution exposure in a given community may not be directly linked to
emissions changes at local power plants. To produce a more accurate estimate of
pollution exposure at the community level, we must go from estimating emissions by
location to estimating air quality by location, which is the next step in our analysis.
5.4. Air Quality Results
Our research goes beyond economic and emissions impacts to identify local air quality
eects of the policy cases we studied. The key dierence between emissions changes
and air quality is that the latter reflects how and where emissions actually accumulate,
after being transported and transformed by weather patterns and by mixing with
other pollutants. Air quality is more relevant for human health because it looks at the
composition of the air people breathe, instead of emissions flows from a source of
pollution.
Our air quality analysis combines the geographic emissions information discussed
above with scientific information about how meteorological patterns and chemical
processes contribute to the distribution of pollutants. The air quality modeling
approach we use (one of the most technically advanced in the field; see Appendix
C) estimates average hourly PM
2.5
concentrations at the 4km
2
grid level. In many
ways, this is a large area for thinking about community-level air quality impacts,
13 Out-of-state emissions in the region may be impacted by changes in program ambition
of the Regional Green House Gas Initiative, which caps carbon emissions from the power
sector, or IRA subsidies, which provide incentives for clean energy.
Prioritizing Justice in New York State Climate Policy 25
but we determined that it was the most geographically granular estimate feasible.
14
Our results cover the overall air quality changes in New York State and specific
community comparisons that highlight outcomes for dierent types of disadvantaged
communities. For the full methodology for the air quality modeling process, see
Appendix C.
5.4.1. Air Quality Changes in Each Policy Case
Figures 7 and 8 show changes in 2030 average hourly concentrations of PM
2.5
at
the census tract level across the state, with a more detailed view of New York City.
15
Darker colors in these figures indicate greater reductions in PM
2.5
concentrations and
hence greater improvements in air quality. For both policy cases, improvements are
highly concentrated in high-density areas, particularly New York City. This reflects
emissions changes discussed in Section 5.3. Location of Emissions Changes (above),
which explains how increased heat pump and electric vehicle adoption tends to
disproportionately benefit areas with high housing density and traic congestion.
Additionally, there tend to be more power-generating units built to service high-density
areas. Though there can be more distance between generators and the demand they
serve—and emissions from power plants can be spread over large distances—we do
observe significant emissions reductions at power plants close to the New York City
area. Although the darkest colors on the map occur mainly there, other urban areas
also see more pronounced improvements relative to their rural neighbors.
One significant dierence between the two policy cases is that the SPC achieves
overall greater air quality improvements than the CPC. The population-weighted
average PM
2.5
concentration change from the BAU to the CPC is 0.03 µg/m
3
, with a
range of 0.05 µg/m
3
increase to 0.10 µg/m
3
decrease. In comparison, the SPC achieves
a population-weighted average reduction below BAU of 0.18 µg/m
3
, with a range
of 0.04 µg/m
3
increase to 0.40 µg/m
3
decrease. The averages and ranges of PM
2.5
concentrations are quite striking, and there is much greater variation in the SPC, which
has a standard deviation of 0.12 µg/m
3
. More information on variation in air quality
improvements can be found in Appendix L. Very high air quality improvements in the
New York City area are largely responsible for the dierence in average air quality
improvement in the two cases.
14 It is possible to model changes at 1km
2
, but our modeling team felt it presented the risk
of “false precision,” where results would indicate a greater amount of accuracy than
scientifically justified. Our estimates at 4km
2
were determined to maximize geographic
granularity without imposing false precision.
15 Our air quality model (see Appendix C) estimates the PM
2.5
concentration in every hour
of 2030 for each 4km
2
grid cell in the state of New York. We then estimate the 2030 av-
erage hourly concentration for each 4km
2
cell, which is mapped onto each census tract.
Therefore, Figures 8, 9, and 10 reflect dierences in 2030 average hourly PM
2.5
concen-
tration levels for each tract.
Resources for the Future and New York City Environmental Justice Alliance 26
An additional important finding is that—as the ranges stated in the previous paragraph
show—both cases cause air quality to worsen (increased PM
2.5
concentrations) in some
census tracts, and although the CPC causes a worsening of air quality in some DACs,
the SPC does not erode air quality in any DACs. In the CPC, 296 tracts (of nearly
5,000) are predicted to have worse air quality than in the BAU, 72 of which are DACs. In
the SPC, only three census tracts experience worse air quality, none of which are DACs.
Figure 8. New York PM
2.5
Concentration Improvements, BAU to SPC, 2030
Note: These maps represent air quality improvements across New York State and in New York City specifically, relative to the
BAU. The blue color represents greater improvements in air quality (higher PM
2.5
concentration reductions in µg/m
3
) while the
yellow color represents smaller improvements in air quality. Tracts are shaded in orange if they experience worse air quality
relative to the BAU.
Prioritizing Justice in New York State Climate Policy 27
Key Findings for Figures 8 and 9
• Air quality improvements under the SPC are greater than air quality
improvements under the CPC.
• Air quality worsens in more tracts under the CPC compared with the SPC;
72 of these tracts are DACs. The SPC does not erode air quality in any
DACs.
• Air quality improvements are most concentrated in urban areas, with high
population densities.
• Air quality improvements are more heterogeneous in the SPC, which has
an even more dramatic spread between New York City and the rest of the
state than the CPC.
Figure 9. New York PM
2.5
Concentration Improvements, BAU to CPC, 2030
Note: These maps represent air quality improvements across New York State and in New York City specifically, relative to the
BAU. The blue color represents greater improvements in air quality (higher PM
2.5
concentration reductions in µg/m
3
) while the
yellow color represents smaller improvements in air quality. Tracts are shaded in orange if they experience worse air quality
relative to the BAU.
Resources for the Future and New York City Environmental Justice Alliance 28
5.4.2. Outcomes for Disadvantaged Communities
Although statewide results are critical to understanding overall impacts of potential
CLCPA implementations, we also consider impacts across dierent types of
communities—specifically, disadvantaged communities, as defined by the draft CJWG
criteria, and non-DACs. Policies implemented as a part of the CLCPA may not harm or
burden DACs, and emissions and pollution reductions in those communities must be
prioritized.
Figure 10. Air Quality Improvements for CJWG Community Types, CPC to SPC
10A. Disadvantaged Communities, Statewide and NYC
Note: These maps represent air quality improvements across New York State and in New York City in the SPC, relative to the
CPC. The blue color represents greater improvements in air quality (higher PM2.5 concentration reductions in µg/m3) while the
yellow color represents smaller improvements in air quality. Tracts are shaded in orange if they experience worse air quality in
the SPC relative to the CPC.
Prioritizing Justice in New York State Climate Policy 29
Key Findings for Figure 10
• Air quality improvements for DACs and non-DACs are more pronounced in
the SPC.
• Greatest dierences are observed in New York City, which has a high
number of DACs.
• Dierences in air quality improvements between communities are
most strongly associated with their location in the state. The greatest
improvements occur in the locations that have the highest baseline
emissions and the poorest baseline air quality.
Figure 10 presents two maps of New York State, highlighting DACs (page 28)
compared to all other communities (page 29). This helps illustrate how air quality
dierences are distributed across community types when comparing the CPC and SPC.
10B. Nondisadvantaged Communities, Statewide and NYC
Note: These maps represent air quality improvements across New York State and in New York City in the SPC, relative to the
CPC. The blue color represents greater improvements in air quality (higher PM2.5 concentration reductions in µg/m3) while the
yellow color represents smaller improvements in air quality. Tracts are shaded in orange if they experience worse air quality in
the SPC relative to the CPC.
Resources for the Future and New York City Environmental Justice Alliance 30
The average dierence in PM
2.5
concentrations between the CPC and SPC for DACs is
0.15 µg/m
3
, favoring the SPC. This is compared with an average dierence of 0.14 µg/
m
3
for non-DACs, still favoring the SPC. These findings indicate that the additional air
quality benefits associated with the SPC are relatively evenly distributed across
DACs and non-DACs, as defined by the CJWG. More information on variation in air
quality improvements for DACs and non-DACs can be found in Appendix L.
Table 2 shows more specific population-weighted improvements by subcategories
of communities. These findings indicate that although the SPC improves air quality
relatively equally among DACs and non-DACs, certain subcategories of disadvantaged
communities see particularly pronounced improvements. We find that implementing
the SPC (compared with the CPC) would provide higher-than-average benefits
to communities with high socioeconomic vulnerability and communities with
historically high PM
2.5
exposure. (Population characteristics and vulnerability index
are defined in Appendix D.)
These findings indicate that a broad definition of disadvantaged communities that
combines a long list of vulnerabilities and exposures could obscure dramatic changes
in communities that are particularly vulnerable in only a few dimensions.
Table 2. Population-Weighted Air Quality Improvements, by Community Type (PM
2.5
µg/m
3
)
Community type
Improvement from BAU
to SPC
Improvement from BAU
to SPC
Improvement from CPC
to SPC
All tracts 0.18 0.03 0.15
Non-DAC tracts (65% of tracts) 0.17 0.03 0.14
DAC tracts (35% of tracts) 0.19 0.03 0.16
High exposure (top 10%) 0.16 0.03 0.13
High vulnerability (top 10%) 0.24 0.03 0.21
High elderly population (top 10%) 0.09 0.03 0.06
High historical PM
2.5
(top 10%) 0.31 0.05 0.26
Prioritizing Justice in New York State Climate Policy 31
A final important finding is that even under the SPC, where air quality improvements
are greater for DACs than for non-DACs, the improvements are not large enough to
eliminate disparities in pollution exposure. Under the BAU, 2030 average hourly pollution
exposure in DACs is 0.20 µg/m
3
higher than in non-DACs, and this dierence declines
by only one one-hundredth (to 0.19 µg/m
3
) under the SPC. This same trend is evident
when comparing non-DACs with high-exposure and high-vulnerability tracts; however,
the relative improvements are greater for these groups. Note that this dierence is
smaller compared with our 2012 baseline, where PM
2.5
concentrations in DACs were
approximately 0.50 µg/m
3
higher than in non-DACs.
5.4.3. Contextualizing Air Quality Changes
Despite very significant reductions in direct PM
2.5
and its precursor emissions
under the policy cases compared with the BAU, the changes in average hourly PM
2.5
concentrations can be characterized as “small”—around 0.18 µg/m
3
in the SPC against a
baseline PM
2.5
concentration around 7 µg/m
3
. Appendix K oers a variety of explanations
for why this number may be more modest than expected. In this section, we provide
context for the health implications of these changes, suggesting they still have
significant benefits, particularly for Black New Yorkers.
Exposure to PM
2.5
is a well-known killer leading to premature mortality in the over-30
population (Di et al. 2017; Krewski et al. 2009; Lepuele et al. 2012). Studies show that for
a 10 µg/m
3
reduction in PM
2.5
concentrations, mortality risks to those 30 and older fall
6 to 14 percent from baseline mortality rates. Of course, 10 µg/m
3
is a huge change in
PM
2.5
: the National Ambient Air Quality Standard for PM
2.5
is 12 µg/m
3
.
To contextualize our findings, we wanted to roughly estimate how the expected air
quality changes could improve health outcomes. To do this, we consider the dierence in
the population-weighted change of PM
2.5
concentrations we see across the state in the
SPC compared with the BAU—around 0.18 µg/m
3
. Holding constant the state’s current
over-30 population of 15 million people and using the national death rate of 800/100,000
people, the total deaths avoided by implementing the SPC would range from 13 to
302 every year in which this change in air quality persists (relative to the BAU).
We can provide additional context if we focus on the data we have on New York City’s
elderly (65 and over) population by race and ethnicity. We focus on the city because
this area would see the largest reductions in PM
2.5
concentrations—around 0.4 µg/m
3
.
In addition, there are profound racial and age dierences in both the mortality rates and
the relationship we noted above between PM
2.5
and the percentage change in mortality
rates (termed a concentration-response factor, or CRF; see Di et al. 2017).
16
16 To make these new calculations, we use population by age and race, CRFs by age and race,
and baseline mortality by age and race, as well as the New York City PM
2.5
change in the
SPC of 0.4 µg/m
3
compared with the BAU. We have age and race population data for the
city’s five counties, we have national-level CRFs from a major PM
2.5
–mortality study (Di
et al. 2017) for people 65 and over and for five racial categories: white (non-Hispanic),
Black (non-Hispanic), Hispanic, Asian, and Native American. The CRFs for the national
population are shown in Table 3 and are far larger for Blacks than for other groups, with
whites having the lowest CRF.
Resources for the Future and New York City Environmental Justice Alliance 32
Table 3. Concentration-Response Factors and Mortality Rates, by Race and Ethnicity
Race, ethnicity
17
CRF
18
Mortality rate range (deaths/1,000 people)
White 6.3 .01 to .11
Black 20.8 .02 to .10
Asian 9.6 .006 to .08
Hispanic 11.6 .01 to .08
Native American 10.0 .005 to .03
We also obtained mortality rates by age and race or ethnicity from the NYC death micro
SAS datasets. Here again, we see that Blacks have another disadvantage compared
with whites, Hispanics, and Asians, since their baseline mortality rates are higher than
for these other groups for all three age categories. With PM
2.5
being reduced, Blacks gain
more health benefits than other groups. We apply these national data for the 65-and-
over group to New York City’s elderly population,
19
which is 1.3 million people.
The bottom line: by implementing the SPC rather than the BAU, 160 deaths among
New York City’s elderly are avoided in 2030, and the disparity by race or ethnicity
disproportionately benefits Black New Yorkers. With Black New Yorkers making up
approximately 22 percent of the city’s over-65 population, the deaths avoided for this
group are 42 percent of the 160 total. In contrast, white New Yorkers make up 41 percent
of the city’s over-65 population, with deaths avoided only 37 percent of all deaths
avoided (Table 4).
17 We include racial and ethnic groups for which we have data.
18 In terms of statistical significance of the CRFs by race, Di et al. (2017) show in their Figure 2 that the CRF for Blacks is
significantly larger than for all other groups and that the CRF for whites is significantly smaller than those for all other
groups. CRFs for Native Americans, Asians, and Hispanics appear to be not statistically dierent from one another.
19 The baseline mortality rates are available by race or by age for NYC counties, but not both. Because of the complication that
“white” and “Black” include Hispanics, we rely on Spiller et al. (2021), who estimated national mortality rates by race and
ethnicity for whites, Blacks, and Hispanics age 65 and older. We use the available New York City data for Asians and Native
Americans and scale the national data for Blacks, whites, and Hispanics to be applicable to New York City by adjusting
these rates for the proportional dierence between the Asian death rate nationally and Asian death rate in the city.
Prioritizing Justice in New York State Climate Policy 33
In addition to estimating mortality eects of policy implementation, we investigate
how a number of other health outcomes (“morbidity indicators”) would be aected by
the SPC. To do this, we simply show the relationship between mortality avoided and
various other health eects avoided, including asthma attacks, lost workdays, and
restricted activity days. This relationship is taken from a table of results (Table 5-5)
in the PM
2.5
NAAQS Regulatory Impact Analysis for 2012 (EPA 2013). Using the ratio
of the number of deaths avoided to the number of other health impacts avoided and
multiplying by the approximately 206 deaths avoided in New York State from the SPC,
we get the results shown in Table 5. The reduced number of asthma attacks, cases of
chronic respiratory diseases, and restricted activity days is 100 times to almost 1,000
times greater than deaths avoided (Table 5).
Table 5. Annual PM
2.5
and GHG Impacts, SPC relative to BAU
Benefits associated with reduced PM
2.5
Adult mortality ~206 avoided deaths ~$2 billion
Asthma exacerbation ~ 7,000 avoided ~$0.7 million
Reduced workdays ~ 11,800 avoided ~$3 million
Reduced activity days ~ 70,000 avoided ~$8 million
Benefits associated with reduced GHGs
Social cost of carbon
20
Avoided damages ~$10 billion
Source: EPA (2013). Values in the table represnt impacts of PM2.5 and GHG reductions for one year, 2030.
20 This is based on the NY social cost of carbon with a 2 percent discount rate ($137 in 2030).
Table 4. Avoided Deaths in New York City, by Race and Ethnicity, SPC Relative to BAU
Race, ethnicity Percentage of population 65+ Percentage of avoided deaths attributable to PM
2.5
Asian 14% 6%
Black 22% 42%
Hispanic 22% 15%
White 41% 37%
Resources for the Future and New York City Environmental Justice Alliance 34
We also quantify the health impacts as monetary savings, as the US Environmental
Protection Agency did in its 2013 report. The mortality impacts are quantified using the
value of statistical life, which is about $10 million (EPA 2023). The other health eects
are quantified through cost savings, such as reduced hospital visits and recovered
workdays. We estimate annual monetary savings associated with the SPC public health
benefits we calculate (there are others) in the rightmost column of Table 5. Finally,
the monetary benefits of the reductions in CO
2
and methane emissions (81 MMT, see
appendix I) are shown in the last row of the table, using New York State’s estimate of
the social cost of carbon ($137/ton CO
2
in 2030). Therefore, comparing the SPC with
the BAU, the social benefits of GHG reductions would exceed $10 billion, and the
health benefits of PM
2.5
concentration reductions would exceed $2 billion.
6. Conclusion
This research project is one of the first to combine behavioral economic modeling
with geospatially granular air quality modeling and stakeholder-driven demographic
mapping to examine the air quality impacts of New York’s Climate Leadership and
Community Protection Act for disadvantaged communities. One of the key goals
for the research project is to assess which set of policies would maximize air quality
benefits to DACs. Two implementations of the CLCPA are considered: a set of policies
based on the New York Climate Action Council scoping plan, and a set of policies
favored by the environmental justice stakeholder community, convened by our
partners at NYC Environmental Justice Alliance. Our research reveals several insights
about how implementation of the CLCPA may aect technology adoption, emissions,
and subsequent air quality across the state between DACs and non-DACs. Generally,
our research showed greater improvements for DACs in the stakeholder policy case
compared with the CAC-inspired case.
Both cases result in substantial reductions in GHG emissions relative to a business-
as-usual scenario modeled for the year 2030. But the stakeholder policy case results
in greater reductions, not only in GHGs but also in emissions of NO
x
, SO
2
, VOCs, and
direct PM
2.5
, which translates into greater reductions in PM
2.5
concentrations across the
state.
These results occur because the stakeholder case features more stringent policies
than the CAC-inspired case, and because in most cases policies that reduce GHGs
also reduce copollutants (including those traditionally referred to under the Clean
Air Act as criteria air pollutants). Key policy drivers of the greater improvements in
the stakeholder case are a higher price on carbon and copollutants, more generous
subsidies for heat pumps targeted at low-income households, and stricter phaseouts of
fossil fuels in the electricity and residential sectors.
In the strict letter of the law, both cases meet the CLCPA regulatory mandate to “not
result in a net increase in co-pollutant emissions or otherwise disproportionately
burden disadvantaged communities,” defining burden in terms of air pollution
concentrations. However, although there is no statewide net increase in copollutant
Prioritizing Justice in New York State Climate Policy 35
emissions, under both cases some communities (at the census tract level) do
experience a worsening of air quality. And although this occurs in about 4 percent of
DACs under the CAC-inspired case, no DACs are aected in the stakeholder case.
In general, DACs do better than non-DACs under the stakeholder case compared with
the CAC-inspired case. This dierential eect occurs in part because New York City
and other cities experience disproportional PM
2.5
concentration reductions compared
with other parts of the state, and cities have a higher concentration of DACs.
Furthermore, if we focus on the communities that are the most vulnerable (according
to a range of metrics), say in the 10th percentile of the state’s measure of social
vulnerability, the stakeholder policies deliver even more favorable air quality
improvements compared with the CAC-inspired policy set.
However, even though both policy cases make air quality improvements in DACs,
neither reverses the historical disparity in air pollution: overall air quality in DACs
continues to be worse compared with non-DACs, and this dierence is even more
pronounced in high-vulnerability communities (where air quality tends to be the worst).
Our focus is on air quality impacts, and it can be diicult to understand the importance
of even small air quality dierences. Therefore, as an aid in contextualizing our results,
we approximate some public health implications of our estimated PM
2.5
concentration
changes, including, most prominently, annual deaths avoided. We also look closely
at the elderly Black population in New York City relative to the elderly of other races
and ethnicities, to account for this group’s increased vulnerability to poor air quality.
We find that the greater air quality improvements in New York City, combined with
its higher Black population and their greater vulnerability, lead to even greater health
improvements than the statewide average would indicate.
Given all the benefits of the stakeholder case over the CAC-inspired case, it should not
be surprising to see that this stakeholder case requires greater investment than the
CAC-inspired case: its policies are more rigorous and oer more generous subsidies.
For example, for residential heating, to achieve the much higher penetration of heat
pumps observed in the stakeholder case, government subsidies need to be far higher
than in the CAC-inspired case. In the power sector, a higher carbon price and a ban on
new fossil fuel generation leads to slightly higher electricity prices. Although a full cost-
benefit analysis was outside the scope of this work, previous regulatory analyses that
evaluate stringency of GHG and air pollution policies often find that the environmental
and health benefits of added stringency outweigh the costs.
21
We see through this analysis that more ambitious policies are eective at decreasing
emissions and improving air quality on a greater scale than more moderate policies.
We also see that some disadvantaged communities benefit from the more aggressive
21 See recent regulatory impact analyses from the US Environmental Protection
Agency, including Table ES-4 in the agency’s assessment of national air pollution
standards for coal plants completed in 2023: https://www.epa.gov/system/files/
documents/2023-04/MATS%20RTR%20Proposal%20RIA%20Formatted.pdf
Resources for the Future and New York City Environmental Justice Alliance 36
policy case and begin to make up their deficit in environmental protection relative to
non-DAC communities, although nearly all communities see benefits. By considering the
impact on environmental justice communities, policymakers can ensure that historically
underserved neighborhoods are protected and are able to experience the intended
benefits of environmental policies.
Our findings show that more ambitious and more targeted climate policies yield the
greatest benefits in climate change mitigation. The higher New York sets its policy
sights, the greater the capacity for decreasing emissions and poor air quality. Because
of these greater improvements in air quality and emissions, NYC Environmental Justice
Alliance encourages the use of more ambitious and targeted approaches, like the
policies modeled in the stakeholder policy case. These options provide a larger capacity
for emissions reductions, which is critical to addressing the climate change crisis
expeditiously.
Disadvantaged communities have so much to lose if these targets are not met or if
mitigation eorts are not directed toward those who are at greatest risk. These at-risk
communities also have so much to gain if their policy requests are implemented and
their safety is prioritized as policymakers and community leaders work to undo the
historical damage that they have been forced to accept as their legacy. By prioritizing
environmental justice communities’ leadership in this space, communities can fully
benefit from the positive impacts their visions of the future can yield when put into
practice.
This work has revealed new insights on community-level improvements associated
with New York’s landmark climate act. The boundaries of this work have also revealed
ripe areas for future research. Interested readers should consult Appendix G on the
limitations of our work, but three areas oer the greatest opportunity for future research.
First, being able to predict economic behavior requires sophisticated models for all parts
of the economy. Our models cover electric utilities, home heating, and heavy-duty and
light-duty vehicles. Agriculture, manufacturing, and commercial heating are missing from
our analysis but contribute significantly to emissions and air pollution. Because some of
the policies being discussed in New York will be economy-wide, we miss the reactions
of those sectors and probably underestimate how much the CLCPA will aect air quality
and emissions. It would be valuable for future work to build and apply these models.
A second limitation is that we model conditions only in 2030 (with and without the
CAC-inspired and stakeholder policies). Many policies, such as incentives to encourage
EV adoption, take time to have a significant eect. Had we run the models for a longer
period (say, to 2040 or 2050), CO
2
and copollutant emissions would have declined more.
However, the further in the future forecasts are made, the more uncertain they become.
We opted to minimize such uncertainty. Future analyses could run the models further
into the century.
Finally, the expense associated with operating a full-complexity air quality model
(compared with reduced-complexity models) limited the number of cases we could
complete in this phase of work. As a result, we could not test the air quality impact of
Prioritizing Justice in New York State Climate Policy 37
individual policies housed in each policy case, or attribute pollution concentration
changes to specific emissions sources (source attribution). In our next phase of work,
we hope to conduct more focused policy analysis highlighting the air quality impacts of
one specific policy: economy-wide carbon pricing through a carbon cap.
This work oers unique insights into the distributional air quality impacts of CLCPA
implementation. It provides a framework for evaluating future policies that aect the
magnitude and location of emissions changes through addressing economic behavior
and methods that can be useful in evaluating how environmental justice communities
in particular will be aected. Though in its early stages, work in this field presents
many opportunities for future research still to explore.
Resources for the Future and New York City Environmental Justice Alliance 38
7. References
Carleton, Tamma, and Michael Greenstone. 2022. A Guide to Updating the US Government’s
Social Cost of Carbon. Review of Environmental Economics and Policy 16(2): 196–218.
César, Ana Cristina Gobb, J. A. Carvalho Jr., and Luís Fernando Costa Nascimento. 2015.
Association between NO
x
Exposure and Deaths Caused by Respiratory Diseases in a
Medium-Sized Brazilian City. Brazilian Journal of Medical and Biological Research 48:
1130–35.
Chen, Yuanlei, Evan D. Sherwin, Elena S. F. Berman, Brian B. Jones, Matthew P. Gordon, Erin
B. Wetherley, Eric A. Kort, and Adam R. Brandt. 2022. Quantifying Regional Methane
Emissions in the New Mexico Permian Basin with a Comprehensive Aerial Survey.
Environmental Science and Technology 56(7): 4317–23.
Di, Qian, Yan Wang, Antonella Zanobetti, Yun Wang, Petros Koutrakis, Christine Choirat,
Francesca Dominici, and Joel D. Schwartz. 2017. Air Pollution and Mortality in the
Medicare Population. New England Journal of Medicine 376(26): 2513–22.
Donaghy, Timothy Q., Noel Healy, Charles Y. Jiang, and Colette Pichon Battle. 2023. Fossil
Fuel Racism in the United States: How Phasing Out Coal, Oil, and Gas Can Protect
Communities. Energy Research and Social Science 100: 103104.
Energy and Environmental Economics (E3). 2022. Appendix G: Integration Analysis Technical
Supplement New York State Climate Action Council. https://climate.ny.gov/resources/
scoping-plan/-/media/project/climate/files/Appendix-G.pdf.
Engineering, Economic, and Environmental Electricity Simulation Tool (E4ST). (2022). https://
www.r.org/topics/data-and-decision-tools/e4st/.
Environmental Protection Agency (EPA). 2013. Regulatory Impact Analysis for the Final
Revisions to the National Ambient Air Quality Standards for Particulate Matter. https://
www.epa.gov/sites/default/files/2020-07/documents/naaqs-pm_ria_final_2012-12.pdf.
———. 2023. Mortality Risk Valuation. https://www.epa.gov/environmental-economics/
mortality-risk-valuation.
———. Reviewing National Ambient Air Quality Standards (NAAQS): Scientific and Technical
Information. https://www.epa.gov/naaqs.
Funke, Christoph, Joshua Linn, Sally Robson, Ethan Russell, Daniel Shawhan, and Steven
Witkin. 2023. What Are the Climate, Air Pollution, and Health Benefits of Electric
Vehicles? https://www.r.org/publications/working-papers/what-are-the-climate-air-
pollution-and-health-benefits-of-electric-vehicles/.
Homan, Jeremy S., Vivek Shandas, and Nicholas Pendleton. 2020. The Eects of Historical
Housing Policies on Resident Exposure to Intra-Urban Heat: A Study of 108 US Urban
Areas. Climate 8(1): 12.
IPCC. 2014. Climate Change 2014: Synthesis Report. IPCC, Geneva, Switzerland. https://www.
ipcc.ch/report/ar5/syr/.
Jacobsen, Mark R., and Arthur A. Van Benthem. 2015. Vehicle scrappage and gasoline policy.
American Economic Review 105, no. 3: 1312-1338.
Kim, Youngseob, Lya Lugon, Alice Maison, Thibaud Sarica, Yelva Roustan, Myrto Valari, Yang
Zhang, Michel André, and Karine Sartelet. 2022. MUNICH v2. 0: A Street-Network Model
Coupled with SSH-Aerosol (v1. 2) for Multi-Pollutant Modelling. Geoscientific Model
Development 15(19): 7371–96.
Prioritizing Justice in New York State Climate Policy 39
Krewski, Daniel, Michael Jerrett, Richard T. Burnett, Renjun Ma, Edward Hughes, Yuanli Shi,
Michelle C. Turner, et al. 2009. Extended Follow-Up and Spatial Analysis of the American
Cancer Society Study Linking Particulate Air Pollution and Mortality. Vol. 140. Boston:
Health Eects Institute.
Lenox, Carol. 2013. EPA US Nine-Region MARKAL Database: Database Documentation. US
Environmental Protection Agency, Oice of Research and Development, National Risk
Management Research Laboratory.
Lepeule, Johanna, Francine Laden, Douglas Dockery, and Joel Schwartz. 2012. Chronic
Exposure to Fine Particles and Mortality: An Extended Follow-Up of the Harvard Six
Cities Study from 1974 to 2009. Environmental Health Perspectives 120(7): 965–70.
Liu, Jiawen, Lara P. Clark, Matthew J. Bechle, Anjum Hajat, Sun-Young Kim, Allen L. Robinson,
Lianne Sheppard, Adam A. Szpiro, and Julian D. Marshall. 2021. Disparities in Air Pollution
Exposure in the United States by Race/Ethnicity and Income, 1990–2010. Environmental
Health Perspectives 129(12): 127005.
Nardone, Anthony, Joan A. Casey, Rachel Morello-Frosch, Mahasin Mujahid, John R. Balmes,
and Neeta Thakur. 2020. Associations between Historical Residential Redlining and
Current Age-Adjusted Rates of Emergency Department Visits Due to Asthma across
Eight Cities in California: An Ecological Study. Lancet Planetary Health 4(1): e24–e31.
National Academies of Sciences, Engineering, and Medicine. 2017. Valuing Climate Damages:
Updating Estimation of the Social Cost of Carbon Dioxide. National Academies Press.
National Renewable Energy Laboratory (NREL). 2017. Electrification Futures Study. https://
www.nrel.gov/analysis/electrification-futures.html.
New York City Department of Health and Mental Hygiene. 2020. Air Quality and Health in
New York City (Data Brief No. 126). https://www1.nyc.gov/assets/doh/downloads/pdf/
epi/databrief126.pdf.
New York State Energy Research and Development Authority (NYSERDA). 2015. Residential
Statewide Baseline Study. Albany.
Peña-Parr, Victoria. 2020. The Complicated History of Environmental Racism. UNM
Newsroom. http://news.unm.edu/news/the-complicated-history-of-environmental-
racism.
Poblete-Cazenave, Miguel, and Shonali Pachauri. 2021. A model of energy poverty and access:
Estimating household electricity demand and appliance ownership. Energy Economics
98: 105266.
Poblete-Cazenave and Rao. Forthcoming. Social and contextual determinants of heat pump
adoption in the US: implications for subsidy policy design. Energy Research & Social
Science.
Rennert, Kevin, Frank Errickson, Brian C. Prest, Lisa Rennels, Richard G. Newell, William Pizer,
Cora Kingdon, et al. 2022. Comprehensive Evidence Implies a Higher Social Cost of CO
2
.
Nature 610(7933): 687–92.
Resnik, David B. 2022. Environmental Justice and Climate Change Policies. Bioethics 36(7):
735–41.
Schell, Christopher J., Karen Dyson, Tracy L. Fuentes, Simone Des Roches, Nyeema C. Harris,
Danica Sterud Miller, Cleo A. Woelfle-Erskine, and Max R. Lambert. 2020. The Ecological
and Evolutionary Consequences of Systemic Racism in Urban Environments. Science
369(6510): eaay4497.
Resources for the Future and New York City Environmental Justice Alliance 40
Shawhan, Daniel, Paul Picciano, and Karen Palmer. 2019. Benefits and Costs of Power Plant
Carbon Emissions Pricing in New York. https://www.r.org/publications/reports/
benefits-and-costs-of-the-new-york-independent-system-operators-carbon-pricing-
initiative/.
Spiller, Elisheba, Jeremy Proville, Ananya Roy, and Nicholas Z. Muller. 2021. Mortality Risk
from PM
2.5
: A Comparison of Modeling Approaches to Identify Disparities across Racial/
Ethnic Groups in Policy Outcomes. Environmental Health Perspectives 129(12): 127004.
Urban Land Institute. 2021. Environmental Justice and Real Estate: A Call to
Action for the Real Estate Industry. https://knowledge.uli.org/-/media/
files/research-reports/2021/environmental-justice-and-real-estate-final.
pdf?rev=487d4e5a7b614843a05e0c6ed4d4e4a6.
US Energy Information Administration (EIA). 2015. Residential Energy Consumption Survey.
https://www.eia.gov/consumption/residential/data/2015/.
———. 2018. Updated Buildings Sector Appliance and Equipment Costs and Eiciencies.
https://www.eia.gov/analysis/studies/buildings/equipcosts/.
———. 2021. Annual Energy Outlook. https://www.eia.gov/outlooks/aeo/tables_side.php.
———. 2023. https://www.eia.gov/state/data.php?sid=NY#Environment.
Zhang, Yang, Steve J. Smith, Michelle Bell, Amy Mueller, Matthew Eckelman, Sara Wylie,
Elizabeth L. Sweet, Ping Chen, and Deb A. Niemeier. 2021. Pollution Inequality 50 Years
after the Clean Air Act: The Need for Hyperlocal Data and Action. Environmental
Research Letters 16: PNNL-SA-160452.
Zhang, Hongliang, Gang Chen, Jianlin Hu, Shu-Hua Chen, Christine Wiedinmyer, Michael
Kleeman, and Qi Ying. 2014. Evaluation of a Seven-Year Air Quality Simulation Using the
Weather Research and Forecasting (WRF)/Community Multiscale Air Quality (CMAQ)
Models in the Eastern United States. Science of the Total Environment 473: 275–85.
Prioritizing Justice in New York State Climate Policy 41
Appendix A. Building the Policy Cases
The original plan for this research was to directly model the policies laid out in the draft
scoping plan, and adjustments desired by the environmental justice policy community in
New York. However, the draft scoping plan lacks the specificity (e.g., level of stringency
of individual policies, their timing) required to model policies, and the integration analysis
conducted by the state models outcomes rather than policies (e.g., “90 percent of vehicle
sales are electric by 2030”). Therefore, to proceed, the research team had to design
policies to deliver the state’s desired outcomes and fill in policy details where they were
missing.
The following general principles guided the development of both policy cases:
• We include only policies in sectors for which we have economic modeling capabilities
(on-road transportation, ports, residential buildings, and electrical power).
• We include only policies that are “modelable” by our group—that is, they predictably
aect inputs in our models that have a known or estimated relationship to outputs.
For example, a tax on gasoline has a known impact on fuel prices, which is an input
to our transportation model, so it can be easily incorporated into our modeling. On
the other hand, a subsidy for bicycle purchases has an unknown impact on miles
traveled in a car, so it cannot be incorporated in our modeling.
• We set stringency or ambition of policies by considering what behavioral responses
are credible for our models to address. For example, our transportation model, which
is parameterized by analysis of historical data, cannot credibly estimate the impact
of an EV subsidy of $100,000 per light-duty vehicle because there is no historical
record for a subsidy of this size.
• We determine type of policy, level of policy ambition, and timing of policy
implementation based on the following:
• precedents from other prominent and comparable jurisdictions, such as
California, which has implemented many of the policies included in the New
York scoping plan and is referred to frequently in the scoping plan;
• precedents from New York policy proposals, especially proposals that have
been introduced in the state legislature; and
• other approaches, as needed.
Because of modeling limitations, we did not include several policies of interest to
stakeholders:
• microgrid construction;
• detailed retail rate design (power sector–wide);
• renewable energy zones;
• state vehicle fleet electrification (including buses);
• technology investments (through research and development);
• congestion pricing;
Resources for the Future and New York City Environmental Justice Alliance 42
• ZEV credit trading;
• dealer incentives;
• residential demand response to peak–o peak prices;
• retrofit rebate options (modeled as fixed assumption);
• renter protections;
• interconnection investment;
• Clean Dispatch Credit Program;
• publicly subsidized financing for LDV ZEVs; and
• state MHDV ZEV procurement.
Prioritizing Justice in New York State Climate Policy 43
Appendix B. Background on Economic
Models
B.1. Economic Models
B.1.1. Power Sector
The Engineering, Economic, and Environmental Electricity Simulation Tool (E4ST) is
power sector modeling software built to project the eects of policies, regulations,
power infrastructure additions, demand changes, and more (E4ST 2022). E4ST
simulates in detail how the power sector will respond to such changes. It models
successive multiyear periods, predicting hourly generator and system operation,
generator construction, generator retirement, and various other outcomes in each
period. The E4ST model of the United States and Canada contains the 19,000 existing
generators with their detailed individual characteristics, tens of thousands of buildable
generators, including location- and hour-specific wind and solar data, and all of the
high-voltage (>200 kV) transmission lines as well as chronically congested lower-
voltage transmission lines. E4ST’s advantages over other models include its high
spatial detail, its realistic representation of power flows and system operation, its
integration of an air pollution and health eects model, its comprehensive benefit-
cost analysis capabilities, its high-quality generator data, its inclusion of Canada, and
its adaptability, transparency, and shareable nature. E4ST has been used to analyze
various policies and investments. It has also been used for multiple peer-reviewed
papers in leading journals. E4ST was developed by researchers at Resources for the
Future, Cornell University, and Arizona State University, with funding, input, and review
by the US Department of Energy, the National Science Foundation, the New York
Independent System Operator, the Power Systems Engineering Research Center, the
Sloan Foundation, Breakthrough Energy, and others.
B.1.2. Light-Duty Vehicles
This description of the light-duty model is adapted from Funke et al. (2022). The
model for light-duty vehicles contains two components: new-vehicle sales and on-
road fuel consumption. The first component characterizes vehicle sales by year (2018
through 2030) and region (California, other ZEV states, and all other states, to enable
an explicit representation of the ZEV program). On the demand side of the market,
consumers choose vehicles that maximize their subjective well-being, which depends
on the vehicle’s price, fuel costs, horsepower, size, and other features, such as all-wheel
drive. Preferences for those vehicle attributes vary across 60 demographic groups,
defined by income, age, urbanization, and geographic region.
Consumer preferences are estimated based on survey responses from 1.5 million
new-car buyers between 2010 and 2018. The survey data include information about
household demographics, such as income, as well as detailed information about the
Resources for the Future and New York City Environmental Justice Alliance 44
vehicle purchased. Vehicles are defined at a highly disaggregated level, with about 1,200
unique vehicle models oered each year. Consumer preferences for fuel costs, fuel type,
and other vehicle attributes are estimated from their vehicle choices.
Each manufacturer chooses vehicle prices and fuel economy (and decides whether to
introduce electric vehicles) to maximize profits while meeting regional ZEV standards
and federal fuel economy and GHG standards. Vehicle prices depend on marginal costs,
consumer demand, ZEV standards, and federal fuel economy and GHG standards.
Manufacturers select a larger markup of prices over marginal costs when consumer
demand is less sensitive to price. Because high-income consumers are typically less
price responsive than low-income consumers, markups tend to be higher for vehicles
purchased by high-income buyers than for vehicles purchased by low-income consumers.
The ZEV, fuel economy, and GHG standards cause manufacturers to reduce prices of
electric vehicles and increase prices of gasoline vehicles. These price changes help
manufacturers achieve the standards.
Each year, manufacturers also decide whether to introduce new electric vehicles to the
market. Vehicle production and entry costs, as well as shadow prices of the standards, are
estimated from observed choices of vehicle prices, fuel economy, and entry between 2010
and 2018, under the assumption that each manufacturer makes these choices to maximize
its own profits.
We simulate the equilibrium in a market (model year and region) given assumptions
about the total number of consumers in the market, fuel prices, battery costs, electric
vehicle subsidies, and standards. For each simulated market, the output includes entry of
new electric vehicles and prices, fuel economy, and sales of each vehicle. The number of
consumers in the market and fuel prices are taken from the EIA AEO 2021. Battery costs
are from 2021 projections by Bloomberg NEF. Marginal costs of electric vehicles decrease
over time in accordance with the vehicle’s battery capacity and the projected battery cost
reduction. Declining battery costs cause manufacturers to reduce electric vehicle prices
over time, all else equal.
The output of the new-vehicle component feeds into the on-road fuel consumption
component of the model. For each county and year, this component of the model
characterizes total gasoline and electricity consumption and tailpipe and upstream
emissions from vehicles owned by households. Vehicles are defined by fuel type (gasoline,
diesel fuel, electric, and plug-in hybrid), class (cars and light trucks), age, and county.
Simulations of the model begin with the stock of on-road vehicles in 2017 that is estimated
from the National Household Travel Survey (NHTS). We compute fuel consumption rates
for gasoline and plug-in hybrid vehicles by vehicle age, class, and state from the NHTS.
The state-level vehicle stocks and fuel consumption rates are disaggregated to the
county level using the Bureau of Transportation Statistics LATCH Survey.
At the beginning of the year, a fraction of vehicles are scrapped, where scrappage
rates depend on vehicle age, class, and vehicle price and are estimated from historical
registrations data from RL Polk. Scrappage rates are adjusted by registration taxes
according to estimates from Jacobsen and van Benthem (2015).
Prioritizing Justice in New York State Climate Policy 45
The on-road vehicle stock is augmented by the new vehicles sold in the vehicle market
component of the model. From that component, we compute new-vehicle sales by fuel
type, class, and region. We compute the average fuel consumption rate (gallons per
mile traveled) for gasoline and plug-in hybrids by region. The regional estimates are
disaggregated to the county level using the LATCH data.
Total national vehicle miles traveled (VMT) data are obtained from the AEO 2021. National
VMT is allocated across counties and vehicles according to the per mile fuel costs and
consumer driving preferences that are estimated from the 2017 NHTS and vary by vehicle
class and age. Compared with the baseline, a scenario with higher fuel costs causes total
VMT to decrease according to the assumed elasticity of VMT to fuel costs of –0.1. Fuel
costs also aect the distribution of VMT across vehicles.
The model is then iterated forward one year, and the entire process is repeated.
The output of the model includes VMT, tailpipe and upstream emissions, gasoline
consumption, and electricity consumption by fuel type, county, and year for 2017–2030.
B.1.3. Medium- and Heavy-Duty Vehicles
Prior to this work, there was no model that could predict MHDV flows for New York at the
resolution required to do air pollution modeling, or to estimate the local impacts from the
freight flows to local communities. The latter information is critical considering the goal of
estimating community impacts, especially to disadvantaged communities. To overcome
this limitation, this project developed a new modeling framework that integrates outputs
and information from a set of publicly available sources of socioeconomic data (e.g.,
Census, ZIP code business patterns), and other truck and economic models. The modeling
framework has three main modules (Figure B1).
The first module (M1) is used to estimate vehicle activity at a network link level. This
is a static representation of truck travel along the primary and secondary highways for
dierent vehicle types for the 2012 baseline and 2030 scenarios. To develop M1, the team
integrated outputs and data from the following sources:
• Freight Analysis Framework (FAF) version 4. FAF was developed by the Bureau
of Transportation Statistics and the Federal Highway Administration to model
aggregate freight flows throughout the nation. M1 uses FAF model outputs to gather
the aggregated multimodal freight flows in and out of the major regions in the state.
• New York Best Practice Model (NYBPM). NYBPM is a travel demand model for
the New York Metropolitan Transportation Council region with high resolution in
the following counties: Manhattan, Queens, Bronx, Brooklyn, Staten Island, Nassau,
Suolk, Westchester, Rockland, Putnam, Orange, and Duchess. Additionally, the
model estimates some flows in the network corresponding to some other regions.
• Freight and freight trip generation models for New York State.
• Public Commodity Flow Survey microdata. This information provides shipment-
level data on commodities, shipment distances, and modes.
Resources for the Future and New York City Environmental Justice Alliance 46
Overall, those various data sources allow estimating aggregated truck flows in the New
York network. Integrating the data sets involved several subprocesses. For example,
FAF and New York Metropolitan Transportation Council had dierent projection
years and vehicle definitions, as well as their geographic resolution. The team used
the various data sets to estimate vehicle type ratios to translate freight flows into
truck traic and estimate short- versus long-haul trip demand, and used indicators of
industry-generated flows to infer the vehicle type characteristics and behaviors. For
the projections, the process uses linear interpolation to estimate freight flows in the
FAF and NYBPM model results for 2030 because FAF projections were available for
only 2012 and 2045, and for the New York Metropolitan Transportation Council, only
2025 and 2035 data were available. Additionally, leveraging the increased resolution of
the NYBPM, the team estimated adjustment factors for the FAF model in urban areas
throughout the state. It was also necessary to create a crosswalk between the vehicle
definitions in FAF (two types), the NYBPM (four types), and the five truck definitions
in MOVES. The resulting five vehicle types include light commercial trucks (primarily
nonpersonal use) (32), single-unit short-haul trucks (52), single-unit long-haul trucks
(53), combination short-haul trucks (61), and combination long-haul trucks (62). The
final outputs of M1 are VMT per day or year on every network link (modeled) for the
baseline and future scenarios for the five vehicle types.
Figure B1. Key Components of MDHV Modeling Framework
M1
Estimate of
Freight/Truck
Activity
M2
Estimate of
Policy Impacts
on Fleet
Composition
M3
Estimate of
Emissions
Rates
Emissions
for MDHD
Vehicles in
New York
State
Prioritizing Justice in New York State Climate Policy 47
Module 2 (M2) integrates a truck vehicle choice model, a transportation transition
(truck turnover) model, and the design of policy scenarios. This was necessary
to evaluate the impact of policies to foster the introduction of ZEVs following the
California Air Resources Board’s Advanced Clean Truck (ACT) rule and the (still under
development) Advanced Clean Fleet program, among others discussed in the draft
scoping plan for New York State. Specifically, M2 uses the Transportation Transitions
Model (TTM), developed at the Institute of Transportation Studies Davis (ITS Davis),
which estimates fleet turnover based on sales target requirements (e.g., ACT)
considering assumptions about vehicle characteristics and travel activity. Because of
lack of New York data, the research team used assumptions drawn from their expertise
and the experience in California, extrapolating to assume that New York would follow
a similar trajectory as California. The main outputs of the TTM are stock turnover by
model year and major vehicle categories (e.g., diesel, ZEV).
M2 also uses the ITS Davis Truck Choice Model (TCM) to estimate the share of ZEV
technologies (e.g., battery electric, hydrogen fuel cell) that satisfy the transition
estimates from the TTM, and the level of incentives required to achieve such sales
targets. The TCM considers variables and factors such as vehicle specifications, price,
fuel or energy eiciency, incentives (e.g., purchase vouchers, infrastructure, feebates,
low-carbon fuel standard credits), operational and maintenance costs, and carbon and
co-pollutant costs, among others. The output of the TCM is the fleet mix per year by
share of technology.
The team then implemented the SPC and CPC through the TTM and TCP. The final
outputs of M2 are then the share of vehicle technologies for various vehicle categories
in the policy scenarios. It is important to note that the TTM and TCP estimates are at
the state level and are assumed to be uniform statewide. Additionally, the models used
in M2 consider the following vehicle categories: heavy-duty long-haul, heavy-duty
short-haul, medium-duty delivery, heavy-duty vocational, medium-duty vocational,
and heavy-duty pickup trucks. Considering the definitions of the vehicle types from
M1 (and MOVES), the team related M2 and M1 outputs to be consistent with the five
vehicle types from M1.
Finally, module 3 (M3) uses the Environmental Protection Agency’s Motor Vehicle
Emissions Simulator (MOVES 3) to generate an emissions profile for a vehicle fleet
in New York State. As the output of M3, the team estimated an average tailpipe
emission rates in grams per mile for various pollutants for the five vehicle types based
on MOVES estimates for the 2012 baseline, and a 2030 business-as-usual scenario.
Figure B2 shows a schematic of the key inputs, processes, and outputs of the various
modeling framework components.
Altogether, the modeling framework then generates a composite emissions rate for
the SPC and CPC, modifying the base rates from M3 and the outputs from M2. The
scenario-based composite rates are then used to estimate total tailpipe emissions at
the link level throughout the state. These emissions are aggregated with the emissions
from the light-duty sector and the port emissions to estimate emissions change factors
at a 36km
2
grid.
Resources for the Future and New York City Environmental Justice Alliance 48
B.1.4. Residential Buildings
A residential energy demand model was developed to predict the adoption and use of
space-conditioning equipment in households in New York State in 2030. The novelty
of the model is that, rather than using a representative household for all of New York,
the model predicts the probability of heating appliance ownership for a broad range of
households identified by a range of socioeconomic characteristics, as well as building
and climate conditions. The model outputs used in this project include state-level
electricity demand for heating and cooling and oil and gas demand for space and water
heating and cooking, by PUMA.
This model implements policies such as heat pump subsidies and building shell
eiciency standards for new construction and retrofits. We implement the fossil-
fuel phaseout in the SPC as a floor on heat pump adoption. Below, the model design
and methods for implementing these policies are described. More details about the
methodology can be found in Poblete-Cazenave and Rao (2023).
B.1.4.1. Model Design
The space-conditioning model used here is an extension and adaptation of the space-
conditioning module of the energy demand model first presented in Poblete-Cazenave
and Pachauri (2021), an indirect utility maximization model, where households choose
Figure B2. Modeling Framework Diagram
Joint Model Results & Estimating
2012 & 2030 Forecast
Input Data
External Model
Policy Input
Process
Output
Fuel/Energy
Costs
Carbon $
LCFS Credit
EIA
Truck Choice
Model
Fleet Mix
Per Year %
of
Technology
ACF
Assumed
as
Necessary
Condition
Sales Targets
(ACT)
Freight Analysis
Framework (FAF)
• 2012–2045
• Statewide
• Major Roads
NYMTC - Best
Practice Model
• 2012, 2025, 2035
• 12 NY Counties
• Major & Local Roads
NY-DOT Traic
Data Trip
Generation
Link-Level Annual Average Daily
Truck Traic (AADTT)
Link-Level VMT
Transportation
Transition
Models TTM
Stock
Turnover 5
Years by
Model Year
Fleet VMT
5 Years by
Model
Year
Other
Total
Incentives to
Achieve Sales
Target Per Year
Incentives
Purchase
Feebates
Infrastructure
Link-Level Annual Average Daily
Emissions for 5 MHHD Vehicle
Types
Emission Rates, Source
Composition
EIA, EPA, BTS,
EMFAC
MOVES
M1
M3
M2
Prioritizing Justice in New York State Climate Policy 49
among dierent appliances and fuels to satisfy their energy needs. The model is
estimated using simulation-based structural econometrics. The advantage of using
simulation-based modeling for our purposes lies in the ability to use dierent data sets
to create simulated households with characteristics obtained from multiple surveys,
and to simulate future populations with additional assumptions on population drivers.
We start with the 73,149 household observations in New York in the American
Community Survey as a base for the simulated households, which includes standard
socioeconomic attributes. Additional attributes about the building condition, such as
vintage and insulation, are imputed to these simulated households from the Residential
Energy Consumption Survey (RECS 2015) using a common set of attributes between
the two surveys. We project the stock of residences to 2030 using housing unit
and income projections, assuming the new stock reflects the current distribution of
household characteristics.
We separately use a multinomial discrete choice model to estimate the probability
of adoption of dierent heating technologies based on the Northeast US sample
of the American Housing Survey for 2015, which contains detailed information on
heating equipment, buildings, and socioeconomic attributes. The predictors include
socioeconomic household characteristics, including income, race, and age. Physical
conditions include floorspace, building shell conditions (e.g., insulation), and building
material type. We also included average building insulation R-values based on vintage,
which were modeled as heating appliance eiciency penalty factors. Climate conditions
(heating degree days) were dierentiated by climate zones.
We combine the estimated coeicients from this discrete choice model with the
parameter values of the simulated New York households in 2030 to obtain their
probability of owning dierent heating appliances.
Since the surveys are representative at the PUMA spatial scale, we present results as
appliance penetration rates and fuel consumption at the PUMA scale.
B.1.4.2. Energy Consumption
The model estimates fuel consumption for all heating appliances, as well as air-
conditioning consumption for cooling and gas consumption for water heating
and cooking, since these are required to determine air pollution estimates. Using
simulation-based estimation methods on Residential Energy Consumption Survey
data, the model obtains a distribution of energy consumption estimates for dierent
appliances, which, joined with estimated sociodemographic eects, are used to
calculate the utility-maximizing total energy consumption for each simulated
household (Poblete-Cazenave and Pachauri 2021). We scale up the heat pump
electricity consumption estimates for the Northeast, given that the underlying
survey data reflect ownership largely in warmer areas (South and Mid-Atlantic). We
use an engineering-based adjustment that takes into consideration building shell
characteristics, climate-adjusted heat pump eiciency, and heating degree days to
reflect theoretically expected consumption values. Finally, total county-level electricity
Resources for the Future and New York City Environmental Justice Alliance 50
consumption numbers are calibrated to match utilities’ monthly consumption data in the
base year for the state from NYSERDA, whereas gas consumption estimates are kept as
obtained from the model, given their proximity to utilities’ values.
B.1.4.3. Data
We use industry-standard rules of thumb for heating demand per square foot of floorspace
for dierent climate zones. Hence, heat pump costs vary with climate and size of
dwelling. For future technology cost and performance, we use US EIA’s Updated Buildings
Sector Appliance and Equipment Costs and Eiciencies (2018). For heat pump cost and
performance, we use the National Renewable Energy Laboratory’s Electrification Futures
Study (2017). The approximate average heat pump cost for the sample is $11,300.
For fuel prices, we use the high oil and gas supply case of EIA’s Annual Energy Outlook 2021.
For residential income growth, we use AEO 2021. For growth in residential units and climate
zone designations, we use the New York State Climate Action Council Draft Scoping Plan,
Integration Analysis Technical Supplement, Section I, Annex 1: Inputs and Assumptions.
For determining fossil fuel phaseout retirement schedules, we use NYSERDA’s Residential
Statewide Baseline Study 2015, Volume 1, based on the Single Family and Tenant Survey,
which has a breakdown of the share of households by age (Table 20).
For building shell R-values for future construction, we use the NYSERDA Stretch Codes. For
existing building shell R-values by vintage, we use data from the National Renewable Energy
Laboratory’s ResStock model.
B.1.4.4. Ports
To estimate port emissions for the 2012 baseline and 2030 future scenarios, the team
relied on a number of sources, notably past port emissions inventories for New York and
New Jersey, to develop a model to extrapolate emissions as a function of cargo-handling
equipment and intraterminal heavy-duty vehicle activity. For cargo handling, the team
considered equipment such as terminal tractors, straddle carriers, forklifts, and other
primary and ancillary equipment. The estimation process relies on two processes: first,
service hours for each type of equipment are based on container movements and hourly use
per year from inventory data, and then emissions factors per hour are used to estimate total
yearly emissions for various pollutants. Similarly, for the heavy-duty vehicle component,
inventory estimates of VMT at auto terminals, container terminals, and between terminal
warehouses are then multiplied by corresponding heavy-duty port and yard trucks’
emissions factors to estimate total emissions for the baseline. For 2030, the team estimated
container movement growth and used this average growth factor to expand the count of
cargo-handling equipment and heavy-duty vehicles’ intraport activity and the associated
emissions. Changes in emissions between 2012 and 2030 are estimated based on the
literature and scaled as a function of the relationship between 2030 and 2012. For the SPC
and CPC, based on experiences in California, the estimates assume a very high share of
electrification for equipment and yard trucks. The drayage movements outside terminals are
Prioritizing Justice in New York State Climate Policy 51
included in the truck flows modeled directly on the network. For drayage, the analyses
follow the same assumptions of the general fleet.
The port’s model considered the following facilities: Brooklyn Port Authority marine
terminal, Port Jersey Port Authority marine terminal, Elizabeth Port Authority marine
terminal, and the Howland Hook marine terminal. The analyses assume the same
emissions factors and policies across these facilities, although some are in New Jersey.
Resources for the Future and New York City Environmental Justice Alliance 52
Appendix C. Background on Air Quality
Modeling
Our air quality modeling produces PM
2.5
concentrations at a grid resolution of 4km
2
across New York State for 2012 and 2030 under the BAU and two policy scenarios.
Two air quality models were used: a 3-D air quality model containing comprehensive
representations of atmospheric transport, physics, and chemistry to simulate PM
2.5
concentrations at a grid resolution of 36km
2
(Weather Research and Forecasting model
with chemistry extension, WRF-Chem), and a computationally eicient statistical model
utilizing the 36km
2
simulation results from the 3-D air quality model to predict PM
2.5
concentrations at a grid resolution of 4km
2
. The 3-D air quality model can be applied at
both grid resolutions, but it is prohibitively expensive to run the year-long simulation at
the 4km
2
resolution.
WRF-Chem is a state-of-science online-coupled weather-chemistry model. The newest
version, WRF-Chem v4.3, is used. The results are compared with those from an older
version of WRF-Chem v3.7 and the Community Multiscale Air Quality (CMAQ) model
v5.0.2, generated previously under a separate project.
Figure C1 shows the US modeling domain at a horizontal grid resolution of 36km
2
(D01)
with an insert showing the 4km
2
modeling domain, centered on New York State.
Figure C1. Domain of Air Quality Modeling
Prioritizing Justice in New York State Climate Policy 53
Figure C2 shows the 36km
2
grid resolution spatial distributions of simulated (a) max
8-h O3 overlaid with surface observations from AQS and CASTNET (b) daily PM
2.5
overlaid with surface observations from CSN and IMPROVE, and normalized mean
bases (NMBs) of annual mean (c) max 8-h O
3
and (d) daily PM
2.5
predicted by WRF-
Chem v4.
Table C1 shows the domain-mean performance statistics for max 8-h O
3
and daily PM
2.5
against data from several surface networks. The NMBs are within ±11 percent and the
NMEs are <14 percent for O
3
, and the NMBs are within ±9 percent and the NMEs are
<50 percent for PM
2.5
, which are well within the thresholds for good performance.
Table C1 also compares the performance statistics for the 3-D simulations of 2012 at
36km
2
over CONUS using three models including WRF-Chem v4.3, WRF-Chem v3.7,
and CMAQ v5.0.2. The model used for this project, WRF-Chem v4.3, shows the best
performance for PM
2.5
against CSN with NMBs of -2.3 percent vs. 21.3 percent by
CMAQ v5.0.2 and 8.8 percent by WRF-Chem v3.7. For daily PM
2.5
against IMPROVE,
WRF-Chem v4.3 gives an NMB of 8.4 percent, which is better than 9.6 percent by WRF-
Chem v3.7. For daily PM
2.5
against AQS, WRFChem v4.3 gives an NMB of –5.0 percent,
which is better than –7.8 percent by CMAQ v5.0.2. Overall, WRFChem v4.3 performs
better for PM
2.5
predictions against surface observations when compared with WRF-
Chem v3.7 and CMAQ v5.0.2.
To translate the predictions to a 4km
2
grid resolution, we interpolate the 36km
2
model
predictions of PM
2.5
concentrations to a modeling domain centering on New York
State at the 4km
2
grid resolution, and validate the interpolated PM
2.5
predictions with
Figure C2. Spatial Distributions of Annual Means Predicted by
WRF-Chem v4.3 at a Grid Resolution of 36km (2012)
Resources for the Future and New York City Environmental Justice Alliance 54
available surface observations from the US EPA AQS data network. The interpolated
PM
2.5
predictions for 2012 show excellent performance, with an NMB of 4 percent and
NME of 40 percent, which is consistent with the performance of WRF-Chem PM
2.5
predictions at 36km
2
.
Table C1. Performance Statistics for 3-D Simulation of 2012 at 36km over CONUS
Species/
network/
units
Mean
obs
WRF-Chem v4.3 WRF-Chem v3.7 CMAQ v5.0.2
Mean
sim
R MB
NMB
(%)
NME
(%)
Mean
sim
R MB
NMB
(%)
NME
(%)
Mean
sim
R MB
NMB
(%)
NME
(%)
Max 8-hr O
3
AQS (ppb)
44.4 49.2 0.4 4.8 10.8 14.6 43.2 0.5 -1.2 -2.8 10.1 52.3 0.6 7.9 17.7 18.0
Max 8-hr O
3
CASTNET
(ppb)
41.7 39.3 0.3 -2.4 -5.8 12.4 35.1 0.5 -6.6 -15.9 17.6 44.7 0.6 3.0 7.1 10.5
PM
2.5
CSN
(μg m
-3
)
10.2 10.0 0.5 -0.24 -2.3 21.0 11.1 0.5 0.9 8.8 22.9 12.4 0.5 2.2 21.3 28.9
PM
2.5
IMPROVE
(μg m
-3
)
4.6 5.0 0.6 0.4 8.4 32.1 5.1 0.7 0.5 9.6 28.8 4.7 0.7 0.1 1.4 37.0
PM
2.5
AQS
(μg m
-3
)
8.9 8.5 0.3 -0.4 -5.0 50.0 9.1 0.3 0.1 1.5 50.5 8.2 0.4 -0.7 -7.8 48.8
Prioritizing Justice in New York State Climate Policy 55
Appendix D. Identifying Disadvantaged
Communities
When our work started, we did not have the final Climate Justice Working Group
definition of disadvantaged communities, but we did have the preliminary list of
indicators being used to calculate the index. The team at the Yale School of Public
Health worked with this list of indicators, sourced their own data, and experimented
with dierent index designs prior to the release of the CJWG index methodology. Their
work on index design is the subject of an upcoming working paper. For this report, we
leveraged the same methodology as CJWG, to create alignment. When “disadvantaged
communities” are referenced in the report, we are referring to the list defined by the
Climate Justice Working Group.
1
The index leverages 44 statewide indicators at the census tract level representing
environmental burdens, climate change risks, population characteristics and health
vulnerabilities. The index is based on two main groups of statewide indicators at the
census tract level:
• Environmental burdens and climate change risks (19 indicators)
• potential pollution exposures
• land use associated with historical discrimination or disinvestment
• potential climate change risks
• Population characteristics and health vulnerabilities (25 indicators)
• income
• education and employment
• race, ethnicity, and language
• health impact and burdens
• housing, energy, and communications
For each indicator, we calculated the percentile rank (0–100) for a given census tract
across all census tracts in the state. The use of percentiles weakens the impact of
extreme values for a given indicator and can represent a relative score for a census tract
for that indicator. For certain types of land use (e.g., remediation sites, power generation
facilities), since a significant number of census tracts have zero values, we directly
allocated a zero percentile to these census tracts and recalculated the percentile ranks
for the remaining census tracts with nonzero values.
Next, we calculated the weighted average of indicator percentile ranks within each group
(0–100) for a given census tract. Certain metrics within groups were given double the
weight: for example, in the population characteristics and health vulnerabilities group,
income less than 80 percent of the area median income, income less than 100 percent of
federal poverty line, Latino/a or Hispanic, and Black or African American (Figure D1).
1 A map and description of the criteria can be found here: https://climate.ny.gov/Resourc-
es/Disadvantaged-Communities-Criteria
Resources for the Future and New York City Environmental Justice Alliance 56
Figure D1. Indicators and Their Respective Weights Used to Construct the CHVI
Environmental burdens and climate change risks
Potential pollution
exposures
Land use associated with historical
discrimination or disinvestment
Potential climate
change risks
1X
1X
1X
1X
1X
Vehicle traic density
Diesel truck and bus traic
Particulate Matter (PM
2.5
)
Benzene concentration
Wastewater discharge
1X
1X
1X
1X
1X
Remediation sites
Regulated Management Plan
(chemical) sites
Major oil storage facilities
Power generation facilities
Active landfills
1X
1X
1X
1X
Municipal waste combustors
Scrap metal processors
Industrial/manufacturing/mining land use
Housing vacancy rate
1X
1X
1X
1X
1X
Extreme heat projections
(>90°F days in 2040–2069)
Projected flooding risk
Low vegetative cover
Agricultural land
Driving time to hospitals
1X
1X
2X
Population characteristics and health vulnerabilities
Income, education,
& employment
1X
Race, ethnicity,
& language
1X
Housing, energy, &
communications
1X
Health impacts
& burdens
1X
2X
1X
1X
1X
<80% area median
income
<100% of federal
poverty line
Without bachelor’s
degree
Unemployment rate
Single-parent
households
1X
2X
Latino/a or Hispanic
Black or African
American
Asian
1X
Native American or
Indigenous
1X
Limited English
proficiency
1X
1X
1X
Asthma ED visits
COPD ED visits
Heart attack
hospitalizations
1X
1X
1X
Premature deaths
Low birthweight
Without health
insurance
1X
1X
With disabilities
Adults age 65+
1X
Historical redlining
score
2X
2X
1X
1X
Renter-occupied
homes
Housing cost burden
(rental costs)
1X
1X
Energy poverty/cost
burden
Manufactured homes
1X
1X
Homes built before
1960
Without health
insurance
Prioritizing Justice in New York State Climate Policy 57
Figure D2. Formula for Calculating the CHVI
Environmental
burdens and
climate change
risks
Population
characteristics
and health
vulnerability
Climate health
vulnerability
index
The score for a given census tract was calculated as the product of its percentile rank
in each of the two main groups (Figure D2). Then the Climate Health Vulnerability
Index (0–100) was calculated as the percentile rank of the final score for a given
census tract among all census tracts in New York State. This final score represents
each census tract’s relative ranking in the state.
Resources for the Future and New York City Environmental Justice Alliance 58
Below, census tract 41 at the Bronx (census tract ID 36005004100) is used to illustrate
the construction of the Climate Health Vulnerability Index.
Figure D3. Percentile Rank of Each Indicator in Both Criterea Groups for Census Tract 41
(The Bronx)
Environmental Burdens and Climate Change Risks
Potential pollution exposures
Indicator Raw value Percentile
Vehicle traic density 700.74 51.1
Diesel truck and bus traic 372.52 42.3
PM
2.5
7.76 50.3
Benzene 0.71 99.5
Wastewater discharge 0.00086 48.4
Overall factor score 58.32
Land use associated with historical discrimination or disinvestment
Indicator Raw value Percentile
Remediation sites 1.03 52.4
Regulated Management Plan (chemical) sites 1.00 85.5
Major oil storage facilities (incl. airports) 1.18 73.1
Power generation facilities 1.73 81.3
Active landfills 0 0.0
Municipal waste combustors 0 0.0
Scrap metal processors 0.10 46.2
Industrial/manufacturing/mining land use 2.36 67.9
Housing vacancy rate 4.60 22.8
Overall factor score 47.68
Potential climate change risks
Indicator Raw value Percentile
Extreme heat projections 40.60 78.06
Coastal and inland flood projections 2.92 94.32
Low vegetative cover 96.17 69.68
Agricultural land 0 0.00
Driving time to hospitals 4.68 11.79
Overall factor score 50.77
Prioritizing Justice in New York State Climate Policy 59
Population Characteristics and Health Vulnerabilities
Income, education, and employment
Indicator Raw value Percentile
<80% area median income 93.73 99.5
<100% of federal poverty line 46.89 99.0
Without bachelor’s degree 88.86 92.2
Unemployment rate 21.5 99.1
Single parent households 26.14 98.8
Overall factor score 98.17
Race, ethnicity, and language
Indicator Raw value Percentile
Latino/a or hispanic 70.5 96.9
Black or African American 38.5 82.2
Asian 0.2 9.2
Native American 3.7 91.8
Limited English proficiency 30.26 93.1
Historical redlining score 4 100.0
Overall factor score 81.54
Health impacts and burdens
Indicator Raw value Percentile
Asthma emergency department visits 480.2 100.0
Chronic obstructive pulmonary disease emergency department Visits 135 86.4
Heart attack (myocardial infarction) hospitalization 12.8 99.3
Premature deaths 41.83 94.7
Low birthweight births 7.84 80.5
Without health insurance 10 86.3
With disabilities 16.9 84.7
Adults aged 65 and above 6.3 4.6
Overall factor score 79.55
Housing, energy, and communications
Indicator Raw value Percentile
Renter-occupied homes 87.48 90.0
Housing cost burden (rental costs) 707 8.6
Energy poverty/cost burden 4 82.4
Manufactured homes 0 0.0
Homes built before 1960 48.45 35.9
Without internet (home or cellular) 26.1 84.4
Overall factor score 50.20
Resources for the Future and New York City Environmental Justice Alliance 60
The figure below shows how the final figure for the tract is calculated.
Environmental burdens and climate
change risks
Population characteristics and health vulnerabilities
Potential
pollution
exposures
Land use
associated
with historical
discrimination or
disinvestment
Potential
climate
change risks
Income,
education,
and
employment
Race,
ethnicity,
and
language
Health
impacts and
burdens
Housing,
energy, and
communications
Factor scores 58.3 47.7 50.8 98.2 81.5 79.5 50.2
Weighted
average of
factor scores
[1 (58.32) + 1 (47.68) + 2 (50.77)]/(1+1+2)
= 51.88
[98.17 + 1 (81.54) + 1 (79.50) + 1 (50.20)]/(1+1+1+1)
= 77.36
Climate health
vulnerabilitiy
index score
51.88 x 77.36 = 4013.44
Climate health
vulnerabilitiy
index
percentile
97.8
Figure D4. Calculation of the Final Climate Health Vulnerability Index Score for Census
Tract 41 (The Bronx)
Prioritizing Justice in New York State Climate Policy 61
Appendix E. Supplementary
Methodologies
E.1. Model Integration and Coordination
Our energy models operate independently of one another, but outputs of one may
inform the inputs of another. For example, retail electricity prices may aect incentives
to install an electric heat pump, or how much that heat pump is used. But how many
heat pumps are operating also may aect electricity prices. Our model is not designed
to find a general equilibrium solution, so we do our best to match electricity price
and demand across model runs without that functionality. Our transportation models
leveraged AEO electricity prices in the BAU case and increase prices proportional
to the increases projected by the power sector for the policy cases. The residential
model considers prices directly from the power sector model, iterating until it finds
the appropriate combination of electricity price and residential demand. Electricity
demands from residential and transportation sectors are passed to the power sector
model for final emissions projections.
E.2. Ancillary Pollutant Valuation
To address the state carbon tax proposal, emissions taxes on conventional pollutants
contributing to PM
2.5
(NO
X
, SO
2
, and direct particulates) are needed by sector. We
assumed that the taxes would equal the dollar benefits to the US per ton of emissions
reduced in New York State. The literature oers such estimates for regions and cities,
as well as by source because sources (e.g., power plants) have a dierent pattern of
dispersal and chemical transformation than ground-level sources (e.g., transportation,
home heating by natural gas). But the literature doesn’t provide these estimates for
New York. To get those, we ran the COBRA model (formally, the CO-Benefits Risk
Assessment Health Impacts Screening and Mapping Tool; https://cobra.epa.gov/),
which assumes that the benefits of NO
X
and SO
2
emissions reductions (to reduce
PM
2.5
concentrations) are additive and separable. The model includes the benefits
of reducing PM
2.5
concentrations on human health and values these benefits using
standard unit values from the environmental economics literature.
The benefits from reducing pollution (2017$/short ton) from COBRA are as follows:
Table E1. Benefits from Reducing Polution (2017$/short ton) from COBRA
Electricity Vehicles Residential fuel
PM
2.5
231,965 465,556 682,730
SO
2
36,382 59,664 55,507
NO
X
9,025 14,355 19,456
Resources for the Future and New York City Environmental Justice Alliance 62
E.3. Methane
Decarbonization goals are defined in terms of carbon dioxide equivalent (CO
2
e) rather
than only CO
2
. Of the many other greenhouse gases, the most important one, and the
only one we track in this project, is methane. Since we are not modeling agriculture
or waste dumps, our focus is solely on upstream methane emissions from oil and gas
wells and how these emissions aect the accounting for meeting the decarbonization
goals. We basically need two sets of information: the leak rate per final product
consumed (gasoline, diesel, electric power) and the global warming potential of
methane to CO
2
to transform the methane emissions into CO
2
e. For the transportation
sector, we used rates of 1.87 kg/gallon of diesel fuel and 1.79 kg/gallon of gasoline. We
assume methane leakage of 0.000434 short tons per million Btu of natural gas use and
0.000174 short tons per million Btu of coal use, taken from Lenox et al. (2013), a source
that includes coal and whose natural gas leakage estimates have stood up well in light
of more recent research about methane leakage associated with natural gas extraction,
transportation, and processing. This methane leakage rate for natural gas implies
that approximately 2.4 percent of natural gas leaks. In line with the CLCPA-related
documentation, we use the 20-year global warming potential, which is 85 (IPCC 2014),
except where otherwise noted.
Prioritizing Justice in New York State Climate Policy 63
Appendix F. Comparison with New
York State’s Analysis
NYSERDA, DEC, and other state agencies worked together to perform their own
analyses of various policy options the state might take to meet its 2050 goals. Our
eort is independent of the state’s eort but was developed in close consultation
and interaction with the state agencies and full disclosure of our approach and
assumptions. Still, we did not want to merely duplicate the state’s approach because
we wanted to maximize our contribution to the debate about the policy pathways
going forward. Here we detail the main points of dierences.
1. When we model certain policies, we use “behavioral” models that confront
electric utilities, vehicle buyers, building residents, trucking companies, and port
operators with specific policies and incentives for action and then, based on past
behavior, record how they respond to those incentives. This is very dierent from
the state’s approach (conducted by the consulting firm E3), which is a “pathway”
model that assumes how the economy will respond and then tracks the eect of
that response on CO
2
emissions and (to a certain extent) air quality levels.
As an example, consider the eect of raising the gasoline tax (not a policy
modeled). Our model would raise that tax a given amount and, through previously
estimated equations tracking how people behave in buying gasoline and electric
vehicles of all types and how their driving changes, estimate CO
2
and vehicle NO
X
and SO
2
emissions from the results. The E3 analysis, in contrast, assumes that x
fewer gasoline vehicles and y more electric vehicles will be purchased without
using a behavioral model.
2. Because of the necessity of having access to behavioral models, we are modeling
only some sectors in New York responsible for emissions: residential buildings,
on-road transportation, ports, and electricity generation. These sectors make
up a significant portion of statewide carbon emissions, but we are not including
commercial buildings, industry, waste, or agriculture emissions in our projections.
E3’s analysis is comprehensive across sectors, but we prioritized modeling
sectors where we had some spatial distribution of emissions in our results, which
is critical to understanding impacts on local air quality.
3. Our air quality modeling is more sophisticated and spatially granular than the
state’s (see Appendix C). We have opted to pursue the most spatially granular
modeling level that balances modeling capabilities, computational resources,
and our desire to identify air quality changes at a fine spatial scale. To model the
entire state, this is the finest scale possible within the timeline and resources
available to us. In contrast, the state’s modeling eort uses COBRA, which
estimates air quality outcomes at the county level.
4. Because of budget limitations, our emissions and air quality projections are only
for 2030. Many decarbonization policies will begin to take eect in the next
several years and significant emissions reductions will need to take place by 2030
for New York to meet the goals set out in the CLCPA. However, for many policies,
the bulk of the impacts will happen after 2030, over the next several decades.
Resources for the Future and New York City Environmental Justice Alliance 64
Some policies we investigate may make little headway on emissions by 2030,
but we include them for completeness. In future work, we would be interested in
projecting these policy cases to 2050 and beyond to better capture their long-
term emissions and air quality impacts.
The state’s analysis includes 2030 results, so there is a point of comparison with ours,
keeping the above caveat in mind. However, the state models emissions changes in
2050 as well.
Prioritizing Justice in New York State Climate Policy 65
Appendix G. Research Limitations and
Caveat
Several limitations to this study result from study design choices, modeling limitations,
data limitations, and the like, indicating areas for further research.
When modeling community exposure to air pollution, it is ideal to have the most
geographically granular analysis possible, given that actual pollution exposure may
vary at a level as granular as a city block.
2
As mentioned above, our analysis is at
the 4km
2
level. Although this is substantially more granular than the county-level
analysis oered by the COBRA model (used in the state’s analysis), it limits our ability
to determine hyperlocal dierences in air pollution exposure. The large scale and
complex air quality models, such as WRF-Chem and CMAQ, can be downscaled to a
grid resolution of 1km
2
or less, but they have limited ability to confidently predict air
quality at this scale because of limitations in model inputs (e.g., hyperlocal emissions)
and representations of some atmospheric processes (e.g., turbulence, mixing,
and chemistry at street intersections and above urban street canyons). Although
hyperlocal air quality modeling methods do exist (Kim et al. 2022), it was determined
that using them for this statewide and cross-sector project would likely lead to false
precision, primarily because of the lack of statewide hyperlocal air quality data
3
required to validate the modeling. Further, applications of those models would require
detailed information at hyperlocal scales (e.g., traic fleets and emissions, urban
street geometry, building dimensions and energy consumption) that would require
considerable time and resources to develop.
To partly address this limitation in our air quality projections, we provide more localized
details (for some sectors) on the emissions projections that drive air quality changes.
For example, our medium- and heavy-duty transportation model estimates emissions
at the road link level, and our power sector model estimates emissions at specific
power plants. We explore these outputs in the Location of Emissions Changes section
in the main text and below (Appendix J).
Another limitation is that we model policy cases as a bundle, rather than as individual
policies. Ideally, we would be able to test sensitivity of dierent policy ambitions (e.g.,
a $100 subsidy versus a $500 subsidy) and test dierent combinations of policies
to better understand the potential air quality impacts of individual policymaking
decisions. However, the modeling process is computationally expensive and time
intensive, limiting our ability to add more dimensions to this analysis. Our approach
2 https://engineering.berkeley.edu/news/2021/09/google-street-view-study-shows-air-
pollution-by-block/
3 The New York Department of Environmental Conservation has begun an initiative to
gather hyperlocal air quality data, covering communities with a cumulative population
of about 5 million people at the time of publication. Eorts like this will provide the
foundation of data to support hyperlocal air quality modeling in the future. For more
information, see https://www.dec.ny.gov/chemical/125320.html.
Resources for the Future and New York City Environmental Justice Alliance 66
restricts our final analysis to comparing cases as a bundle of policies, rather than
individual policy changes. We believe attributing air quality changes directly to
individual policies is a rich area for continued research. Detailed explanations of how
policies aect economic decisions and emissions are in the Economic Modeling Results
section in the main text and below (Appendix H).
Similarly, because of cost and time constraints in our modeling project, we model only
a single policy year, 2030. This decision limits our ability to show the full eects of
the CLCPA, which runs through 2050. This decision is especially limiting for policies
that take time to aect aggregate emissions, such as policies to decarbonize the
transportation fleet through mandates and subsidies for the sale of new vehicles
(which represent a small percentage of the on-road fleet). Furthermore, we do not
model all sources of New York emissions. Most notably, our current modeling does
not incorporate commercial buildings or industrial facilities other than electric power
generation. These two features of our analysis limit our ability to estimate the full eect
of CLCPA implementation on New York’s air quality.
Despite those limitations, our research is an ambitious undertaking to understand the
variable PM
2.5
pollution impact of decarbonization polices in New York communities.
This is just the first step in investigating this relationship, and many research
opportunities remain, including studying the eects of additional pollutants. Again,
limitations in budget made this impossible, but PM
2.5
is the pollutant of most concern
and of higher impact than ozone, NO
X
, SO
2
, CO, and lead—the other pollutants covered
by National Ambient Air Quality Standards under the Clean Air Act (EPA.gov).
4
The caveat concerns an error discovered very late in our project involving direct
PM
2.5
emissions and NO
x
emissions from the transportation sector passed on to the
air quality model. Tests on the eect of this error on our air quality simulations reveal
minor to trivial dierences between the SPC and CPC, which are the focus of this
project. The emissions results reported in the Greenhouse Gas, PM
2.5
, and Precursor
Emissions Results section are unaected, as are our main findings and conclusions.
To pass on the modeled emissions associated with our BAU and policy cases to the
air quality model, we aggregated the emissions from the two separate transportation
models—one for LDVs and the other for HDVs. We discovered that this aggregation
process, which is complex because of diering spatial resolution across the two
models, led to an underestimate of direct PM
2.5
reductions and an overestimate of
NO
x
reductions. The error was substantial in calculating emissions changes between
2012 and 2030, but the error occurred in a relatively uniform manner across policy
cases (2030 BAU, CPC, SPC). The dierences in emissions between the policy cases
(measured as a percentage of the 2012 baseline) were much smaller.
4 Evidence suggests NO
x
on its own can have significant impacts on respiratory disease
(César et al. 2015), and it may have even more dramatic disparities between communities
(Liu et al. 2021)
Prioritizing Justice in New York State Climate Policy 67
Specifically, a 36km
2
grid of the transportation emissions change factors (measured as
the percentage change from the 2012 baseline) reveals that the maximum dierence in
direct PM
2.5
emissions in any grid cell between the CPC and SPC is about 3.5 percent
of the 2012 baseline with the incorrect emissions. The maximum dierence in the
corrected emissions is also 3.5 percent of the 2012 baseline. However, in the corrected
emissions there is more variation in emissions changes across grid cells.
Table G1 shows results from an example grid cell where the error is particularly
pronounced. The first two columns show the reductions in each policy case relative
to the 2012 baseline. There are large dierences between the incorrect and correct
results. The third column shows the dierence between the policy cases. This is what
we are most interested in. The impacts are clearly much smaller (on the order of a few
percentage points). All values are percentage reductions from the 2012 baseline.
Resources for the Future and New York City Environmental Justice Alliance 68
Appendix H. Economic Modeling
Results
Here we describe how each policy case (CPC and SPC) aects technology adoption
and behavioral choices that can influence emissions levels. The policies we modeled
have a wide range of ambition and vary in their timelines for implementation. The
results illustrate how far each policy case goes in pushing behavior that will lead to
decarbonization and air quality improvements.
H.1. Electricity Sector
H.1.1. Electricity Demand and Price
Both policy cases increase total New York electricity consumption (inclusive of
transmission and distribution losses) because of the high rates of electrification in the
residential and transportation sectors. Under the CPC, electricity demand increases by
17 percent, and under the SPC, by 29 percent, compared with the BAU. Table H1 shows
overall electricity demand, the electricity price, and the share of demand met by in-
state generation under the two policy cases and BAU.
Table H1. New York State Electricity Consumption and Wholesale Prices
BAU 2030 CPC 2030 SPC 2030
Electricity demand 155 million MWh 182 million MWh 200 million MWh
Electricity price $98/MWh $107/MWh $116/MWh
Share of demand met
in-state
89% 96% 96%
Prioritizing Justice in New York State Climate Policy 69
Both policy cases also lead to modest increases in wholesale electricity prices
compared with the BAU (as driven especially by increased electricity demand and
the economywide carbon-pricing policy)—a 10 percent increase for CPC and an 18
percent increase for SPC. Both policy cases also increase the proportion of electricity
consumption met from within the state. Several policies contribute to this trend:
• the border electricity price adjustment mechanism;
5
• the 70 percent renewable portfolio standard (which we assume must eectively
be satisfied by generation within the state);
• the large required (per CLCPA) amounts of distributed solar, oshore wind, and
electricity storage capacity within the state; and
• greatly increased electric vehicle requirements in the other ZEV states, three
of which are adjacent to New York. They increase electricity demand in those
states, partially osetting the eect of New York’s increased electric vehicle
requirements by reducing the availability of generation from those states to sell
electricity into New York and increasing those states’ demand for generation from
New York.
In-state generation in the CPC is also bolstered by the continued operation of the
Ginna and Nine-Mile-Point nuclear generators (in the BAU and SPC, these large,
nonemitting generators are retired by 2030). Together, these policy elements more
than oset the downward eects that increased electricity demand and the power
plant emissions fees in the CPC and SPC would otherwise exert on the in-state
generation share.
H.1.2. Generation Mix
Both policy cases prompt a dramatic increase in clean energy generation, relative to
the BAU. Table H2 shows the level of each generation type in each policy case, along
with the percentage of total generation from each generation type.
5 The border mechanism sets a fee on electricity imports and an equal rebate on electricity
exports to partially oset the competitive disadvantage that New York’s CO
2
e fee
creates for New York generators, as described in the scenario descriptions section
above. The fee and rebate are $11 in the CPC and $27 in the SPC.
Resources for the Future and New York City Environmental Justice Alliance 70
SPC policies boost renewable generation and storage capacity above the CPC (30
percent boost for solar, 40 percent boost for wind, and a nearly 200 percent increase
in storage capacity) and cut nuclear, natural gas, and waste-fueled generation (roughly
35 percent less for nuclear and 30 percent less for natural gas and waste-fueled
generation).
The natural gas generation reductions are especially large for the higher-emitting
natural gas generator types, which use steam cycles alone or combustion turbines
alone. The policy dierences that account for this further reduction in fossil-fueled
generation are the fees or caps on NO
X
, PM
2.5
, and SO
2
emissions, the higher fee on
CO
2
and methane emissions, the higher border electricity price adjustment ($27
instead of $11), the prohibition on new natural gas–fueled generation capacity, and the
Table H2. New York State Generation Mix (MWh and Percentage)
Generation
Source
BAU 2030 Percentage CPC 2030 Percentage SPC 2030 Percentage
Total
generation
138,471,392 100% 174,251,411 100% 191,065,047 100%
Nuclear 17,302,119 12% 27,069,379 16% 17,302,118 9%
Coal — 0% — 0% — 0%
Natural gas 40,975,962 30% 14,589,190 8% 10,571,238 6%
With CCUS — 0% 10 0.00% — 0%
Solar 33,708,907 24% 73,055,753 42% 94,337,759 49%
Distributed
solar
3,600,547 3% 19,808,517 11% 19,593,577 10%
Wind 15,578,218 11% 29,129,736 17% 40,257,449 21%
Hydro 27,870,324 20% 27,863,663 16% 27,849,494 15%
Geothermal — 0% — 0% — 0%
Storage –68,199 –0.05% –518,083 –0.30% –1,511,674 –1%
Hydrogen — 0% 1 0.00% — 0%
Waste &
biomass
3,076,401 2% 3,044,666 2% 2,258,635 1%
Other 27,660 0% 17,107 0% 28 0%
Prioritizing Justice in New York State Climate Policy 71
requirement that all of the state’s fossil-fueled peakers retire, not just those subject to
the current New York City peaker retirement requirement.
SPC policies increase solar, wind, and storage generation relative to CPC policies.
These increases are not directly mandated but instead result from the expected price
changes caused by the policy dierences in the two cases. In particular, those policy
dierences raise electricity prices while creating stronger disincentives for energy
generation methods that cause emissions, which allows more solar, wind, and battery
storage capacity to be profitable.
CPC policies have very little eect on waste-fueled generation, reducing it by just 0.03
TWh. SPC policies, specifically the SO
2
and NO
X
emissions fees, do reduce waste-fueled
generation. We have not allowed either case to change waste-fueled capacity (building
or closing facilities). In reality, this could occur, in which case the eects of the policies
on waste-fueled generation would be larger.
The SPC explicitly bans new hydrogen and CCUS plants, but even in the CPC, where
they are permitted, there is essentially no hydrogen or CCUS buildout by 2030.
H.2. Residential Building Sector
In the residential sector modeling, electricity demand (for heating and cooling) and
natural gas and diesel (“heating oil”) consumption are driven by future estimates of
how readily heat pump technologies are adopted instead of alternatives, such as diesel
boilers. In Table H3, we summarize heat pump adoption results for the BAU, CPC, and
SPC (the percentage of total households in the state that have adopted heat pumps).
In addition to the statewide adoption rates listed, we find a range of adoption rates
across counties in each case. For instance, the adoption rate varies by county from 77
to 96 percent in the SPC, from 27 to 78 percent in the CPC, and from 2 to 15 percent
in the BAU case. The higher adoption rates tend to be in the southeastern part of the
state, such as Staten Island and Long Island.
These adoption rates are largely determined in the model by the relative cost
(including incentives to encourage adoption) of heat pumps compared with fossil fuel
alternatives like boilers. Adoption rates are also influenced by the regional climate
across the state, which determines how much a given heating or cooling technology
is used and therefore how much cost savings a household receives from the more
eicient heat pump.
Table H3. New York Homes with Heat Pumps
BAU 2030 CPC 2030 SPC 2030
Heat pump adoption 8% 54% 90%
Resources for the Future and New York City Environmental Justice Alliance 72
As shown in Table 1 in the main text, the SPC has the highest heat pump subsidies
for low- and middle-income (LMI) households; the CPC subsidy level is more modest.
Other factors, such as age of household head, floorspace, race, and level of insulation,
have an influence on the extent of uptake to a lesser degree, which results in dierent
behavior across households. Since these are statistical results, the causal mechanisms
are not determined.
H.2.1. Electricity Demand
Table H4 shows that the higher penetration of heat pumps shifts heating and cooling
energy demand from gas and oil to electricity and enables more use of air conditioning,
increasing electricity consumption in residential buildings. This includes shell eiciency
upgrades, which do not vary across cases.
Table H4. New York Residential Electricity Demand (TWh)
BAU 2030 CPC 2030 SPC 2030
Heating and cooling
electricity
16,080 27,273 44,448
Prioritizing Justice in New York State Climate Policy 73
H.3. Transportation Sector
H.3.1. Light-Duty Vehicles
Passenger vehicle emissions depend on the emissions rates of on-road vehicles and
vehicle miles traveled. Variation in the adoption rate of plug-in electric vehicles,
6
the
fuel economy of gasoline vehicles, and VMT explain the dierences in 2030 emissions
across the two policy cases and the BAU. Table H5 shows how PEV adoption, fuel
economy, and VMT vary across the cases.
The first row shows the number of on-road EVs, which include plug-in hybrid vehicles
such as the Chevrolet Volt as well as all-electric vehicles like the Tesla Model 3. In the
BAU, New York has about 241,000 EVs—about 2 percent of all on-road vehicles. Note
that the share of EVs in new-vehicle sales is substantially higher (about 8 percent)
in the BAU in 2030, but the on-road share is less than the new share because new
vehicles replace older vehicles gradually over time.
The policy cases (CPC and SPC) yield roughly four times more EV sales than the BAU.
The main cause of this dierence is that the policy cases include more ambitious
zero-emissions vehicle (ZEV) standards than the BAU. The ZEV standards incentivize
6 In our use of the term “electric vehicle” we include plug-in hybrid vehicles such as the
Chevrolet Volt, as well as all-electric vehicles such as the Tesla Model 3.
Table H5. New York Light-Duty Vehicle Usage, By Case
BAU 2030 CPC 2030 SPC 2030
On-road EVs 240,648 861,920 984,507
Electricity consumption
from EV battery charging
(million MWh)
0.92 3.84 4.41
Average fuel economy
(miles per gallon)
34 38 40
Vehicle miles traveled
(billions)
134.46 130.99 126.42
Gasoline consumption
(billion gallons)
3.83 3.58 3.36
Resources for the Future and New York City Environmental Justice Alliance 74
vehicle manufacturers to sell EVs. Other policies included in the modeling, such as
EV purchase subsidies, eectively make it easier for manufacturers to attain the ZEV
standards. In principle, subsidies could be suiciently large to render ZEV standards
irrelevant if they cause manufacturers to exceed the ZEV standards. However, for the
cases we consider, the subsidies are smaller than that trigger, and the ZEV standards
essentially determine the level of EV sales and hence the on-road vehicle counts reported
in the table.
Table H5 also shows the amount of electricity consumed from charging EVs. These
amounts are roughly proportional to the vehicle counts in the first row, and this
consumption is an input to the electricity sector modeling discussed above.
Compared with the BAU, the policy cases increase the average fuel economy of on-road
vehicles by about 15 percent. The ZEV standards, again, are the main explanation for the
higher average fuel economy in the policy cases (fuel economy is computed assuming
zero fuel consumption for all-electric vehicles). In the modeling, vehicle manufacturers
achieve federal standards for corporate average fuel economy (CAFE) and GHGs. For a
particular manufacturer, the average fuel economy and GHG emissions rate in a state can
dier from the national average; that is, the manufacturer can undercomply in one state
and overcomply in another state as long as it achieves the national standard. Because of
this flexibility, when New York adopts the ZEV standard, the higher plug-in sales cause
manufacturers to overcomply with the standards in New York (and other ZEV states).
Consequently, adopting tighter ZEV standards in the policy cases causes average fuel
economy of new vehicles to be higher than in the BAU, which in turn causes average fuel
economy in non-ZEV states to be lower.
7
The carbon prices in the policy cases raise the cost of purchasing both liquid fuels and
electricity, which creates a disincentive to consume those products (and to drive) and
therefore reduces VMT. The eect is somewhat larger in the SPC than the CPC because
of the higher carbon price in the former.
The bottom row of the table shows the total fuel consumption, which is the product of the
inverse of the average fuel economy (which yields the average on-road fuel consumption
rate in gallons per mile) and VMT. Fuel consumption is about 6 percent lower in the
CPC and 12 percent lower in the SPC, compared with the BAU. The SPC reduces fuel
consumption more than the CPC because of its bigger eect on VMT.
H.3.2. Medium- and Heavy-Duty Vehicles
The factors that aect energy use and emissions for medium- and heavy-duty vehicles
include VMT, vehicle type (e.g., semitrailer, urban delivery truck), vehicle eiciency and
powertrain (e.g., internal combustion diesel engine, electric), duty cycles, and network
conditions. As described in more detail in Appendix B, the MHDV modeling framework
7 The copollutant emissions taxes also contribute to the higher average fuel economy in the
SPC. The taxes cause households to retire older vehicles sooner than they would have
otherwise, which increases average fuel economy because those retired vehicles tend to be
relatively old and have low fuel economy.
Prioritizing Justice in New York State Climate Policy 75
developed for this project comprises a number of models, which in summary combine
truck flow VMT data with estimates of how vehicle and fuel costs influence vehicle
purchase decisions. Vehicle purchase decisions then influence the mix of vehicles in the
fleet in any given year, which when combined with VMT yields energy use and emissions
estimates.
8
For example, electricity costs and carbon pricing aect vehicle operation
costs, and purchase incentives aect vehicle purchase prices. Additionally, MHDV
manufacturers face a new-sales ZEV mandate (modeled after California’s Advanced Clean
Truck rule), similar to the ZEV mandate for LDVs.
Similar to findings from the LDV modeling, the MHDV ZEV mandate is a primary driver of
the shift to a cleaner MHDV fleet. Because of the MHDV ZEV rule (see Table 1 for details),
by 2030, it is expected that about 14 percent and 13 percent of the fleet will be ZEV
(mostly battery electric) in the SPC and CPC, respectively (See H6 for a projection of ZEV
sales through 2045). Key drivers of the dierence between cases (although small) are the
SPC’s more ambitious ZEV sales mandate and carbon price and its copollutant fees; the
low-carbon fuel standard program is only in the CPC. Although the SPC and CPC assume
dierent sales targets, with the SPC requiring a larger share of Class 7–8 EVs during the
transition period, the consideration of the LCFS will provide additional credit incentive
that can help fleets reduce the cost gap between ZEVs and the incumbent vehicle
technologies. Alone, the LCFS is not expected to be the main driver of ZEV adoption.
Electricity consumption from direct battery charging for these vehicles is estimated
to be 2.13 and 1.93 million MWh in 2030 for the SPC and CPC, respectively. Although
these numbers are only about half of the electricity demand estimated for LDVs in 2030,
without the two policy packages, penetration of ZEV MHDVs (and associated electricity
demand) by 2030 is negligible in the BAU case.
8 We do not vary VMT across cases under the MHDV model for two primary reasons: (1) we
use truck activity flows estimated from the FAF model, adjusted with estimates from the
New York Metropolitan Transportation Council’s NYBPM, and thus our model does not
conduct vehicle routing or assignment; and (2) the strategies considered in the scenarios
do not directly aect freight demand, and although there could be some changes in
VMT due to charging and fueling detouring, these could potentially be minimal if the
infrastructure is located at or near freight facilities.
Figure H1. ZEV Sales Targets Over Time
Resources for the Future and New York City Environmental Justice Alliance 76
The models indicate that electric vehicles will represent the largest share of ZEVs for
short-haul heavy-duty and medium-duty trucks, while fuel cell vehicles will be the
primary technology adopted in long-haul heavy-duty trucks. This is mainly because of
estimated range limitations from battery sizes and weights. It remains unclear whether
the market will be able to ramp up between now and 2030 to supply the number of
vehicles required by the ZEV mandate as identified in our modeling (note that truck
production capacity was not included in the modeling).
Finally, significant financial incentives will be required to achieve the desired ZEV
adoption. The models estimate that for the most part, the various fees considered
(carbon, copollutant, fee on internal combustion engine vehicles), and the existing BAU
vehicle incentives programs (e.g., New York City Clean Truck, New York Truck Voucher
Incentive Program) will not be enough. It is important to mention that the consideration
of the LCFS credits and an increased fee (e.g., 10 percent) on internal combustion
engines significantly reduce the level of incentives required in the CPC, though the
sales target for Classes 2b–3 and 7–8 are also lower than in the SPC. Additional
incentives will be needed for public and private charging and fueling infrastructure.
Prioritizing Justice in New York State Climate Policy 77
Appendix I. Greenhouse Gas, PM
2.5
,
and Precursor Emissions Results
Emissions of multiple types are expected to decline as a result of the CLCPA. That
said, the dierent policy cases lead to significantly dierent emissions outcomes. For
example, in 2030, CPC carbon emissions reductions are about 30 percent below BAU,
while SPC carbon reductions are estimated to be about 54 percent below the BAU. The
2030 percentage reductions below the BAU for methane are even more dramatic in the
SPC (91 percent reduction) compared with the CPC (31 percent reduction).
PM
2.5
precursors are also significantly aected by the dierent policy cases. The CPC
creates estimated reductions below the BAU of 25 percent for SO
2
, 18 percent for NO
X
,
and 42 percent for direct PM
2.5
; the corresponding reductions under the SPC are 52
percent, 32 percent, and 75 percent. Table I1 lists statewide 2030 emissions for the
three sectors we model under each case.
Table I1. New York Emissions Estimates, 2030, by Case and Sector
BAU 2030 CPC 2030
CPC percentage
reduction from BAU
SPC 2030
SPC percentage
reduction from BAU
Electric power
GHGs
CO
2
MMTCO
2
e 15.70 5.10 –68% 3.20 –80%
Methane MMTCO
2
e* 10.08 3.36 –67% 1.68 –83%
PM
2.5
and precursors
SO
2
MT 1,190.00 858.00 –28% 525.00 –56%
NO
X
MT 6,930.00 5,094.00 –26% 3,573.00 –48%
PM
2.5
(direct) MT 1,423.00 554.00 –61% 280.00 –80%
Residential buildings
GHGs
CO
2
MMTCO
2
e 36.60 22.20 –39% 3.10 –92%
Methane MMTCO
2
e 23.52 16.80 –29% 0.00 –100%
Resources for the Future and New York City Environmental Justice Alliance 78
BAU 2030 CPC 2030
CPC percentage
reduction from BAU
SPC 2030
SPC percentage
reduction from BAU
PM
2.5
and precursors
SO
2
MT 181.00 109.00 –40% 16.00 –91%
NO
X
MT 2,883.00 1,598.00 –45% 286.00 –90%
PM
2.5
(direct) MT 2,140.00 1,298.00 –39% 185.00 –91%
On-road transportation
GHGs
CO
2
MMTCO
2
e
(LDV)
30.80 28.80 –6% 27.00 –12%
MMTCO
2
e
(MHDV)
12.90 10.90 –16% 10.60 –18%
Methane
MMTCO
2
e
(LDV)
0.00** 0.00** — 0.00** —
MMTCO
2
e
(MHDV)
1.85 1.51 –18% 1.43 –23%
PM
2.5
and precursors
SO
2
MT (LDV) 236.70 227.30 –4% 218.10 –8%
MT (MHDV) 42.00 35.80 –15% 34.60 –18%
NO
X
MT (LDV) 757.80 732.90 –3% 711.00 –6%
MT (MHDV) 18,000.00 16,000.00 –11% 15,000.00 –17%
PM
2.5
(direct)
MT (LDV) 111.10 108.20 –3% 104.60 –6%
MT (MHDV) 542.70 496.90 –8% 487.60 –10%
Total***
GHGs
CO
2
MMTCO
2
e 96.00 67.00 –30% 43.90 –54%
Methane MMTCO
2
e 35.45 21.67 –39% 3.11 –91%
Prioritizing Justice in New York State Climate Policy 79
I.1. Power Sector Detail
Reduced natural gas generation in both policy cases relative to the BAU leads to
significant electricity sector emissions reductions in 2030. CPC policies reduce
New York power plant NO
X
and SO
2
emissions by smaller proportions than the other
emission types because waste–fueled generation accounts for large portions of New
York power plants’ NO
X
and SO
2
emissions, and the CPC policies do not appreciably
change waste–fueled generation. Even in the baseline scenario results, waste-fueled
generation accounts for more than half of New York power plants’ NO
X
and SO
2
emissions despite producing less than 10 percent generation as natural gas. The
reduction of waste-fueled generation in the SPC is a significant contributor to the
emissions reductions in that scenario.
Of all the emissions types in Table I1, PM
2.5
tends to have the most localized eects.
Most harm from NO
X
and SO
2
is from their formation of ozone (NO
X
only) and
particulate matter (NO
X
and SO
2
), but these “secondary” pollutants take time to form,
so to a large extent they form miles (or hundreds of miles) downwind. The impact on
their formation of PM
2.5
in New York State is covered by our air quality model.
I.2. Residential Buildings Sector Detail
Similarly, in the residential building sector, both GHGs and local air pollutants would
decline significantly under all future scenarios because of reductions in fossil fuel
(natural gas and diesel) use for heating. SPC reductions in both GHGs and local air
BAU 2030 CPC 2030
CPC percentage
reduction from BAU
SPC 2030
SPC percentage
reduction from BAU
PM
2.5
and precursors
SO
2
MT 1,649.70 1,230.10 –25% 793.70 –52%
NO
X
MT 28,570.80 23,424.90 –18% 19,570.00 –32%
PM
2.5
(direct) MT 4,216.80 2,457.10 –42% 1,057.20 –75%
* In this document, we assign each ton of methane a CO
2
equivalent of 85 (except where otherwise noted). This is approximately
seven times as large as the methane CO
2
equivalent of 12.9 implied by the US Environmental Protection Agency’s new draft
guidance (2022) on the social cost of greenhouse gases. That new draft guidance is based on extensive research (e.g., Carleton
and Greenstone 2022; Rennert et al. 2022) guided by an expert panel convened by the National Academies of Sciences,
Engineering, and Medicine (2017). This dierence might be partially oset by the fact that recent studies focusing on small parts
of the country suggest that natural gas methane leakage rates might be considerably larger than the national estimates that we
and other modelers use (Chen et al. 2022; Lenox 2013). If so, those high leakage rates might persist through 2030.
** These upstream values appear as zero because of rounding; we include upstream methane from light-duty vehicles in our
analysis.
*** The totals are the sum of emissions across the three sectors we model, not New York economywide totals.
Resources for the Future and New York City Environmental Justice Alliance 80
pollutants are more than double those of the CPC (with 90–100 percent reductions
from the BAU across the various emissions types).
These emissions changes are the result of reductions in the use of natural gas and
heating oil (diesel), which are the result of dierent rates of uptake of electric heat
pumps (see above). In the CPC, about half of New York households continue to
use natural gas and diesel for heating, while in the SPC, more than 90 percent of
households use heat pumps.
GHG emissions are driven by the extent of natural gas consumption. Methane
emissions include leaks both in the distribution system and in homes (from gas stoves)
as well as the upstream leakages associated with delivered natural gas.
I.3. Transportation Sector Detail
I.3.1. Light-Duty Vehicle Fleet
For LDVs, the CPC reduces CO
2
emissions by 6 percent and SPC by 12 percent
below the BAU—the same as the fuel consumption reductions reported above.
Methane (from incomplete combustion and upstream fugitive emissions associated
with gasoline production and distribution) accounts for a trivial share of LDV GHG
emissions. Consequently, GHG emissions are proportional to the carbon content of fuel
and the fuel consumption that was reported above. Since the carbon content of fuel
does not vary across the BAU and policy cases,
9
the GHG reductions are proportional
to the fuel consumption reductions.
Compared with the BAU, the CPC and SPC reduce direct PM
2.5
, NO
x
, and SO
2
emissions
by small amounts (3 to 8 percent across the two cases). The SPC does reduce
emissions by about double the CPC, although again it is a small amount (from 4
percent in the CPC to 8 percent in the SPC for SO
2
, for example; see Table 7 for more
detail). The main policy driving this dierence is the ZEV standards, since EVs do not
emit these pollutants directly when running on electricity (the power sector modeling
accounts for emissions caused by battery charging). The SPC achieves greater
emissions reductions than the CPC because of the additional EVs, and to a lesser
extent because of the copollutant taxes.
I.3.2. Medium- and Heavy-Duty Vehicle Fleet
For MHDVs, the CPC reduces CO
2
emissions by 16 percent and the SPC by 18 percent
below the BAU (there are similar reductions in methane; see Table I1), primarily
resulting from the penetration of ZEVs, with the SPC having a slightly larger share of
ZEVs by 2030. Compared with the BAU, the CPC achieves a further reduction of 15
percent for SO
2
, 11 percent for NO
x
, and 8 percent for direct PM
2.5
; the SPC reduces
these emissions by 18, 17, and 10 percent, respectively.
9 The CPC includes a LCFS. However, by 2030 there is no change to the carbon content of
gasoline, the primary fuel consumed by LDVs (the LCFS does change the carbon content
of diesel fuel).
Prioritizing Justice in New York State Climate Policy 81
Appendix J. Location of Emissions
Changes
Emissions changes are generally concentrated in populous areas, particularly for
vehicles and residential sectors. Beyond population density, certain policies may aect
certain geographies. For example, policies that are thoroughly means-tested may lead
to greater emissions reductions in low-income areas than what is observed in the BAU.
This section covers details for each sector about where emissions changes take place,
to the greatest level of spatial detail possible. For simplicity of presentation, we restrict
our discussions to direct emissions of PM
2.5
, even though the models predict changes in
NO
X
, SO
2
, and VOCs (and other pollutants). We focus on direct PM
2.5
emissions because
they have the greatest impact on local air quality. The extent to which other pollutants
combine to form secondary PM
2.5
is covered in the Air Quality Results section.
J.1. Electricity Sector
This study focuses on emissions changes in New York State. However, some models,
including the electricity sector modeling, also calculate the eects of the policy
dierences between the CPC and SPC on power plant emissions outside New York.
“Emissions leakage” is a consequence of some emissions reduction policies, including
policies applied to the electricity sector. Emissions leakage occurs when emissions
increase outside the state or country where the policy is adopted as a result of the
policy. For example, an emissions price in one state can cause an increase in other
states as production moves to where it is not subject to an emissions price.
However, the policies in SPC cause the opposite of emissions leakage: they cause power
plant emissions in other states to decrease. They reduce non–New York electricity
sector CO
2
emissions by 600,000 short tons, or by 28 percent as much as they reduce
New York electricity sector CO
2
emissions. They reduce non–New York electricity sector
SO
2
emissions by 3.6 million pounds, or nearly five times as much as they reduce New
York electricity sector SO
2
emissions. And they reduce non–New York electricity sector
NO
x
emissions by 2.6 million pounds, or by 80 percent as much as they reduce NY
electricity sector NO
x
emissions.
This study is not the first to find that New York’s electricity sector emissions reduction
policies would reduce emissions outside the state as well (Shawhan et al. 2019),
although this eect is a function of the type of policy. Again, relative to the CPC, the
SPC has considerably higher emissions prices (accompanied by a concomitantly
higher electricity border carbon adjustment), fewer New York nuclear generators,
fewer fossil fuel peaker generators, and no new fossil fuel generators. These policies
reduce emissions outside the state for two reasons. First, they increase New York’s
reliance on solar and wind generation, which in turn causes New York’s generation and
wholesale electricity prices to vary more from hour to hour across the year, even outside
New York. This higher electricity price variability outside New York favors natural
gas–fueled generators over coal–fueled generators. Second, the eects just described
Resources for the Future and New York City Environmental Justice Alliance 82
increase dispatchable, fossil fuel generation and generation capacity near New York,
in New Jersey. That in turn reduces the need for dispatchable, fossil fuel capacity and
generation in the next state to the west, Pennsylvania (see Figure 3). New Jersey does
not have coal-fueled generators, but Pennsylvania does, so the shift from Pennsylvania
to New Jersey reduces total emissions.
Figure J1 shows the location of estimated electricity PM
2.5
emissions changes in New
York and surrounding states, under various policy scenarios. For the electricity sector,
we include estimates for adjacent states for two reasons: (1) New York policy has a
greater eect on out-of-state emissions in the electricity sector than in other sectors;
and (2) electricity emissions get dispersed over a broader geographic area than are
emissions from the other sectors we model because of the tall smokestacks at power
plants. Given this, New York electricity emissions changes would aect air quality both
in New York and in other states and Canadian provinces, and vice versa (depending on
prevailing wind directions).
Figure J1. Direct PM
2.5
Emissions Dierences, by Power Generator, 2030
Figure J1A. Change in Direct PM
2.5
Emissions, BAU vs. SPC
Prioritizing Justice in New York State Climate Policy 83
Figure J1B. Change in Direct PM
2.5
Emissions, BAU vs. CPC
Figure J1C. Change in Direct PM
2.5
Emissions, CPC vs. SPC
Resources for the Future and New York City Environmental Justice Alliance 84
Figure J1 shows the power generating unit emissions changes in New York, New
Jersey, Pennsylvania, Ohio, Connecticut, Massachusetts, and Vermont as a result of the
dierences between the policy cases. A green dot means emissions are lower in the
policy case than in the BAU; a red dot means emissions are higher. The larger the dot, the
greater the changes in emissions.
Just comparing the number of green and red dots, we can see that although most
power-generating units in New York State decrease emissions (green dots) in both policy
cases relative to BAU, the CPC results in a larger number of power-generating units that
increase emissions (red dots)—only seven of the 348 generating units increase emissions
in the SPC, compared with 64 in the CPC.
10
Emissions increases at some generating units allow greater emissions reductions at
other units. One of the eects of emissions prices and caps is to shift generation from
generating units with higher per kWh emission rates to generating units with lower per
kWh emission rates. As a result, the generating units that are used more as a result of
emissions prices (or higher emissions prices) tend to have low per kWh emissions rates.
Examining the size of the dots, we find that in both policy cases, the largest decreases in
emissions are at power-generating units close to or in New York City. The concentration
of emissions reductions in southeastern New York is beneficial for public health, since that
is the most densely populated part of the state. New York City is also upwind of densely
populated Connecticut and Rhode Island as well as the Boston metropolitan area. That
further increases the benefits of emissions reductions in the New York City area.
10 Since there can be more than one generating unit at a given site, there may be emissions
increases at fewer than seven and 64 sites. This is partially because some of the units with
emissions increases may be adjacent to each other, and partially because a unit with an
emissions increase may be next to one or more units with a larger emissions decrease.
Key Findings for Figure J1
• Although most of the power-generating units in New York decrease their
emissions in both policy cases relative to the BAU (the green dots), the CPC
results in a larger number of power-generating units that increase emissions
(the red dots).
• In both policy cases, the largest decreases are at generating units close to or in
New York City.
• In both policy cases, emissions increase and decrease at many generating units
outside the state. Some of the largest increases (e.g., in New Jersey) occur at
generating units that are close to and upwind of New York City. The largest out-
of-state decreases occur at Pennsylvania coal generating units.
• Compared with the CPC, the SPC produces lower emissions at nearly every New
York generating unit and also leads to greater out-of-state reductions (Figure
J1C).
Prioritizing Justice in New York State Climate Policy 85
Looking at the dots outside New York State, we see that in both policy cases, emissions
both increase and decrease at many power-generating units outside the state. The
largest increases are at generating units close to and upwind of New York City, and the
largest decreases occur at Pennsylvania coal generating units that are also upwind of
the city but farther away. Note that, as explained above, the only dierences between
the cases are New York policies and the EV sales mandates that several states plan to
adopt together. As a result, any changes in emissions in other states are the result of
New York policy and the collective action on ZEV mandates.
Table J1 shows the total PM
2.5
emissions increases and decreases in DACs and within
a 10km buer of DACs. For both cases, less than 6 percent of the PM
2.5
emissions
increases within the 10km buer zones originate from New York sources.
Finally ,in Figure J1C, the dots represent dierences between SPC and CPC emissions.
We see that for nearly every New York generating unit, emissions are lower for the
SPC than the CPC. Only 14 of 348 generating units in the state have higher predicted
PM
2.5
emissions in the SPC than the CPC. This striking result occurs because of the
strict regulation on new fossil fuel generation in the SPC, as well as the higher price on
carbon and copollutants.
Table J1. PM
2.5
Emissions Eects and New York DACs, SPC vs. CPC, 2030
SPC CPC
Emissions decreases in short tons (number of electricity-generating units)
Direct PM
2.5
emissions decreases in DACs –312.98 (126) –166.63 (105)
Direct PM
2.5
emissions decreases within 10 km of DACs (NYS generators only) –1154.67 (283) –848.77 (231)
Direct PM
2.5
emissions decreases within 10 km of DACs (all states’ generators) –1156.05 (322) –853.23 (270)
Emissions increases in short tons (number of electricity-generating units)
Direct PM
2.5
emissions increases in DACs 11.71 (3) 16.34 (24)
Direct PM
2.5
emissions increases within 10 km of DACs (NYS generators only) 27.34 (5) 16.34 (57)
Direct PM
2.5
emissions increases within 10 km of DACs (all states’ generators) 472.56 (28) 284.74 (80)
Resources for the Future and New York City Environmental Justice Alliance 86
J.2. Residential Buildings Sector
Direct PM
2.5
emissions from residential heating are spatially modeled by public use
microdata area.
11
In New York, PUMAs tend to be similar in size to counties, but in some
areas—particularly around the densely populated metropolitan areas—PUMAs are
smaller (more granular) than counties. Given the spatial scale, there is limited ability
to derive direct PM
2.5
emissions estimates for the residential sector at the census tract
level.
That said, we are able to observe broader trends that provide useful insights about
local pollution exposure. The location of residential emissions has greater nexus with
the health of colocated communities (compared with electricity generation) because
these emissions tend to occur closer to the level where people are living (although
some residential high-rise buildings may be a slight exception).
Our modeling shows that direct PM
2.5
emissions from residential heating decline in all
New York PUMAs, under both the CPC and the SPC compared with the BAU. In the
SPC, emissions reductions from the BAU are roughly uniform across the state. In the
CPC, the eastern counties have relatively higher percentage reductions from the BAU
compared with other counties (not shown).
11 PUMAs are US Census Bureau–defined geographic delineation of population containing
at least 100,000 people. For more information, see https://www.census.gov/programs-
surveys/geography/guidance/geo-areas/pumas.html.
Figure J2. Residential Home Heating Direct PM
2.5
Emissions from CPC to SPC (2030)
Prioritizing Justice in New York State Climate Policy 87
Figure J2 compares SPC with CPC residential PM
2.5
results, by PUMA. Darker colors
indicate greater emissions reductions. Note the modest change in direct PM
2.5
emissions in New York City relative to other parts of the state. This result is caused by
the relatively high penetration of heat pumps in that region in the CPC, implying that
there are few additional opportunities for heat pump penetration even with the larger
subsidies in the SPC. In contrast, the generous subsidies for heat pumps in the SPC
raise heat pump adoption rates upstate, where adoption in the CPC was relatively low.
J.3. Transportation Sector
As noted above, we use two models to estimate emissions from light-duty vehicles
and medium- and heavy-duty vehicles. Results are reported for LDVs and MHDVs.
As with residential emissions, the location of transportation emissions (both LDV and
MHDV) has significant nexus with the health of colocated communities (compared
with electricity generation) because these emissions largely occur at the ground level,
where people are living and breathing.
J.3.1. Light-Duty Vehicle Fleet
The light-duty vehicle modeling is performed at the county level, which is slightly less
granular than PUMAs. There is therefore limited ability to assess unique impacts in
each census tract (the level at which DACs are designated). That said, we are able to
assess broader trends. Unlike the electricity sector, where emissions increase in some
locations as a result of SPC and CPC policies, all counties in New York experience
reductions in LDV PM
2.5
relative to the BAU as a result of the SPC and CPC policies.
We can also look at dierences in emissions across the two policy scenarios. In Figure
J3, the darker the color, the larger the emissions reductions for the SPC relative to
the CPC. Not surprisingly, the darkest areas are in cities, where vehicles and their
emissions are concentrated. And again, not surprisingly, the dierences in SPC versus
CPC emissions reductions are greatest in New York City.
Key Findings for Figure J2
• PM
2.5
emissions from residential home heating are consistently lower in the
SPC than in the CPC.
• Because the geographic unit for modeling purposes is the PUMA, high-
density areas do not necessarily see greater absolute reductions in emissions
than low-density areas.
• The largest emissions dierences between the CPC and SPC are upstate,
where heat pump penetration is low until large subsidies are provided in the
SPC.
Resources for the Future and New York City Environmental Justice Alliance 88
J.3.2. Medium- and Heavy-Duty Vehicle Fleet
For the MHDV fleet, emissions are estimated for each major road segment (“network
link”) along the primary and secondary highway system in New York State. The length
or geographic scope of individual network links varies. In this analysis of the location
of emissions changes, PM
2.5
is displayed by census tract (by grouping network links
within a given census tract). Most emissions concentrate in denser regions, especially
the New York City metropolitan area.
Figure J4 compares direct PM
2.5
emissions in the CPC and SPC. The dierence in
emissions ranges between 0.4 and 4 percent reduction (with dark blue indicating a
larger reduction). The SPC (mostly because of its higher ZEV sales requirement and
faster sales ramp-up) generates a larger reduction across the state. Except for New
York City, the map shows that these increases happen on intercity infrastructure. One
of the main reasons for these results is the fact that the share of long-haul heavy-duty
Figure J3. Light-Duty Vehicle Direct PM
2.5
Emissions, 2030
Key Findings for Figure J3
• PM
2.5
emissions from light-duty vehicles are consistently lower in the SPC,
relative to the CPC.
• The areas of greatest improvement in the SPC have the highest population
density, particularly the New York City area.
• The extent to which DACs experience increased benefits of PM
2.5
reductions
is directly related to the concentration of DACs in urban areas.
Prioritizing Justice in New York State Climate Policy 89
combination vehicles is larger on these network links; this is the segment that will
experience a significantly lower penetration of ZEVs by 2030, and because of activity
growth, it is thus not able to mitigate the emissions increase.
Table J2 goes into more detail about emissions dierences between the two
policy cases by specific census tract group. We go beyond the DAC and non-DAC
designations to identify communities that experience the greatest improvements from
implementing the SPC over the CPC. We find that average PM
2.5
emissions dierences
between the two policy cases are consistently in favor of the SPC but are relatively
modest, averaging around 1.5 percent. In terms of absolute PM
2.5
emissions dierences,
we find that non-DAC tracts experience greater improvement under the SPC. This is
likely due to the fact that DACs as defined here do not take up the majority of space
along major highways, where pollution reductions from medium- and heavy-duty
vehicles are concentrated.
Figure J4. Medium- and Heavy-Duty Vehicle Direct PM
2.5
Emissions, 2030
Key Findings for Figure J4
• PM
2.5
emissions from medium- and heavy-duty vehicles are consistently
lower by 1 to 4 percent in the SPC, relative to the CPC.
• The areas of greatest improvement in the SPC have the highest congestion,
including major highways and New York City.
• Tracts with the greatest dierence in emissions appear to be the non-DAC
tracts in the New York City area.
Resources for the Future and New York City Environmental Justice Alliance 90
Table J2. Direct PM
2.5
EMissions Dierence by Tract Type, CPC vs. SPC, 2030
Community type
Average PM
2.5
emissions
Dierence from CPC to SPC
(short tons)
Average PM
2.5
emissions
dierence from CPC to SPC
(percentage)
All tracts –.003 –1.53%
Non–DAC tracts (65% of tracts) –.004 –1.56%
DAC tracts (35% of tracts) –.001 –1.48%
High exposure (top 10%) –.003 –1.55%
High vulnerability (top 10%) –.001 –1.47%
High elderly population (top 10%) –.003 –1.48%
High historical PM
2.5
(top 10%) –.001 –1.54%
Prioritizing Justice in New York State Climate Policy 91
Appendix K. Additional Results
Context: Nonlinearities and Excluded
Emissions
K.1. Nonlinearities
The relationship between NO
x
and SO
2
and the creation of PM
2.5
is nonlinear,
sometimes highly so. For our case, this relationship means that under the “wrong”
conditions, reductions in NO
x
emissions can lead to very small or even zero reductions
in PM
2.5
concentrations and, in even more specialized conditions, increase PM
2.5
emissions. Again, we cannot precisely model this relationship for this project, but our
investigation of the composition of PM
2.5
concentrations do reveal some nonlinearities.
Although reducing SO
2
and NO
x
emissions indeed reduces SO
4
2–
, NO
3
–
, and associated
NH
4
+
, we see some increases in total secondary organic aerosol (SOA), showing the
nonlinearity in inorganic aerosol and SOA formation. Lower SO
2
and NO
x
would free
more OH available for the oxidation of volatile organic compounds (VOCs), leading to
more SOA formation.
As expected, SO
2
and NO
x
decrease, but NH
3
increases because of less NH
4
+
formation
(because of less SO
4
2–
and NO
3
–
to neutralize NH
4
+
). Although both NO
x
and VOCs
decrease, the decrease in the former is greater, leading to increases in O
3
(which
shows disbenefit of NO
x
reduction), because O
3
is VOC-limited in New York State and
O
3
titration is weaker because of lower NO
x
emissions. In the VOC-limited regime,
reducing NO
x
would increase O
3
rather than decreasing it, leading to increases in PM
2.5
when the reduction in NO
3
- is compensated by increases in other PM
2.5
composition,
such as SOA.
This confirms that the nonlinearity pollution formation eect partly explains the
smaller-than-expected reduction in PM
2.5
when the SPC is compared with the BAU.
Eective reduction in PM
2.5
requires understanding of PM
2.5
-precursor relations and
PM
2.5
formation chemical regime, since PM
2.5
formation may be limited by any of the
precursors (e.g., NO
x
or VOCs or NH
3
or SO
2
).
K.2. Modeling Choices
We model emissions changes in only a few of the sectors contributing to PM
2.5
concentrations. Beyond power plants, LDV, MHDV, ports, and home heating, the direct
PM
2.5
and precursor emissions also come from industry, aviation, shipping (other than
ports), commercial heating, waste management, and more. Agricultural emissions are
another important sector, particularly for ammonia, which combines with SO
2
and with
NO
x
to form sulfate and nitrate aerosols, which count as PM
2.5
concentrations.
Resources for the Future and New York City Environmental Justice Alliance 92
We model only 2030 emissions reductions. Many of the policies in the SPC and CPC
take time to be fully implemented or to fully realize their emissions-reducing potential.
An example is transport policies that aect only new-car purchases. Yearly turnover
of the vehicle stock is only a fraction of that stock, so until the new-vehicle policies
have been in eect for at least seven to 10 years, they will not make a huge dent in the
transportation emissions.
We find that air emissions are already significantly reduced in the BAU case by 2030
compared with emissions in 2012, our baseline. This is partly for economic reasons
and partly because of policy. The most prominent economic reason has to do with
natural gas prices. During this almost 20-year period, coal plants continue to retire
and are replaced by cheaper and cleaner natural gas generation, made possible by
fracking, and by renewable generation, made possible by price declines and technology
breakthroughs (themselves aided by tax credit policies).
Policies reducing PM
2.5
direct emissions and PM
2.5
concentration precursors have
been a mainstay of federal air pollution policy since the Clean Air Act was passed in
1970. Since 2012 (our base year) there have been numerous policies to further reduce
these types of emissions. In the auto and truck sectors, very tight emissions standards
and the continual retirement of older, more polluting vehicles lead to significantly
lower tailpipe emissions by 2030, even in the absence of additional policy. In the
residential sector, the transition to natural gas furnaces in lieu of traditional oil furnaces
significantly reduces PM
2.5
and SO
2
by 2030, even without additional policy. State-
level policy, such as New York’s decision to shut down peaker plants in high-density
areas, also contributes. In our 2012 emissions inventory, the power sector emitted an
estimated 39.3 ktons of SO
2
, compared with an estimated 1.8 ktons in 2030 with no
additional policy.
Table K1. Economy-Wide Emissions Changes in New York
Baseline (2012) to BAU 2030 BAU 2030 to SPC 2030
Pollutant Kt of pollutant Percentage change Kt of pollutant Percentage change
CO –722 –27% –24 –1%
SO
2
–50 –35% –5 –5%
NO
X
–123 –34% –16 –7%
VOCs –90 –17% –1 ~0%
PM
2.5
–10 –16% –1 –2%
PM
10
0 0% –2 0%
Prioritizing Justice in New York State Climate Policy 93
Because certain emissions changes are left out of our analysis, the actual percentage
reduction in emissions becomes more misleading and diicult to interpret. Say that our
modeling of emissions changes is capturing only half the relevant emissions inventory.
As the other sectors are held equal, the share of emissions changes we cover by 2030
will be smaller than it was in 2012 (Table K1). For this reason we focus on absolute
changes in economywide emissions and air pollution concentrations.
Our own air quality modeling is not suited to identifying the largest contributing
sectors to PM
2.5
outside our modeled sectors. The Environmental Protection Agency’s
2011 emissions inventory reveals that our modeled emissions changes cover
approximately 50 percent of SO
2
emissions, 37 percent of PM
2.5
direct emissions, and
54 percent of NO
x
emissions in 2011. Major contributors to emissions that we do not
model include ambient dust, o-road vehicles, waste disposal, and industrial fuel
combustion.
New York State’s air quality analysis of the CLCPA impacts provides some helpful
insights on future emissions (Energy and Environmental Economics 2022). It estimates
that approximately 75 percent of the projected reference case PM
2.5
emissions are from
“noncombustion sources,” including dust or biogenic sources. Nearly all of the PM
2.5
emissions associated with combustion sources come from residential or industrial
wood combustion, which we do not model. Figure K1 estimates the PM
2.5
emissions
sources in 2025. Furthermore, the New York State analysis finds that many of the
benefits associated with reduced power sector emissions are realized in 2040, which is
beyond our modeling timeline.
Industrial (fossil fuel)
Industrial (wood)
Commercial/ Residential
(fossil fuel)
Commercial/ Residential
(wood)
On-road
Non-road
Electricity Generation
Combustion
Non-combustion
Figure K1. New York Integration Analysis Sector-Level PM
2.5
Reference Case Emissions,
2025
Note: This data is available in Appendix G of the New York State integration analysis (E3 2022).
Resources for the Future and New York City Environmental Justice Alliance 94
K.3. Traveling Air Pollution
Many factors contributing to PM
2.5
concentrations are beyond the ability of New
York State policies to control. The largest factor might be termed background
concentrations—the concentrations that do not have a cause in economic activity in
the state. They are caused, for example, by fine dust picked up by the wind and by
emissions from economic activity and natural causes in other states and Canada. Even
emissions from China have shown up in the United States, so these emissions can
travel long distances.
We can examine some of these traveling emissions through our power sector modeling.
Whether chemically transformed or not, emissions move with the wind direction and
speed. This means that emissions reductions upwind from a border with another state
could improve air pollution in those states, but not necessarily in the emitting state,
particularly when the emissions come from tall stacks, as in the power sector.
Figure K2 shows the wind “rose” (wind direction, frequency, and speed) for JFK airport.
The winds from the south and northwest are the most frequent and strongest. This has
two implications. At least some emissions from power plants along the eastern border
are swept into Massachusetts and Connecticut, with emissions reductions benefiting
those states. Second, emissions from power plants west and, most importantly, south
of the border are swept into New York. This means emissions from Ohio, Pennsylvania,
New Jersey, Delaware, and Maryland could all contribute to New York pollution in the
summer. Although emissions reductions in those states benefit New York residents,
energy generation and associated pollution increases in these bordering states can
ameliorate PM
2.5
improvements in New York.
Figure K2. JFK Windrose
Prioritizing Justice in New York State Climate Policy 95
Given the resources available to this project we cannot precisely estimate how much
of the benefits of NYS emissions reductions are being realized in NYS on net, but we
can give a sense of the size of these emissions/concentrations cross-border flows. NYS
power plants on the eastern border (See Figure 4) account for about 50-56 percent
of total power sector emissions throughout the state (depending on the pollutant).
And NYS power plant emissions are only between 3 and 9 percent of the emissions
counting all the border states’ power emissions. This indicates that some emissions
improvements near the border of New York may be transported out of state, and some
increased emissions in PA, OH, and NJ may be transported into the state.
Resources for the Future and New York City Environmental Justice Alliance 96
Appendix L. Distribution of PM
2.5
Concentration Reductions by Scenario
and DACs vs. Non-DACs
In the main text, we presented detailed maps (Figure 10) showing how PM
2.5
concentrations vary across the state for DACs and non-DACs associated with the two
scenarios (CPC-BAU and SPC-BAU) and use the mean of the concentration reductions
to quantitatively describe impacts across the scenarios for DACs and non-DACs. While
the means are easy to interpret, they miss other features of the distribution of PM
2.5
concentrations changes across communities and scenarios. For instance, we could use the
median of the various distributions, or various percentiles of the distributions. Rather than
these limited measures, Figures L1 and L2 below simply show the entire distributions.
These distributions illustrate the frequency (or percentage) for which DAC (or non-
DAC) communities experience a given concentration reduction in PM
2.5
as a result of
the CPC scenario relative to BAU. In Figure L1, we depict two distributions, one for the
disadvantaged communities (red) and the other the non-disadvantaged (blue). The
frequency (the percentage of the DAC or the percentage of the non-DACs) is on the Y-axis
and the concentration reductions in PM
2.5
from BAU to CPC is on the X-axis. Note that
the X-axis legend shows a range of negative reductions from -0.05 μg/m
3
(at the origin).
Negative numbers represent increases in PM
2.5
concentrations as a result of a scenario. At
the other end of the X-axis, the largest PM
2.5
concentration reduction experienced by any
community for the BAU to CPC scenario is 0.10 μg/m
3
.
Figure L1. Distribution of PM
2.5
Concentration (μg/m³) Improvements
from BAU to CPC
Prioritizing Justice in New York State Climate Policy 97
Turning to the red line for the DAC distribution, we see that most communities experience
PM
2.5
reductions in the range of 0.02 to 0.05 μg/m
3
(the big hump in the red distribution).
The non-DAC communities (blue line) are also experiencing reductions primarily in this
range but there’s also a hump representing a high percentage of non-DAC communities
experiencing higher PM
2.5
reductions.
What we care most about is the dierence in these two distributions for any given
PM
2.5
concentrations. What we see is that the red line is lower than the blue line for PM
increases. This shows that a greater percentage of non-DAC communities experience
increases in pollution than the percentage of DAC communities. This is a good result
for environmental justice. Of course, most communities experience reductions in PM
2.5
concentrations. Where the state scenario does well for DACs is where the big red hump
is higher than the blue hump in the range of 0.02 to 0.05 μg/m
3
. The big advantage to
non-DAC communities is where the blue line is higher than the red line for the bigger PM
2.5
concentration reductions on the right-hand side of the figure.
Now let’s turn to what happens with the stakeholder scenario. What is dierent about
Figure L2 compared to L1? First, notice the scale on the X-axis. There are no communities
experiencing PM
2.5
concentration increases and some communities experience reductions
in PM
2.5
concentrations from BAU to SPC that are far larger than in the CPC scenario—a
bit more than 0.4 μg/m
3
. Thus, the SPC policies do more for both types of communities.
Second, the shape of the distributions are dierent. Both are double-humped, with the
humps about equal in size and shape for the DACs and the double humps for non-DACs
looking similar to those in the state scenario.
Figure L2. Distribution of PM
2.5
Concentration (μg/m³)
Improvements from BAU to SPC
Resources for the Future and New York City Environmental Justice Alliance 98
These dierences lead to the most important dierence, which is that the red hump
on the right side is so much higher than the blue hump (in the range of 0.2 to 0.3 μg/
m
3
. In this area a far higher percentage of DACs are experiencing these large PM
2.5
reductions compared to the percentage of non-DACs. Finally, as we saw for the state
case, the very largest reductions in PM are experienced more frequently in the non-
DAC communities, but the dierence in frequency compared to DACs is not very large.
Figure L3 looks at the distribution of PM
2.5
concentration changes comparing the
SPC to the CPC casers to make the above discussion graphically more explicit. Here
we see very minor dierences in the percentages of DAC and non-DAC communities
experiencing PM
2.5
concentration reductions for the two scenarios. The exception is
the relatively large PM
2.5
concentration reductions ranging from 0.20 to 0.27 μg/m
3
,
where a far higher percentage of DAC communities are represented compared to non-
DAC communities for the SPC case.
Figure L3. Distribution of PM
2.5
Concentration (μg/m³)
Improvements from CPC to SPC
