Request from States for Removal
of Gasoline Volatility Waiver
Technical Support Document
and Cost Analysis
rnA United States
Environmental Protection
Agency
-------
Request from States for Removal
of Gasoline Volatility Waiver
Technical Support Document
and Cost Analysis
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
United States
Environmental Protection
Agency
EPA-420-R-24-002
February 2024
-------
Table of Contents
1. Overview 1
2. Fuel Production and Distribution 2
A. Overview 2
B. Refinery Actions to Control Gasoline Volatility 4
C. Pipelines 9
D. Terminals 10
3. Evaluating the Supply of Gasoline 12
A. Assessing the Refining Industry's Ability to Supply Low-RVP Gasoline 12
B. Impact of the Action on Supply-Demand Balance 20
C. Accommodating a Reduced Gasoline Supply in 2025 31
D. Recent and Current Supply and Demand Balance of Gasoline Inventories 32
E. Oil Industry Estimated Impact on Supply 38
F. 2022 ICF Report Analyzing Supply 39
4. Cost of Removing the 1-psi Waiver in the Petitioning States 41
A. Cost Studies 41
B. Distribution Cost 44
C. Discussion of Costs 45
5. Potential Price Impacts When Implementing the Removal of the 1-psi Waiver 48
6. Benefits of Removing the 1-psi Waiver 53
7. Gasoline Supply Situation Changes for 2025 54
8. Screening Analysis for Potential Impacts on Small Entities 56
A. Background 56
B. Screening Analysis Approach and Results 57
C. Conclusions 58
D. Refiner CBI Data 59
l
-------
1. Overview
This document provides additional detail supporting EPA's final rule to remove the 1-psi
fuel volatility waiver for E10 in Illinois, Iowa, Minnesota, Missouri, Nebraska, Ohio, South
Dakota, and Wisconsin (hereinafter the "petitioning states"). Section 2 provides background on
how the fuel production and distribution system operates and discusses modifications to this
system that are likely to be or are necessary as a result of the action. Section 3 quantifies the
impacts on refinery gasoline supply and discusses fuel distribution system impacts on supply.
Section 4 discusses the cost of this action and Section 5 provides analysis of potential consumer
pricing impacts as a result of program cost and supply reductions. Section 6 discusses the
potential benefits of this action and Section 7 discusses how the supply of gasoline may change
in 2025. Finally, Section 8 discusses EPA's screening analysis evaluating the potential impacts
of this action on small entities. While the Clean Air Act (CAA) does not permit EPA to consider
costs and prices in its decision-making on this action, we are nevertheless providing the
information as a means of informing stakeholders of the impacts of this action.
The removal of the 1-psi waiver only affects summer conventional gasoline, so the
discussion in this document focuses solely on the impacts on summer conventional gasoline,
primarily in the petitioning states. While reformulated gasoline (RFG) is also sold within the
petitioning states, the 1-psi waiver does not apply to RFG, nor does it apply to winter gasoline.
As such, there is comparatively little discussion of impacts on RFG and winter gasoline in this
document.
1
-------
2. Fuel Production and Distribution
A.
Overview
This section presents a high-level overview of how gasoline is produced and distributed
in the United States, to provide background for the remainder of this document.
The petitioning states are all located in the Midwest in a refining region called Petroleum
Administration for Defense District (PADD) 2. A number of adjacent non-petitioning states are
also located in PADD 2, as well as PADDs 1, 3, and 4. The various U.S. refining districts are
shown in Figure 2.A-1.
Figure 2.A-1: Refinery Petroleum for Defense Districts (PADDs)
Petroleum Administration for Defense Districts
I
PADD 5:
West Coast,
AK, HI
San FrancisM
•
MT
PADD 4:
Rocky
Mountain
UT Dc2~r
CO
MM
PADD 2:
Midwest
IA
KS MO
,M i
OK
eia
PADD 1 a. A
New .
paddib ^ mwT
Central lA
' jfi
PADD 3: Gulf Coast
7-
aA Lowei
H§ Atlanta
Mi
W
PADD 1:
East
Coast
Source: U.S. Energy Information Administration.
Before assessing the fuel industry's ability to supply a new lower-volatility gasoline to
the petitioning states, it is important to understand how gasoline is supplied, from production to
retail outlets. Figure 2.A-2 shows a simplified form of the current fuel distribution system for
conventional gasoline in the petitioning states and the volatility of gasoline at various points in
the system.
2
-------
Figure 2.A-2: Interdependent Gasoline Production and Distribution System
10 psi
Ethanol
Petroleum Refinery "e^nerV
Storage
9 psi CBOB
10 psi i.e.,
Indiana
Ethanol
10 psi - i.e.,
Ohio
ft
Retail
Outlets
ft
Terminal
Some Barge and Rail Distribution
The first step of this system is fuel production: petroleum refineries refine crude oil using
various processing units to produce blendstocks and then blend the various blendstocks in
gasoline blending tanks. Once the gasoline meets required specifications, it is certified and
transported to the refinery storage tanks where it awaits scheduling through the fuel distribution
system to the retail fuels market. For the most part, refiners produce "suboctane" gasoline
blendstocks, which are then blended with ethanol downstream in the fuel distribution system to
meet the final gasoline specifications before distribution to retail outlets.1
The next step is the fuel distribution system, which transports the gasoline out of the
refinery. The last step of refining—storage of gasoline in dedicated tanks—is also the first step
of the fuel distribution system, since the gasoline in these tanks is then transported through the
fuel distribution system in batches. There are multiple ways for gasoline to be distributed from
refineries to retail outlets (e.g., pipelines, barges, and even rail), but gasoline can also be
distributed directly to local retail outlets off the refinery's terminal racks.
Most gasoline is transported by pipeline and the pipeline system is capable of
transporting multiple fuels, including different gasoline types and grades, diesel fuel, jet fuel, and
other various fuel blendstocks. When a pipeline reaches a juncture where a single pipeline
branches out to two different pipelines serving different gasoline markets, a set of short-term
storage tanks ("breakout tanks") are necessary to offload the fuel from the upstream pipeline to
enable scheduling the various fuels through the two downstream pipelines. Pipeline systems
often have many branches from the upstream to downstream pipelines to enable transporting the
1 There are two different types of gasoline blendstocks: conventional gasoline before oxygenate blending (CBOB),
which is used in conventional gasoline areas; and reformulated gasoline before oxygenate blending (RBOB), which
is used inRFG areas. Additionally, there are two octane grades: regular, which is 87 octane after blending with 10%
ethanol; and premium, which is 91-94 octane (generally 93 octane) after blending with 10% ethanol.
3
-------
fuel to downstream markets. These breakout tanks are particularly important for the supply of
different fuel types to different markets.
The next step in the fuel distribution system comprises the downstream terminals that
receive the fuel from the fuel distribution system and store the gasoline. The downstream
terminals are equipped with truck racks that enable tank trucks to load up with gasoline to
transport the fuel to retailers for dispensing to consumers. Storage tanks located at refineries can
also act as downstream terminals since gasoline stored in these tanks can be loaded at the
refinery's truck racks to distribute the gasoline locally. These terminals also blend ethanol into
the CBOB and RBOB to produce finished gasoline.
The final step of the fuel distribution system comprises the retail outlets (gas stations)
that receive truck shipments of finished gasoline from downstream and refinery terminals.
B. Refinery Actions to Control Gasoline Volatility
As a result of this rulemaking, certain refiners will need to reduce the volatility of their
CBOB to produce an -8.0 psi RVP CBOB (i.e., "low-RVP CBOB"), which when blended with
ethanol, becomes a 9.0 psi RVP gasoline (i.e., "low-RVP gasoline") for the petitioning states.
Today, refiners that distribute gasoline to PADD 2 produce a -9.0 psi RVP CBOB, which
becomes a 10.0 psi RVP finished gasoline once blended with 10% ethanol.2 Thus, when the 1-psi
waiver is removed for the petitioning states, certain refiners will need to produce a CBOB
approximately 1.0 psi lower in RVP. Some refineries that distribute gasoline to PADD 2 also
produce a 6.4 psi RVP RBOB, which becomes a 7.4 psi RVP finished RFG once blended with
10% ethanol. The RFG gasoline is sold in the Chicago-Milwaukee and St. Louis metropolitan
areas in PADD 2 and is unaffected by this rulemaking.
Refineries impacted by the removal of the 1-psi wavier are those located in the
petitioning states but may also include many, if not all, refineries located in states adjacent to the
petitioning states. Figure 2.B-1 highlights the petitioning states and shows the location of
refineries in and around the petitioning states that may be impacted by the removal of the 1-psi
waiver. What is not shown in the figure are the refineries located outside this region (e.g., Gulf
Coast refineries) that also produce gasoline to supply this region. We note also that 4 petitioning
states (Iowa, Missouri, Nebraska, and South Dakota) do not have refineries.
2 While refiners also produce RBOB for the Chicago-Milwaukee and St. Louis gasoline markets within the
petitioning states, the production and distribution of RFG should largely be unaffected by this rulemaking.
4
-------
Refineries typically control the volatility of gasoline at three different points: (1) From
the crude oil; (2) At the fluid catalytic cracker (FCC) unit and other refinery units; and (3)
Through removal of light hydrocarbons that would otherwise be blended into gasoline. For
refineries that utilize light and medium American Petroleum Industry (API) gravity crude oil
(i.e., less viscous) that contains butanes and lighter hydrocarbons, a refinery removes such
butanes and lighter hydrocarbons either by using a stabilizer unit (a type of distillation column)
prior to sending the crude oil to the refinery's atmospheric crude oil distillation tower or by using
the atmospheric crude oil distillation tower. A refinery running heavier, low-API gravity crude
oil (i.e., more viscous) is less likely to need to remove butane from crude oil because such crude
oil contains little or no butane that would need to be removed. We do not anticipate that these
refineries refining very heavy crude oil would need to adjust this initial processing step to
produce low-RVP CBOB.
Refineries are also able to control RVP by adjusting the product stream of their FCC unit
and other refinery units that produce butane. For most refineries, the FCC unit produces a
significant portion of the blendstock material needed to produce CBOB. The FCC unit converts
heavy hydrocarbons into lighter products such as FCC naphtha—a major gasoline blending
component—that usually has an RVP of approximately 9.0 psi. Refineries that produce gasoline
blends usually also have a debutanizer that removes at least some of the butane hydrocarbon
from the FCC naphtha product prior to blending the FCC naphtha into gasoline. The removed
butane is then typically sent to a gas plant where a butane tower separates the various butane
compounds from each other.
While the FCC unit likely provides the largest portion of butane to a refinery's gasoline
pool, other units also produce butane and other light hydrocarbons that contribute to the volatility
of the refinery's gasoline. The reformer—which reforms hydrocarbons into high-octane aromatic
compounds—cracks some heavier hydrocarbons into smaller, more volatile hydrocarbons,
3 Data source: EIA, Energy Infrastructure and Resources Maps, https://atlas.eia.gov/pages/energy-maps.
5
-------
including butane. However, due to the widespread blending of ethanol into gasoline, refiners
tend to operate their reformers at a lower severity, which contributes a smaller amount of butane
and other light hydrocarbons to gasoline. Cokers and hydrocrackers, like the FCC unit, crack
heavier refinery streams into lighter petroleum products such as gasoline, and therefore produce
some butane and other light hydrocarbons to the gasoline pool. While most refineries do not have
cokers and hydrocrackers, those that do could be faced with the need to remove the butane from
more than one gasoline blendstock to allow them to produce low-RVP CBOB.
Butane is a highly volatile hydrocarbon that quickly increases the RVP of gasoline. Even
if a refinery does not intend to produce low-RVP CBOB, some butane material is typically
removed from the FCC naphtha to provide feedstock to an alkylation unit. Alkylation units
produce alkylate, a gasoline blendstock that is low in RVP and high in octane, so making use of
these units is an economical option for providing octane and producing low-RVP CBOB.
Alkylation units react normal butene and iso-butylene—which are olefins or unsaturated analogs
of the butane family—with isobutane to produce alkylate, the product stream of the alkylation
unit. Once the olefinic and branched chain butane compounds are removed for alkylation, what is
left over is normal butane (n-butane; a straight-chain variety of butane). Alkylation units do not
use normal butane as feedstock material, although to upgrade it to alkylation feedstock, normal
butane could be isomerized to isobutane with the addition of an isomerization unit. However,
normal butane is left in the gasoline pool when refineries are producing gasoline for blending
with 10% ethanol to meet a 10.0 psi RVP standard.
To meet a 1.0 psi lower RVP standard, typical refineries would need to remove more
butane—mainly normal butane—to allow them to produce low-RVP CBOB. However, a
refinery's ability to do so depends on the capability of its existing debutanizer column at the FCC
unit, or the presence and capacity of debutanizer units at reformers, cokers, and hydrocrackers. If
an existing debutanizer column has excess capacity, the refinery may only need to adjust the
operation of its existing debutanizer column to enable it to remove the necessary amount of
butane. On the other hand, if its debutanizer column is at or near maximum capacity, the refinery
may still be able to remove the necessary amount of butane, but to do so would require it to
remove some pentanes as well, causing it to remove more gasoline material than desired. If the
refinery must modify one or more debutanizer columns to remove the necessary amount
additional butane, it could mean replacing the distillation column with a larger, taller column,
replacing the heater and associated boiler, and increasing the pumps and associated piping. We
anticipate that the need for modifications could be a significant limit on the ability of some
refineries to produce low-RVP CBOB since it could take two or more years for a refinery to
procure and install the necessary equipment and could require a refinery maintenance shutdown
to install and start up the equipment. Any shutdown, although temporary, would impact fuel
supply from the refinery. Figure 2.B-2 provides a representation of the FCC unit and associated
downstream units, including the debutanizer that removes butane compounds from FCC naphtha;
other refinery units that also produce butane would likely have a similar configuration.
6
-------
Figure 2.B-2: Volatility Control of FCC Naphtha
Fluidized Catalytic Cracker Unit
Debutanizer; must be large
enough to remove n-butane
Isobutyleneand Isobutanefor
Vacuum Gas Oil 650F to 1000F
Refiners need adequate butane sphere
capacity, and sufficient rail car capacity
to store/sell the excess butane
Removing n-butane,
about 2% of gasoline,
is needed to produce
an 8 psi C806
Replacing a debutan^er likely requires 4 years and a refinery shutdown to
implement
If refineries solely removed butane to satisfy a 1.0 psi lower RVP standard, the amount of
butane needed to be removed amounts to approximately 1.7% of the gasoline. However, we
believe that in removing butane, refineries will end up also removing a very small amount of
pentane along with the butane. Thus, we estimate that ultimately the removal of light
hydrocarbon material to produce low-RVP CBOB will result in a 2% reduction in butane and a
small amount of pentane from the gasoline pool.
An additional challenge for refiners associated with removing butane is the needed outlet
for the butane. Refineries that do not remove butane from gasoline prior to removal of the 1 -psi
waiver may not have the means to: (1) Remove butane from gasoline (which requires adequate
debutanizer capacity) and separate it from the rest of the butane compounds going to the
alkylation unit (which requires adequate gas plant capacity or capability); (2) Pipe the butane
from the gas plant; and (3) Either store the butane or load it onto a railcar for shipment out of the
refinery. Since refiners typically purchase butane for blending into winter gasoline, they likely
have some equipment—including piping and railcar offloading equipment—to be able to blend
butane into winter gasoline. However, this does not necessarily mean that they have the
equipment needed for creating an outlet for butane from gasoline, including a butane storage
tank or outlet to railcars.
Some refineries are already removing all the butane possible for their current product mix
(e.g., if its product mix includes a significant amount of RFG). For them to further reduce the
RVP of their product mix would require them to also remove pentanes—the next most volatile
hydrocarbons—which would have a much larger impact on gasoline production. Pentanes are
much lower in RVP than butane, so to achieve a 1-psi reduction in RVP would require roughly
8-10% of the gasoline pool in pentanes. For even other refineries, their feedstock supplies are so
light that even removing pentanes may not be sufficient. In such situations, the refinery may
have to remove light straight run naphtha (LSR)—the next most volatile hydrocarbons—with an
7
-------
even larger impact on gasoline production.4 Yet another set of refineries are in a different
situation but with a similar outcome wherein they may somehow have to remove large volumes
of light hydrocarbons from their gasoline supply. Such refineries are currently blending in
natural gas liquids (NGLs) to bring their gasoline up to the RVP limit.5 For example, refineries
processing Canadian tar sands crude oil receive NGLs along with the heavy crude oil and blend
the NGLs into their gasoline up to the levels allowed by the RVP standard. Based on information
provided in the Baker and O'Brien Study6 and EPA conversations with individual refiners, these
heavy crude oil refineries—some of which are located in the PADD 2—will likely need to
remove some NGLs from their gasoline blending. Thus, for some specific refineries within the
petitioning states, the impact on supply will likely be significantly larger than just 2%.
In addition to contributing to gasoline's volatility, butane also contributes to its octane;
normal butane and isobutane have blending octane values of 92.5 and 99, respectively, higher
than regular gasoline's 87 octane. Thus, when removing butane, a refinery must also make other
changes to replace the lost octane to keep its product consistent with and in compliance with
market gasoline specifications.
A refinery has several choices for making up the lost octane. Producing more alkylate is
one possibility that would increase both octane and gasoline volume. However, the alkylation
units are typically already optimized for maximum production during the summer season, so unit
expansion would likely be needed to produce low-RVP CBOB. Such expansions would be
expected to take several years of planning, permitting, and construction. Most refiners have spare
capacity to increase the severity of their reformers to offset the octane loss associated with
butane removal. Reformers convert low octane naphtha into high-octane aromatic compounds.
However, this process tends to also increase the production of benzene—which is also controlled
in gasoline under EPA's fuel quality regulations—so the refinery may need to add additional
controls for benzene. This could also take considerable time to put in place, but in the meantime
the refinery could purchase benzene reduction credits. Reformers also reduce gasoline yield, and
thus, further compound the supply shortfall. Furthermore, increasing the severity of reformers
would also produce more butane and other light hydrocarbons, which would further hinder the
refinery in its production of low-RVP CBOB.
Alternatively, or in addition, refineries could also change how they operate their
hydrocrackers, which can alter the relative refinery production of gasoline versus distillate fuel.
As an alternative to replacing lost volume and octane, particularly in the short term, a refinery
could choose to not make up for the reduced octane and instead produce less premium gasoline.
Premium gasoline is required to meet a higher octane specification, and thus producing less
premium gasoline could allow the refinery to make up for some of the loss in octane. Doing so,
however, would reduce the supply of premium gasoline.
4 LSR has an RVP of 12-17 psi and primarily contains pentanes and hexanes.
5 NGLs are hydrocarbons separated at natural gas processing plants where the residual liquids carried with the
natural gas are separated from the methane/ethane contained in natural gas. NGLs typically purchased and used by
refineries have an RVP of 12-17 psi and primarily contain pentanes and hexanes.
6 Baker and O'Brien, "Midwest States Gasoline RVP - 1 psi Waiver Study, Report for American Fuel and
Petrochemical Manufacturers," February 24, 2023. Submitted as part of comments from the American Fuel and
Petrochemical Manufacturers (AFPM), Docket Item No. EPA-HQ-OAR-2022-0513-0077.
8
-------
C. Pipelines
Petroleum pipelines allow for transportation of gasoline products throughout the U.S. and
are generally a safer, more reliable, and more efficient method of transporting petroleum
products from refineries to terminals compared to other modes of product movement. Figure
2.C-1 shows all pipelines transporting gasoline in the U.S., while Figure 2.C-2 shows pipelines
transporting gasoline through the western Midwest, where seven of the eight petitioning states
are located.
of Petroleum Pipelines in the United States7
Figure 2.C-2: Map of Pipelines in the Petitioning States8
Source: ICF analysis of ArcGIS, Energy Information Administration
LEGEND
• Sueplpng »•
Rtfllon Mattm ONy
ORilnim Supprnv «•
mm Ool Riow
O —
R»fin«y Capacity
(Thousand
O00*
O 200 - 399
O 100-199
O <100
K«y Pipolmai
EXPLORER
MAGELLAN
NUSTAR ENERGY
BUCKEYE
MARATHON
PHILLIPS 66
—West shore
KOCH
— BPPIPEUNES(NA)
ENTERPRISE TEPPCO
— PLAINS ALL AMERICAN
7 Data source: EIA, Energy Infrastructure and Resources Maps, https://atlas.eia.gov/pages/energy-maps.
8 Figure 3-2 was generated as part of ICF's report to the Renewable Fuels Association (RFA) and was completed
prior to the addition of the Ohio and Missouri petitions to remove the 1-psi waiver and prior to Kansas and North
9
-------
Most pipelines also have breakout tanks that allow for transferring gasoline from the
pipeline to another pipeline (or other transportation means), enabling it to serve multiple gasoline
markets. As shown in Figure 2.C-2, pipelines transport petroleum products through both
petitioning and non-petitioning states. There are an estimated 110 breakout tank terminals
located in the petitioning states, and more located outside this area that serve the petitioning
states.9 Some of these breakout tank terminals solely serve the petitioning states and could
simply convert their existing gasoline breakout tanks over to low-RVP CBOB; however, other
terminals provide gasoline to both petitioning and non-petitioning states. Many of the breakout
tank terminals that currently serve both markets will need additional breakout tankage to supply
both 9.0 psi and low-RVP CBOB.10 Lacking this additional tankage, a pipeline will likely need
to be dedicated to a single gasoline type.
Additionally, when transporting gasoline, different grades and types intersect as they are
shipped through the pipelines to their final or interim destinations. The intermixing of two grades
or types at the place they abut is referred to as interface. Pipeline companies attempt to sequence
batches of fuel such that the interface can be blended or cut into the batch with the least stringent
regulatory specifications. For example, when this interface occurs between premium and regular
grade CBOBs, the interface is cut from the premium grade portion of the shipment and blended
into the regular grade portion. If this is not possible, then the mixture is referred to as transmix
(e.g., an interface between gasoline and diesel fuel), which must be cut out of the pipeline,
stored, and then sent to refineries for reprocessing. With the addition of the new low-RVP
CBOB, additional smaller batches of gasoline will now have to be shipped. The pipelines will
have to reoptimize their shipment schedules to minimize the amount of downgraded product and
transmix. They will likely try to schedule batches such that 9.0 psi and low-RVP CBOB abut,
allowing the interface to be cut into the 9.0 psi RVP CBOB and minimize the production of
transmix. Nevertheless, we expect increased downgrading of gasoline to occur, which will
reduce the supply of low-RVP CBOB and increase the supply of 9.0 psi RVP CBOB.
I). Terminals
Refined product terminals take receipt of motor vehicle fuels (mostly from pipelines, but
also from barges, railcars, ships, etc.), store them, and then dispense them to tank trucks that
deliver the fuels to the ultimate consumers, mainly through retail outlets. There are an estimated
250 large gasoline refined product terminals in the petitioning states, each with a storage
capacity of at least 50,000 barrels. There are likely another 100 or so smaller fuels
storage/transfer facilities—often referred to as bulk plants—that serve more sparsely populated
areas. Figure 2.D-1 shows the location of refined product terminals in and around the petitioning
states.
Dakota rescinding their petitions. ICF, "Impact of Potential 8-State RVP Waiver Exclusion on Midwest Gasoline
Markets," prepared for the Renewable Fuels Association, September 2022 (hereinafter the "2022 ICF Report").
9 U.S. Department of Transportation, "National Map of pipeline breakout tanks," National Pipeline Mapping
System.
10 See, e.g., Baker and O'Brien Study; Magellan Midstream Partners, "Petition to Delay the Elimination of the 1 psi
RVP Waiver for E10 during the Summer Months - Insufficient Supply of Gasoline," September 16, 2022; Magellan
Midstream Partners, "Petition to Delay the Elimination of the 1 psi RVP Waiver for E10 Until the Summer of 2025
- Insufficient Supply of Gasoline," August 18, 2023.
10
-------
Refined product terminals generally have installed the number of fuel storage tanks
needed to serve their local fuels market. If a new fuel must be distributed through the terminal in
addition to the existing fuels already being serviced, the terminal will often not have additional
tankage available and would likely need to install a new product storage tank(s) to facilitate the
distribution of the new fuel. This is likely to be the case for terminals that are located near the
border of petitioning and non-petitioning states and serve markets in both types of states,
particularly in less-populated areas. In the case where a medium or large city is located near a
border between a petitioning and non-petitioning state, and the city is served by multiple
terminals, the various terminals may be able to choose between 9.0 psi and low-RVP CBOB in a
way that can efficiently distribute both gasoline types.
When a tank needs to be taken out of service for inspection or repair, some refined
product terminals have extra tankage or adjustability with regard to how they operate their
existing tanks in order to remain in operation. This is likely the case at larger terminals, as
smaller outlying terminals likely have less tankage and support a larger area. In these cases, a
terminal could decide to use this flexibility in tank usage to provide some supply of low-RVP
CBOB as a short-term solution. However, there would likely only be enough tankage to supply a
single grade of low-RVP CBOB at the terminal (i.e., regular grade low-RVP CBOB), leaving
premium grade low-RVP CBOB still in short supply. Use of this flexibility would also put the
terminal at risk of a supply shortage should a tank need to be taken out of service during the
summer season. Although most terminals likely do not have sufficient spare tankage to handle
both regular and premium grade low-RVP CBOB in addition to their current gasoline slate, some
terminals located in or near former 7.8 psi RVP areas may have had additional tankage that could
be utilized.12
11 Data source: EI A, Energy Infrastructure and Resources Maps, https://atlas.eia. gov/pages/energy-maps.
12 See Section 3.A for a list of such former 7.8 psi RVP areas.
11
-------
3. Evaluating the Supply of Gasoline
A.
Assessing the Refining Industry's Ability to Supply Low-R VP Gasoline
To explain how the integrated fuel production and distribution system will need to adjust
to the removal of the 1-psi waiver in the petitioning states, it is helpful to lay out a particular
example. Figure 3.A-1 shows the conditions required after the removal of the 1-psi waiver and
assuming that a pipeline is serving gasoline markets in both Ohio and Indiana. Although we use
Ohio and Indiana as an example of petitioning and non-petitioning states, any other adjacent pair
of a petitioning and non-petitioning state would suffice. Without the 1-psi waiver, gasoline in
Ohio must meet a 9.0 psi RVP standard, while the 1-psi waiver for Indiana remains unchanged,
allowing 10.0 psi RVP gasoline to continue to be sold in Indiana.
Figure 3.A-1: Integrated Fuel Production and Distribution System After Removal of the 1-
psi Waiver
Ethanol
Petroleum Refinery storage
9 psi CBOB
10 psi i.e.,
Indiana
9 psi - i.e.,
Ohio
i
Retail
Outlets
i
Some Barge and Rail Distribution
To continue to serve the gasoline markets in both states, both gasoline types must be
accommodated in production and throughout the fuel distribution system. This will require, for
example, separate gasoline storage tanks at refineries and pipeline breakout terminals for each
fuel type (i.e., 9.0 psi and low-RVP CBOB) and each fuel grade (i.e., regular and premium).
Up to this point, we have shown how the fuel distribution system could provide both 10.0
psi and low-RVP gasoline to PADD 2 markets. However, most of the current fuel production and
distribution system in and adjacent to PADD 2 was designed and optimized around a single 9.0
psi RVP CBOB with the 1-psi wavier throughout PADD 2.13 Therefore, many of the refined
product terminals currently serving both petitioning and non-petitioning states depicted in Figure
3.A-1—including pipeline breakout terminals and downstream product terminals—may not be
capable of handling the production and distribution of another gasoline type. We received
comments on the proposed rule that most of the gasoline storage facilities at pipeline breakout
13 The system also produces and distributes RFG and some boutique fuels for certain PADD 2 markets.
12
-------
terminals and refined product terminals are not immediately capable to supply both fuel types.14
Thus, in Figure 3.A-1, if a pipeline cannot distribute two gasoline types due to the lack of spare
tankage at its breakout terminal, it would have to provide low-RVP CBOB to both Indiana and
Ohio. Because of that limitation, a refinery would also be limited to only providing low-RVP
CBOB to the pipeline, increasing the volume of low-RVP CBOB that needs to be produced and
its associated impact on production volume. The refinery, however, may be able to provide 10.0
psi RVP gasoline off its own terminal rack if it has a second set of tanks for its terminal rack.
Alternatively, it could also provide 9.0 psi RVP CBOB to the pipeline and low-RVP gasoline off
its terminal rack.
The Baker and O'Brien Study proposed a way that the refining and the pipeline systems
could function in the near term after removal of the 1-psi waiver in the petitioning states. As
capital investments such as tank construction are unavailable for a short-term rollout, many
pipelines will likely need to dedicate their assets to distribute either 9.0 psi or low-RVP CBOB.
Figure 3.A-2 shows an example of how pipelines could manage distribution of both 9.0 psi and
low-RVP CBOB within its constraints. It also highlights the circumstances that arise with such a
system, including the parts of the fuel production and distribution system that could experience
supply issues should an unexpected outage occur.
Figure 3.A-2: Concept of Pipeline Movements After Removal of the 1-psi Waiver15
As shown in Figure 3.A-2, areas in western South Dakota and western Nebraska are not
depicted as having access to a major pipeline for providing low-RVP CBOB under this concept.
However, as shown in the more detailed pipeline map in Figure 2.C-2, there are short segments
of pipelines supplying these areas. These pipelines are shown in Figure 3.A-2 as being supplied
out of Colorado and Wyoming and are indicated as being dedicated to distributing 9.0 psi RVP
14 The Baker and O'Brien Study included a survey of downstream terminals that revealed that terminals in non-
petitioning states do not have the spare tankage required to receive, store, and distribute a second gasoline type.
15 Source: Baker and O'Brien Study. Red lines are movements of low-RVP CBOB and black lines are movements of
9.0 psi RVP CBOB.
13
-------
CBOB. If in fact these pipelines cannot also distribute low-RVP CBOB, the supply of low-RVP
CBOB into western South Dakota and western Nebraska would either have to be trucked in, or
the pipeline (or a portion thereof) would have to converted to low-RVP CBOB, causing at least
parts of Colorado and Wyoming to have to use low-RVP CBOB, increasing the volume of low-
RVP CBOB that would need to be produced and its associated impact on production volume.
Additionally, the conversion of the Denver area from a 7.8 psi RVP gasoline area to RFG
starting in 2024 will create a greater challenge to providing gasoline to western South Dakota
and western Nebraska.16
Similarly, while northern Minnesota retail outlets currently lack direct pipeline access for
gasoline, they are supplied with some gasoline from Fargo, which receives gasoline via pipeline
from the west (i.e., Montana). Figure 3.A-2 shows that Fargo would be serviced with a dedicated
9.0 psi RVP CBOB pipeline. With this loss of supply to the region, gasoline would need to be
supplied from another source, such as the two refineries in the Minneapolis region, although that
would take gasoline away from the areas to the south that they currently supply. The gasoline
supply to Minnesota is further complicated by the closure of a products pipeline that brings
gasoline in from the southwest.17
If a refinery that supplies low-RVP CBOB has an unexpected shutdown during the
summer season, both petitioning and non-petitioning states that are supplied by the refinery
would be more likely see the impacts of the lost supply, as it is more difficult to resupply than if
the gasoline in the region is fungible.
In certain situations, downstream terminals may also be impacted by the removal of the
1-psi waiver, which will affect gasoline supply. While a terminal that solely distributes gasoline
to either petitioning or non-petitioning states will likely continue its current practice of simply
selling the gasoline it receives, a terminal located near the border of a petitioning and non-
petitioning state may end up in a more challenging situation. Such terminals likely serve gasoline
markets in both states because it is the most efficient way to distribute gasoline to the markets in
both states—it minimizes the travel distance from the terminal to retail outlet and requires fewer
tank trucks and tank truck drivers. As an example, we consider terminals located near the border
of Indiana and Ohio. Figure 3.A-3 shows how these terminals can provide the same gasoline type
to both states prior to the removal of the 1-psi waiver.
16 87 FR 60926 (October 7, 2022).
17 Pipeline County Star, "Pipeline running through Pipestone County to be decommissioned," May 5, 2022,
https://www.pipestonestar.com/articles/pipeline-running-through-pipestone-countv-to-be-decommissioned.
14
-------
Currently, terminals can sell fuel over state lines
State Line; both states have waivers
Indiana; 10 psi gasoline
Ohio; 10 psi gasoline
Figure 3.A-3: Gasoline Distribution From Terminals Near State Borders Prior to Removal
of the 1 -psi Waiver
An estimated75 terminals are located
close to the bordetsbetween petitioning
and nonpetitioning states
If a terminal near the border of a petitioning and non-petitioning state does not have extra
tanks to accommodate the addition of low-RVP gasoline, it will likely have to adjust its
distribution area to serve either one state or the other, but this is still challenging. Again, we
consider the example of terminals located near the border of Indiana and Ohio. Figure 3.A-4
shows how such terminals that do not have spare tankage to distribute both 10.0 psi and low-
RVP gasoline would need to choose to sell only one type of gasoline after the removal of the 1-
psi waiver in Ohio.
Figure 3.A-4: Gasoline Distribution From Terminals Near State Borders After Removal of
the 1-psi Waiver
State with waiver 10 psi i.e., Indiana
State without waiver
9 psi-i.e., Ohio
In this case, if the terminal located in Indiana primarily served the Indiana gasoline
market prior to the removal of the 1-psi waiver, we expect it would continue to sell only 10.0 psi
RVP gasoline even after the removal of the 1-psi waiver in Ohio, such that the terminal would be
limited to distributing its gasoline only to retailers in Indiana. In this case, another terminal
would need to pick up distributing low-RVP gasoline to the retailers in Ohio. However,
compared to Figure 3.A-3, the distribution of gasoline in Figure 3.A-4 would be less efficient
15
-------
and thus more costly (in both states) because the gasoline would need to be distributed over
longer distances. We expect that either more tank trucks would be needed, or the existing tank
trucks would need to be used more often (i.e., more and longer trips per day) to cover the further
distances that gasoline would need to be distributed. Either way, more drivers and probably more
trucks would be needed to transport the gasoline longer distances, which would require some
time to put into place. Note that while the terminal in Indiana could not distribute its gasoline
into Ohio, the terminal in Ohio could distribute its gasoline into Indiana since the low-RVP
gasoline would meet the 10.0 psi RVP standard in Indiana, which could mitigate some of the
gasoline distribution challenges near these terminals.
We also believe that at least some downstream terminals may have spare tankage that
would allow those terminals to store another gasoline type. There have been previous lower RVP
and RFG programs in PADD 2 that required refiners and fuel distributors to provide gasoline
types other than 10.0 psi RVP gasoline (i.e., 9.0 psi RVP CBOB). In recent years, a number of
these programs have been eliminated, as listed below. However, the capabilities that allowed
these additional fuel types to be produced and distributed may still be available in some locations
to provide some flexibility to produce and distribute both 9.0 psi and low-RVP CBOB.
In petitioning states:
• Missouri
o On March 12, 2021, EPA published a final rule (86 FR 14007) that removed
Missouri's 7.0 psi RVP standard that applied to 3 counties in the Missouri portion
of the Kansas City, KS-MO area from the approved SIP. The final rule was
effective on April 12, 2021.
• Ohio
o On April 7, 2017, EPA published a final rule (82 FR 16932) that removed Ohio's
7.8 psi RVP standard that applied to 8 counties in the Cincinnati and Dayton areas
from the approved SIP. The final rule was effective on April 7, 2017.
In adjacent non-petitioning states:
• Kansas
o On March 12, 2021, EPA published a final rule (86 FR 14000) that removed
Kansas' 7.0 psi RVP standard that applied to 2 counties in the Kansas portion of
the Kansas City, KS-MO area from the approved SIP. The final rule was effective
on April 12, 2021.
• Kentucky
o On May 15, 2018, EPA approved an opt-out petition (83 FR 22593) that removed
the RFG requirement for 3 counties in the Kentucky portion of the Cincinnati-
Hamilton, Ohio-Kentucky-Indiana maintenance area for the 2008 ozone NAAQS
(Northern Kentucky Area). The opt-out was effective on July 1, 2018.
• Tennessee
o On June 7, 2017, EPA published a final rule (82 FR 26354) that removed the 7.8
psi federal RVP standard in 5 counties in the Middle Tennessee area (Nashville).
The final rule was effective on June 7, 2017.
16
-------
o On December 22, 2017, EPA published a final rule (82 FR 60675) that removed
the 7.8 psi federal RVP standard for 1 county in the Memphis area. The final rule
was effective on January 22, 2018.
To minimize the cost and other impacts to enable production and distribution of low-RVP
CBOB, refiners and fuel distributors will need time to make capital investments to optimize the
refinery and fuel distribution system to enable replacing the gasoline solely in the petitioning
states with low-RVP gasoline. We expect that fuel distributors will need approximately two
years to finance, design, permit, and construct additional tankage and put it into service to
efficiently distribute low-RVP gasoline solely to the petitioning states. Neither the
implementation schedule requested by the petitioning states (summer of 2023) nor the summer of
2024 provide that amount of lead time.
Several refiners stated in their comments that the removal of the 1-psi waiver in the
petitioning states should be delayed until at least the summer of 2025 to allow for sufficient lead
time. Smaller projects (e.g., adjusting piping and pumps) could potentially be completed in less
than a year with minimal operations interruptions, which could help reduce supply losses on the
margin (e.g., moving more butane to storage that was previously inaccessible, allowing for
shipment and storage offsite if needed). However, many refineries will require more significant
capital projects to offset or at least minimize the supply losses. Refiners estimate that a major
capital construction project (e.g., new tankage) requires at least 2 years to complete.18 With this
timeline in mind, implementing these types of major projects cannot be completed ahead of the
summer of 2024, but could be finished in time for the summer of 2025 if the investment process
has already begun. However, with the uncertainty in the final implementation date of the removal
of the 1-psi waiver, most refineries have expressed difficulty in obtaining financing for these
types of investments. These refineries are looking at a much tighter construction window and
may not have their projects completed before the summer of 2025, particularly if the projects are
of sufficient scale to require a refinery shutdown for the work to occur (e.g., re-traying or
converting the debutanizer to packing). Some refineries may not be able to make such large,
more complex modifications until their next regularly scheduled full refinery maintenance
shutdown, although, depending on their size and scope, some projects (e.g., debutanizer
debottlenecking) may only require a partial refinery shutdown.
The total amount of time needed to invest in, put in place, and make operational new
capital investments is not the only factor affecting refiners' and fuel distributors' ability or desire
to put in place new capital investments in time. The lack of a firm effective date for the removal
of the 1-psi waiver is another factor. The CAA prescribes the effective date of the removal of the
1-psi waiver for E10 and allows EPA to extend that date by a year upon a determination of
insufficient gasoline supply. That extension can then be renewed in additional one-year
increments. This allowance creates uncertainty in the ultimate effective date of the removal of
the 1-psi waiver, thereby limiting refiners' and fuel distributors' ability to plan for investments
that take more than a year to implement.
In their comments, refiners expressed concern that if they invest in upgrades to enable
production of low-RVP CBOB, two different actions could occur that could strand their capital
18 Petition from HF Sinclair (November 15, 2023).
17
-------
investments. The first action would be Congress promulgating legislation to extend the 1-psi
waiver to El5, which they assert would allow El5 to be produced using the same CBOB
currently produced for E10.19 The second action would be where a Governor requests that the 1-
psi waiver for E10 be reinstated after the rule is finalized.20 The commenters further suggested
that price spikes or spot shortages could be reasons for such a request to reinstate the 1-psi
waiver. If refiners and fuel distributors are concerned that any capital investments could be
stranded, they are less likely to invest in them even if they have sufficient time to plan for such
investments.
Due to the lack of detailed information on every component of the fuel production and
distribution system in PADD 2—and the resulting significant uncertainty—it is not possible to
conduct a detailed assessment of the extent that non-petitioning states would use low-RVP
gasoline instead of 10.0 psi RVP gasoline. Various uncertainties include:
• Refineries
o Which refineries have sufficient butane removal equipment and the ability to store
the butane or move it out of the refinery. These are likely the refineries that can
produce low-RVP CBOB at a lower cost and with only a 2% loss of gasoline
production.
o Which refineries would need to remove less-volatile hydrocarbons at a much
higher cost and a greater loss of gasoline volume, and if the refinery can remove
at least some butane in addition to less-volatile hydrocarbons, what proportion of
that is butane.
o Which refineries have extra gasoline blending/storage tanks or other tankage that
could be switched into gasoline service, which would allow them to produce both
9.0 psi and low-RVP CBOB.
o What gasoline distribution options each refinery has access to—for example, what
portion of its gasoline can it sell off its terminal rack versus other downstream
distribution means (e.g., pipeline, barge, railcar).
• Gasoline Distribution (Including Pipelines, Barges, Railcar and Tank Truck)
o Which breakout terminals have extra gasoline blending/storage tanks or other
tankage that could be switched into gasoline service, which would allow the
pipelines to transport both 9.0 psi and low-RVP CBOB.
o For pipelines that are unable to transport both 9.0 psi and low-RVP CBOB, are
there any other options, or would the entire downstream market need to convert to
low-RVP gasoline.
o Which pipelines (or segments thereof) would not be able to transport both 9.0 psi
and low-RVP CBOB, and how much would it affect the sale of low-RVP gasoline
in non-petitioning states.
o What capacity exists to rely on barges, railcars and tank trucks—in lieu of
pipelines—to move gasoline into either the petitioning or non-petitioning states,
19 See, e.g., Senate Bill 2707, Nationwide Consumer and Fuel Retailer Choice Act of 2023.
20 Indeed, we have provided such a regulatory provision in this action at 40 CFR 1090.297. Additionally, we have
had several governors rescind their petitions for removal of the 1-psi waiver prior to this final action.
18
-------
and the ability for those alternative transportation means to respond once a
gasoline supply issue becomes apparent.
• Terminals
o How many terminals serve both petitioning and non-petitioning states and what
portion of the gasoline they distribute is currently distributed to either state,
o Which terminals have extra gasoline blending/storage tanks or other tankage that
could be switched into gasoline service, which would allow them to distribute
both 10.0 psi and low-RVP gasoline,
o How many terminals located near the border of petitioning and non-petitioning
states have the storage capacity to receive and distribute both 10.0 psi and low-
RVP gasoline.
o How many terminals have access to alternative gasoline distribution means (e.g.,
railcar, barge) if the pipeline they currently rely on is limited in the type of
gasoline it can distribute,
o Whether there would be sufficient availability of additional tank trucks and
drivers in the case of terminals without spare tankage need to distribute gasoline
further distances.
It is clear from the above examples that the fuel production and distribution system will
need to sort out which refineries will supply low-RVP gasoline and how it will be moved to
various markets in the petitioning states while still supplying 10.0 psi RVP gasoline to non-
petitioning states. Ideally this sorting out would occur as contracts are established for supplying
different gasoline markets in advance of the summer season. It is unlikely, however, that the
supply of low-RVP gasoline to every single gasoline market in the petitioning states can be
sorted out in advance. If the supply of low-RVP gasoline is insufficient to any gasoline market
and causes the drawing down of its gasoline inventories, the price of gasoline will increase in
that market and the fuel distribution system will respond to the higher gasoline prices by moving
more gasoline into that market. In cases of refinery outages today, it is this pricing mechanism
that causes the fuel distribution system to move gasoline from different markets to fill the void
left by the refinery outage.
After considering the capacity constraints of the fuel production and distribution system,
it is clear that some amount of low-RVP gasoline—potentially a significant amount—will need
to be supplied to at least part of the area of the non-petitioning states immediately adjacent to the
petitioning states. Although selling low-RVP gasoline in non-petitioning states will significantly
increase the volume of low-RVP gasoline needed to be produced and distributed to satisfy
demand, it will dramatically ease the burdens on the fuel distribution system and reduce the
chance for supply issues. Over time, the fuel production and distribution system will invest in the
necessary capital to optimize fuel production and distribution to more efficiently target low-RVP
gasoline solely to the petitioning states. In the next section we estimate the volumetric supply
impact of producing and distributing low-RVP gasoline and provide an estimate of the amount of
gasoline in non-petitioning states that could be impacted.
19
-------
B. Impact of the Action on Supply-Demand Balance
Estimating the reduction in gasoline supply after the removal of the 1-psi waiver is
challenging. There are two principal components: (1) The amount of the reduction in gasoline
supply by the refiners producing low-RVP CBOB; and (2) The amount of gasoline—including
gasoline used in non-petitioning states—that will be low-RVP gasoline. Although our evaluation
of insufficient supply described in the preamble is solely concerned with the supply of gasoline
in the petitioning states, we are still interested in the other impacts of the removal of the 1-psi
waiver, including gasoline supply impacts in non-petitioning states. This concern is because as
more light hydrocarbons are removed from the PADD 2 gasoline pool, it further decreases the
quantity of gasoline available in PADD 2, including the petitioning states.
As discussed in Section 2.B, we identified two different means for refiners to produce
low-RVP CBOB. The first is by solely removing butane, which is the most desirable way
because it results in only a 2% loss in gasoline production. The second, less-desirable way to
produce low-RVP CBOB is to remove somewhat less-volatile hydrocarbons (e.g., pentanes, LSR
or NGLs). However, because of the lower volatility of these hydrocarbons relative to butane, as
much as 10% of these hydrocarbons would have to be removed to achieve the same 1-psi
reduction in RVP. The refineries that fall into this second category are those that may not have
the ability to remove or store all the butane in their gasoline, as well as refineries that refine
heavy crude oil and therefore rely on these less-volatile hydrocarbons to blend their gasoline up
to the RVP limit.
To estimate the impact on gasoline supply, it is necessary to estimate what portion of
refinery gasoline production would achieve the 1-psi reduction in RVP by removal of butane
versus by removal of the less-volatile hydrocarbons. We began by reviewing the information
provided in the Baker and O'Brien Study and in conversations with individual refiners. The
Baker and O'Brien Study included a survey indicating that 30% of refineries would have
difficulty producing low-RVP CBOB, to the point that these refineries may need to reduce their
crude oil throughput volume in order to produce low-RVP CBOB. This would reduce not only
gasoline supply, but also the supply of diesel fuel, jet fuel, and other refined products. However,
in our discussions with some refiners, they stated that the challenges they would face in
producing low-RVP CBOB would result in a much larger decrease in gasoline production than if
they only needed to reduce butane content in their gasoline, but none said that they would need
to reduce crude oil throughput volume at their refineries.
Because of the higher supply impact and associated higher cost compared to refineries
that can produce low-RVP CBOB by solely removing butane, some of these refineries may want
to continue producing 9.0 psi RVP CBOB to sell in non-petitioning states. If a refinery has
separate tankage to allow this, it could sell one gasoline type to a pipeline (e.g., low-RVP
CBOB) and the other gasoline type off its terminal racks (e.g., 10.0 psi RVP gasoline). However,
some of these refineries are located in petitioning states and may not be able to easily sell their
gasoline in non-petitioning states, so they may need to find a way to produce low-RVP CBOB
despite its much larger impact on gasoline production and higher cost.
20
-------
The Baker and O'Brien Study listed several ways that a refinery could reduce the RVP of
its gasoline to help reduce the loss of gasoline supply, to which we added additional ideas. These
include:
• If the refinery is isomerizing its LSR to increase the RVP and octane of the LSR, it could
stop isomerizing the LSR. To make up for the resulting loss in octane, the refinery could
run its reformers at a higher severity, which would produce more aromatics that are lower
in RVP, but also produce more light hydrocarbons from the cracking reactions.
• Although refineries already tend to send their heavy gasoline swing cuts to gasoline
during the summer, if the refinery is still sending either the heavy straight run naphtha or
heavy FCC naphtha to distillate, it could move these streams back to gasoline. This
would increase the volume of low-RVP gasoline blendstocks, thereby lowering the
overall RVP of the refinery's gasoline pool and making up for any loss of gasoline
volume by removing any high-RVP gasoline blendstocks.
• The refinery could refine a somewhat heavier crude oil slate, which would help to reduce
the low-RVP material the refinery produces. However, this option is often limited by the
refinery configuration and doing so could cause a reduction in refinery's gasoline and
diesel fuel production.
• The refinery could purchase some heavy gasoline blendstock material (e.g., heavy FCC
naphtha, heavy reformate, or alkylate) to blend into its gasoline, which would lower the
RVP of the refinery's gasoline pool.
While the reduction in gasoline volume based solely on butane removal is easily
calculatable, the impact for these other refineries facing a larger gasoline loss is more difficult to
estimate. Each of these refineries is configured differently and the necessary data to calculate
their gasoline loss is not publicly available. Aside from not knowing which refineries can take
advantage of the options from the above list, there are two critical pieces information that would
allow us to better quantify the impacts:
(1) Does the refinery have the butane removal and storage/export capacity to allow it to
produce low-RVP CBOB by solely removing butane, and for those refineries that cannot, what
RVP level is achievable by solely removing butane?
(2) Does the refinery have the blending/storage tank capacity to produce a different
gasoline type to sell off its terminal rack (e.g., 10.0 psi RVP gasoline) compared to the gasoline
it sends to a pipeline (e.g., low-RVP CBOB)? As shown in Figure 2.B-1, many of the refineries
that produce gasoline for the petitioning states are located in or near non-petitioning states, so an
outlet of both 10.0 psi and low-RVP gasoline is possible.
The first component needed to estimate the impact on gasoline supply as a result of the
removal of the 1-psi waiver is an estimate of the volume of gasoline that will be low-RVP
gasoline. Table 3.B-1 summarizes the month-by-month conventional gasoline volume in the
petitioning states in the summer of 2021 and shows that the average gasoline demand in the
petitioning states was 41.9 million gallons per day. This estimate from the prime supplier sales
volume—which provides volumes on a state-by-state basis—is lower than the product supplied
volume, which according to EIA is the most accurate estimate of gasoline consumption but
21
-------
unfortunately is not provided on a state-by-state basis. When comparing the product supplied
volume to the prime supplier volume, the product supplied volume exceeds the prime supplier
sales volume for the U.S. by 5%. Based on discussions with EIA, the product supplied volume is
more accurate, so we adjusted the state-by-state volumes prime supplier sales volume higher by
5%. The prime supplier sales volume also includes the volume of ethanol in the gasoline, which
should not be included for purposes of our analysis since the reduction in refinery output applies
only to the volume of CBOB produced at refineries. We assumed that the gasoline volumes
contained an average of 10% ethanol. Adjusting the total average gasoline volume by the 5%
increase and 10% decrease results in a final average gasoline volume of 39.6 million gallons per
day for the petitioning states.
Table 3.B-1: Conventional Gasoline Volume in the Petitioning States in 2021 (kgpd)
State
March
April
May
June
July
August
Average
Illinois
3,987
4,182
4,275
4,300
4,373
4,354
4,245
Iowa
3,326
3,699
3,786
3,871
3,850
3,815
3,724
Minnesota
5,344
5,814
6,423
6,781
6,730
6,640
6,289
Missouri
5,405
5,711
5,793
6,009
6,071
5,880
5,811
Nebraska
2,158
2,345
2,404
2,570
2,497
2,485
2,410
Ohio
12,623
13,146
13,584
13,634
13,718
13,880
13,431
South Dakota
1,114
1,188
1,253
1,425
1,481
1,394
1,309
Wisconsin
3,908
4,334
4,733
5,047
5,136
4,973
4,689
Total
37,900
40,400
42,300
43,600
43,900
43,400
41,900
Adjusted Total
35,800
38,200
39,900
41,200
41,400
41,000
39,600
a Total volume adjusted by the 5% increase and 10% decrease discussed earlier in this section to account for: (1) The
underestimate of the prime supplier sales volume data compared to the product supplied volume data; and (2) The
volume of ethanol in the reported gasoline volume.
However, as discussed above, the limitations in the fuel distribution system—especially
in the first year after the removal of the 1-psi waiver—are expected to result in some of the
gasoline outside the petitioning states to also be low-RVP gasoline to avoid supply shortfalls in
the petitioning states and adjacent non-petitioning states. Table 3.B-2 summarizes the
conventional gasoline demand in adjacent non-petitioning states in the summer of 2021.
21 Data source: EIA, Petroleum & Other Liquids, Prime Supplier Sales Volumes (2021),
https://www.eia.gov/dnav/pet/pet cons prim dcu nus m.htm.
22
-------
Table 3.B-2: Conventional Gasoline Volumes in Adjacent Non-Petitioning States in 2021
(kgpd)22
March
April
May
June
July
August
Average
Arkansas
3,903
4,144
4,147
4,160
4,237
4,074
4,111
Colorado
5,703
5,896
6,099
6,456
6,675
6,571
6,233
Indiana
6,786
7,296
7,260
7,467
7,542
7,499
7,308
Kansas
4,330
4,494
4,596
4,812
4,743
4,617
4,599
Kentucky
4,649
4,924
4,973
5,007
5,100
4,970
4,937
Michigan
11,207
11,281
12,084
12,415
12,626
12,776
12,065
Montana
1,814
2,036
2,191
2,611
2,715
2,504
2,312
North Dakota
914
972
1,013
1,142
1,101
1,044
1,031
Oklahoma
5,556
5,664
5,710
5,825
5,842
5,673
5,712
Pennsylvania
6,698
6,952
7,215
7,508
7,549
7,512
7,239
Tennessee
9,255
9,816
9,828
9,828
9,947
9,898
9,762
West Virginia
1,799
1,809
1,829
1,823
1,874
1,834
1,828
Wyoming
824
901
1,006
1,235
1,268
1,175
1,068
In light of the significant uncertainty in evaluating the extent to which the removal of the
1-psi waiver might lead to an insufficient supply of gasoline, we evaluated a low-, medium-, and
high-impact scenario for how the market might respond. The scenarios reflect different volumes
of low-RVP gasoline that would end up being supplied to non-petitioning states to reflect our
uncertainty for what that impact could be. Table 3.B-3 provides very rough estimates of the
percentage of gasoline impacted in each adjacent non-petitioning state, which vary based on the
type of border the state shares with petitioning states.
Table 3.B-3: Percentage of Gasoline Impacted in Adjacent Non-Petitioning States
Impact
Scenario
Two Borders
Single Border
State Corner
Indiana, Kansas,
Michigan, North Dakota
Arkansas, Kentucky,
Pennsylvania, West
Virginia, Wyoming
Colorado, Montana,
Oklahoma, Tennessee,
Low
25%
10%
0%
Medium
50%
25%
10%
High
75%
50%
25%
There is likely a temporal component to what the low-, medium-, and high-impact
scenarios could represent, as well as a feasibility component. The temporal component is that
early on there is less ability for the fuel distribution system to supply low-RVP gasoline to the
petitioning states without also supplying low-RVP gasoline to non-petitioning states. We expect
that early on—especially if we had removed the 1-psi wavier for 2023 or 2024—the volume of
low-RVP gasoline supplied to non-petitioning states would have been towards the high end of
the impact scenario. Although we still expect an impact on gasoline supply in 2025, we
anticipate that it will be less severe than what would have been observed with a 2023 or 2024
implementation date. A 2025 implementation date provides refineries with more time for at least
22 Data source: EIA, Petroleum & Other Liquids, Prime Supplier Sales Volumes (2021),
https://www.eia.gov/dnav/pet/pet cons prim dcu nus m.htm.
23
-------
minor projects to be completed (e.g., piping or debottlenecking). Over time, as fuel distributors
add the capital equipment needed to optimize their operations and better target distributing low-
RVP gasoline solely to the petitioning states, we expect the impacted volumes to move towards
the low-impact scenario.
The feasibility of producing and distributing gasoline is also an important consideration
for how much low-RVP gasoline will be sold in non-petitioning states. If fuel distributors find it
infeasible to distribute low-RVP gasoline solely to the petitioning states without also distributing
low-RVP gasoline to non-petitioning states, and refiners can produce more low-RVP gasoline
even if doing so is at a higher cost, we expect the impacted volumes will be towards the high-
impact scenario. If, however, fuel distributors find it feasible to distribute low-RVP gasoline
mostly to the petitioning states, but the cost for refiners to produce additional low-RVP gasoline
is significant, then we expect the impacted volumes will be towards the low-impact scenario.
Considering the uncertainties in producing and distributing low-RVP gasoline—which
varies by individual facility and by groups of facilities in parts of PADD 2, both initially and
over time—it is very challenging to estimate how much low-RVP gasoline will be sold in non-
petitioning states. Also, while we use consistent low, medium, and high percentages in Table
3.B-4 to estimate the amount of low-RVP gasoline sold in non-petitioning states, it is possible
that there might be more low-RVP gasoline sold in non-petitioning states in one part of PADD 2,
while there might be less low-RVP gasoline sold in non-petitioning states in another part of
PADD 2.
We then applied the percentages in Table 3.B-3 to the volumes in Table 3.B-2 to estimate
the volume of low-RVP gasoline in non-petitioning states, as summarized in Table 3.B-4.
24
-------
Table 3.B-4: Estimated Low-RVP Gasoline Volumes in Adjacent Non-Petitioning States
tod)
State
Average Conventional
Gasoline Demand
Impact Scenario
Low
Medium
High
Arkansas
4,111
411
1,028
2,055
Colorado21
3,117
0
312
779
Indianab
7,308
1,827
3,654
5,481
Kansas
4,599
1,150
2,299
3,449
Kentucky
4,937
494
1,234
2,468
Michigan0
6,032
1,508
3,016
4,524
Montana
2,312
0
231
578
North Dakota
1,031
258
516
773
Oklahoma
5,712
0
571
1,428
Pennsylvania
7,239
724
1,810
3,619
Tennessee
9,762
0
976
2,440
West Virginia
1,828
183
457
914
Wyoming
1,068
107
267
534
Total
6,661
16,371
29,045
Adjusted Totald
6,300
15,500
27,400
a Gasoline demand volume in Colorado was reduced by 50% to exclude existing 7.8 psi RVP (and future RFG)
program in the Denver area based on Denver's population relative to the population of the entire state.
b Gasoline demand volume in Indiana was not reduced to exclude the 7.8 psi RVP program in the Indiana portion of
the Louisville area, as this area makes up only 3% of the gasoline consumption in Indiana.
0 Gasoline demand volume in Michigan reduced by 50% to exclude existing 7.0 psi RVP program in the Detroit area
based on Detroit's population relative to the population of the entire state.
d Total volume adjusted by the 5% increase and 10% decrease discussed earlier in this section to account for: (1)
The underestimate of the prime supplier sales volume data compared to the product supplied volume data; and (2)
The volume of ethanol in the reported gasoline volume.
The estimated total daily volume of low-RVP gasoline in both the petitioning and non-
petitioning states for the three scenarios is summarized in Table 3.B-5.
Table 3.B-5: Estimated Total Daily Volume of Low-RVP Gasoline in Petitioning and Non-
Impact Scenario
Low
Medium
High
Thousand Gallons per Day (kgpd)
45,900
55,100
67,100
Thousand Barrels per Day (kbpd)
1,090
1,300
1,600
The next component needed to estimate the impact on gasoline supply as a result of the
removal of the 1-psi waiver is the portion of gasoline production that would experience this
higher gasoline production loss due solely to butane removal versus those that would need to
remove less-volatile hydrocarbons (e.g., pentanes, NGLs, and LSR). These impacts are likely to
be greater at the start of the summer gasoline production season, as pipelines require that
gasoline meet a more-stringent interim RVP specification starting March 1st each year (the
"transition period") to ensure that the measured RVP of the gasoline in all downstream terminal
tanks will comply with the applicable RVP standard by the May 1st start of the summer season.
25
-------
This extra-low RVP gasoline mixes with the high-RVP winter gasoline leftover in the various
distribution storage tanks to more quickly and efficiently "turn-over" the gasoline from winter to
summer specifications. While the duration of the transition period is likely to be different for
different pipelines, regions, and refineries, at least one pipeline company plans to keep this
interim RVP specification in place during March and April for most of its pipeline system. We
therefore separate our analysis below into two sets of calculations: One for May to September
and one for the transition period (March and April).
We begin with our analysis of impacts during May to September. In our conversations
with refiners, the refineries that might need to remove less-volatile hydrocarbons in addition to
butane as a result of producing low-RVP CBOB accounted for much less than 50% of the
gasoline pool. The survey of refiners included in the Baker and O'Brien Study indicated that
30% of refineries were categorized as potentially needing to reduce their crude oil throughput
volume. While we do not think that refineries will need to reduce their crude oil throughput, we
do anticipate that these are the refineries that will need to remove less-volatile hydrocarbons and
experience a greater gasoline production loss.23 For the purposes of our analysis, we assumed
that 30%) of gasoline production would experience a 5—10% loss commensurate with needing to
remove less-volatile hydrocarbons,24 while 70% of gasoline production would experience only a
2% loss commensurate with only butane removal.
Table 3.B-6: Estimated Gasoline Production Loss From Producing Low-RVP CBOB (May
to September)
Butane
Less-Volatile
Removal Only
Hydrocarbon Removal
Overall Loss
Gasoline Production
70%
30%
Gasoline Production Loss - Low
2%
5%
2.9%
Gasoline Production Loss - High
2%
10%
4.4%
The high gasoline production loss scenario is more representative of the supply impact if
the implementation date of the removal of the 1-psi waiver had been 2023 or 2024. The low
gasoline production loss scenario is more representative of the supply impact for later years (e.g.,
a 2025 or later implementation date), where more time has been allowed for capital projects to be
completed. Over time, as refineries add and complete these capital equipment projects, allowing
them to remove butane instead of less-volatile hydrocarbons, we expect the gasoline production
loss to decrease towards the low gasoline production loss scenario.
23 This analysis estimates that during the initial year or two after the removal of the 1-psi waiver, there will be a
larger effect on gasoline supply by refiners needing to remove more, less-volatile hydrocarbons (e.g., pentane,
NGLs, and LSR) instead of butane. As refiners make capital investments and optimize their refining operations to
enable removing butane, the supply impacts of producing low-RVP gasoline will decrease.
24 In discussions with refiners who reported challenges with producing low-RVP CBOB, their estimated impact on
gasoline production at these refineries ranged from 4-9%. To capture this range in reduced gasoline production at
these refineries for our supply analysis, we rounded the range to a low- and high-impact scenario of 5% and 10%,
respectively.
26
-------
We then combined the range of gasoline production loss estimates in Table 3.B-6 with
the range of low-RVP gasoline volumes in Table 3.B-5 to estimate gasoline supply impacts
during May to September, as shown in Table 3.B-7.
Table 3.B-7: Estimated Gasoline Supply Impacts During May to September
Gasoline Production
Loss Scenario
Impact Scenario
Low
Met
ium
High
kgpd
kbpd
kgpd
kbpd
kgpd
kbpd
Low
1,330
32
1,600
38
1,940
46
High
2,020
48
2,420
58
2,950
70
Next, we move to our analysis of impacts during the transition period. One pipeline
company stated that while low-RVP CBOB would normally be required to meet a 7.8 psi RVP
specification, the interim RVP specification would be 7.3 psi.25 Pipelines currently require 9.0
psi RVP CBOB to meet an interim 8.5 psi RVP specification instead of the normal 8.8 psi RVP
specification, so the net effect for refineries is now a 1.2-psi reduction in RVP, going from 8.5
psi to 7.3 psi (instead of from 8.8 psi to 7.8 psi).26
While we previously estimated the impacts of a 1 -psi reduction in RVP in Table 3.B-7,
going from 8.5 psi to 7.3 psi RVP requires a change in the mix of hydrocarbons that are
removed. The more-stringent RVP standard is expected to increase the amount of less-volatile
hydrocarbons that need to be removed in comparison to during the May to September timeframe,
increasing the impact on gasoline supply. To estimate the supply impact of the interim 7.3 psi
RVP specification, we need to estimate: (1) The percentage of gasoline that must be removed;
and (2) The portion of gasoline impacted.
We first estimate the mix of light hydrocarbons that would be removed to meet the
interim 7.3 psi RVP specification. In Table 3.B-6, we estimated that refiners would need to
remove 2.9-4.4% of butane and other less-volatile hydrocarbons from their gasoline to achieve a
1-psi reduction in RVP to produce low-RVP CBOB. This estimate assumed that 70% of gasoline
production could be achieved by solely removing butane, while the other 30% would need to
remove less-volatile hydrocarbons. However, for a 1.2-psi reduction in RVP during the transition
period, we extrapolated this 70/30 ratio to estimate that 55% of gasoline production could be
achieved by solely removing butane, while the other 45% would need to remove less-volatile
hydrocarbons. We again estimate that, per 1-psi RVP reduction, solely removing butane would
reduce the gasoline pool by 2%, while removing less-volatile hydrocarbons would reduce the
gasoline pool by 5—10%, or an average loss of 7.5%. As shown in Table 3.B-8, we estimate that
refiners that produce 7.3 psi RVP CBOB would experience an average loss of 5.4% during the
transition period.
25 Magellan Midstream Partners, "Petition to Delay the Elimination of the 1 psi RVP Waiver for E10 during the
Summer Months - Insufficient Supply of Gasoline," September 16, 2022.
26 Magellan Midstream Partners, "Schedule of Origin Volatility Requirements," Revision Date July 1, 2017.
27
-------
Table 3.B-8: Estimated Gasoline Production Loss During the Transition Period
Butane
Removal
Only
Less-Volatile
Hydrocarbon
Removal
Overall Loss
per 1-psi RVP
Reduction
RVP
Reduction
Overall
Loss
Gasoline Production
Loss
2%
7.5%
Gasoline Production
(7.8 psi RVP)
70%
30%
3.7%
1 psi
3.7%
Gasoline Production
(7.3 psi RVP)
55%
45%
4.5%
1.2 psi
5.4%
To estimate the impact on gasoline supply during the transition period, we estimated that
between 30-70% of the low-RVP gasoline in petitioning and non-petitioning states would need
to meet the interim 7.3 psi RVP specification. The balance of low-RVP gasoline in these states
would come from sources that are not forced to meet this more-stringent RVP specification (e.g.,
refineries that sell gasoline off their terminal racks). At the low end of the range (30%), we
assumed that more gasoline is sold from refinery racks and blended and stored in separate
refinery tanks. Additionally, we assumed that some pipeline segments would not require the
more-stringent RVP specification, either due to these segments being shorter or having fewer
branches with fewer breakout terminals that need to be turned over. Conversely, at the high end
of the range (70%), we assumed that less gasoline is sold from refinery racks and that all
gasoline distributed by pipeline is required to meet the more-stringent RVP specification. We
then combined this range with the range of gasoline production losses in Table 3.B-8 to estimate
the average gasoline production loss across the entire gasoline pool during the transition period,
as shown in Table 3.B-9.
Table 3.B-9: Estimated Average Gasoline Loss During the Transition Period
7.8 psi RVP CBOB
7.3 psi RVP CBOB
Gasoline Production
Gasoline
Gasoline
Average
Loss Scenario
Portion
Loss
Portion
Loss
Gasoline Loss
Low
70%
3.7%
30%
5.4%
4.2%
High
30%
3.7%
70%
5.4%
4.9%
We then combined this range with the range of low-RVP gasoline volumes in Table 3.B-
5 to estimate additional gasoline supply impacts during the transition period, as shown in Table
3.B-10.
Table 3.B-10: Estimated Gasoline Supply Impacts During the Transition Period
Gasoline Production
Loss Scenario
Impact Scenario
Low
Met
ium
High
kgpd
kbpd
kgpd
kbpd
kgpd
kbpd
Low
1,928
46
2,314
55
2,818
67
High
2,249
53
2,700
64
3,288
78
There are also other future market changes that could further stress gasoline supply in the
petitioning states. For example, the Denver area will convert from a 7.8 psi RVP area to an RFG
28
-------
area in 2024.27 This will require gasoline sold in the area to meet a more-stringent 7.4 psi RVP
standard (without the 1-psi waiver for E10) rather than the current 7.8 psi RVP standard (with
the 1-psi waiver for E10).28 To produce RFG, refineries must remove 5% of light gasoline
blending components relative to the 8.8 psi RVP gasoline currently produced for the Denver
area. Based on the estimated volume of gasoline sold in the Denver area, RFG production for this
area is estimated to cause the removal of 5 kbpd or more of gasoline blendstock. The fuel
distribution system in the Mid-Continent area will also be significantly impacted as it will need
to be modified to accommodate another type of gasoline (both regular and premium grades) to be
distributed to this area. While Denver is located in PADD 4, there are connections both to and
from the petitioning states for gasoline supply; thus, there will be some ripple effects in PADD 2.
We have considered the potential impact of increased El 5 sales volumes as a result of the
removal of the 1-psi waiver in the petitioning states. Although CAA section 211(h)(5) requires
removal of the 1-psi RVP waiver for E10, the governors, in their petitions, cite a desire to
support year-round sales of E15, and this action will allow E10 and E15 to be produced using the
same blendstock, thus removing one of the hurdles to El 5 sales in the summer.29 However, other
hurdles remain, including vehicle compatibility, fuel offerings, liability concerns, and especially
compatibility with existing retail outlet infrastructure. This action does not authorize the use of
El 5. While removal of the 1-psi waiver may have some impact on El 5 sales volumes, any
impact is difficult to project and quantify as data on E15 consumption is limited.
Nevertheless, in prior actions, we have assessed future volumes of El 5. In the Renewable
Fuel Standard (RFS) Set Rule, EPA used data from USDA's Biofuels Infrastructure Partnership
(BIP) to complete its analysis on potential future volumes of E15.30 This analysis demonstrated
an overall increase in El5 volumes over the studied years (with the exception of 2020 due to the
effects of the COVID-19 pandemic), as seen in Figure 3.B-1. When future years were
extrapolated from the data, a continued upward trend in El5 volumes was projected, primarily as
a result of increased El 5 retail outlet growth.
27 87 FR 60926 (October 7, 2022).
28 40 CFR 1090.215(a)(3).
29 We note that in the Energy Policy Act of 2005, Congress instituted a renewable fuel program requiring increasing
volumes of renewable fuel be used in gasoline through 2022. And in recognition of the expected increase in ethanol
use resulting from these provisions, Congress added the state relief provision in CAA section 211(h)(5) to allow
states to obtain an exclusion from the less-stringent RVP limit under CAA section 211(h)(4) for air quality reasons.
Additionally, CAA section 211(a) authorizes EPA to designate fuels and fuel additives and requires manufacturers
of such fuels and fuel additives to register them with EPA prior to introduction into commerce. EPA allowed for the
introduction of E15 into commerce in 2010 and 2011. 75 FR 68094 (November 4, 2010), 76 FR 4662 (January 26,
2011).
30 Chapter 6.5, "RFS Program: Standards for 2023-2025 and Other Changes: Regulatory Impact Analysis," EPA-
420-R-23-015, June 2023 ("RFS Set Rule RIA").
29
-------
Figure 3.B-1: Total Nationwide Annual Sales of E15
300,000,000
250,000,000
£ 200,000,000
_o
= 150,000,000
100,000,000
50,000,000
• BIP Reported
Data
• Extrapolated
2014 2016 2018 2020 2022 2024 2026 2028
Year
There are several different programs at the federal and state levels that have been
supporting the buildout of E15 retail outlets, which is influencing the majority of E15 sales
volume growth. Minnesota and Iowa both have regulations and incentives to promote the use of
E15 and provide annual E15 sales data, which allows for further analysis. Figures 3.B-2 and 3
show the data reported by these two states and, similar to the BIP data, we extrapolated future
El5 use in these states based on the data provided. When future years were extrapolated from the
data, a continued upward trend in El 5 volumes was projected, primarily as a result of increased
El5 retail outlet growth.
Figure 3.B-2: Iowa Annual Reported E15 Sales and EPA Extrapolations
250
200
c
o
"ro
c
o
in
150
100
50
0
• • •
. «
i BIP Data
> Extrapolated Low
> Extrapolated Mid
i Extrapolated High
2014 2016 2018 2020 2022 2024 2026 2028
Year
30
-------
Figure 3.B-3: Minnesota Annual Reported E15 Sales and EPA Extrapolations
-•—•
2018
2020
2022 2024
Year
2026
> BIP Data
i Extrapolated Low
> Extrapolated Mid
Extrapolated High
2028
A series of EPA actions from 2019-2023 has allowed parties blending El 5 to utilize the
same blendstock used to blend E10 during the summer; thus, any impact of removing the 1-psi
waiver in the petitioning states should already be reflected in the baseline El5 sales volume
data.31 Given existing barriers, and our understanding of the factors that contribute to increased
growth, we believe that removal of the 1-psi waiver will have relatively little impact on the
growth of El 5 sales volumes in the future.
Finally, even if an increase in El 5 sales volumes were to occur, the overall impact on
gasoline supply would be muted due to the fact that each gallon of El 5 displaces a gallon of
E10—with just a 5% impact on gasoline volume—and that El 5 sales are only a very small
fraction of overall gasoline sales. We estimated that the impact of the 1-psi waiver on E15 sales
is approximately 16% of annual per-station sales of E15.32 At current E15 sales penetration, the
resulting change in ethanol sales volume associated with a 16% change in El5 sales would
equate to a change of less than 0.1% of summertime gasoline sales, and only 1-2% of the
gasoline supply reduction projected earlier in this section.
C. Accommodating a Reduced Gasoline Supply in 2025
As estimated in the previous section and summarized in Tables 3.B-7 and 10, 30-80 kbpd
of light gasoline material will be removed from the gasoline pool as the fuel production and
distribution system adjusts to comply with removal of the 1-psi waiver in the petitioning states.
This light gasoline material is less energy dense than gasoline; therefore, the gasoline-equivalent
volume of the removed light gasoline material will be less than its actual volume by -15%.
31 If the 1-psi waiver is not extended to E15 during the summer of 2024, E15 sales in the petitioning states could
decrease slightly for 2024 and then be restored for 2025. Our estimate of the impact of the 1-psi waiver on E15 sales
is approximately 16% of annual per-station sales of E15. Chapter 1.7.2, "RFS Program: RFS Annual Rules:
Regulatory Impact Analysis," EPA-420-R-22-008, June 2022 ("2020-2022 RFS Rule RIA").
32 Chapter 1.7, 2020-2022 RFS Rule RIA.
31
-------
There are several ways that, in combination, this reduced gasoline supply can be accommodated
in 2025.
First, downstream terminals that distribute low-RVP gasoline to non-petitioning states
due to the inflexibilities in the fuel distribution system could blend butane into low-RVP gasoline
to bring the RVP up to 10.0 psi. This option is only available to downstream terminals that have
the capability to receive and blend butane.
Second, refiners could choose to delay a full or partial maintenance shutdowns of
refineries that supply gasoline to PADD 2 and instead continue to supply PADD 2 gasoline
markets in anticipation of a supply shortfall due to the removal of the 1-psi waiver. This choice
would depend on the severity of the maintenance planned for the shutdown.
Third, any spare gasoline supply in the PADDs 1, 3, and 4 could be distributed to PADD
2 to make up for its gasoline supply shortfall. In particular, additional volumes could be moved
up from Gulf Coast refineries via pipelines, barge, and rail. In the past, a considerable volume of
gasoline flowed from PADD 3 into PADD 2. However, with the higher arbitrage received for
export to PADD 1 or overseas following the war in Ukraine, excess gasoline production from
PADD 3 has flowed there instead. Higher gasoline prices resulting from implementation of the
removal of the 1-psi waiver could cause considerable volumes of gasoline from PADD 3 to flow
once again to PADD 2. Furthermore, pipelines can adjust the products shipped in other ways.
Based on conversations with pipeline operators, however, we are aware that due to increasing
Canadian tar sands crude oil production—which require NGLs as diluent to permit shipping the
tar sands—the pipeline capacity out of the Gulf Coast into PADD 2 has been increasingly taken
up by these northward shipments of NGLs, reducing the capacity to move gasoline into PADD 2.
Refiners that remove NGLs and LSR to produce low-RVP CBOB could send these removed
materials to Canada as diluent, which would free up pipeline space to allow more gasoline to be
shipped from the Gulf Coast to PADD 2.
Fourth, some or potentially all of the gasoline supply shortfall could also be covered by
the gasoline stored in inventory at the many gasoline storage tanks in the fuel distribution
system. We review the impact of the 2022 gasoline supply shortfalls in PADD 2 in the next
section, and it appears that they were primarily covered by PADD 2 gasoline inventories.
Finally, we also note that some of the gasoline supply shortfall could be made up through
a reduction in gasoline demand. EIA forecasts that gasoline demand will increase by 60 kbpd
from 2023 to 2024, and then subsequently decrease by 140 kbpd from 2024 to 2025.33
Furthermore, while gasoline demand is fairly inelastic with respect to gasoline prices, higher
gasoline prices resulting from implementing the removal of the 1-psi waiver would be expected
to result in a small decrease in gasoline demand in PADD 2.
I). Recent and Current Supply and Demand Balance of Gasoline Inventories
The impact of removing the 1-psi waiver on gasoline supply and the extent to which it
might result in insufficient supply needs to be understood in the context of the existing supply
33 EIA, Annual Energy Outlook (AEO) 2023, Table 11, https://www.eia.gov/outlooks/aeo.
32
-------
and demand balance of gasoline in the U.S. and more specifically PADD 2 where the petitioning
states are located. While the supply of gasoline is primarily from the refineries that produce the
gasoline, inventory of gasoline contained in the fuel distribution system can also be a source of
supply. The fuel distribution system usually contains more gasoline inventories than the
minimum volume required for the system to be functional, and this excess volume can be utilized
as a supplemental source of gasoline supply. Changes in refinery gasoline supply and the
drawing down of gasoline inventories in the fuel distribution system can both have price impacts.
This review of gasoline refinery supply and inventory is particularly warranted due to: (1) The
supply issues that occurred and affected many industries across the U.S. economy coming out of
the COVID-19 pandemic and subsequent war in Ukraine; (2) The recent closure of two large
Gulf Coast refineries and conversion of several other U.S. refineries to produce renewable diesel;
and (3) Refinery outages in 2022 that impacted PADD 2 gasoline inventories.
To assess the supply and demand balance, we reviewed the total supply volume of the
primary refined products of refineries, including gasoline and gasoline blendstocks, distillate,
and jet fuel. While each of these refined products are distinctly different from each other,
refineries can move some refinery blendstocks between different product pools and change
refinery unit operations. For example, heavy naphtha material can be blended into either gasoline
or distillate, and hydrocrackers can be adjusted to produce different amounts of refinery gasoline
or distillate blendstocks. During much of 2022, there was a shortfall in distillate supply that led
to low inventories and elevated distillate prices. Therefore, most refiners were operating their
refineries to produce a maximum amount of distillate during this time. The supply of all refined
products is estimated by adding together monthly "product supplied" volumes of gasoline
blendstock, finished gasoline, distillate, and jet fuel, as depicted in Figure 3.D-1.
Figure 3.D-1: U.S. Fuel Supplied and Fuel Inventories
£ 350,000
Month, Year
34 Data source: EI A, Petroleum & Other Liquids, Product Supplied,
https://www.eia.gov/dnav/pet/pet cons psup dc nus mbbl in.lUin: EIA, Petroleum & Other Liquids, Stocks by
Type, https://www.eia.gov/dnav/pet/pet stoc typ d nus SAE mbbl m.htm.
33
-------
Refined product supplied averaged 461 million barrels (15.4 million bpd) in 2019, the
year prior to the COVID-19 pandemic. From March to December of 2020, when the COVID-19
pandemic emerged, refined product supply decreased to an average of 383 million barrels (12.8
million bpd), a 17% reduction compared to 2019. Refined product supply increased to 445
million barrels (14.6 million bpd) over the second half of 2021 through the beginning of 2023.
While refined product supply has increased substantially since the COVID-19 pandemic, as of
April 2023 it is still 5% lower than the average supply in 2019.
It is important to understand the change in refinery disposition as this is the source of
reduced refined product supply to the U.S. There were a number of refinery closures and
conversions, most due to impacts from the COVID-19 pandemic. Shell closed its Convent,
Louisiana refinery at the end of 2020, which had a crude oil refining capacity of 211 kbpd.
Increased product demand, along with the closure of the Shell refinery, caused refined product
stocks to decrease by an average of 240 kbpd in 2021. At the end of 2021, Phillips 66 decided to
shut down its Belle Chasse, Louisiana refinery, which had a crude oil refining capacity of 255
kbpd. Additionally, several refiners have opted to fully or partially convert their petroleum
refineries to produce renewable diesel in recent years, including full conversions of the Marathon
refinery in Dickinson, North Dakota, and the Holly Frontier refineries in Artesia, New Mexico
and Cheyenne, Wyoming, and a partial conversion of the CVR refinery in Wynnewood,
Oklahoma. Although only the Dickinson and Wynnewood refineries are in PADD 2, the other
refineries can indirectly affect the volume of gasoline available to PADD 2. These refineries had
a combined crude oil refining capacity of approximately 200 kbpd. While still in operation and
producing a distillate blendstock, they no longer contribute to gasoline production.
From Figure 3.D-1, it is clear that the reduced demand for refined products during the
COVID-19 pandemic caused an increase in refined product inventories. Refined product
inventories for the entire U.S. averaged 400 million barrels for most of 2019, but then increased
to 450 million barrels in 2020 after the pandemic hit and caused a reduction in fuel demand.
After the pandemic and refinery closures and conversions, total fuel inventories in 2022 were 50
million barrels lower compared to 2019. In 2019, wintertime fuels inventories were 450 million
barrels, and the following summer showed fuels inventories of 400 million barrels. In contrast,
wintertime fuels inventories in 2022 were 420 million barrels, and the subsequent summer were
370 million barrels. Fuels inventories decreased further to 400 million barrels in the winter of
2023. It appears that despite the lower fuel demand in 2021 and 2022 relative to 2019, national
fuel inventories seem to be declining, perhaps due to the aforementioned refinery closures and
conversions.
Since removal of the 1-psi waiver will directly impact the gasoline supply, it is important
to understand the inventory of gasoline. Figure 3.D-2 compares U.S. national gasoline
inventories for 2022 and 2023 to the five-year historical range of gasoline inventories.
34
-------
Figure 3.D-2: U.S. National Gasoline Inventories
Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
Figure 3.D-2 shows that U.S. national gasoline inventories have been below their typical
historical range since early 2022, when the war in Ukraine began. With less gasoline in
inventory, there would be less gasoline available in the fuel distribution system to cover reduced
supply. However, it is most important to understand the inventory of gasoline in the region
where the petitioning states are located (i.e., PADD 2).
The inventory of gasoline specifically in PADD 2 is depicted in Figure 3.D-3. This data
shows that the impacts of the pandemic on PADD 2 were more muted, as gasoline inventories in
2020 and 2021 were relatively stable, with no large increase in 2020 and no significant decrease
in 2021. However, PADD 2 gasoline inventories suffered a dramatic decline in 2022, decreasing
at a relative high rate of 100 kbpd during the early part of the year. Since May 2022, PADD 2
gasoline inventories have been below their historical minimums and did not recover over the
winter of 2022/2023. As a result, PADD 2 gasoline inventories at the start of 2023 were 10
million barrels lower than typical levels for that time of year and remained below the five-year
historical range of gasoline inventories throughout the summer of 2023.
35
-------
Figure 3.D-3: PADD 2 Gasoline Inventories35
Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
Figure 3.D-4 depicts the factors that impacted gasoline supply to PADD 2 in explain why
PADD 2 gasoline inventories began falling dramatically in April 2022 and generally remained
below the five-year historical range of gasoline inventories for most of 2022 and the first half of
2023.
Figure 3.D-4: Factors Impacting PADD 2 Gasoline Inventories
100 kbpd less gasoline from PADD 3
r 1 1
BP whiting refinery
shutdown ~200 kbpd
BP-Husky Toledo Refinery]
Shutdown ~70 kbpd
2017 - 2021 Inventory Range
5 Year Average
—2022
5% lower PADD 2
refinery utilization
~200 kbpd - Keystone
pipeline shutdown?
In response to the war in Ukraine that began in March 2022, gasoline supply to the East
Coast was impacted as fewer imports from Europe were being received. As a result, PADD 3
(Gulf Coast) refineries routed gasoline normally destined for PADD 2 to the East Coast to make
35 Data source: EIA, Petroleum & Other Liquids, Weekly Stocks,
https://www.eia.gov/dnav/pet/pet stoc wstk dcu r20 w.htm.
36
-------
up for the gasoline loss there. Over the three-month period (April to June 2022), an average of
100 kbpd of gasoline supply from the Gulf Coast was not available to PADD 2. In July 2022, the
gasoline movements from the Gulf Coast returned to their normal historical levels and gasoline
inventories recovered somewhat relative to historical levels. On August 24, 2022, the BP refinery
in Whiting, Indiana, which is the largest refinery in PADD 2, experienced an electrical fire that
shut down a significant portion of the refinery. The refinery was reportedly starting up several
weeks later, and an additional report stated that some of the refinery units were still not
operational at the end of September.36 The BP Whiting refinery's maximum crude oil throughput
capacity is 440 kbpd.37 Assuming that the refinery's gasoline production is about half of its crude
oil throughput volume, its gasoline production likely ranges between 180-220 kbpd when it is
operating at full capacity.38 Thus, this short-term shutdown of the BP Whiting refinery caused
PADD 2 gasoline inventories to decrease once again.
PADD 2 gasoline inventories began to recover at the end of September 2022, likely due
to the restart of the BP Whiting refinery, along with the end of the summer RVP season which
allows the blending of butane. However, on September 20, 2022, the BP-Husky refinery (now
owned by Cenovus) in Toledo, Ohio experienced an explosion that forced a shutdown of the
refinery. The refinery restarted in June 2023, significantly after the date when refiners begin
producing summer gasoline in 2023. The Cenovus Toledo refinery's maximum crude oil
throughput capacity is 160 kbpd and its gasoline production likely ranges between 60-80 kbpd
when it is operating at full capacity. Cenovus also owns a refinery in Superior, Wisconsin that
has been shut down since an explosion in 2018. Cenovus has been rebuilding the refinery for
several years and is still in the process of starting it back up.39 Although small, this refinery's
gasoline production likely ranges between 17-25 kbpd when it is operating at full capacity,
providing gasoline to the difficult-to-supply region of northern Wisconsin.
Despite the Cenovus Toledo refinery shutdown, gasoline inventories rose to a point
above the 5-year minimum at the end of October 2022 and appeared to be on a path to recover
over the winter. However, in early December 2022, PADD 2 refinery utilization decreased by
5%. The reason for the decline is not clear, although it may have been due to a shutdown of the
Keystone pipeline, which brings Canadian tar sands crude oil to the lower Midwest and Gulf
Coast refineries. Whatever the cause, a 5% reduction in PADD 2 refinery utilization is estimated
to have reduced gasoline supply by around 200 kbpd. This decreased PADD 2 gasoline
inventories during the winter when gasoline inventories were expected to increase by 10 million
barrels.
36 Reuters, "BP bringing Whiting, Indiana, refinery back to normal operation -company," September 2, 2022,
https://www.reuters.com/business/energy/bp-bringing-whiting-indiana-refinerv-back-normal-operation-companv-
2022-09-02.
37 EIA, Refinery Capacity Report, Table 3, https://www.eia.gov/petroleum/refinervcapacitv/table3.pdf.
38 Gasoline production at U.S. refineries averaged 47% of the volume of all the crude oil they refined. EIA,
Petroleum & Other Liquids, Refinery Yield, https://www.eia.gov/dnav/pet/pet pnp pet dc nus pet a.htm.
39 Wisconsin Public Radio, "Superior refinery still hasn't resumed full operations," July 31, 2023,
https://www.wpr.org/superior-refinerv-cenovus-energv-huskv-gas-full-operations.
37
-------
E. Oil Industry Estimated Impact on Supply
The Baker and O'Brien Study included a summary of their analysis of the impact of the
removal of the 1-psi waiver on gasoline supply.40 As part of the study, Baker and O'Brien
modeled each individual refinery that would likely be impacted by the removal of the 1-psi
waiver. However, it appears that the study's findings on gasoline supply impacts were mostly
informed by a survey that Baker and O'Brien conducted of refiners and fuel distributors on their
perceived impact of the removal of the 1-psi waiver on gasoline and distillate production and
distribution. The Baker and O'Brien Study listed the following major findings of their survey:
• Many of the refineries currently operate near a physical or economic limit for removing
light ends from the summertime gasoline pool.
• Some refiners may need to:
o Reduce crude runs (i.e., reducing production of gasoline and all other products) in
order to control the amount of high RVP gasoline components blended in the
gasoline pool—estimated to be 30% of refineries,
o Augment a mode of butane or LSR sales (e.g., truck, rail, or pipeline deliveries),
o Reduce high octane gasoline production or purchase high octane blendstocks.
o Invest in fractionation, piping, and storage to enable more efficient production of
low-RVP CBOB—estimated to be $50-$75 million per refinery.
• Capital projects take 18-24 months to implement after a final go-ahead decision, but such
decisions will likely be delayed until there is clarity regarding a possible nationwide
extension of the 1-psi waiver to El 5.
• Due to logistical constraints, some 10.0 psi RVP gasoline areas will only be supplied with
low-RVP gasoline, especially in the first two years.
• Terminals that serve 10.0 psi RVP gasoline markets will not be available for low-RVP
gasoline storage.
The Baker and O'Brien Study divided refineries into 4 different groups and estimated the
gasoline and distillate supply impact in the petitioning states for each group, as shown in Table
3.E-1:
Group A - Refineries within the petitioning states
Group B - Refineries within adjacent non-petitioning states
Group C - Refineries in Oklahoma
Group D - Refineries in the Gulf Coast
40 The Baker and O'Brien Study was conducted based on the removal of the 1-psi waiver in seven states—before
Missouri submitted their petition.
38
-------
Table 3.E-1: E
taker and O'Brien Study Estimated Gasoline and Distillate Supply Impacts
Group
Impact on Gasoline
(%)
Reduced Gasoline Volume
(kbpd)
Reduced Distillate
Volume (kbpd)
A
6.2
63-72
12-20
B
4.7
21-41
6-10
C
5.0
4-12
2-3
Total
88-125
20-33
The Baker and O'Brien Study assumed that the supply shortfalls in Table 3.E-1 would be
made up by additional gasoline and distillate supplied from refineries in Group D (i.e., Gulf
Coast refineries), while at the same time highlighting some of the associated challenges and
changes to the distribution system to allow that to happen. Additional information from the study
shows that approximately two-thirds of the estimated supply impact is from refineries producing
less gasoline, presumably by removing light hydrocarbons. The other one-third of estimated
supply impacts is due to reduced crude oil throughput, impacting both gasoline and distillate
production.
While the Baker and O'Brien Study provided a summary of their survey and made other
statements relevant to their supply analysis, there was insufficient information provided to
explain how the supply impacts in Table 3.E-1 were estimated. For example, the study does not
explain how the percent impact on gasoline values in Table 3.E-1 were calculated. Also, while
some refiners apparently reported that they may need to reduce crude oil throughput volumes, the
Baker and O'Brien Study conservatively assumed that these refineries would in fact reduce their
crude oil throughput. This conservative assumption adopted by the Baker and O'Brien Study
may explain why it estimated a higher supply impact compared to EPA's estimate.
F. 2022ICF Report Analyzing Supply
In a 2022 report conducted for the Renewable Fuels Association (RFA), ICF analyzed the
fuel supply in the petitioning states in 2023.41 In general, the ICF Report assessed similar issues
as EPA, including the supply and demand balance, gasoline production, butane logistics, refinery
changes, refinery and pipeline logistics, and ability to respond to disruptions. The report
analyzed the supply impacts for the original eight states that petitioned to remove the 1-psi
waiver. Thus, it includes Kansas and North Dakota, which have subsequently rescinded their
petitions to remove the 1-psi waiver, and it excludes Ohio and Missouri, which petitioned to
remove the 1-psi waiver in June and December 2022, respectively. Consequently, some
reassessment is warranted. Additionally, while the report assessed refinery production and
market demand, it did not assess gasoline inventories in the region. This factor contributes to the
region's ability to withstand fluctuations in the supply of gasoline.
The report acknowledged some potential changes that may have been unable to be
implemented prior to the summer of 2023, including refinery processing changes to reduce RVP,
addition of infrastructure, and changes at pipelines, terminals, and refineries for additional
41 ICF, "Impact of Potential 8-State RVP Waiver Exclusion on Midwest Gasoline Markets," prepared for the RFA,
September 2022 ("ICF Report").
39
-------
segregation tankage. The report did not assess whether such changes would be necessary prior to
an implementation date in 2023.42
The report further acknowledged that maintaining supply of gasoline to the region should
be "manageable," but also noted a potential difficulty responding to unexpected outages of
gasoline supply to the region, as only low-RVP gasoline can be used to in the petitioning states.43
It also acknowledged that "states facing gasoline shortages could request a temporary waiver to
allow temporary reinstatement of the 1-psi RVP waiver for El 0 or a general RVP waiver"
without addressing the timing or implications of such actions.44
42 The ICF Report was conducted prior to the proposed rule with a focus on implementation in 2023 and therefore
did not speak specifically to 2024 or 2025. Nevertheless, the information in the report is also relevant for assessing
implementation in 2024 or 2025.
43 ICF Report at 22.
44 Id. at 24.
40
-------
4. Cost of Removing the 1-psi Waiver in the Petitioning States
Potentially every part of the fuel distribution system could incur some cost when
providing low-RVP gasoline to the petitioning states. The cost incurred by refineries to produce
low-RVP gasoline is the largest and most predictable portion of the total cost. The cost to
refineries is driven by the opportunity cost of selling the light hydrocarbons removed from
gasoline (a high value product) and selling them in much lower-priced hydrocarbon markets.
Because the fuel distribution system is complicated, it is much more difficult to estimate the
ultimate cost of compliance for the fuel distribution system, although we provide some cost
information that provides some context for estimating the ultimate fuel distribution system cost.
We reference several different cost studies in this section. First, two separate refinery
modeling studies conducted by MathPro examined the long-term refining cost for removing the
1-psi waiver nationwide—one conducted for RFA45 and another conducted for the International
Council on Clean Transportation (ICCT)46 (collectively the "MathPro Studies"). We also
reference the Baker and O'Brien Study in this section, which included cost estimates specifically
for removal of the 1-psi waiver in the petitioning states. Notably, the Baker and O'Brien Study
compared two different approaches for estimating the cost of removing the 1-psi waiver in the
petitioning states: (1) Baker and O'Brien's proprietary "PRISM" refinery analysis software,
which is a "typical RVP cost model" that uses and assumes "'ideal' operation and 'average'
properties" for a refinery;47 and (2) An "extended cost model" that provides a "range of costs
based on each refinery's specific capabilities plus any infrastructure and logistics costs
associated with bringing Low RVP CBOB from each refinery to the affected states."48 These
models are discussed in more detail below.
A. Cost Studies
There are several factors that contribute to the cost of producing low-RVP gasoline. The
largest portion of the cost is the lost revenue associated with having to sell the removed butane at
market prices for butane that are much lower than the high-value gasoline it would otherwise be
blended into.49 There are also additional capital and operating costs that will need to be recouped
over time.
We first discuss the two MathPro refinery modeling studies. The portion of the studies
summarized here was performed for the entirety of PADD 2, not just the eight petitioning states.
Nevertheless, they provide a reasonable estimate of the long-term per-gallon cost associated with
the removal of the 1-psi waiver in the petitioning states. Table 4.A-1 summarizes the key cost
information from the two studies.
45 MathPro, "Assessment of a 1-psi reduction in the RVP of Conventional Gasoline Blendstock (CBOB) in the
Summer Gasoline Season," prepared for RFA, December 1, 2021.
46 MathPro, "Refining Economics of a National Low Sulfur, Low RVP Gasoline Standard," prepared for the
International Council for Clean Transportation (ICCT), October 25, 2011.
47 Baker and O'Brien Study at 9.
48 Id. at 8.
49 Butane is typically priced at less than half of the wholesale price of gasoline.
41
-------
Table 4.A-1: MathPro Estimates of Long-Term Refining Cost for Removal of the 1-psi
Waiver in PADD 2
RFA Study
(2021)
ICCT Study
(2011)
Crude Oil Price ($/bbl)
57-94
90
Capital Cost ($MM)
214-261
250
Refining Cost (0/gal)
2.2-2.6
2.4
Fuel Economy Savings (0/gal)
(0.7M1.0)
(0.2)
Net Consumer Cost (0/gal)
1.5-1.6
2.2
Total Annual Cost ($MM)
258-309
261
Based on the MathPro Studies, the estimated refining cost for reducing the RVP of
gasoline by 1.0 psi ranged from 2.2-2.60 per gallon, with the cost of RVP control increasing at
higher crude oil prices. However, the net cost to consumers is slightly lower due to an expected
increase in the energy density of low-RVP gasoline, which allows vehicles to travel further on
each gallon of gasoline. The magnitude of the fuel economy effect varies by the study case—
ranging from 0.2-1.00 per gallon—and reduces the net cost of low-RVP gasoline to 1.5-2.20 per
gallon.
Table 4.A-1 also summarizes the MathPro Studies' estimates of the capital costs
investments that would be required to replace the volume and octane content of the removed
butane for all of PADD 2. Since this action affects only a portion of PADD 2, the capital costs
for refineries as a result of the removal of the 1-psi waiver in the petitioning states is expected to
be half of the MathPro estimates. The MathPro models, however, tend to underestimate a portion
of the capital costs because they do not consider the capital costs for revamping or adding
debutanizers, gas plants, or butane handling and storage. If the available excess volumetric and
octane production capacity among all PADD 2 refineries, as well as the PADD 3 refineries that
supply PADD 2, is sufficient to make up for this removed butane, then the refining sector may
not need to invest any capital dollars to cover the octane loss due to the removal of butane and
may allow for a more rapid transition to low-RVP gasoline in the petitioning states. However, it
may also mean that some refineries currently providing gasoline to the petitioning states would
cease to do so or reduce their supply, while others with fewer hurdles would enter or expand into
these markets. Of course, these changes would complicate the supply of gasoline to the
petitioning states.
In addition to assessing the gasoline supply impact, the Baker and O'Brien Study also
estimated potential cost impacts. Baker and O'Brien surveyed refiners and pipeline operators in
and around the petitioning states and were able to estimate the cost to produce low-RVP gasoline
in both the near- and long-term using these survey results. Similar to the MathPro Studies, the
Baker and O'Brien Study examined the cost of removing butane from gasoline, but also analyzed
the cost of removing pentanes and additional effects of low-RVP gasoline production. The
PRISM model from the Baker and O'Brien Study (i.e., the "typical RVP cost model") estimated
that refiners' production cost would be about 30 per gallon based on solely butane removal from
gasoline production.50 This is slightly higher but still close to the estimated refining cost from the
50 Baker and O'Brien Study at 9.
42
-------
MathPro Studies. The higher cost estimate from Baker and O'Brien compared to MathPro is
likely due to Baker and O'Brien including some additional costs (e.g., additional production,
blendstock purchases, and distribution capabilities).
Table 4.A-2 shows the estimated refining costs with each model from the Baker and
O'Brien Study. These estimates depend upon the implementation date of the removal of the 1-psi
waiver and location of fuel supply. Implementation in 2024 would likely lead to price impacts on
the higher end of their cost ranges, whereas implementation in 2025 or later could allow for
lower price impacts. The cost also varied based on proximity to the petitioning states. The Baker
and O'Brien Study estimated the projected cost not only to the petitioning states, but also to
adjacent non-petitioning states due to the production and distribution limitations described in
Section 3. The Baker and O'Brien Study estimated that adjacent non-petitioning states are likely
to see similar price increases as the petitioning states. For example, the Baker and O'Brien
Study's extended cost model estimates that if removal of the 1-psi waiver was implemented in
2024, petitioning and adjacent non-petitioning states could have seen a cost increase of 3-120
per gallon. States beyond the petitioning and adjacent non-petitioning states were also analyzed
and projected to have a much lower, if any, price impact from the low-RVP gasoline production.
Table 4.A-2: Baker and O'Brien Study Estimates of Refining Cost
Refining Cost
(Wgal)
Model
Timeframe
Region
Low
High
Typical RVP
2022
Petitioning
2
10
Cost Model
Adjacent
2
3
2023-2024
Petitioning
3
12
Extended Cost
Adjacent
3
12
Model
2025+
Petitioning
3
11
Adjacent
3
8
As described in the Baker and O'Brien Study, the timeframe for implementation of
removal of the 1-psi waiver effects potential price increases. In the near term (e.g., 2023 and
2024), any investment requiring a permit would require more lead time and cause a need for
other solutions to produce low-RVP gasoline. This includes storage tank alternatives (e.g.,
railcars), adjusting crude slate purchases for RVP control, or using tank trucks to ship products
further to markets in need, as discussed in Section 3. A longer-term implementation in 2025 or
later would allow for more changes and preparation for refineries and pipelines. Although
projects such as storage tank construction would likely take longer than a 2025 implementation
date, other projects could likely be completed (e.g., piping changes, pump installation, or
operational unit debottlenecking). This timeframe would also allow for more railcars to be
constructed and made available for use, which would aid in the sale and storage of excess butane.
Baker and O'Brien also studied the historical wholesale price difference between RBOB
and CBOB in Chicago as an additional evaluation of the cost of RVP control.51 While price data
can capture market effects that can bias the costs, wholesale price data are essentially refinery
51 Id. at 46.
43
-------
gate prices and generally are assumed to reflect marginal production costs, not other market
effects. Baker and O'Brien found that the wholesale price difference between RBOB and CBOB
can be estimated based on a cost estimate of 87% butane removal and 13% NGLs removal. Since
RFG must meet a 7.4 psi RVP specification and does not receive the 1-psi waiver, RBOB is
estimated to be 2.6 psi lower in RVP than CBOB; thus, the RBOB/CBOB price difference is
divided by 2.6 to estimate the refining cost per 1-psi reduction in RVP. Table 4.A-3 summarizes
the Baker and O'Brien Study's estimated cost of RVP control based on this RBOB and CBOB
wholesale price data.
Table 4.A-3: Baker and O'Brien Study Estimated Cost of RVP Control
Month
RVP
(^/gal per 1
Cost
psi decrease)
2019
2022
April
6.9
5.5
May
10
10.5
June
10.6
10
July
8.4
11.4
August
4.5
10.9
Average
8.1
9.7
The Baker and O'Brien Study found that this cost estimate based on historical RBOB and
CBOB wholesale price data is in line with their near-term costs for low-RVP gasoline. However,
estimating the cost of a 1-psi reduction in RVP based on the cost of complying with the RFG
program risks overestimating the cost because as the RVP standard increases in stringency, the
cost of compliance increases. In other words, it costs more per 1-psi reduction in RVP to produce
RFG than it does to produce low-RVP gasoline. Another important observation is that crude oil
prices were higher in 2022 due to the political uncertainty that occurred that year as a result of
the war in Ukraine; thus, the higher average cost in 2022 likely represents RVP costs at higher
crude oil prices, while the lower average cost in 2019 likely represents RVP costs at more
moderate crude oil prices.
B. Distribution Cost
The Baker and O'Brien Study estimated distribution costs associated with the removal of
the 1-psi waiver, although these costs were not listed separately from other costs. The study
estimated the installation cost for a gasoline storage tank to be $7-10 million.52 Assuming that
the storage tank has a 50,000-barrel capacity and amortizing the capital cost over the gasoline
stored in the tank assuming a 3-day storage time, installing this storage tank would add 0.30 per
gallon to the cost of distributing gasoline. If more than one storage tank needs to be installed in
the gasoline distribution chain from refinery to downstream terminal, the total distribution cost
would be higher.
Many terminals that currently provide gasoline to both petitioning and non-petitioning
states will likely not be able to continue to do so due to the lack of available tankage. This will
52 Id. at 31.
44
-------
likely change how gasoline is distributed to these gasoline markets, increasing distribution costs.
Distributing gasoline from product terminals to retail outlets is estimated to typically cost 60 per
gallon, although this typical cost is an average over a large range due to varying transportation
distances.53 Presumably, the gasoline shipped to some of these markets would need to be shipped
from further away, which would increase the gasoline distribution cost for these markets and
demand more from the fleet of tank trucks and drivers, exacerbating the existing truck driver
shortage.54
There would likely be other costs associated with distributing an additional type of
gasoline. Since conventional gasoline consumed in PADD 2 would largely be divided between
10.0 psi and low-RVP gasoline, gasoline batch sizes would be smaller in many cases, increasing
the cost of distributing both gasoline types. Furthermore, if a refinery serving PADD 2 only
produces one of the two gasoline types, it could mean that another refinery would have to
produce a portion of the gasoline previously served by the first refinery, and the gasoline sold by
both of these refineries would likely need to be moved further distances than before, increasing
the distribution cost for both refineries' gasoline. Similarly, if a downstream terminal decided to
carry only one of the two gasoline types, it would have to sell solely into either petitioning or
non-petitioning states. This in turn would likely mean that the trucks that distribute gasoline from
that terminal would have to travel further distance than they currently do.
C. Discussion of Costs
MathPro and Baker and O'Brien both estimated the cost of removing the 1-psi waiver
using linear program refinery models. The Baker and O'Brien Study's refinery modeling costs,
using an approach similar to that used by MathPro, were only somewhat higher than those
presented in the MathPro Studies. However, after including a more detailed refinery-by-refinery
analysis using their extended cost model and incorporating a wholesale price analysis of Chicago
RFG, the Baker and O'Brien Study concluded that removing the 1-psi waiver would cost
considerably more than the cost estimate from their typical RVP cost model.
After reviewing the refinery modeling analyses in the MathPro Studies, we identified
several reasons why their cost studies likely underestimate the cost of complying with the
removal of the 1-psi waiver in the petitioning states:
(1) By using an aggregated model, the MathPro Studies did not identify or account for
the higher costs incurred by some refineries, such as those refining heavier crude oils,
which would have to produce low-RVP gasoline by removing less-volatile
hydrocarbons (e.g., NGLs).
(2) The MathPro Studies did not account for additional costs due to the implementation
of the removal of the 1-psi waiver prior to the installation of capital projects needed
53 National Association of Convenience Stores, "Who Makes Money Selling Gas?" November 12, 2021,
https://www.convenience.org/Media/conveniencecorner/Who-Makes-Monev-Selling-Gas.
54 CNBC, "Why driving big rig trucks is a job fewer Americans dream about doing," July 5, 2022,
https://www.cnbc.eom/2022/07/05/whv-driving-big-rig-trucks-isnt-a-iob-americans-want-to-do-anvmore.html.
45
-------
to optimize production and distribution of low-RVP gasoline by refiners and fuel
distributors.
(3) In assuming that that the entire conventional gasoline pool would be converted to
low-RVP gasoline, the Mathpro Studies did not assess or quantify the additional costs
incurred by limitations in the fuel distribution system to distribute low-RVP gasoline
when it was not designed for widespread distribution of an additional gasoline type.
This review is not to criticize MathPro's refinery model, nor their ability to model low-
RVP programs; they simply modeled costs under the optimal circumstances of a nationwide
removal of the 1-psi waiver, rather than in just the petitioning states.
The Baker and O'Brien Study, on the other hand, specifically modeled costs of the
removal of the 1-psi waiver in the petitioning states for both short- and long-term compliance
cases based at least partially on a survey conducted of refiners and fuel distributors. Furthermore,
the Baker and O'Brien Study analyzed the compliance costs of refineries on an individual basis.
For these reasons, the Baker and O'Brien Study overcame the limitations inherent in the
MathPro Studies and better estimates the costs of the removal of the 1-psi waiver. However, we
are concerned that the Baker and O'Brien Study may have overestimated these costs for several
reasons:
(1) After learning that some refineries were concerned that they may need to reduce their
crude oil throughput to produce low-RVP gasoline, Baker and O'Brien conservatively
assumed that these refineries would in fact reduce their crude oil throughput.
However, when the survey was conducted, many refineries likely had not completed
their detailed review of how they could or would produce low-RVP gasoline.
(2) If the Baker and O'Brien Study did model low-RVP gasoline being sold in non-
petitioning states—which they indicated would need to occur due to limitations in the
fuel distribution system—it did not mention, nor did it likely model, butane blending
at downstream terminals in non-petitioning states to bring the RVP of the gasoline up
to 10.0 psi RVP, which would reduce the cost of compliance.
(3) The Baker and O'Brien Study appeared to at least partially rely on a review of
Chicago RFG price data to represent the cost of removing the 1-psi waiver. Although
the Baker and O'Brien Study compared this price data to cost data to help ensure that
the prices reasonably represented costs, we are still concerned that the price data
included market factors that would overestimate the cost of removing the 1-psi
waiver.
After reviewing the MathPro and Baker and O'Brien Studies, we conclude that the cost
of removing the 1-psi waiver is likely above that estimated by MathPro, but also that the Baker
and O'Brien cost estimates are likely too conservative. We conclude that the cost of removing
the 1-psi waiver is most likely somewhere in-between the estimates of these cost studies.
46
-------
Using the MathPro and Baker and O'Brien Studies as the potential range of per-gallon
costs and the low-, medium-, and high-impact scenarios discussed in Table 3.B-5, we were able
to bound the potential range of the total annual cost of the removal of the 1-psi waiver in the
petitioning states, as shown in Table 4.C-1. For the low-impact scenario, we used a cost of
20/gal, which is the average cost estimated by the two MathPro Studies. For the high-impact
scenario, we used a cost of 120/gal, which is the highest cost estimated by the Baker and O'Brien
Study. For the medium-impact scenario, we used a cost of 70/gal, which is the midpoint between
the low- and high-impact scenarios. For estimating total gasoline volume, we assumed that
refiners would produce low-RVP gasoline for 180 days on average.55
Table 4.C-1: Estimated Total Annual Cost of Removing the 1-psi Waiver in the Petitioning
States
Impact
Daily Volume
Total Volume
Per-Gallon
Total Annual Cost
Scenario
(kgpd)
(million gallons)3
Cost (0/gal)
(million $)
Low
45,900
8,260
2
170
Medium
55,100
9,910
7
690
High
67,100
12,070
12
1,450
a Total Volume = Daily Volume * 180 days 1,000
Given that we expect the actual per-gallon cost to be somewhere in-between the MathPro
and Baker and O'Brien Studies, we believe that the medium-impact scenario is most likely to
represent costs in the first year or two after the removal of the 1-psi waiver. After this time, we
expect that refiners will have completed the capital changes needed to optimize the production of
low-RVP gasoline by concentrating on removing butane instead of more-expensive less-volatile
hydrocarbons. We also expect that after the first several years, the fuel distribution system will
have made the necessary capital changes and optimized their operations to more cost effectively
distribute low-RVP gasoline to the petitioning states. As a result of these investments and
operational improvements, we expect compliance costs to decrease to a value closer to the low-
impact scenario.
55 While refiners distributing low-RVP CBOB through regional pipelines—with multiple onloading and offloading
points—are expected to produce low-RVP CBOB for about 200 days each year (March 1 to September 15), other
refiners will not need to produce low-RVP CBOB as long. For example, a refiner that owns its own pipeline
connecting its refinery to its retail market likely has a much shorter summer gasoline production period. A refiner
that sells low-RVP gasoline off its own terminal racks may be able to reduce its summer gasoline production period
to as little as 145 days, starting their production just prior to the start of the May 1 summer season.
47
-------
5.
Potential Price Impacts When Implementing the Removal of the 1-psi Waiver
There is a temporal element to how fuel prices reflect the changing costs associated with
producing that fuel. For example, while gasoline prices generally reflect production costs in the
competitive gasoline market, this may not be the case when removal of the 1-psi waver is first
implemented, as gasoline supply is reduced and not yet recovered. Given the challenges in
estimating market price impacts, we have not attempted to do so, as it would be difficult to
estimate which refiner will set the gasoline price and what that price might be. However, we
have analyzed the various factors that contribute to fuel prices and discuss them in this section
and provide some information on historical price impacts for comparison. While fuel prices are
primarily a function of crude oil prices and the fuel production and distribution costs, they are
also often a function of the relative balance of fuel supply and demand.
If refineries increase the volume of their crude oil runs to produce more low-RVP
gasoline to make up for the reduction in gasoline output, there could be an increase in crude oil
prices. If refineries increase their crude oil runs by 30-80 kbpd, which is the same volume of
gasoline material estimated to be removed from the gasoline pool due to the removal of the 1-psi
waiver, it would only impact 0.03-0.08% of the roughly 100 million bpd of world crude oil
demand. The short-term oil price elasticity of demand is estimated to be 0.1,56 and therefore for a
midpoint 0.05% impact on supply, the price impact is 0.5%. Assuming crude oil is priced at $80
per barrel, the price impact would be $0.40 per barrel, or a 10/gal increase in the price of crude
oil, which would affect all products refined from crude oil. Since there are many factors that
affect crude oil prices—including refinery startups or closures or changes in crude oil
production—the impact of the removal of the 1-psi waiver on crude oil prices is just one factor
among many impacting crude oil prices.
As discussed in Section 4, the removal of light hydrocarbons to produce low-RVP
gasoline will incur a cost of 2-120 per gallon. Refiners will seek to pass that cost onto
consumers, so the price of gasoline in the petitioning states will likely increase by at least that
amount.
When fuel supply falls short of demand and inventories drop, fuel prices typically rise.
As previously described, removal of the 1-psi waiver may cause a reduction in supply that could
increase fuel prices beyond the cost impacts discussed in Section 4.57 As summarized in Section
3, we estimate the gasoline supply impact of the removal of the 1-psi waiver to range from 30-80
kbpd. To understand the potential impact of this reduced gasoline supply on gasoline prices, we
conducted an analysis to estimate the price impacts associated with the PADD 2 supply shortfalls
in 2022, which we described in Figure 3.D-4.
56 Caldara, Dario, Michele Cavallo, and Matteo Iacoviello (2016). Oil Price Elasticities and Oil Price Fluctuations.
International Finance Discussion Papers 1173. http://dx.doi.org/10.17016/IFDP.2016.1173. A percent crude oil price
change multiplied by -0.1 estimates the percent change in supply. In the case for the percent change in supply,
divide the percent change in supply by -0.1 to estimate the percent change in the price of crude oil.
57 The cost estimates of the removal of the 1-psi waiver reflect a cost to society. The price impacts discussed here are
what consumers pay at the pump. To the extent that fuel prices exceed the average costs, it would result in a wealth
transfer from consumers to the refining industry.
48
-------
In the discussion that follows, we discuss historical price impacts in PADD 2. This
assessment requires an initial comparison of the price changes in PADD 2 vs. PADD 3 in order
to account for any broad market factors (e.g., crude oil prices) that are independent of supply-
induced price impacts.58 This then allowed the broad market factors to be backed out from the
overall PADD 2 price changes to better isolate just the supply-induced price impacts in PADD 2.
To validate this price comparison between PADD 2 and PADD 3, we need to verify that PADD
3 was not also experiencing a gasoline supply shortfall in 2022 that would complicate an
estimate of the price impacts caused by the supply issues in PADD 2. Figure 5-1 summarizes
PADD 3 gasoline inventories.
Figure 5-1: PADD 3 Gasoline Inventories59
100
95
90
"3
c 85
0
1
80
75
70
Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
As shown in Figure 5-1, PADD 3 gasoline inventories during 2022 were consistent with
that of previous years, which should make the gasoline prices in PADD 3 a valid baseline for
comparison with those of PADD 2.
Comparing the prices in PADD 2 and PADD 3 eliminates price impacts due to broad
market factors (e.g., the war in Ukraine) that affect gasoline prices in both PADDs, so we expect
this effect to be largely, if not completely, zeroed out. Figure 5-2 compares PADD 2 RFG and
conventional gasoline prices to PADD 3 conventional gasoline prices.
58 PADD 3 consists of Alabama, Arkansas, Louisiana, Mississippi, New Mexico, and Texas.
59 Data source: EIA, Petroleum & Other Liquids, Weekly Stocks,
https://www.eia.gov/dnav/pet/pet stoc wstk dcu r20 w.htm.
49
-------
Figure 5-2: Comparison of PADD 2 and PADD 3 Gasoline Prices60
120
"c"
0
ro
3s* 80
1
%
*
\
•
•
%•
L
•
1«
•
•
• • •
•
1"
9 s *
9
•
- •*
>
•
w.
v
<
•
ft
bJ
»~1
>\
• •
•
~
~
r
r
0
¦20 !
(
A
H
* •••>•• rv J*
~ > •( 2 • 4^ • 4. v
V .V*
• •
•• •*
V.V*
¦m
>•
v ?' *
• >
«
ro
~
vtVV
TltrnwiT
a;
v *
~
, ~
~
«
tnhtiifiiiiiifiMi
; a s s s 8 s a £ s" s a £ a s' a s§ s 3's
s!1!* 1sIII111s11! 111
• PADD 2 RFG vs PADD3
8 ^ S S S 8 ! S 5 S 3 S 5 S « 8 S 3 " 5 o ! S ?
limfHSIIipiSUimiH
Conventional Gasoline Prices ~ PADD 2 vs PADD 3 Convention;
8hgBfiSH3SBilBia
* 3 S 5 3 S S S S 5 s s s s s s
31J3|S2S|fi!0!|
il Gasoline Prices
The first thing that stands out in Figure 5-2 is the very large RFG and conventional
gasoline price spike in 2022, which likely reflects the price effect due to the gasoline supply
shortfall. However, there is a fairly consistent price difference in previous years that is unrelated
to the supply shortfall in 2022. There are several differences between PADD 2 and PADD 3 that
can account for this difference, which include: (1) Differences in crude oil prices that vary by
PADD; (2) Differences in gasoline production and distribution costs; and (3) Different state tax
rates. Not accounting for this normal price difference between the two PADDs would bias the
supply shortfall price analysis. Therefore, we evaluated the price difference between PADD 2
and PADD 3 prior to 2022 to determine how to estimate this price difference. The price
difference between the two PADDs for 2021 was lower than that for 2019 and 2020, but higher
than that before 2019. Since the price differences in 2021 were more typical of those in previous
years, we chose 2021 as the comparison year.
The gasoline price increase in PADD 2 due to the gasoline supply shortfall in 2022 was
estimated in two steps. First, the price difference between PADD 2 and PADD 3 was estimated
for 2022. Then the price difference was also estimated for 2021 and subsequently subtracted
from the 2022 price difference. The price differences were estimated using weekly price data for
both PADD 2 RFG and conventional gasoline in comparison to conventional gasoline in PADD
3.61
The results of the estimated price impacts attributed to the supply shortfall in PADD 2 are
summarized in Figure 5-3. The price impacts are shown by the red and blue lines referenced to
60 Data source: EIA, Petroleum & Other Liquids. Weekly Retail Gasoline and Diesel Prices,
https://www.eia.gov/dnav/pet/pet pri end dcus nus w.htm.
61 Id. For example, the average PADD 2 RFG price the first week of May 2022 was $4.30/gal, while the average
PADD 3 conventional gasoline price for the same week was $3.85/gal, for a difference of $0.45/gallon. During the
first week of May 2021, the average PADD 2 RFG price was $2.79/gal, while the average PADD 3 conventional
gasoline price was $2.43/gal, for a difference of $0.36/gal. Subtracting the $0.36/gal price difference in 2021 from
the $0.45/gal price difference in 2022 yields a price increase of $0.09/gal attributed to the supply shortfall in 2022.
50
-------
the lefthand vertical axis. The figure also shows the impact of the gasoline supply shortfall,
which is the grey line referenced to the righthand vertical axis.
Figure 5-3: PADD 2 Supply Factors in 2022 and Their Price Effects
CG Price Increase RFG Price Increase Gasoline Supply
Figure 5-3 shows the supply shortfall in PADD 2 for April, May, and June to be 50, 160,
and 110 kbpd, respectively, estimated by comparing the gasoline supply from PADD 3 to PADD
2 in 2022 to that in 2021.62 The 50 kbpd shortfall in April 2022 seemed to cause a modest price
increase, particularly for RFG.
However, the large 160 kbpd supply shortfall in May 2022—and corresponding decrease
in gasoline inventories depicted in Figure 3.D-4—is associated with an overall price increase of
over 400/gal in PADD 2. Despite the additional 110 kbpd supply shortfall in June 2022, the RFG
price was relatively flat while conventional gasoline prices declined by 70/gal. The RFG price
increased further to over 500/gal in July 2022, before both RFG and conventional gasoline prices
started to decrease with the return to balanced supply and increasing gasoline inventories.
As discussed in Section 3.D, the BP Whiting refinery suffered an emergency shutdown at
the end of August 2022 and was still at least partially shutdown for several weeks. RFG prices
spiked again, this time above 600/gal, and eventually above 700/gal, at a time when PADD 2
gasoline inventories declined to their lowest point of the year. Notably, however, conventional
gasoline prices did not increase, perhaps because the BP Whiting refinery is located adjacent to
Chicago, the primary RFG area in PADD 2.
Also as discussed in Section 3.D, the Cenovus (formerly BP-Husky) Toledo refinery
experienced an emergency shutdown at the end of September 2022 and remained shut down for
about 6 months. The Cenovus Toledo refinery shutdown did not result in a decrease in PADD 2
62 In 2021, the volume of gasoline supplied from PADD 3 to PADD 2 was more typical of the volume supplied in
most years. Thus, when we compared the volume of gasoline supplied from PADD 3 to PADD 2 in 2022 to that in
2021, we were able to estimate the gasoline supply reduction from PADD 3 to PADD 2 during April to June 2022.
51
-------
gasoline inventories, likely because the summer gasoline season was over and the gasoline pool
expanded due to increased blending of butane. However, conventional gasoline prices increased
by about 150/gal after the Cenovus Toledo refinery shutdown. It is likely, however, that the 6-
month shutdown of the Cenovus Toledo refinery played an important role in preventing PADD 2
gasoline inventories from recovering over the wintertime.
The estimated range of 30-80 kbpd of reduced gasoline supply caused by the removal of
the 1-psi waiver is smaller than the average 100 kbpd supply shortfall experienced by PADD 2
from April to June 2022. This lower impact on gasoline supply suggests a smaller price increase
is likely as a result of the removal of the 1-psi waiver. However, the supply shortfall has at least
some potential to last longer—perhaps over the entire summer.
A separate, but related factor are PADD 2 gasoline inventories when the removal of the
1-psi waiver takes effect. At the beginning of the 2022 summer gasoline production season—
which was prior to the 100 kbpd supply shortfall in PADD 2—PADD 2 gasoline inventories
were 3 million barrels higher than the 5-year average for that time of year. The price impacts
from the removal of the 1-psi waiver will likely be larger if PADD 2 gasoline inventories are still
low at the beginning of 2024 like it was in 2023, in addition to the previously discussed supply
shortfall.63 The opposite will likely be true if PADD 2 gasoline inventories recover and match
and even exceed those of early 2022.
An additional factor affecting gasoline price impacts is uncertainty regarding whether and
how gasoline markets will be supplied when the removal of the 1-psi waiver is implemented. If
refiners that currently supply a certain gasoline market decide not to participate in that market
due to the increased cost of producing low-RVP gasoline or limits in their ability to supply low-
RVP gasoline to that market, it could create uncertainty about whether that gasoline market will
be supplied. This uncertainty could have a price effect of its own and would potentially be
additive to other price effects.64 In Figure 5-3, we analyzed the price effects of the reduced
supply of gasoline and decreasing PADD 2 gasoline inventories for 2022, but this price effect
was solely due to reduced supply and shrinking gasoline inventories. In the case of the removal
of the 1-psi waiver, there are a combination of factors that could increase prices: (1) The reduced
supply of gasoline caused by removing light gasoline material to produce low-RVP gasoline; and
(2) Changes in how gasoline is produced and distributed throughout PADD 2, creating
uncertainty about how gasoline markets will be satisfied. These two price effects may be
additive, causing an even greater increase in gasoline prices than either factor alone. In addition,
the latter factor would tend to lead to different price impacts in different markets, depending on
the localized distribution issues faced.
63 EIA, "Gasoline explained: Gasoline price fluctuations," https://www.eia.gov/energyexplained/gasoline/price-
fluctuations.php.
64 Id. General Accountability Office (GAO), "Special Gasoline Blends Reduce Emissions and Improve Air Quality,
but Complicate Supply and Contribute to Higher Prices," GAO-05-421, June 2005, https://www.gao.gov/assets/gao-
05-421.pdf.
52
-------
6.
Benefits of Removing the 1-psi Waiver
As discussed in the preamble, under the relevant CAA provisions, upon receiving a
petition from a state governor that is accompanied by a successful demonstration of emissions
increases as a result of the 1-psi waiver, EPA is required to remove the 1-psi waiver in the areas
requested by the governor. In deciding whether to grant the petition, the statute does not provide
EPA with the authority to consider the benefits of the removal of the 1-psi waiver. Therefore, we
have not considered benefits in this action; we merely present here some assessment of the
potential benefits for awareness.
Modeling performed by the petitioning states in support of their petitions indicated
reductions in emissions of VOC, NOx, and CO. Specifically, those results show reductions
between 0.66-2.9% for VOC, 0.19-0.53% for NOx, and 0.05-0.14% for CO.65 The modeling
results also demonstrated increases between 0.08-0.30%) for PM2.5 and 0.08-0.32%) for PM10.
However, quantifying and monetizing air pollution-related health benefits related to these
reductions was not possible, and even if we had carried out such an analysis, we believe that the
results would not indicate meaningful benefits.
To put these changes in context, we can refer to recent work published by EPA assessing
the emissions and air quality impacts (i.e., ambient ozone, PM, and NO2 levels) of fuel
formulation changes resulting from the RFS Program, including increases in ethanol blend level
and volatility.66 This "anti-backsliding study" (ABS), required under CAA section 21 l(v)(l),
examined the impacts on air quality that might result from changes in vehicle and engine
emissions associated with renewable fuel volumes of ethanol under the RFS program relative to
approximately 2005 levels. Hoekman, et al., (2018) also reviewed available literature on
potential air quality impacts for E10 versus E0 across the entire lifecycle.67 Both studies found
potential increases and decreases in ambient concentration levels of pollutants, but none of them
were large despite having much larger emission inventory impacts than those associated with
removing the 1-psi waiver. Thus, we similarly expect any air quality impacts from the removal
of the 1-psi waiver to also be small.
Additionally, any benefits (or cost impacts) as a result of increased El 5 sales volumes are
negligible or nonexistent. As discussed in Section 3.B, we do not anticipate significant increases
in El 5 in the marketplace as a result of the removal of the 1-psi waiver. This action removes one
hurdle to El 5, but others remain, including vehicle compatibility, fuel offerings, liability
concerns, and especially compatibility with existing retail outlet infrastructure. Even if El 5 sales
volumes were to increase, as discussed in the recent RFS Set Rule,68 El5, while often priced
lower than E10 at retail, currently costs more to produce and distribute than E10, and thus we do
not expect any cost savings from increased El 5 sales volumes.
65 We believe that the reductions of CO and NOx would, in addition to VOC emissions impacts, satisfy the
requirements of the statute and justify granting the petitions.
66 "Clean Air Act Section 211(v)(l) Anti-backsliding Study," EPA-420-R-20-008, May 2020; "Final Determination
for Renewable Fuels and Air Quality Pursuant to Clean Air Act Section 21 l(v)," EPA-420-R-21-002, January 2021.
67 Hoekman, S. K., Broch, A., & Liu, X. (2018). Environmental implications of higher ethanol production and use in
the U.S. Renewable and Sustainable Energy Reviews, 81, 3140-3158.
68 Chapter 7.4, RFS Set Rule RIA.
53
-------
7. Gasoline Supply Situation Changes for 2025
As discussed in the preamble and supported throughout this document, there are a
number of considerations supporting our determination of insufficient supply for 2024. These
include:
(1) Continued low gasoline inventories in PADD 2.
(2) The limited time available after the promulgation of this action for coordination
between various parties to make the necessary physical changes to the gasoline
production and distribution infrastructure.
(3) Greater reduction in supply as a result of the removal of the 1-psi waiver than
estimated at the time of the proposal.
(4) The lack of sufficient time to make the capital investments and physical changes to
refineries and the fuel distribution system.
(5) Less flexibility within the fuel distribution system than had been anticipated to
adequately mitigate the supply reduction until such time as the capital and physical
changes can be made.
Delaying the removal of the 1-psi waiver to 2025 will allow more time for the necessary
coordination between the various parties. Allowing more time for the gasoline supply and
demand balance to improve and for refiners and fuel distributors to make necessary changes to
supply increased volumes of low-RVP gasoline will also alleviate supply constraints in 2025.
As described in Section 3, PADD 2 gasoline inventories were low in 2023. Due to an
expected large increase in the number of refinery maintenance projects in the fall of 2023 and
first quarter of 2024, gasoline inventories were expected to remain low going into 2024.69
Furthermore, EIA estimates that U.S. gasoline demand will increase by 60 kbpd in 2024
compared to 2023, which will further strain PADD 2 gasoline inventories.70 Further, as described
in Section 3.B, the start of the RFG program for the Denver area in 2024 will place an additional
strain on the gasoline supply and demand balance in 2024. As described in Section 3.D, PADD 2
gasoline inventories are an important source of gasoline supply during times of disruption and
shortfall in new production and import supply. The confluence of all these impacts on the supply
and demand balance support our determination of insufficient supply for 2024; however,
additional time to allow gasoline inventories to recover, refinery maintenance-related outages to
be completed, gasoline demand to fall, and the market to adjust to supplying RFG to Denver
should provide a much-improved gasoline supply and demand balance in PADD 2 in 2025.
The amount of time required for some refineries to produce low-RVP gasoline—or to
produce low-RVP gasoline without a large impact on their gasoline supply—and the difficulties
associated with distributing low-RVP gasoline to the petitioning states also informs our
determination of insufficient supply of gasoline in 2024. The various limitations to produce and
distribute low-RVP gasoline to the petitioning states—which can be overcome by making capital
69 Bloomberg News, "Nearly 2.5 Million Barrels a Day of US Refining Capacity to Shut for Fall Maintenance,"
October 2, 2023, https://www.bnnbloomberg.ca/nearlY-2-5-million-barrels-a-daY-of-us-refining-capacitv-to-shut-
for-fall-maintenance-1.1979186.
70 EIA, AEO 2023, Table 11, https://www.eia.gov/outlooks/aeo.
54
-------
investments—are outlined in Section 3.A. However, such investments take time to plan,
engineer, permit, and construct. Some easy-to-implement capital projects can be completed by
2024. However, other more-involved projects requiring more extensive design, permitting, and
construction (e.g., debottlenecking debutanizers, installing new gasoline storage tanks) will take
more time to implement. These projects often require at least 2 years to complete.
Consequently, EPA typically provides significant lead time and/or phase-in of its fuel
standards. In theory, refiners and fuel distributors could have begun making these investments
after the first petition was submitted in April 2022. Alternatively, they could have started their
planning after EPA proposed to remove the 1-psi waiver in March 2023. Regardless of whether
refiners and fuel distributors earnestly started their planning and engineering design mid-2022 or
in early 2023, there would not have been sufficient time to design, permit, and complete the
construction of these more-involved capital projects before the start of the 2024 summer gasoline
production season (e.g., most refiners will start producing summer gasoline in March 2024).
Many of these same supply and demand balance concerns may still exist for 2025, in
large part depending on the progress that the fuel production and distribution system is able to
make in putting in place their capital investments. However, the magnitude of these concerns is
expected to diminish not only due to the additional time available, but also due to changing
circumstances in 2025. First, the gasoline supply and demand balance is expected to improve in
2025. As discussed in Section 3.C, EIA forecasts that nationwide gasoline demand in 2025 will
decrease by 140 kbpd relative to 2024, which is 80 kbpd less than 2023.71 Thus, due to reduced
gasoline demand, the supply and demand balance is expected to improve significantly in 2025
relative to 2024, and even improve relative to 2023. The forecasted reduction in gasoline demand
in 2025 would help to offset much of the estimated loss of gasoline production caused by
producing low-RVP gasoline. Refiners are also expected to catch up with their refinery
maintenance in 2023 and 2024; thus, they will be able to maintain higher gasoline production at
the end of 2024 heading into 2025, allowing PADD 2 gasoline inventories to recover closer to
normal prior to the summer of 2025. Refiners supplying RFG to Denver will also have optimized
their gasoline supply to Denver in 2024; thus, these refiners will be better positioned to continue
to supply RFG to Denver in 2025 while also supplying low-RVP gasoline to nearby petitioning
states.
Second, the types of capital investments that can be made by refiners and fuel distributors
by 2025 will improve their ability to produce and distribute low-RVP gasoline. Refiners will be
able to produce low-RVP gasoline at a lower overall loss of gasoline production. We have heard
from several refiners and fuel distributors that some planned investments are able to be
completed by the summer of 2025. Both refiners and fuel distributors will be able to further
improve their ability to distribute low-RVP gasoline, particularly those refineries and pipeline
segments that serve both petitioning and non-petitioning states.
71 EIA, AEO 2023, Table 11, httos://www.eia.gov/outlooks/aeo. In making its estimate, EIA considers economic
growth, oil price, oil and gas supply, and zero-carbon technology cost (i.e., electric vehicles). EIA, "Annual Energy
Outlook 2023 Release at Resources for the Future," March 16, 2023,
https://www.eia.gov/outlooks/aeo/pdf/AEQ2023 Release Presentation.pdf.
55
-------
8. Screening Analysis for Potential Impacts on Small Entities
This section discusses EPA's screening analysis evaluating the potential impacts of the
removal of the 1-psi waiver on small entities. The Regulatory Flexibility Act (RFA), as amended
by the Small Business Regulatory Enforcement Fairness Act of 1996 (SBREFA), generally
requires an agency to prepare a regulatory flexibility analysis of any rule subject to notice and
comment rulemaking requirements under the Administrative Procedure Act or any other statute,
unless the agency certifies that the rule will not have a significant economic impact on a
substantial number of small entities (referred to as a "No SISNOSE finding"). Pursuant to this
requirement, EPA has prepared a screening analysis for this rule.
Section 8.A provides background on the RFA and this rule, including the regulated small
entities. Section 8.B describes EPA's calculations of the costs of the rule and the resulting cost-
to-sales ratios. Section 8.C concludes.
A. Background
i. Overview of the Regulatory Flexibility Act (RFA)
The RFA was amended by SBREFA to ensure that concerns regarding small entities are
adequately considered during the development of new regulations that affect those entities. The
RFA requires us to carefully consider the economic impacts that our rules may have on small
entities. The elements of the initial regulatory flexibility analysis accompanying a proposed rule
are set forth in 5 U.S.C. § 603, while those of the final regulatory flexibility analysis
accompanying a final rule are set forth in section 604. However, section 605(b) of the statute
provides that EPA need not conduct the section 603 or 604 analyses if we certify that the rule
will not have a significant economic impact on a substantial number of small entities.
ii. Need for the Rulemaking and Rulemaking Objectives
A discussion on the need for and objectives of this action is in Preamble Section I. CAA
section 211(h)(5) requires EPA to remove the 1-psi waiver for El0 via regulation upon a
demonstration by a governor that the 1-psi waiver increases emissions in their state.
iii. Definition and Description of Small Entities
Small entities include small businesses, small organizations, and small governmental
jurisdictions. For the purposes of assessing the impacts of the rule on small entities, a small
entity is defined as: (1) A small business according to the Small Business Administration's
(SB A) size standards; (2) A small governmental jurisdiction that is a government of a city,
county, town, school district or special district with a population of less than 50,000; or (3) A
small organization that is any not-for-profit enterprise that is independently owned and operated
and is not dominant in its field.
Small businesses (as well as large businesses) would be regulated by this rulemaking, but
not small governmental jurisdictions or small organizations as described above. As set by SB A,
56
-------
the categories of small entities that would potentially be directly affected by this rulemaking are
described in the table below.
Small Business Definitions
Industry
Defined as small entity by SBA if less
than or equal to:
NAICS" code
Gasoline and diesel fuel refiners
1,500 employees'3
324110
a North American Industrial Classification System.
b EPA has included in past fuels rulemakings a provision that, in order to qualify for small refiner flexibilities, a
refiner must also produce no greater than 155,000 barrels per calendar day (bpcd) crude capacity. See, e.g., 40 CFR
80.225(a)(1) (2019), 40 CFR 80.550(a) and (b) (2019), 40 CFR 80.1142(a)(1) (2019), 40 CFR 80.1338(a) (2019), 40
CFR 80.1442(a)(1), 40 CFR 80.1620(a) (2019).
EPA used the criteria for small entities developed by the Small Business Administration
under the North American Industry Classification System (NAICS) as a guide. Information about
the characteristics of refiners comes from sources including the Energy Information
Administration (EIA) within the U.S. Department of Energy, oil industry literature, and previous
rulemakings that have affected the refining industry. In addition, EPA used publicly available
employment information to determine which companies meet the SB A definition of "small
entity." These refiners fall under the Petroleum Refineries category, 324110, as defined by
NAICS.
Small entities that are subject to this rulemaking include domestic refiners that produce
and distribute gasoline to the petitioning states. While in the proposed rulemaking EPA did not
identify any affected small refiners that would be affected by this action, two commenters
identified three potential small refiners that would be affected by this action. After evaluating the
information submitted by the commenters, EPA now believes that there is currently one refiner
(CountryMark) located in a non-petitioning state that produces and distributes gasoline to the
petitioning states that meets the small entity definition of having 1,500 employees or fewer.
However, while EPA disagrees that the other two refiners (Ergon-West Virginia and Wyoming
Refining Company) are eligible to qualify as small entities,72 for purposes of this screening
analysis we have nonetheless evaluated the impact of this rulemaking on these companies as
well.
iv. Reporting, Recordkeeping, and Other Compliance Requirements
Registration, reporting, and recordkeeping are necessary to track compliance with EPA's
fuel quality regulations. However, these requirements are already in place under the existing fuel
quality regulations. Therefore, we do not anticipate that there will be any significant cost on
directly regulated small entities.
B. Screening Analysis Approach and Results
This section concerns EPA's screening analyses performed for the removal of the 1-psi
waiver. In general, we expect that refiners, including small refiners, will be able to recover the
72 See RTC Section 7.1.
57
-------
cost associated with the removal of the 1-psi waiver through higher gasoline prices in the
petitioning and surrounding states. Nevertheless, we estimated the cost-to-sales ratios for each of
the three purported small refiners that distribute gasoline to the petitioning states using refinery-
specific data under the assumption that they could not recover their increased production costs.
Using recent RFS compliance data, we first estimated the annual gasoline production
volume for each refinery assuming the total gasoline production for each of these refineries
remains unchanged. Using information from recent small refinery exemption (SRE) petitions and
other publicly available information, we then estimated the amount of each refinery's gasoline
production that would be distributed to the petitioning states, assuming each of these refineries
continues to distribute gasoline to the petitioning states and that only that portion would be
required to be low-RVP CBOB. We then multiplied that volume of gasoline by 120 per gallon,
which is the upper end of the range of projected fuel costs in Table 4.A-2 and represents the
worst-case scenario for refinery production costs as reported in the Baker and O'Brien Study.
The actual calculations for each refiner are provided in Section 8.D; a non-CBI example of these
calculations is shown below in Table 6.B-1.
Table 6.B-1: Example Refiner Costs Calculation
Company
Total Gasoline
Production
(gal)
Gasoline
Distribution to
Petitioning States
Low-RVP CBOB
Production
(gal)
Cost
($/gal)
Total Cost
Example
150,000,000
50%
75,000,000
$0.12
$9,000,000
Using information from recent SRE petitions and other publicly available information,
the final step in our analysis is to divide the total estimated costs for each refiner by its total
estimated annual sales. The resulting range of cost-to-sales ratios for these refiners are shown in
Table 6.B-2, along with a non-CBI example of these calculations using the data in Table 6.B-1.
Table 6.B-2: Estimai
ted Cost-to-Sales Ratios
Company
Total Cost
(Million Dollars)
Total Sales
(Million Dollars)
Cost-to-Sales
Ratio
Refiners (Actual)a
—
—
0.13-0.15%
Example
$9.0
$2,500
0.36%
a The actual calculations for each refiner are provided in Section 8.D.
C. Conclusions
We conducted a screening analysis by looking at the potential impacts on the three
specific purported small refiners that distribute gasoline to the petitioning states. While we
believe that refiners will recover the cost associated with the removal of the 1-psi waiver through
higher gasoline prices in the petitioning and surrounding states, we have nonetheless evaluated
the impacts of this rule assuming a worst-case scenario wherein the cost of producing low-RVP
CBOB was 120/gal and refiners could not recover their costs. Under these extreme assumptions
we were able to estimate costs of this rule using the methodology described in the previous
58
-------
section and then use a cost-to-sales ratio test (a ratio of the estimated annualized compliance
costs to the value of sales per company) to assess whether the costs were significant.73
Even if refiners are not able to recover the cost associated with the removal of the 1-psi
waiver through higher gasoline prices, based on our cost-to-sales analysis, the refiners would be
affected at less than 1% of their sales as a result of this action (i.e., the estimated costs of this
rule would be less than 1% of their sales); the actual cost-to-sales percentages ranged from
0.13% to 0.15%. Therefore, based on our outreach, fact-finding, and analysis of the potential
impacts of this rule on small businesses, EPA finds that the removal of the 1-psi waiver in the
petitioning states will not have a significant economic impact on a substantial number of small
entities.
I). Refiner CBI Data
[Information Redacted - Claimed as CBI]
73 A cost-to-sales ratio of 1% represents a typical agency threshold for determining the significance of the economic
impact on small entities. See "Final Guidance for EPA Rulewriters: Regulatory Flexibility Act as amended by the
Small Business Regulatory Enforcement Fairness Act," November 2006.
59
------- |