CALIFORNIA AIR RESOURCES BOARD v. UNITED STATES ENVIRONMENTAL PROTECTION AGENCY et al

Filing 1

COMPLAINT against NATIONAL HIGHWAY TRAFFIC SAFETY ADMINISTRATION, UNITED STATES ENVIRONMENTAL PROTECTION AGENCY ( Filing fee $ 400 receipt number 0090-6043290) filed by CALIFORNIA AIR RESOURCES BOARD. (Attachments: #1 Civil Cover Sheet Civil Cover Sheet, #2 Summons Summons to U.S. Environmental Protection Agency, #3 Summons Summons to National Highway Traffic Safety Administration, #4 Summons Summons to U.S. Attorney General, #5 Summons Summons to U.S. Attorney for the District of Columbia)(Hoffman, Ryan)

IN THE UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF COLUMBIA

CALIFORNIA AIR RESOURCES BOARD,
1001 I Street
Sacramento, CA 95814

Case No. 1:19-cv-965

Plaintiff,
v.
UNITED STATES ENVIRONMENTAL
PROTECTION AGENCY,
1200 Pennsylvania Avenue, N.W.
Washington, D.C. 20460
NATIONAL HIGHWAY TRAFFIC SAFETY
ADMINISTRATION,
1200 New Jersey Avenue, S.E.
Washington, D.C. 20590
Defendants.

COMPLAINT
INTRODUCTION
1. Plaintiff California Air Resources Board (“CARB”) brings this action under the
Freedom of Information Act (“FOIA”), 5 U.S.C. § 552, for declaratory, injunctive, and other
appropriate relief against the United States Environmental Protection Agency (“EPA”) and the
National Highway Traffic Safety Administration (“NHTSA”) for ongoing violations of Sections
552(a)(6)(A)(i), 552(a)(6)(A)(ii), and 552(a)(6)(B)(i) of FOIA, as well as the agencies’
regulations implementing FOIA.
2. This case concerns information underlying sweeping federal proposals to (1) alter
greenhouse gas emission and fuel economy standards for automobiles, and (2) preempt authority

1

held by the states to set automobile emissions standards. These proposals contradict previous,
thorough technical analyses conducted by EPA, NHTSA, CARB and others. Moreover, in a
stark departure from prior rulemakings, critical information underlying EPA’s and NHTSA’s
analyses was not disclosed. As very serious flaws in the agencies’ analyses and conclusions are
evident – and the resulting proposals threaten public health – CARB submitted FOIA requests to
both agencies for documents concerning vehicle fleet composition, new car sales, vehicle safety,
battery technology, and other information that NHTSA and EPA used in proposing to roll back
vehicle emission and fuel economy standards. This complaint seeks release of that critical
information.
3. CARB issued identical FOIA requests to EPA and NHTSA in early September 2018.
EPA has failed to issue timely determinations regarding the requests. NHTSA responded to the
requests, but withheld information requested by CARB based on inadequate justifications, and in
some instances without justification or explanation. Some of NHTSA’s unexplained and
unjustified withholdings amount to failure to make and convey determinations as required by
FOIA. CARB appealed NHTSA’s determinations and failure to issue determinations in midDecember 2018, and NHTSA has failed to timely respond to CARB’s appeal. CARB seeks
immediate release of the requested agency records from EPA and NHTSA, and other relief as set
forth below.
PARTIES
4. Plaintiff CARB is a California state agency charged with “coordinating efforts to attain
and maintain ambient air quality standards, to conduct research into the causes of and solution to
air pollution, and to systematically attack the serious problem caused by motor vehicles, which is
the major source of air pollution in many areas of the state.” Cal. Health & Saf. Code, § 39003.

2

5. Defendant EPA is an agency of the United States federal government with
responsibility for environmental protection, including adopting and enforcing motor vehicle
greenhouse gas emissions standards. EPA is an agency of the United States federal government
within the meaning of 5 U.S.C. § 552(f)(1). EPA has possession of and control over agency
records that CARB seeks, which CARB has properly requested pursuant to FOIA and EPA’s
implementing regulations.
6. Defendant NHTSA is a component agency of the United States Department of
Transportation responsible for, inter alia, adopting and enforcing motor vehicle fuel economy
standards. NHTSA is an agency of the United States federal government within the meaning of
5 U.S.C. § 552(f)(1). NHTSA has possession of and control over agency records that CARB
seeks, which CARB has properly requested pursuant to FOIA and NHTSA’s implementing
regulations.
JURISDICTION AND VENUE
7. This Court has subject-matter jurisdiction over this action and personal jurisdiction
over the parties for purposes of this action pursuant to 5 U.S.C. § 552(a)(4)(B) and 28 U.S.C. §§
1331 and 1361.
8. Venue is proper in this district pursuant to 5 U.S.C. § 552(a)(4)(B), which grants the
district court of the United States in the District of Columbia jurisdiction to enjoin federal
agencies from withholding agency records and to order the production of any agency records
improperly withheld from a complainant.
9. This Court has authority to grant declaratory relief pursuant to 28 U.S.C. § 2201.
10. This Court has authority to grant injunctive relief pursuant to 28 U.S.C. § 2202 and 5
U.S.C. § 552(a)(4)(B).

3

STATUTORY AND REGULATORY BACKGROUND
11. FOIA requires that federal agencies release records to any person, upon request,
unless one of nine statutory exemptions from disclosure applies. 5 U.S.C. § 552(a)-(b). “Any
reasonably segregable portion of a record shall be provided to any person requesting such record
after deletion of the portions which are exempt.” 5 U.S.C. § 552(b).
12. Within twenty business days of an agency’s receipt of a FOIA request, the agency
must issue a determination resolving the request, and must “immediately notify” the requester of
“such determination and the reasons therefor.” 5 U.S.C. § 552(a)(6)(A)(i)(I); 40 C.F.R. §
2.104(a); 49 C.F.R. § 7.31(a)(2).
13. An agency may only delay its response to a request if “unusual circumstances” (as
described by FOIA) exist. Even in this event, the agency’s time to respond is extended by no
more than ten days. 5 U.S.C. § 552(a)(6)(B)(i); 40 C.F.R. § 2.104(d); 49 C.F.R. § 7.34(a). Any
such extension must be “by written notice” to the requester, “setting forth the unusual
circumstances for such extension and the date on which a determination is expected.” Id.
14. If an agency’s determination as to a request is appealed, the agency must make and
convey a determination with respect to the appeal within twenty business days after receipt. 5
U.S.C. § 552(a)(6)(A)(ii); 49 C.F.R. § 7.33(a)(2).
15. An agency must “promptly” release non-exempt records (or reasonably segregable
portions of records) requested in accordance with FOIA. 5 U.S.C. § 552(a)(6)(C)(i).
16. If an agency fails to comply with the statutory time limits for issuing and
communicating determinations as to requests and appeals, the requester is deemed to have
exhausted their administrative remedies and may immediately file suit. 5 U.S.C. §§ 552(a)(4)(B),
(a)(6)(C)(i).

4

17. FOIA grants federal district courts the authority to enjoin an agency from withholding
agency records and “to order the production of any agency records improperly withheld.” 5
U.S.C. § 552(a)(4)(B).
18. FOIA permits the courts to grant “reasonable attorney fees and other litigation costs
reasonably incurred in any case under this section in which the complainant has substantially
prevailed.” 5 U.S.C. § 552(a)(4)(E)(i).
FACTS
Background
19. In August 2018, EPA and NHTSA issued a joint proposal entitled “The Safer
Affordable Fuel-Efficient (‘SAFE’) Vehicles Rule for Model Years 2021-2026 Passenger Cars
and Light Trucks” (the “Proposal”), 83 Fed. Reg. 42,986 (Aug. 24, 2018). The Proposal departs
substantially from extensive technical analyses conducted to support the existing regulatory
regime, including claims that lower emission vehicle technologies are costlier than previously
understood. The agencies ground these conclusions in substantial part on a series of models and
analyses that have not been fully disclosed.
20. Contrary to prior commitments from EPA and NHTSA to collaborate with California
on these standards, and the agencies’ past practice in developing motor vehicle emission
standards (see, e.g., 77 Fed. Reg. 62,624, 62,632, 62,784-62,785 (Oct. 15, 2012) [discussing
coordination with CARB to develop the standards at issue and for changes to said standards]),
CARB has not been materially involved in discussions developing the Proposal.
21. In the “Compliance and Enforcement” portion of the Proposal, 83 Fed. Reg. at
43,476, the agencies discuss Executive Order 13,132 (“Executive Order”), which establishes
requirements for federal agencies to address federalism concerns in formulating and

5

implementing policies. Exec. Order No. 13,132, 64 Fed. Reg. 43,255 (Aug. 4, 1999). The
Executive Order requires federal agencies, including EPA and NHTSA, to “have an accountable
process to ensure meaningful and timely input by State and local officials in the development of
regulatory policies that have federalism implications.” 64 Fed. Reg. 43,257. When promulgating
regulations that have federalism implications or preempt state law, agencies are required to
consult with State and local officials early in the process of developing the proposed regulation.
Id. at 43,258.
22. In the Proposal, EPA and NHTSA stated that they complied with the Executive
Order’s requirements, but provided no details about that purported compliance or any
consultations with States about the Proposal before it was issued. 83 Fed. Reg. 43,476.
23. On August 27, 2018, CARB (along with Attorneys General and agencies of more
than fifteen States) requested that EPA and NHTSA extend the comment period for the Proposal
by at least 60 days. This request expressly noted that some information necessary to evaluate and
comment meaningfully on the Proposal was not publicly available. This request was denied by
both EPA and NHTSA.
24. As part of its comments in the rulemaking docket, and also pursuant to FOIA, CARB
requested information necessary to evaluate the Proposal from both EPA and NHTSA by letter
sent to each agency on September 11, 2018. A true and correct copy of this letter is attached
hereto as Exhibit A.
25. CARB’s requests focused narrowly on critical models, technology analyses, and
related information that are central to the federal agencies’ claims. But, as described below,
CARB did not receive much of the information it requested. EPA and NHTSA have withheld

6

this information from CARB and the public. Without this information, CARB was unable to
fully evaluate and comment on the Proposal and its asserted bases.
CARB’s FOIA Request Letter to Agencies
26. CARB sent a letter dated September 11, 2018, to both EPA and NHTSA, which
contained 12 requests for records, with subparts elaborating on some of the requests. The
requests stated as follows:
1. Information about the models and data used to estimate battery costs for
electrified vehicles.
a. The proposal and the Preliminary Regulatory Impact Analysis (PRIA),
NHTSA-2018-0067-1972, reference the Argonne National Laboratories'
(ANL) website for the BatPaC model for estimating battery costs for
vehicles, and state that the agencies used "an up-to-date version" of the
model, but do not identify the version. (See, e.g., 83 Fed.Reg. 42,985,
43,002 (August 24, 2018).) U.S. EPA and NHTSA have posted to the
dockets for this action a document describing how BatPaC was developed,
but this document appears to be from 2012. It does not state which version
of BatPaC NHTSA and U.S. EPA used to estimate battery costs. See
"Modeling the Performance and Cost of Lithium-Ion Batteries for
Electric-Drive Vehicles, Argonne National Laboratory, ANL-12/55,"
NHTSA Docket ID#: NHTSA-2018-0067-1692; EPA Docket ID#: EPAHQ-OAR-2018-0283-0764.
b. U.S. EPA and NHTSA should make available the information
specifying the full battery sizes, in kilowatt-hours (kWh), battery pack
configuration, and costs used for each vehicle iteration in the CAFE
model. See PRIA, Electrification Technologies, Technology Overview,
section 6.3.8.1, p. 357. The PRIA states that NHTSA posted ANL vehicle
files that have battery pack sizes and costs for each vehicle, but there is no
additional information about battery pack configuration (e.g., the number
of cells, and the electrical topology of how those cells are arranged in the
battery pack), nor do they directly reference where the files are posted.
NHTSA and U.S. EPA have not posted the BatPaC model file(s) that were
used. ANL cost and battery size data referenced in the PRIA, p. 358,
footnote 325, but the footnote refers to a docket identification number that
is not available. Previously, in support of the draft Technical Assessment
Report and Proposed Determination on the Appropriateness of the Model
Year 2022-2025 Light-duty Vehicle Greenhouse Gas Emissions Standards
under the Midterm Evaluation (Proposed Determination), EPA-420-R-16020, November 2016, U.S. EPA posted the BatPaC files that it used.

7

c. The proposal and PRIA provide conflicting information about which
battery chemistries the agencies considered. For instance, the proposal and
PRIA refer to NMC441-Gr chemistry for both plug-in hybrid-electric
vehicles and battery-electric vehicles, but the ANL summary refers to
NMC333. See, e.g., PRJA, pp. 372, Table 6-27, 373 ["We selected
NMC441 as choice of chemistry for PHEVs and BEVs. NMC441 more
suitable for high energy batteries capable of discharge rates."]. The Excel
file titled "ANL-Summary of Main Component Performance Assumptions
NPRM" has a tab labeled 'Description - BatPac' with the same table listed
as in the PRIA, except the chemistry listed for PHEVs and BEVs is
NMC333-G instead of NMC441-Gr. See Docket ID Nos. EPA-HQ-OAR201 8-02 83-0054 and NHTSA-2018-0067-0003. The proposal and PRIA
do not directly reference this file.
This information is required to replicate and evaluate whether the modeling
underlying the proposal is appropriate, considering the various vehicle and
technology types.
2. The PRIA references Polk registration data, including survival rates aggregated
by model year, calendar year, and body style. These data are needed to verify the
coefficients of the new model predictions for vehicle retirement (scrappage), but
have not been made available. See, e.g., PRIA at pp. 1008, 1014, 1023, fig. 8-23,
1025, fig. 8-24, and 1027, fig. 8-25.
3. New vehicle sales and price data referenced in the proposal. This includes:
a. Data provided by the National Automobile Dealers Association
(NADA) and others. See 83 Fed.Reg. at 43,095; PRIA, pp. 1017 -1018.
b. Data describing historical transaction prices, and quarterly new vehicle
sales data used to develop the dynamic new sales model. See PRIA, pp.
954-961.
c. Economic data used to develop the autoregressive distributed-lag
(ARDL) model that predicts new vehicle sales and is used in the CAFE
model. See 83 Fed.Reg. at 43,074.
This data is necessary to evaluate the proposal's predictions for fleet population,
sales, and fatalities.
4. Report of analysis of the standard errors and significance of the ARDL sales
model coefficients, F-statistic and R2 of the overall model, and variable
stationarity and co-integration indicators. This information is needed to verify the
statistical significance and errors of the coefficients used in the Volpe model. The
coefficients for the ADRL sales model listed on p. 957, Table 8-1 of the PRIA,

8

are not consistent with those implemented in the model. See CAFE Model
Documentation, PRIA, p. 78, Table 17, available at:
https://www.nhtsa.gov/corporate-average-fuel-economy/compliance-and-effectsmodeling-system ["2018 NPRM for Model Years 2021-2026 Passenger Cars
and Light Trucks," Model Documentation].
5. The coefficients for the dynamic fleet share equation described in the CAFE
Model Documentation on p. 79. These are not listed anywhere. Additionally,
according to the PRIA on p. 955, the model was based on EIA's National Energy
Modeling System (NEMS), but no reference is provided for the NEMS model.
This information is necessary to evaluate the equation used in the model.
6. Cited reference Anders Lie, Swedish Transport Administration, Peer Review of
NHTSA's November 2011 Preliminary Report "Relationships Between Fatality
Risk, Mass, and Footprint in Model Year 2000-2007 Passenger Cars and LTVs."
This document reviews the 2011 study by Kahane that NHTSA relies upon,
regarding the effects of mass reduction on fatality risk. See 83 Fed.Reg. at 43,112,
n. 307. It is also item 0035 in the NHTSA-2010-0152 docket. However,
attempting to access the document on the docket website results in a server error.
See https://www.regulations.gov/document?D=NHTSA-2010-0152-0 035,
attempted August 28, 2018.
7. CAFE model output file Annual Societal Effects Report. See CAFE Model
Documentation, infra. This data describes predicted fatalities by regulatory class,
body style, and weight threshold of vehicle. This detailed information is necessary
to evaluate the fatalities computation in the CAFE model.
8. The agencies' detailed explanation and derivation of their point estimates for
the increase in fatalities per hundred pounds of mass reduction over a constant
footprint based on historical crash data, for model years 2004-2011 and calendar
years 2006-2012. Previously, these details were provided in a separate report such
as the "2016 Puckett and Kindelberger report." No such report is available this
time. The PRIA only provides a summary table of the results of this analysis, yet
states an "updated analysis" exists. See PRIA, p. 1357, section 11.4.
9. Data used by the agencies to derive the new statistical model that predicts
fatality rates by vehicle age. See PRIA Table 11-21, p. 1397. The coefficients of
the model are provided, but without the data it is not possible to evaluate whether
the coefficients were properly derived. Additionally, the coefficients provided in
the PRIA are different (significant digits and sign changes) than those identified
in the actual model source code (which are also commented out such that they are
nonfunctional) and are different from the model year based coefficients used in
the input files. This renders unclear what coefficients the analysis in the NPRM is
based upon.

9

10. So-called "off model" analytical tools the agencies used to summarize and
tabulate the results of the CAFE model. See 83 Fed.Reg. at 43,256, n. 595. These
tools and calculations analyzed various components of social and private costs
and benefits, as well as other factors. These analytical tools are supposedly
available on NHTSA's website but we have not located them.
11. Input files used for the Autonomie model for various engine technologies that
U.S. EPA and NHTSA rely on to calculate the efficiency improvements of
various technologies, such as turbocharging and high-compression-ratio
(Atkinson) engines. The input and output files are required to be able to
understand what U.S. EPA and NHTSA relied on as representative of these
engine technologies, and to confirm correct efficiency levels were calculated for
the various technologies, considering the current state of the art.
12. Modeling tools developed by U.S. EPA including:
a. All files necessary to utilize - with the Advanced Light-Duty Powertrain
and Hybrid Analysis (ALPHA) and the Optimization Model for reducing
Emissions of Greenhouse gases from Automobiles (OMEGA) - the
response surface equations developed by U.S. EPA as identified or
referenced in: "Peer Review of EPA's Response Surface Equation Report"
(Docket ID No. EPA-HQ-OAR-2018-0283 -0025); and SAE paper 201801-1273 authored by U.S. EPA (Docket ID No. EPA-HQ-OAR-20180283-0028).
b. All new or modified input files, source code, and executable files for
U.S. EPA's OMEGA model developed since the release of the Proposed
Determination in late 2016.
c. All current and new input files, source code, and executable files for
ALPHA used for the Proposed Determination in late 2016 and/or modified
since then.
d. All current and new pre-processors and their inputs used for the
Proposed Determination in late 2016 or modified since then to categorize,
sort, and rank technology packages and costs for use with OMEGA.
27. CARB’s letters to EPA and NHTSA properly requested waiver of searching and
copying fees for CARB’s requests.

10

28. CARB’s FOIA requests to EPA and NHTSA complied in all respects with all
applicable laws and agency regulations, including but not limited to 5 U.S.C. § 552 and EPA’s
and NHTSA’s regulations implementing FOIA.
EPA’s Response
29. By letter dated October 23, 2018, EPA acknowledged receipt of CARB’s request
letter of September 11, 2018, and assigned it FOIA tracking number EPA-HQ-2018-011521. A
true and correct copy of this letter is attached hereto as Exhibit B.
30. EPA declined to make any determination as to CARB’s requests numbers 1 through
11, stating only EPA’s belief that NHTSA would provide its own responses to these requests.
31. As to CARB’s request number 12, EPA stated that it “is reviewing records that may
be responsive to this request to determine whether they are appropriate for production, or
whether they should be withheld pursuant to statutory exemptions to disclosure. . . . Once EPA
makes a determination under FOIA number EPA-HQ-2018-011521, you will be provided appeal
rights accordingly.” (See Exhibit B.)
32. EPA is in possession of records responsive to each of CARB’s twelve requests, and
each subpart of each request.
NHTSA’s Response
33. By letter dated October 23, 2018, NHTSA responded to CARB’s request letter and
assigned it FOIA tracking number ES18-003395. A true and correct copy of this letter is
attached hereto as Exhibit C.
34. While NHTSA provided some of the information CARB requested, its responses to
many of CARB’s requests were inadequate and many of the records sought were not provided.

11

35. NHTSA withheld the records sought by requests numbers 1(a), 1(c), 4, and 12
without explanation or justification.
36. NHTSA determined that it would withhold the records sought by requests numbers 2
and 9 based on FOIA Exemption 4 (trade secrets and commercial or financial information), and
the records sought by request number 8 based on FOIA Exemption 5 (pre-decisional agency
deliberation, opinions or recommendations). These determinations are not supported by the
factual record.
37. NHTSA’s deficient responses are addressed in detail in the following sections.
CARB’s Request No. 1
38. CARB’s first request in its FOIA letter to EPA and NHTSA sought information
pertaining to BatPaC—a Battery Pack and Costing tool developed by Argonne National
Laboratory (“ANL”). The tool is critical to estimating the costs of electric vehicle technologies –
a consideration that is central to the overall set of vehicle standards. Neither EPA nor NHTSA
properly replied to the request.
39. The BatPaC modeling tool simulates the performance and cost of batteries; because
that technology is rapidly evolving, the precise specifications of the model and its inputs are of
considerable importance to its results. BatPaC is contained within an Excel workbook, in which
the user inputs various data. The input data include: established battery chemistries, often
denoted by the letters “NMC” followed by several numbers (e.g., NMC441, NMC622, etc.);
battery pack requirements or configurations, including the number of cells in the battery pack,
the energy capacity of each cell, the voltage of each cell, the fade over the battery’s lifetime, and
so on; various key constraints, such as the maximum electrode thickness; and costs of materials,
such as lithium or nickel. Battery pack configurations and materials costs can be adjusted by the

12

user as desired, so it is especially important for this data to be provided. Based on the inputs
provided, BatPaC calculates manufacturing requirements at scale, cost, and other physical and
electrical parameters for a particular battery pack configuration.
40. EPA and NHTSA used BatPaC in the modeling they did to develop the Proposal.
The battery cost information BatPaC produces is one consideration of many in assessing the
cost-effectiveness of vehicle standards and can have important effects on that determination.
41. Subparts (a) and (c) of CARB’s first request sought records reflecting the version of
BatPaC and the battery chemistries that were used by the agencies to develop the Proposal.
42. ANL updates BatPaC periodically.
43. Each new version of BatPaC includes new developments in vehicle battery
technology and costs, among other possible updates. Thus, the version used is of real importance
to the public in evaluating EPA’s and NHTSA’s conclusions in the Proposal.
44. In its response to subparts (a) and (c) of CARB’s first request, NHTSA asserted that
BatPaC version 3.0 had been used in developing the Proposal. This assertion is inconsistent with
statements and documentation provided by NHTSA and EPA in the Proposal and its rulemaking
docket.
45. EPA and NHTSA stated that they used “an updated version” of BatPaC in developing
the Proposal, but did not identify the version used. See, e.g., 83 Fed. Reg. 42,986, 43,002 (Aug.
24, 2018).
46. In an April 2018 interagency review of NHTSA’s draft of the Proposal, EPA
indicated that it could not discern which version of BatPaC NHTSA had used. EPA-HQ-OAR2018-0283-0453 (see “Email 5 – Email from William Charmley to Chandana Achanta – June 18,
2018”).

13

47. EPA and NHTSA posted a document describing how BatPaC was developed to the
rulemaking dockets for the Proposal. This document was from 2012 and did not identify which
version of BatPaC was used in developing the Proposal. Argonne National Laboratory, ANL12/55, Modeling the Performance and Cost of Lithium-Ion Batteries for Electric-Drive Vehicles
(2012), EPA-HQ-OAR-2018-0283-0764, NHTSA-2018-0067-1692.
48. At the time that the notice of proposed rulemaking (“NPRM”) for the Proposal was
published, the 2012 version of BatPaC was not available to the public on ANL’s website.
49. ANL updated BatPaC several times between 2012 and the release of the Proposal on
August 24, 2018.
50. ANL released BatPaC version 3.0 on December 25, 2015.
51. ANL released BatPaC version 3.1 on October 9, 2017.
52. NHTSA and EPA did not complete their battery modeling for the Proposal before
October 9, 2017.
53. Different battery chemistries correlate to different versions of the BatPaC model.
54. In the Preliminary Regulatory Impact Analysis (“PRIA”) for the Proposal, NHTSA
and EPA stated that they used battery chemistry NMC441 for electric vehicles and some plug-in
hybrid electric vehicles in their modeling for the NPRM (see p. 373 table 6-27, and p. 374).
55. However, in one document (an Excel file) included in the rulemaking docket for the
Proposal, a tab states that NHTSA and EPA used battery chemistry NMC333 for electric
vehicles and some plug-in hybrid electric vehicles in their modeling for the NPRM. ANLSummary of Main Component Performance Assumptions NPRM, EPA-HQ-OAR-2018-02830054, NHTSA-2018-0067-0003.

14

56. There is no version of BatPaC (or, at a minimum, no version released to the public)
that uses both NMC441 and NMC333 as battery chemistries.
57. Battery chemistry NMC441 is not an option in BatPaC versions 3.0 or 3.1 but is an
option in BatPaC versions 2.0 and 2.1.
58. ANL dropped battery chemistry NMC441 from BatPaC after version 2.1 in response
to statements by companies in the auto industry that no company in the auto industry would use
battery chemistry NMC441.
59. Battery chemistry NMC333 is an option in BatPaC versions 3.0 and 3.1 but is not an
option in BatPaC versions 2.0 and 2.1.
60. Documents in the rulemaking dockets thus referred to battery chemistries available in
different versions of BatPaC, making it impossible for CARB (or the public) to determine which
version of BatPaC the agencies used. These documents suggest that the agencies used different
versions of BatPaC or made their own modifications to a version of BatPaC, but nothing in the
rulemaking docket expressly says this is so, identifies the versions used or modifications made,
or explains how or why different versions were used or modifications made.
61. NHTSA also did not clarify why documents in the rulemaking dockets reference
battery chemistry NMC441, which is not available in version 3.0.
62. Battery chemistry has significant impact on the cost of the battery in a vehicle.
Without providing the battery chemistry the agencies actually used, neither CARB nor the public
can fully evaluate the cost estimates of the Proposal.
63. As described above in paragraphs 54-55 and in CARB’s FOIA request (Exh. A, p. 3),
NHTSA and EPA have not clearly specified which battery chemistry they used in developing the
Proposal.

15

64. Moreover, the version of BatPaC appears to have varied throughout the regulatory
process. For instance, EPA used battery chemistry NMC622 in its modeling for the 2017 Final
Determination under EPA’s Midterm Evaluation regulations. EPA, Proposed Determination of
the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle Greenhouse Gas
Emissions Standards under the Midterm Evaluation: Technical Support Document 2-108 (2016);
EPA, NHTSA, & CARB, Draft Technical Assessment Report 5-104 to 5-105 (2016).
65. Yet, NHTSA and EPA did not use battery chemistry NMC622 in developing the
Proposal.
66. Nor do EPA and NHTSA appear to have considered more advanced battery
chemistries, although this cannot be determined with certainty without the records sought by
CARB’s request number 1. For instance, NHTSA and EPA did not use battery chemistry
NMC811, which is an advancement from NMC622, in developing the Proposal.
67. NHTSA and EPA did not consider incorporating battery chemistry NMC811 into the
BatPaC version the agencies used for the Proposal.
68. NHTSA and EPA did not request ANL incorporate battery chemistry NMC811 into
the BatPaC version the agencies used for the Proposal.
69. NHTSA and EPA also have not disclosed the BatPaC cost inputs they used in
developing the Proposal. Because these costs are a variable input subject to change by the user,
CARB and the public cannot know upon what cost assumptions the Proposal is based unless the
agencies produce that information.
70. The records sought by subparts (a) and (c) of CARB’s first request were relied upon
by NHTSA and EPA in developing the Proposal. Without this information, CARB was unable
to fully evaluate and comment on the Proposal and its asserted bases. NHTSA’s failure to

16

address CARB’s request for this information constitutes a failure to make and convey a
determination as required by FOIA.
71. Subpart (b) of CARB’s request sought inputs EPA and NHTSA used to run BatPaC
for the Proposal, including “the full battery sizes, in kilowatt-hours (kWh), battery pack
configuration, and costs used for each vehicle iteration in the CAFE model.” CARB’s request
explained that battery pack configuration would include the number of cells and the electrical
topology of how those cells are arranged in the battery pack.
72. As indicated in CARB’s request, EPA had made its BatPaC input files publicly
available in at least one previous analysis of GHG emissions standards.
73. In responding to subpart (b) of CARB’s first request, NHTSA failed to provide all the
files or data the agencies input into BatPaC, and failed to provide any legal basis for withholding
them or to address the withholding in any way. NHTSA’s failure to address CARB’s request for
this information constitutes a failure to make and convey a determination as required by FOIA.
74. In responding to subpart (b) of CARB’s first request, NHTSA also failed to provide
battery pack configurations, stating that “there are no battery pack configurations.” Exh. C, p. 2.
This statement cannot be correct, as battery pack configurations are a critical input for BatPaC;
the model cannot produce battery cost estimates without inputting battery pack configurations.
75. The records sought by subpart (b) of CARB’s first request were relied upon by
NHTSA and EPA in developing the Proposal. Without this information, CARB was unable to
fully evaluate and comment on the Proposal and its asserted bases.
CARB’s Request No. 2
76. The records known to be related to CARB’s second request are vehicle registration
data obtained from IHS Markit (formerly R.L. Polk & Company) as well as aggregations and

17

modifications of this data prepared by NHTSA. In essence, these records concern which
vehicles are present within the national auto market; they are critical to understand and forecast
how the national vehicle fleet composition changes with emissions and fuel economy standards.
77. The records actually sought by CARB pursuant to its second request are the
aforementioned aggregations and modifications of vehicle registration data obtained from IHS
Markit that were prepared by NHTSA.
78. Despite having relied on the records sought by CARB’s second request in developing
the Proposal, NHTSA and EPA have not included them in the rulemaking docket or otherwise
made them available to the public.
79. In responding to CARB’s request number 2, NHTSA determined that it would not
provide the requested records on the asserted basis that they are “related to trade secrets and
commercial or financial information pursuant to FOIA Exemption 4. 49 U.S.C. § 552(b)(4).”
Exh. C, p. 2.
80. IHS Markit permits aggregations of its data to be published, provided the purchaser of
the data has also acquired publication rights.
81. NHTSA purchased data related to CARB’s second request from IHS Markit.
82. NHTSA could acquire the right to release the data related to CARB’s second request
to the public, at least in aggregated form, if it has not already acquired the right to do so.
83. EPA purchased data related to CARB’s second request from IHS Markit.
84. EPA could acquire the right to release the data related to CARB’s second request to
the public, at least in aggregated form, if it has not already acquired the right to do so.
85. Neither NHTSA nor EPA have made any attempt to segregate and produce nonexempt records responsive to CARB’s second request.

18

86. Neither NHTSA nor EPA have notified IHS Markit (or any other submitter) of
CARB’s request number 2 for aggregations and modifications of IHS Markit’s data prepared by
NHTSA.
87. Neither NHTSA nor EPA have sought or obtained any written objections to release of
its aggregations and modifications of IHS Markit’s data that are responsive to CARB’s second
request.
88. The records sought by CARB’s second request were relied upon by NHTSA and EPA
in developing the Proposal. Without this information, CARB was unable to fully evaluate and
comment on the Proposal and its asserted bases.
CARB’s Request No. 4
89. CARB’s fourth request sought a report of analysis of statistical significance
(including standard errors) concerning the sales model NHTSA developed, and on which the
EPA and NHTSA relied, for the Proposal. This sales model is critical because it projects how
newer, cleaner vehicles will enter the market and the pace at which consumers buy them (or
retain older vehicles). The model has not, however, been fully tested, and the agencies did not
document critical statistical values accompanying the model, which are necessary to fully
evaluate the model and its robustness.
90. In responding to CARB’s request number 4, NHTSA failed to provide the requested
report, including the variable stationarity and co-integration indicators. NHTSA failed to
provide any legal basis for withholding it or address the withholding in any way, instead
referring CARB to a table of data related to the report.
91. Without the requested records, the model’s ability to provide reasonable and realistic
predictions cannot be fully validated.

19

92. NHTSA’s failure to address CARB’s request number 4 for this record constitutes a
failure to make and convey a determination as required by FOIA.
93. The records sought by CARB’s fourth request were relied upon by NHTSA and EPA
in developing the Proposal. Without this information, CARB was unable to fully evaluate and
comment on the Proposal and its asserted bases.
CARB’s Request No. 8
94. CARB’s request number 8 sought records pertaining to relative impacts – including
fatalities – the agencies claimed would occur under the existing federal standards and under the
Proposal. The agencies relied heavily on their fatality projections in setting these standards and
frequently discussed them in the Proposal, so a full understanding of this information is of
considerable importance to the public.
95. In responding to CARB’s request number 8, NHTSA determined that it would not
provide the requested records on the asserted basis that they are “related to pre-decisional agency
deliberation, opinions or recommendations pursuant to FOIA Exemption 5. 5 U.S.C. §
552(b)(5).” Exh. C., p. 6.
96. The records sought by CARB pursuant to its eighth request that NHTSA has withheld
are a detailed statistical analysis, including a logistic regression derivation, of NHTSA’s point
estimates for the increase in fatalities per hundred pounds of mass reduction over a constant
footprint based on historical crash data, for model years 2004-2011 and calendar years 20062012.
97. The decision to which the records sought by CARB pursuant to its eighth request are
relevant (the Proposal) has already been made by NHTSA, as the Proposal has been noticed and

20

presented for public comment. NHTSA has not identified any other contemplated decision to
which these records are relevant.
98. The records sought by CARB pursuant to its eighth request are factual in nature, and
do not contain suggestions or recommendations.
99. In developing the regulations that NHTSA and EPA have proposed to replace with
the Proposal, the agencies made publicly available data equivalent to that sought by CARB’s
eighth request as part of the rulemaking docket.
100. Neither NHTSA nor EPA have made any attempt to segregate and produce nonexempt records responsive to CARB’s eighth request.
101. The records sought by CARB’s eighth request were relied upon by NHTSA and
EPA in developing the Proposal. Without this information, CARB was unable to fully evaluate
and comment on the Proposal and its asserted bases.
CARB’s Request No. 9
102. The records known to be related to CARB’s ninth request are data used by NHTSA
and EPA, including data obtained from IHS Markit and aggregations and modifications of this
data prepared by NHTSA, to derive a statistical model that predicts fatality rates by vehicle age
and model year. This data is needed so that CARB and the public can determine whether the
agencies correctly derived fatality estimates and verify that those estimates are well-supported.
103. The records actually sought by CARB pursuant to its ninth request are NHTSA’s
aforementioned aggregations and modifications of the data obtained from IHS Markit as well as
any other data used to derive the agencies’ estimated fatality rates.

21

104. Despite having relied on the records sought by CARB’s ninth request in developing
the Proposal, NHTSA and EPA have not included these records in the rulemaking docket or
otherwise made them available to the public.
105. In responding to CARB’s request number 9, NHTSA determined that it would not
provide the requested records on the asserted basis that they are “related to trade secrets and
commercial or financial information pursuant to FOIA Exemption 4. 49 U.S.C. § 552(b)(4).”
Exh. C, p. 6.
106. NHTSA purchased data related to CARB’s ninth request from IHS Markit.
107. NHTSA could acquire the right to release the data related to CARB’s ninth request
to the public, at least in aggregated form, if it has not already acquired the right to do so.
108. EPA purchased data related to CARB’s ninth request from IHS Markit.
109. EPA could acquire the right to release the data related to CARB’s ninth request to
the public, at least in aggregated form, if it has not already acquired the right to do so.
110. Neither NHTSA nor EPA have made any attempt to segregate and produce nonexempt records responsive to CARB’s ninth request.
111. Neither NHTSA nor EPA have notified IHS Markit (or any other submitter) of
CARB’s request number 9 for aggregations and modifications of IHS Markit’s data prepared by
NHTSA.
112. Neither NHTSA nor EPA have sought or obtained any written objections to release
of its aggregations and modifications of IHS Markit’s data that are responsive to CARB’s ninth
request.

22

113. The records sought by CARB’s ninth request were relied upon by NHTSA and EPA
in developing the Proposal. Without this information, CARB was unable to fully evaluate and
comment on the Proposal and its asserted bases.
CARB’s Request No. 12
114. CARBS’s twelfth request sought records comprising and pertaining to the Advanced
Light-Duty Powertrain and Hybrid Analysis (“ALPHA”) and the Optimization Model for
reducing Emissions of Greenhouse gases from Automobiles (“OMEGA”), which are analytical
models developed by EPA specifically to forecast automaker response to and compliance with
emissions standards. EPA used ALPHA and OMEGA in developing the existing regulations that
EPA has now proposed to replace. These models thus are of great value to the public in testing
and understanding the Proposal.
115. In a departure from prior practice, the Proposal was not developed using either
ALPHA or OMEGA.
116. Neither EPA nor NHTSA is presently using ALPHA in developing regulatory
proposals.
117. Neither EPA nor NHTSA is presently using OMEGA in developing regulatory
proposals.
118. In responding to CARB’s request number 12, NHTSA failed to provide the
requested records and failed to provide any legal basis for withholding them, instead improperly
referring the request to EPA and declining to make and convey a determination as required by
FOIA. EPA has also failed to make a determination and provide its current modelling systems to
the public.
119. EPA intermittently revises OMEGA source code.

23

120. EPA intermittently revises OMEGA input data.
121. EPA last publicly released the OMEGA source code in July 2016.
122. EPA has revised the OMEGA source code since the July 2016 public release.
123. The current OMEGA source code is not publicly available.
124. EPA last publicly released significant OMEGA input data in November 2016. EPA
released a limited set of input files in April 2017 and October 2018, but it did not release a
complete set of input data at those times.
125. Since the November 2016 public release of input data, EPA revised some OMEGA
input data in addition to the data released in April 2017 and October 2018.
126. EPA utilized OMEGA in preparing an April 2018 presentation for OIRA regarding
NHTSA’s draft of the Proposal.
127. The OMEGA version that EPA utilized to prepare this April 2018 presentation was
not the same version as the OMEGA model publicly released in July 2016.
128. EPA intermittently revises ALPHA source code.
129. EPA intermittently revises ALPHA input data.
130. EPA last publicly released the ALPHA source code in January 2017.
131. EPA has revised the ALPHA source code since the January 2017 public release.
132. The current ALPHA source code is not publicly available.
133. EPA utilized ALPHA in preparing an April 2018 presentation for OIRA regarding
NHTSA’s draft of the Proposal.
134. The ALPHA version that EPA utilized to prepare this April 2018 presentation was
not the same version as the ALPHA model publicly released in January 2017.

24

135. EPA’s April 2018 presentation for OIRA included EPA’s critique of NHTSA’s
CAFE model.
136. EPA has previously stated that it could not conclude that the CAFE model reflects
the conclusions of “research performed by EPA over the last five years.”
137. “Research performed by EPA over the last five years” includes updates made to the
ALPHA and OMEGA models since 2016.
138. Without the records sought by CARB’s twelfth request, CARB was unable to fully
evaluate and comment on the Proposal and its asserted bases or to fully understand EPA’s April
2018 critique of NHTSA’s modeling.
CARB Has Exhausted All Administrative Remedies
EPA Exhaustion
139. In a letter sent on December 19, 2018, CARB notified EPA that it had failed to
comply with the time limit under FOIA for making and conveying determinations as to CARB’s
requests.
140. As of the date that this complaint was filed, CARB has not received a determination
from EPA regarding any of CARB’s FOIA requests.
141. As of the date that this complaint was filed, EPA has not made or conveyed a
determination regarding any of CARB’s FOIA requests.
142. CARB has exhausted all administrative remedies as to its requests to EPA because
EPA has failed to comply with the time limit under FOIA for making and conveying
determinations as to CARB’s requests.

25

NHTSA Exhaustion
143. As of the date that this complaint was filed, CARB has not received a determination
from NHTSA regarding CARB’s FOIA requests numbers 1(b), 4, and 12.
144. As of the date that this complaint was filed, NHTSA has not made or conveyed a
determination regarding CARB’s FOIA requests numbers 1(b), 4, and 12.
145. By letter dated December 19, 2018, CARB administratively appealed NHTSA’s
responses to CARB’s request numbers 1, 2, 4, 8, 9, and 12. A true and correct copy of this letter
is attached hereto as Exhibit D.
146. CARB’s appeal to NHTSA complied in all respects with all applicable laws and
agency regulations, including but not limited to 5 U.S.C. § 552 and NHTSA’s regulations
implementing FOIA.
147. As of the date that this complaint was filed, CARB has not received a response from
NHTSA regarding CARB’s appeal.
148. CARB has exhausted all administrative remedies as to its requests to NHTSA
because NHTSA has failed to comply with the time limit under FOIA for making and conveying
determinations as to some of CARB’s requests, and has failed to comply with the time limit
under FOIA for responding to CARB’s appeal.
149. EPA and NHTSA continue to move ahead to finalize the Proposal, even though the
public has not been provided critical information required to evaluate it fully.
CLAIMS FOR RELIEF
First Claim – EPA’s Violation of FOIA
150. Paragraphs 1-149, inclusive, are realleged and incorporated herein by reference.

26

151. EPA’s failure to make and convey determinations as to CARB’s September 11,
2018 requests for records violates FOIA (5 U.S.C. §§ 552(a)(6)(A)(i) and 552(a)(6)(B)(i)) and
EPA’s own corresponding regulations (40 C.F.R. § 2.104(a) and (d)).
152. CARB has a statutory right to have EPA process its requests in a manner that
complies with FOIA. 5 U.S.C. § 552(a)(3). EPA violated CARB’s rights in this regard when it
unlawfully failed to undertake a search that is reasonably calculated to locate all records that are
responsive to CARB’s September 11, 2018 requests for records numbers 1-11.
153. EPA’s failure to produce all non-exempt records (or reasonably segregable nonexempt portions of any responsive records deemed to be exempt) responsive to CARB’s
September 11, 2018 requests for records violates FOIA (5 U.S.C. § 552(a)(3)(A)).
Second Claim – NHTSA’s Violation of FOIA
154. Paragraphs 1-153, inclusive, are realleged and incorporated herein by reference.
155. NHTSA’s failure to make and convey determinations as to CARB’s September 11,
2018 requests for records numbers 1(b), 4, and 12 violates FOIA (5 U.S.C. §§ 552(a)(6)(A)(i)
and 552(a)(6)(B)(i)), and NHTSA’s own corresponding regulations (49 C.F.R. § 7.31(a)(2)).
156. NHTSA’s failure to respond to CARB’s December 19, 2018 appeal of NHTSA’s
October 23, 2018 response to CARB’s September 11, 2018 letter requesting records violates
FOIA (5 U.S.C. §§ 552(a)(6)(A)(ii) and 552(a)(6)(B)(i)), and NHTSA’s own corresponding
regulations (49 C.F.R. § 7.33(a)(2)).
157. CARB has a statutory right to have NHTSA process its requests in a manner that
complies with FOIA. 5 U.S.C. § 552(a)(3). NHTSA violated CARB’s rights in this regard when
it unlawfully failed to undertake a search that is reasonably calculated to locate all records that
are responsive to CARB’s September 11, 2018 requests for records numbers 1(b), 4, and 12.

27

158. NHTSA’s failure to produce all non-exempt records (or reasonably segregable nonexempt portions of any responsive records deemed to be exempt) responsive to CARB’s
September 11, 2018 requests for records numbers 1(a), 1(c), and 4 violates FOIA (5 U.S.C. §
552(a)(3)(A) and NHTSA’s own corresponding regulations (49 C.F.R. § 7.23(b), (d)).
159. NHTSA’s withholding of records responsive to CARB’s September 11, 2018
requests numbers 2, 8, and 9 based on Exemptions 4 and 5 was improper and violates FOIA (5
U.S.C. §§ 552(a)(3)(A) and 552(b), and NHTSA’s own corresponding regulations (49 C.F.R. §
7.23(b), (d)).
PRAYER FOR RELIEF
WHEREFORE, CARB respectfully requests that the Court grant relief as follows:
1. Declare that EPA’s and NHTSA’s failures to timely make and convey determinations
regarding CARB’s respective FOIA requests are unlawful;
2. Declare that NHTSA’s failure to timely respond to CARB’s appeal is unlawful;
3. Declare that EPA’s and NHTSA’s failures to produce non-exempt records (and
reasonably segregable non-exempt portions of any responsive records deemed to be exempt)
responsive to CARB’s respective FOIA requests are unlawful;
4. Order EPA and NHTSA to each immediately conduct searches that are reasonably
calculated to locate all records responsive to CARB’s requests;
5. Order EPA and NHTSA to each make available to CARB all non-exempt agency
records that are responsive to CARB’s requests, as well as all reasonably segregable non-exempt
portions of any responsive records deemed to be exempt, on a rolling basis as the records are
located, without charging search or copying fees;

28

6. Order EPA and NHTSA to each complete their respective productions of records to
CARB by a date certain;
7. Order EPA and NHTSA to each produce indexes identifying any responsive agency
records (or portions thereof) being withheld as exempt from disclosure, and the basis for the
withholding, promptly upon determining to withhold such records;
8. Retain jurisdiction over this action to rule on any assertions by EPA or NHTSA that
any responsive records, in whole or in part, are exempt from disclosure;
9. Award CARB its costs and reasonable attorneys’ fees in this action as provided by 5
U.S.C. § 552(a)(4)(E); and
10. For such other relief as the Court may deem just and proper.

Dated: April 5, 2019

Respectfully submitted,
XAVIER BECERRA
Attorney General of California
MYUNG PARK
Supervising Deputy Attorney General
ELAINE MECKENSTOCK
Deputy Attorney General

/s/ Ryan R. Hoffman
RYAN R. HOFFMAN
Deputy Attorney General
California Department of Justice
Office of the Attorney General
455 Golden Gate Avenue, Suite 11000
San Francisco, CA 94102
[phone]
[e-mail]
California State Bar Number 283297
Attorneys for Plaintiff California Air
Resources Board

29

INDEX TO EXHIBITS

Exhibit A – CARB’s September 11, 2018 FOIA request letter to EPA and NHTSA
Exhibit B – EPA’s October 23, 2018 response letter to CARB
Exhibit C – NHTSA’s October 23, 2018 response letter to CARB
Exhibit D – CARB’s December 19, 2018 appeal letter to NHTSA

Exhibit A

�'1

CALIFORNIA

f � AIR RESOURCES BOARD

Mary D. Nichols, Chair
Matthew Rodriquez, CalEPA Secretary
Edmund G. Brown Jr., Governor

September 11, 2018

Andrew K. Wheeler
Acting Administrator
United States Environmental Protection Agency
Office of the Administrator
Mail Code 1101A
1200 Pennsylvania Avenue, NW
Washington, D.C. 20460
Heidi King
Deputy Administrator
National Highway Traffic Safety Administration
United States Department of Transportation
1200 New Jersey Avenue, SE
Washington, D.C. 20590

RE: REQUEST FOR DOCUMENTS IN SUPPORT OF:
REQUEST FOR EXTENSION OF COMMENT PERIOD AND ADDITIONAL PUBLIC
HEARINGS REGARDING JOINT PROPOSED RULE TO ROLL BACK VEHICLE
GREENHOUSE GAS EMISSIONS AND CORPORATE AVERAGE FUEL ECONOMY
STANDARDS FOR MODEL YEARS 2021-2026 LIGHT-DUTYVEHICLES
Docket IDs: NHTSA-2018-0067 / EPA-HQ-OAR-2018-0283
Dear Acting Administrator Wheeler and Deputy Administrator King:
On August 27, 2018, the California Air Resources Board (GARB), along with the
Attorneys General of the State of California and several other states, and several state
agencies, requested that the United States Environmental Protection Agency (U.S. EPA)
and the National Highway Traffic Safety Administration (NHTSA) extend the comment
periods for the joint proposed rule referenced above, 1 and the associated draft
Environmental Impact Statement. The proposal departs from legal precedent and from
the factual record as GARB understands it, and needs careful evaluation. CARB writes
separately in its capacity as a co-regulator with U.S. EPA, specifically recognized in the
Clean Air Act to protect public health from motor vehicle air pollution and engines, to
request information necessary to evaluate the proposal.
1

Published at 83 Fed.Reg. 42,986 (August 24, 2018).

arb.ca.gov

1001 I Street• P.O. Box 2815 • Sacramento, California 95812

[phone]

Mr. Andrew Wheeler, Ms. Heidi King
September 11, 2018
Page2
NHTSA and U.S. EPA must identify and make available the technical studies and data
on which it is relying to propose these rules.2 As we wrote in the request for a
reasonable comment period, the modeling, assumptions, and analysis underlying these
proposals are dramatically different from that of previous, similar rulemakings. The
explanation for this proposal, despite encompassing hundreds of pages, is not clearly
presented or adequately supported, as elaborated on below. Further, unlike previous
federal proposals for regulations of emissions from motor vehicles, CARB has not been
involved in discussions developing the proposal, contrary to prior commitments from
U.S. EPA and NHTSA to collaborate with California on these standards, and past
practice developing motor vehicle emission standards. 3 Thus, CARB is unable fully to
evaluate the federal proposal based on the information made available to date and other
stakeholders are likely to encounter similar difficulties.
Under the Freedom of Information Act (FOIA) as well as the Administrative Procedures
Act, CARB requests U.S. EPA and NHTSA make the information identified below
publicly available in the dockets for the proposed rule within20 days of your receipt of
this letter. Please make the requested information available on a rolling basis as soon as
it is identified, without awaiting identification or production of other requested
information. Absent the requested information, along with the requested extensions to
the comment periods for the proposal and associated draft Environmental Impact
Statement, neither CARB nor other interested persons are being provided a reasonable
opportunity to evaluate the proposed rule and its asserted basis.
If any of the requested information is asserted to be exempt from disclosure, please
provide an index of those records as required by Vaughn v. Rosen, 484 F.2d820 (D.C.
Cir. 1973). This index should describe each document claimed as exempt from
disclosure with sufficient information to allow a reasoned judgment as to whether it is
properly exempt under FOIA.4
1. Information about the models and data used to estimate battery costs for electrified
vehicles.
a. The proposal and the Preliminary Regulatory Impact Analysis (PRIA),
NHTSA-2018-0067 -1972, reference the Argonne National Laboratories'
(ANL) website for the BatPaC model for estimating battery costs for vehicles,
and state that the agencies used "an up-to-date version" of the model, but do
not identify the version. (See, e.g., 83 Fed.Reg. 42,985, 43,002 (August 24,
1
Connecticut Light & Power Co. v. Nuclear Regulatory Com., 673 F.2d 525, 530-531 (D.C. Cir. 1982); 42 U.S.C. §
7607(d)(3) [notice of proposed rulemaking "shall be accompanied by a statement of its basis and purpose" including
"the factual data on which the proposed rule is based; the methodology used in obtaining and in analyzing the data; and
the major legal interpretations and policy considerations underlying the proposed rule."].
3
See, e.g., 77 Fed.Reg. 62,624, 62,632, 62,784-62,785 (Oct. 15, 2012) [discussing coordination with CARB to develop
the standards at issue and for changes to standards].
4
Founding Church a/Scientology of Washington, D.C. v. Bell, 603 F.2d 945,949 (D.C. Cir. 1979).

Mr. Andrew Wheeler, Ms. Heidi King
September 11, 2018
Page 3
2018).) U.S. EPA and NHTSA have posted to the dockets for this action a
document describing how BatPaC was developed, but this document
appears to be from 2012. It does not state which version of BatPaC NHTSA
and U.S. EPA used to estimate battery costs. See "Modeling the
Performance and Cost of Lithium-Ion Batteries for Electric-Drive Vehicles,
Argonne National Laboratory, ANL-12/55," NHTSA Docket ID#: NHTSA2018-0067-1692; EPA Docket ID#: EPA-HQ-OAR-2018-0283-0764.
b. U.S. EPA and NHTSA should make available the information specifying the
full battery sizes, in kilowatt-hours (kWh), battery pack configuration, and
costs used for each vehicle iteration in the CAFE model. See PRIA,
Electrification Technologies, Technology Overview, section 6.3.8.1, p. 357.
The PRIA states that NHTSA posted ANL vehicle files that have battery pack
sizes and costs for each vehicle, but there is no additional information about
battery pack configuration (e.g., the number of cells, and the electrical
topology of how those cells are arranged in the battery pack), nor do they
directly reference where the files are posted. NHTSA and U.S. EPA have not
posted the BatPaC model file(s) that were used. ANL cost and battery size
data referenced in the PRIA, p. 358, footnote 325, but the footnote refers to a
docket identification number that is not available. Previously, in support of the
draft Technical Assessment Report and Proposed Determination on the
Appropriateness of the Model Year 2022-2025 Light-duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation
(Proposed Determination), EPA-420-R-16-020, November 2016, U.S. EPA
posted the BatPaC files that it used.
c. The proposal and PRIA provide conflicting information about which battery
chemistries the agencies considered. For instance, the proposal and PRIA
refer to NMC441-Gr chemistry for both plug-in hybrid-electric vehicles and
battery-electric vehicles, but the ANL summary refers to NMC333. See, e.g.,
PRJA, pp. 372, Table 6-27, 373 ["We selected NMC441 as choice of
chemistry for PHEVs and BEVs. NMC441 more suitable for high energy
batteries capable of discharge rates."]. The Excel file titled "ANL-Summary of
Main Component Performance Assumptions NPRM" has a tab labeled
'Description - BatPac' with the same table listed as in the PRIA, except the
chemistry listed for PHEVs and BEVs is NMC333-G instead of NMC441-Gr.
See Docket ID Nos. EPA-HQ-OAR-2018-0283-0054 and NHTSA-2018-00670003. The proposal and PRIA do not directly reference this file.
This information is required to replicate and evaluate whether the modeling underlying
the proposal is appropriate, considering the various vehicle and technology types.
2. The PRIA references Polk registration data, including survival rates aggregated by
model year, calendar year, and body style. These data are needed to verify the

Mr. Andrew Wheeler, Ms. Heidi King
September 11, 2018
Page 4
coefficients of the new model predictions for vehicle retirement (scrappage), but
have not been made available. See, e.g., PRIA at pp. 1008, 1014, 1023, fig. 8-23,
1025, fig. 8-24, and 1027, fig. 8-25.
3. New vehicle sales and price data referenced in the proposal. This includes:
a. Data provided by the National Automobile Dealers Association (NADA) and
others. See 83 Fed.Reg. at 43,095; PRIA, pp. 1017-1018.
b. Data describing historical transaction prices, and quarterly new vehicle sales
data used to develop the dynamic new sales model. See PRIA, pp. 954-961.
c. Economic data used to develop the autoregressive distributed-lag (ARDL)
model that predicts new vehicle sales and is used in the CAFE model. See
83 Fed.Reg. at 43,074.
This da\a is necessary to evaluate the proposal's predictions for fleet population, sales,
and fatalities.
4. Report of analysis of the standard errors and significance of the ARDL sales model
coefficients, F-statistic and R2 of the overall model, and variable stationarity and co­
integration indicators. This information is needed to verify the statistical significance
and errors of the coefficients used in the Volpe model. The coefficients for the ADRL
sales model listed on p. 957, Table 8-1 of the PRIA, are not consistent with those
implemented in the model. See CAFE Model Documentation, PRIA, p. 78, Table 17,
available at: https://www.nhtsa.gov/corporate-average-fuel-economy/compliance-and­
effects-modeling-system ["2018 NPRM for Model Years 2021-2026 Passenger Cars
and Light Trucks," Model Documentation].
5. The coefficients for the dynamic fleet share equation described in the CAFE Model
Documentation on p. 79. These are not listed anywhere. Additionally, according to
the PRIA on p. 955, the model was based on EIA's National Energy Modeling
System (NEMS), but no reference is provided for the NEMS model. This information
is necessary to evaluate the equation used in the model.
6. Cited reference Anders Lie, Swedish Transport Administration, Peer Review of
NHTSA's November 2011 Preliminary Report "Relationships Between Fatality Risk,
Mass, and Footprint in Model Year 2000-2007 Passenger Cars and LTVs." This
document reviews the 2011 study by Kahane that NHTSA relies upon, regarding the
effects of mass reduction on fatality risk. See 83 Fed.Reg. at 43,112, n. 307. It is
also item 0035 in the NHTSA-2010-0152 docket. However, attempting to access the
document on the docket website results in a server error. See
https://www.regulations.gov/document?D=NHTSA-2010-0152-0035, attempted
August 28, 2018.

Mr. Andrew Wheeler, Ms. Heidi King
September 11, 2018
Page 5
7. CAFE model output file Annual Societal Effects Report. See CAFE Model
Documentation, infra. This data describes predicted fatalities by regulatory class,
body style, and weight threshold of vehicle. This detailed information is necessary to
evaluate the fatalities computation in the CAFE model.
8. The agencies' detailed explanation and derivation of their point estimates for the
increase in fatalities per hundred pounds of mass reduction over a constant footprint
based on historical crash data, for model years 2004-2011 and calendar years
2006-2012. Previously, these details were provided in a separate report such as the
"2016 Puckett and Kindelberger report." No such report is available this time. The
PRIA only provides a summary table of the results of this analysis, yet states an
"updated analysis" exists. See PRIA, p. 1357, section 11 .4.
9. Data used by the agencies to derive the new statistical model that predicts fatality
rates by vehicle age. See PRIA Table 11-21, p. 1397. The coefficients of the model
are provided, but without the data it is not possible to evaluate whether the
coefficients were properly derived. Additionally, the coefficients provided in the
PRIA are different (significant digits and sign changes) than those identified in the
actual model source code {which are also commented out such that they are non­
functional) and are different from the model year based coefficients used in the input
files. This renders unclear what coefficients the analysis in the NPRM is based
upon.
10. So-called "off model" analytical tools the agencies used to summarize and tabulate
the results of the CAFE model. See 83 Fed.Reg. at 43,256, n. 595. These tools and
calculations analyzed various components of social and private costs and benefits,
as well as other factors. These analytical tools are supposedly available on
NHTSA's website but we have not located them.
11. Input files used for the Autonomie model for various engine technologies that U.S.
EPA and NHTSA rely on to calculate the efficiency improvements of various
technologies, such as turbocharging and high-compression-ratio (Atkinson) engines.
The input and output files are required to be able to understand what U.S. EPA and
NHTSA relied on as representative of these engine technologies, and to confirm
correct efficiency levels were calculated for the various technologies, considering
the current state of the art.
12. Modeling tools developed by U.S. EPA including:
a. All files necessary to utilize - with the Advanced Light-Duty Powertrain and
Hybrid Analysis (ALPHA) and the Optimization Model for reducing Emissions
of Greenhouse gases from Automobiles (OMEGA) - the response surface
equations developed by U.S. EPA as identified or referenced in: "Peer
Review of EPA's Response Surface Equation Report" (Docket ID No. EPA­
HQ-OAR-2018-0283-0025); and SAE paper 2018-01-1273 authored by U.S.

Mr. Andrew Wheeler, Ms. Heidi King
September 11, 2018
Page 6
EPA (Docket ID No. EPA-HQ-OAR-2018-0283-0028).
b. All new or modified input files, source code, and executable files for U.S.
EPA's OMEGA model developed since the release of the Proposed
Determination in late 2016.
c. All current and new input files, source code, and executable files for ALPHA
used for the Proposed Determination in late 2016 and/or modified since then.
d. All current and new pre-processors and their inputs used for the Proposed
Determination in late 2016 or modified since then to categorize, sort, and
rank technology packages and costs for use with OMEGA.
We understand that these models evaluate the cost and effectiveness of available
technologies to meet greenhouse gas emissions targets. We also understand that
NHTSA and U.S. EPA have previously relied on these models. These files and data are
thus necessary for CARS and the public to be able to verify the agencies' claims that
that the CAFE model has advantages to the U.S. EPA models and to consider the
relevance and make sense of the docketed materials, including the peer review and SAE
paper identified above.
It is unreasonable for U.S. EPA and NHTSA to expect interested persons to evaluate the
massive changes in outcomes, models, approaches, inputs, and analyses in a 60-day
comment period (and in even less time for the draft environmental impact statement).
This is especially apparent when not even CARS, despite its considerable expertise in
these matters, is able to perform a complete review.
CARS also requests a waiver of searching and copying fees for this request. CARS is a
non-commercial, governmental organization, and should not be subject to fees for this
request. The information requested is likely to significantly contribute to public
understanding of NHTSA and U.S. EPA's proposed rules. CARS is a co-regulator of motor
vehicle emissions with U.S. EPA, and has coordinated with U.S. EPA and NHTSA on the
regulations that the agencies are proposing to change. CARS has a significant interest in
the proposed action, and the requested information will enable CARS to evaluate the
potential impacts of the proposed rule.
Under FOIA, agencies must waive such fees in instances like this where disclosure is
likely to contribute to public understanding of the operations and activities of the
government and disclosure is not primarily in the commercial interest of the requester. 5
You may contact me at [phone] or [e-mail], or Senior Attorney
Pippin C. Srehler at [phone] or [e-mail], to discuss any of these
issues. If this request for a fee waiver is denied, please contact Mr. Brehler before
5

See 5 U.S.C. § 552(a)(4)(A)(iii); 40 C.F.R. §2.107(1)(1) [U.S. EPA adoption of these requirements].

Mr. Andrew Wheeler, Ms. Heidi King
September 11, 2018
Page 7
incurring any costs in responding to this request. We look forward to receiving your
response to these requests within twenty business days, as required by FOIA.

�//_))�
Ellen M. Peter
Chief Counsel
Executive Office
California Air Resources Board
CC:
Christopher Lieske
Office of Transportation and Air Quality,
Assessment and Standards Division
U.S. Environmental Protection Agency
2000 Traverwood Drive
Ann Arbor, Michigan 48105
[e-mail]
National Freedom of Information Officer
U.S. Environmental Protection Agency
1200 Pennsylvania Ave., NW (2822T)
Washington, D.C. 20460
James Tamm
Office of Rulemaking,
Fuel Economy Division
National Highway Traffic Safety Administration
1200 New Jersey Avenue SE
Washington, D.C. 20590
Freedom of Information Act Public Liaison
National Highway Traffic Safety Administration
1200 New Jersey Ave., SE
West Building, 41-304
Washington, D.C. 20590

Exhibit B

Exhibit C

October 23, 2018
Ms. Ellen Peter
California Air Resources Board
1001 I Street
Sacramento, CA 95812
Via email to: [e-mail]
Re: Freedom of Information Act (FOIA) Request ES18-003395
Dear Ms. Peter:
This responds to your September 11, 2018 FOIA request that was submitted to
www.regulations.gov. Your request seeks records related to the Safer Affordable Fuel-Efficient
(SAFE) Vehicles Rule that amends the existing Corporate Average Fuel Economy (CAFE)
standards.
Information related to your FOIA request concerning the SAFE Vehicles Rule is publicly available
on NHTSA’s website at https://www.nhtsa.gov/corporate-average-fuel-economy/safe. The website
and links available there have data related to the 2018 Notice of Proposed Rulemaking for Model
Years 2021-2026 Passenger Cars and Light Trucks (proposal) that are responsive to your request.
To assist you in locating the publicly available records that you seek, please see the responses to the
twelve items listed in your request below:
1. Models and data used to estimate battery costs for electrified vehicles:
a. Information regarding the BatPaC that the NTHSA and EPA used to estimate battery
costs.
Response: The BatPaC version 3.0 model was used to estimate battery costs.
NHTSA does not maintain the BatPaC model. The model is maintained by the U.S.
Department of Energy’s Argonne National Laboratories (ANL). To obtain a copy of
the BatPaC version 3.0 model, please contact ANL directly.
b. Information specifying the full battery sizes, in kilowatt-hours (kWh), battery pack
configuration, and costs used for each vehicle iteration in the CAFE model.
Response: The full battery sizes in kWh and costs are available in files in the
docket for each “tech class.” Information for peak battery power, battery total
energy in kWh, and battery pack direct manufacturing cost can be found in

2

columns AN-AR for each of the files listed below. Please note that there are no
battery pack configurations.
ANL Compact Non Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1855
ANL Compact Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1856
ANL Small SUV NonPerfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1486
ANL Small SUV Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1485
ANL Pickup NonPerfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1661
ANL Pickup Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1487
ANL Midsize SUV Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1662
ANL Midsize SUV NonPerfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1492
ANL Midsize NonPerfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1494
ANL Midsize Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1663
c. Information regarding the battery chemistries referenced in the proposal and
Preliminary Regulatory Impact Analysis.
Response: NHTSA and EPA used the battery chemistries associated with the
BatPaC version 3.0 model.
2. Polk registration data, including survival rates aggregated by model year, calendar year, and
body style.
Response: The Polk registration data is proprietary information and is being withheld in
its entirety from disclosure because it is related to trade secrets and commercial or
financial information pursuant to FOIA Exemption 4. 49 U.S.C. § 552(b)(4). To

4

c. Economic data used to develop the autoregressive distributed-lag (ARDL) model
that predicts new vehicle sales and is used in the CAFE model.
Response: Please see Attachment B.
4. Report of analysis of the standard errors and significance of the ARDL sales model
coefficients, F-statistic and R2 of the overall model, and variable stationarity and
cointegration indicators. The coefficients for the ADRL sales model listed on page 957,
Table 8-1 of the PRIA, are not consistent with those implemented in the model.
Response: NHTSA has identified an error in Table 8-1. The agency published a revised
PRIA correcting the error in Table 8-1, which is reproduced below and available on page
949 of the revised PRIA at
https://www.nhtsa.gov/sites/nhtsa.dot.gov/files/documents/ld cafe co2 nhtsa 2127al76_epa_pria_181016.pdf.

5. The coefficients for the dynamic fleet share equation described in the CAFE model
documentation on page 79.
Response: Records for the dynamic fleet share equation is provided in the Energy
Information Administration’s 2016 National Energy Modeling System (NEMS)
documentation beginning on page 48 and can be found at
https://www.eia.gov/outlooks/aeo/nems/documentation/transportation/pdf/m070(2016).p
df.
A copy of the Fortran code for the NEMS implementation is detailed below. The NEMS
implementation in the CAFE model is available in the public source code and described
in the CAFE Model Documentation on page 78 at ftp://ftp.nhtsa.dot.gov/CAFE/20212026_CAFE_NPRM/CAFE_Model/CAFE_Model/CAFE_Model_Documentation_NPR
M_2018.pdf.

5

!...Estimate percent of total light vehicles <8,500 GVW that are
cars
!...coefficients used to determine new car sales
CCONSTANT = 3.4468
CRHO
= 0.8903
CINC
= 0.0
! disposable income
CFUEL
= 0.1441 ! fuel cost
CHP
= -0.4436 ! vehicle horsepower
CWGT
= -0.0994 ! vehicle weight
CMPG
= -0.5452 ! vehicle fuel economy
CDUMM
= -0.1174 ! dummy variable
!...coefficients used to
TCONSTANT = 7.8932
TRHO
= 0.3482
TINC
= 0.0
TFUEL
= -0.4690
THP
= 1.3607
TWGT
= -1.5664
TMPG
= 0.0813
TDUMM
= 0.6192

determine new light truck sales

!
!
!
!
!
!

disposable income
fuel cost
vehicle horsepower
vehicle weight
vehicle fuel economy
dummy variable

if(n.le.25)then
CARSHARE(N) = CARLTSHR(N)
TRKSHARE(N) = 1.0-CARLTSHR(N)
else
CARSHRT(N) = EXP(CCONSTANT*(1-CRHO)+(CRHO*LOG(CARSHARE(N1))) +
&
CINC *(LOG(INC00$16(11,N))(CRHO*LOG(INC00$16(11,N-1)))) + &
CFUEL*(LOG(PMGTR00$C(N)) (CRHO*LOG(PMGTR00$C(N-1)))) +
&
CHP *(LOG(AHPCAR(N-1))
(CRHO*LOG(AHPCAR(N-2)))) +
&
CWGT *(LOG(AWTCAR(N-1))
(CRHO*LOG(AWTCAR(N-2)))) +
&
CMPG *(LOG(TRUEMPG(1,N-1)) (CRHO*LOG(TRUEMPG(1,N-2)))) +
&
CDUMM*(log(DUMM(N)) - (CRHO*log(DUMM(N-1)))))
TRKSHRT(N) = EXP(TCONSTANT*(1-TRHO)+(TRHO*LOG(TRKSHARE(N1))) +
&

6

6. Cited reference to Anders Lie, Swedish Transport Administration, Peer Review of NTHSA’s
November 2011 Preliminary Report “Relationships Between Fatality Risk, Mass, and
Footprint in Model Year 2000-2007 Passenger Cars and LTVs.”
Response: The link located on www.regulations.gov appears to be functioning. The
Swedish Transport Administration’s peer review is available at
https://www.regulations.gov/document?D=NHTSA-2010-0152-0035.
7. CAFE model output file Annual Societal Effects Report.
Response: The CAFE model does not use output societal effects by regulatory class,
body style, or weight threshold of vehicle. The basis and coefficients for the calculations
are provided in Section II.F of the 2018 NPRM beginning on page 43106 and can be
found at https://www.gpo.gov/fdsys/pkg/FR-2018-08-24/pdf/2018-16820.pdf. The exact
equations used for the computations are in Section 6 of the NPRM’s CAFE Model
Documentation and can be found at ftp://ftp.nhtsa.dot.gov/CAFE/20212026_CAFE_NPRM/CAFE_Model/CAFE_Model/CAFE_Model_Documentation_NPR
M 2018.pdf.
8. The agencies’ detailed explanation and derivation of their point estimates for the increase
in fatalities per hundred pounds of mass reduction over a constant footprint based on
historical crash data, for model years 2004-2011 and calendar years 2006-2012.
Response: The “updated analysis” referenced in the PRIA at p. 1357, refers to
information available in the PRIA in Section 11.4, pps 1345-51. NHTSA intends to
publish a technical summary of the logistic regression analysis and its results in the near
future. In addition, NHTSA intends to publish a report similar to the “2016 Puckett and
Kindelberger report” that will describe the methodological process by which the results
were derived. Accordingly, I am withholding these records as exempt from the statutory
disclosure requirement that contains information related to pre-decisional agency
deliberation, opinions or recommendations pursuant to FOIA Exemption 5. 5 U.S.C. §
552(b)(5).
9. Data used by the agencies to derive the new statistical model that predicts fatality rates by
vehicle age.
Response: The data used to derive the new statistical model for fatality rates was
obtained from IHS Markit (formerly R.L. Polk & Company) is proprietary information.
Thus, I am withholding the data in its entirety from disclosure because it is related to
trade secrets and commercial or financial information pursuant to FOIA Exemption 4. 49
U.S.C. § 552(b)(4). To request a copy of the data, please contact IHS Markit directly.

7

10. “Off model” analytical tools the agencies used to summarize and tabulate the results of the
CAFE model.
Response: The Supplementary R scripts for table calculations was used to summarize
and tabulate the results of the CAFE model and can be found at
https://www.nhtsa.gov/corporate-average-fuel-economy/compliance-and-effectsmodeling-system.
11. Input files used for the Autonomie model for various engine technologies that the EPA and
NTHSA rely on to calculate the efficient improvements of various technologies, such as
turbocharging and high-compression-ratio (Atkinson) engines.
Response: The Autonomie simulation results and assumptions files are available in the
docket in the supporting document section at
https://www.regulations.gov/docket?D=NHTSA-2018-0067. An overview of the
Autonomie model is described in detail in section 6.1 of the PRIA on page 188.
ANL – All Assumptions Summary
https://www.regulations.gov/document?D=NHTSA-2018-0067-0005
ANL – Data Dictionary for Input Files
https://www.regulations.gov/document?D=NHTSA-2018-0067-0004
ANL – Model Documentation
https://www.regulations.gov/document?D=NHTSA-2018-0067-1490
ANL – Summary of Main Component Performance Assumption
https://www.regulations.gov/document?D=NHTSA-2018-0067-0003
Output files:
ANL Compact NonPerfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1855
ANL Compact Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1856
ANL Small SUV NonPerfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1486
ANL Small SUV Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1485
ANL Pickup NonPerfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1661

8

ANL Pickup Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1487
ANL Midsize SUV Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1662
ANL Midsize SUV Non Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1492
ANL Midsize NonPerfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1494
ANL Midsize Perfo
https://www.regulations.gov/document?D=NHTSA-2018-0067-1663
12. Modeling tools developed by the EPA including:
a. All files necessary to utilize – with the Advanced Light Duty Powertrain and Hybrid
Analysis (ALPHA) and the Optimization Model for reducing Emission of
Greenhouse gases from Automobiles (OMEGA) – the response surface equations
developed by the EPA as identified or referenced in “Peer Review of EPA’s
Response Surface Equation Report” and SAE paper 2018-01-1273.
b. All new or modified input files, source code, and executable files for the EPA’s
OMEGA model developed since the release of the Proposed Determination in late
2016.
c. All current and new input files, source code, and executable files for ALPHA used
for the Proposed Determination in late 2016 and/or modified since then.
d. All current and new pre-processors and their inputs used for the Proposed
Determination in late 2016 or modified since then to categorize, sort, and rank
technology packages and costs for use with OMEGA.
Response: Records related to the EPA’s modeling tools fall under that agency’s
jurisdiction and must be requested from the EPA directly.
In your letter, you requested a waiver (or reduction) of fees associated with processing of the
request. However, the fees to process your request did not exceed $20.00, which is the minimum
amount that NHTSA assesses fees. Pursuant to the provisions of 49 CFR Part 7, there is no charge
for this response.
I am the person responsible for this determination. If you wish to appeal this decision, you may do
so by writing to the Chief Counsel, National Highway Traffic Safety Administration, 1200 New
Jersey Avenue, S.E., West Building, W41-227, Washington, DC 20590, pursuant to 49 C.F.R. §

9

7.32(d). Alternatively, you may submit your appeal via electronic mail to
[e-mail]. An appeal must be submitted within 90 days from the date of this
determination. It should contain any information and argument upon which you rely. The decision
of the Chief Counsel will be administratively final.
You also have the right to seek dispute resolution services from NHTSA’s FOIA Public Liaison,
Mary Sprague, who may be contacted at [phone] or by electronic mail at
[e-mail]. Further dispute resolution is available through the Office of Government
Information Services (OGIS). You may contact OGIS on [phone] or by electronic mail at
[e-mail].
Very Truly Yours,

Senior Attorney
Attachments

Exhibit D

r�

�/--,\ CAL IF ORN IA
AIR RESOURCES BOARD

Mary D. Nichols, Chair
Matthew Rodriquez, CalEPA Secretary
Edmund G. Brown Jr., Governor

Sent via Overnight Delivery
December 19, 2018
Christopher Lieske
U.S. Environmental Protection Agency
EPA Docket Center (EPA/DC),
EPA West Room 8102
1301 Constitution Avenue NW
Washington, D.C. 20460
James Tamm
National Highway Traffic Safety Administration
U.S. Department of Transportation
West Building, Ground Floor, Room W12-140,
1200 New Jersey Avenue SE
Washington, D.C. 20590
Attention: NHTSA Docket ID Nos. NHTSA-2018-0067 and NHTSA-2017-0069
U.S. EPA Docket ID No. EPA-HQ-OAR-2018-0283
Freedom of Information Act (FOIA) Request Nos. ES18-003395, EPA-HQ2018-011521; FOIA Appeal
Re:

Safer Affordable Fuel-Efficient (SAFE) Vehicles Rule for Model Years 2021-2026
Passenger Cars and Light Trucks

Dear Mr. Lieske and Mr. Tamm:
The California Air Resources Board (CARB) is writing to identify substantial procedural
deficiencies regarding information used to support proposed federal relaxations in the
existing passenger car and light-duty truck greenhouse gas (GHG) emissions and
corporate average fuel economy (CAFE) standards in the Safer Affordable Fuel-Efficient
(SAFE) Vehicles Rule for Model Years 2021-2026 Passenger Cars and Light Trucks. 1
This letter follows our substantive comments on the proposed rules, filed in October, 2
and our information request letter (submitted in part under the Freedom of Information
83 Fed.Reg. 42,986 (August 24, 2018).
Please also see our detailed comments on the SAFE Vehicles Rule, Docket No. EPA-HQ-OAR-20180283-5054 (CARB Detailed Comments), and our comments filed on the Draft Environmental Impact
Statement (DEIS) for the proposal, Docket No. NHTSA-2017-0069-0625 ("DEIS Comments").
1

2

arb.ca.gov

1001 I Street• P.O. Box 2815 • Sacramento, California 95812

[phone]

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 2
Act (FOIA)) of September 11, 2018. 3 The information we requested has not been fully
provided, even though the National Highway Transportation Safety Administration
(NHTSA) and the U.S. Environmental Protection Agency (U.S. EPA), collectively
referred to here as the Agencies, assert their proposals are based on this information.
Moreover, the Agencies may be considering relying upon inaccurate information
provided by some commenters. This letter highlights these deficiencies, and includes
an appeal of several of NHTSA's initial FOIA determinations. 4
The proposed SAFE Vehicles Rule is confounding, substantively and procedurally. As
explained in our previously submitted comments (and the comments of many others),
the proposal cannot be justified based on the information made available by the
Agencies and, in fact, appears to be unjustifiable. 5 Further, the opportunity to evaluate
the proposal remains unreasonable and inadequate. We reiterate that additional
information and additional opportunity for public comment are necessary to consider this
proposal.
As an initial matter, U.S. EPA's and NHTSA's responses to our request for additional
information were received just three days before the end of the comment period, which
was brought to a close quickly despite requests to extend the comment period from
many diverse parties. This late response compounds the failure to provide a
reasonable opportunity to comment on the proposal itself with a failure to provide a
reasonable opportunity to consider and comment on the Agencies' response. The
Agencies' decision to revise the preliminary regulatory impact analysis supporting the
proposal during the same period worsened the information gap.
We note that the Agencies would further exacerbate their failures to provide adequate
information and comment opportunities if, in the final rule or supporting documents
(including, but not limited to, NHTSA's final environmental impact statement), they rely
on new or supplemental information or analysis not fully disclosed with an opportunity
for public comment. In this regard, we note that the comments of the Auto Alliance ask
the Agencies to consider an alternative analysis of the proposal that was prepared by its
consultants, NERA Economic Consulting and Trinity Consultants (collectively "NERA­
Trinity"). As discussed below, the information provided about this alternative analysis is
woefully inadequate to permit review by CARB or the public and, accordingly, the
Agencies may not rely on it without first providing public notice of their intent to do so,
substantial additional information so that the public may understand it, and an
3

California Air Resources Board, Request for Extension Of Comment Period and Additional Public
Hearings Regarding Joint Proposed Rule to Roll Back Vehicle Greenhouse Gas Emissions and Corporate
Average Fuel Economy Standards for Model Years 2021-2026 Light-Duty Vehicles, September 11, 2018,
Docket No. EPA-HQ-OAR-2018-0283-0883.
4 We have copied appropriate FOIA staff, as well
as agency officials.
5 We further note that, after the close of the comment period, additional analyses have been published,
including in peer-reviewed journals, highlighting deficiencies in the Agencies' rushed analysis. See, e.g.,
Bento, et al., Flawed analyses of U.S. auto fuel economy standards (December 7, 2018) Science, v. 362,
iss. 6419, p. 1119.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 3
opportunity to comment. This would, of course, be true of any alternative analytical
paths the Agencies might use to justify or support any final rule.
Below, we summarize the information that the Agencies still have not made available
about their own analyses, the information that would prevent meaningful consideration
of the NERA-Trinity alternative analysis, and the patent deficiencies in the NERA-Trinity
analysis that can be discerned from the limited information provided. Please place this
letter in the dockets for both rules, correct the inadequate disclosures and comment
periods, and ensure appropriate steps are taken by your FOIA officers.
I. NHTSA and U.S. EPA Continue to Fail to Provide Requested, Relevant
Information Necessary to Meaningful Public Comment; CARB Appeals Certain
FOIA Responses

The Agencies' responses to CARB's request for additional information concerning the
proposed rollback were inadequate under the laws that govern these rulemakings and
under FOIA U.S. EPA flatly declined to share information requested, or even to make a
timely FOIA determination. NHTSA shared a limited amount of information, while
withholding materials critical to the public's evaluation of the proposed rules. Neither
response was proper and both are contrary to law. In this portion of the letter, we
discuss these deficiencies with regard to each of CARB's relevant data requests.6 We
look forward to a prompt written response to our appeal, as well as to additional
disclosures and further opportunities to comment.
1. CARB's Request: Information about the models and data used to estimate battery
costs for electrified vehicles. [This request comprised three categories of information,
identified below.] This information is required to replicate and evaluate whether the
modeling underlying the proposal is appropriate, considering the various vehicle and
technology types.
CARB's Request
a. The proposal and the Preliminary Regulatory Impact Analysis (PRIA),
NHTSA-2018-0067-1972, reference the Argonne National Laboratories'
(ANL) website for the BatPaC model for estimating battery costs for
vehicles, and state that the agencies used "an up-to-date version" of the
model, but do not identify the version. (See, e.g., 83 Fed.Reg. 42,985,
43,002 (August 24, 2018).) U.S. EPA and NHTSA have posted to the
dockets for this action a document describing how BatPaC was developed,
but this document appears to be from 2012. lt does not state which
6
Requests are numbered as they are in CARB's September 11, 2018 letter. We omit further discussion of
CARB data requests 6, 7, 10, and 11. Although this information should have been made available in the
docket initially, and was provided far too late to CARB, CARB is not appealing NHTSA's responses at this
time (though CARB reserves the right to file an additional appeal).

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 4
version of BatPaC NHTSA and U.S. EPA used to estimate battery costs.
See "Modeling the Performance and Cost of Lithium-Ion Batteries for
Electric-Drive Vehicles, Argonne National Laboratory, ANL-12/55," NHTSA
Docket ID #: NHTSA-2018-0067-1692; EPA Docket ID#: EPA-HQ-OAR2018-0283-0764.
NHTSA's Response: The BatPaC version 3.0 model was used to estimate battery
costs. NHTSA does not maintain the BatPaC model. The model is maintained by the
U.S. Department of Energy's Argonne National Laboratories (ANL). To obtain a copy
of the BatPaC version 3. 0 model, please contact ANL directly.
CARB's Rebuttal: GARB appeals this response, pursuant to FOIA (5 U.S.C. § 552)
and Department of Transportation FOIA regulations (see, e.g., 49 C.F.R. § 7.32(d)).
As NHTSA is aware, FOIA requires it make its records "promptly available" to any
person. (5 U.S.C. § 552(a)(3)). NHTSA recognizes its statutory duty (see 49 C.F.R. §
7.23(b)) and has a standard policy to make records available to the "greatest extent
possible" (49 C.F.R. § 7.23(a)). NHTSA should provide the full model used to
evaluate battery costs. NHTSA has not cited any relevant FOIA exemption to justify
withholding the model as NHTSA used it. Indeed, NHTSA concedes it used a
particular version of the BatPaC model, and that version must be in NHTSA's
possession. Explicitly providing the model version that the agencies used enables
review to determine whether the agencies modified it in any way.
Contrary to NHTSA's assertion it used the latest version of BatPaC, ANL released a
revision, version 3.1, in October 2017, as noted in our previous comments.7 Since
releasing that version, ANL has not made the prior version 3.0 available on its
website.
It is unreasonable to fail to make available the modeling tools used in support of the
proposal, and to compel the public without notice or instruction to seek information
from third parties. In doing so, U.S. EPA and NHTSA fail to make clear what they are
relying in support of their proposal.
CARB's Request:
b. U.S. EPA and NHTSA should make available the information specifying
the full battery sizes, in kilowatt-hours (kWh), battery pack configuration,
and costs used for each vehicle iteration in the CAFE model. See PRIA,
7 See CARB, Analysis in Support of Comments of the California Air Resources Board on the Safer
Affordable Fuel-Efficient (SAFE) Vehicles Rule for Model Years 2021-2026 Passenger Cars and Light
Trucks (GARB Detailed Comments), October 26, 2018, Docket Nos. NHTSA-2018-0067-11873, EPA-HQ­
OAR-2018-0283-5054, p. 140.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 5
Electrification Technologies, Technology Overview, section 6.3.8.1, p. 357.
The PRIA states that NHTSA posted ANL vehicle files that have battery
pack sizes and costs for each vehicle, but there is no additional information
about battery pack configuration (e.g., the number of cells, and the
electrical topology of how those cells are arranged in the battery pack), nor
do they directly reference where the files are posted. NHTSA and U.S.
EPA have not posted the BatPaC model file(s) that were used. ANL cost
and battery size data referenced in the PRIA, p. 358, footnote 325, but the
footnote refers to a docket identification number that is not available.
Previously, in support of the draft Technical Assessment Report and
Proposed Determination on the Appropriateness of the Model Year 20222025 Light-duty Vehicle Greenhouse Gas Emissions Standards under the
Midterm Evaluation (Proposed Determination), EPA-420-R-16-020,
November 2016, U.S. EPA posted the BatPaC files that it used.
NHTSA's Response: The full battery sizes in kWh and costs are available in files in
the docket for each "tech class." Information for peak battery power, battery total
energy in kWh, and battery pack direct manufacturing cost can be found in columns
AN-AR for each of the files listed below. Please note that there are no battery pack
configurations. [Specific links offered are omitted from this CARB letter for brevity.]
CARB's Rebuttal: CARB appeals this response, pursuant to FOIA (5 U.S.C. § 552)
and Department of Transportation FOIA regulations (see, e.g., 49 C.F.R. § 7.32(d)).
All information responsive to this request has not been provided, and no relevant
exemption has been cited. In this request, CARB requested information "required to
replicate and evaluate whether the modeling underlying the proposal is appropriate"
but NHTSA did not provide all relevant records. After reviewing the files provided,
NHTSA appears to provide only a small part of the Agencies' analysis. CARB was
therefore unable to use BatPaC v3.0 to replicate the federal agencies' results.
This information is important because the precise modeling files used go to critical
questions in this rulemaking. We requested the model inputs and results and the
battery pack configurations (number of cells, cells in parallel, nominal pack voltages,
etc.) for each unique battery pack because they are essential to understanding how
the Agencies used the model to reach their conclusions.
Specifically:
(1)
Very few of the inputs that are needed to run the BatPaC model were
disclosed. We discuss below the inputs that the Agencies provided in support
of the SAFE Vehicles Rule, and contrast them with the degree of information

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 6
provided in previous analyses by the Agencies. The information provided in
this proceeding is not sufficient even to ascertain whether the inputs that were
used are reasonable.
(2)
There is no electrical configuration information about the battery packs
for any technology combination. The electrical configuration encompasses
several physical parameters about a battery pack that are essential to
understanding if the battery packs are being modeled appropriately and
reflective of reality. Examples of electrical configuration information include, but
are not limited to, the following:
• Total number of cells in the battery pack
• Total number of battery cells in wired in series
• Total number of cells in wired in parallel'
• Nominal voltage of each cell
• Energy capacity, in kilowatt-hours (kWh), of each cell
• Number of cells in each module
• Number of modules in each battery pack
• Energy capacity, in kWh, of each module
• Battery pack nominal voltage
• Battery pack energy capacity, in amp-hours (Ah)
• Number of finished battery packs in a single vehicle, because BatPaC
allows for multiple packs per vehicle
Without the above information, we cannot determine whether the Agencies' modeling
was reasonable or if it contained mistakes, because electrical configurations drive cell
design in the BatPaC model. The one example in the Preliminary Regulatory Impact
Analysis (PRIA) indicated that the analysis may have used configurations with low
nominal pack voltages. These voltages may have been unreasonably low, leading to
unreasonable results.
Both these pieces of information were reasonably included in CARB's initial request
and should have been provided as part of the rulemaking record available for public
comment and as part of the response to our FOIA request. Without this information,
CARB was unable to meaningfully comment on the modeling of electric vehicle battery
packs and their costs for the SAFE Vehicles Rule. The files cited by NHTSA provide
only one input for the BatPaC model, and only two pieces of output information from
the BatPaC modeling runs. In its analyses reflected in the midterm evaluation and
Proposed Determination, U.S. EPA provided the BatPaC modeling and battery size
files. Those disclosures enabled meaningful comment and participation in the
midterm evaluation process.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018

Page 7

These failures and omissions are illustrated by the two figures below. Figure 1 is a
screenshot of one of the ANL files that NHTSA pointed to, specifically
'ANL_CompactNonPerfo_07202017.xlsb', that were provided in the public record for
the SAFE Vehicles Rule. All other ANL files that NHTSA pointed to are structured the
same, but for different vehicle classes. Each row in the file represents a different
technology combination and the resulting Autonomie and BatPaC output data. The
arrows point to the only three BatPaC related pieces of information that are provided
in the files. All other information listed in the other columns for the files do not connect
to information needed to understand input or output information for either the BatPaC
model, or other critical information about the modeled battery packs for the modeled
vehicles.
Figure 1

-- --- -·- - --- -·- -- -· --- --- ---- -- -- --·- -- -- -.....
-- -- ·--- -- -"- -"- --- -,,

�

,.

-.......
.......
-...........

.

.....
.....
.....
-

'<J
AT
�AS
AQ
..
"'
,..
Al
"'
C
I
A
ll,!;u.� � � '-Jt1
I 'teNdeOiii55 Vi!I.-PakOtlllpiif �fnfnl'NbOfCyllnlkq � �,Nd&iP®urmry'tal.ll&tfllYrSOCOt'ttiDDS Dff'....i.5,0COIHWf az,acx.t
1--.....,
971
'61:17.JM
1DMUC.--.-.,:
NonPeno
1EVa

,_,.,.,.._

_,......,
,

..._.

,,,..,,,_
_,,,__
..,.,__
_,,,__

.... °""""'

...,.

Fuel:Cel t!EV

IIEVlDO

......,,

F-'CdHEV
,moo

...,._.,.

.......
.,.,,_
.,.,_,,
........
"""""
..........
........

"'°'c.l!HEV

.........,,.
,mm

--

��l..v

fLINCdHf\l'

""""

......"""
.......
............
.......,,,.
""""
...........

DO£VMV50
k.lttcdMEV
,mm

furicelMEV

-·- -- . ---- --- -

_,,'""-'

'*-"'tri'o

-'""-"

.....

�

__.i

st*ME.Y

IVf'lOIHlV
�PHEY•

s,ll!HlV

J#-V.NV'J,A

,..,

"""

,..,

....

-

""'

,.
....
uoo

....

,,..

"'"
"""

....

...
""'

,

""'

llllO

1

......... ,,,.,_ .........,
,.....
""""'
_,..,.,
,.,,...,
,,_
""""""
.......

,,__

--

'""""'
..........
..........
........
-"""""
-........
-,......
,.,,,..,.,

14W.lDM
llllti
e.111.1M

....
......

,,,.,_

....,_.,..

,,....

.....

ii7l1.'JM

,,._...

""'-'""

,,.,.......
,..,,....
,,__
"'""'"
,,.........

144JJ.Cl3M

,,_

,,.,,_
,.,,....
,,..,._

.....,.

"6JUW

,,_

_,,,,

,,...,

�711.114
ill6t.Ml6

,_..

""'·'

:i&U2.D1M

,,,.,_
....�
,..,.,.....
161'17.7'1

1'W.OIM
�JJJ.114

,,,.,__
,,.._.
,,.,..

li67J.7.1M
)lJll......

......
...,,

.....
.,,,.
.....
"""
..,,.,

-......... -·...- --- ""'
"'"
..........
·- -- .....
-- ......
......... ......
- ,.,,,.
..
""
-·.....
.....
�-,..
....__
---- ......... ......__ "'"
·�........... -- ........
........ ....... - """
tm.l
,om

a-..

,,,_.,

.....

1022]0

..,,.

.........,,

>KlU1J1

,,__
,.,.....,.
...,._,.
...,......

-..,,,

...,,

.CJ.Cll!I06lnl

,__

a""'"

'·""""'

a..,,....

...,_.

-a_ooooin11
•......
71

.,..,...,

,,,..,.,
,_
-..,,,

,,...

,....

"""
...,,

..,..
,,.
''"'

,,,..,,,.

am&

,....
..,..

lll5&

..,..

......u

..,.,.,,,

.,.n

_,

...,

..,..,,....

,,

><W

,,,,,.
..,,,
...,.

""'

"""

-.,n,
muns

-,_,m_

..,,.
,..,,
,.,,,.
..,,.
"'21
.,,..,,
,.,,,.
..,,.
'''"
,,.,..

.....
...
-

•

Ill l".l

Figure 2, below, is a screenshot of one of the files,
'Battery_Sizer_PD_NMC_OWR.xlsx', that U.S. EPA provided as part of the Mid-Term
Evaluation process that led to the 2017 Final Determination and is representative of
what the Agencies released as part of that earlier process regarding these same
standards. The file contains both the iterative battery solver used to find an optimized
battery size for different electrified vehicle technology combinations for a given
chemistry and targeted mass reduction, and the BatPaC model used to model battery

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 8
costs for those optimized battery packs from the battery sizer. Each column of data
represents a single optimized battery solution and the resulting data associated with
that solution. The arrows point to the same three pieces of information that would be
used to populate the three columns in the file referenced for Figure 1 .
Figure 2
R

'...

=

,o

"
"
"
,..
"
"
"
"

.

�;aC'll«l<odolW,l'tt£11.l'l[V·ltPj
8'11".Ccao: . .,.,.lalhrf�
Mt!U1C0"8C'bl

"-·

��IMS

U...til�SOC
11,,,lhC ,.....,,llfV ,OM�

,..

lartf'•C!U.l,Ulf:•f

.,

·"'-

f�
fl1�

111�

l'l1j

c.-:-.Jpk

..

"

�.

-

'

""

.

""

,

'

.

.v

.•

"'

""

"'

... ...

""

l:W

tWi,
,nt
Int
Int
,r,

O•l

lflt

t,;fpA.I,
i;.,;1Ah

.I),.
M

P1dil;f:
Wl',/tc

kil
W�

�m,, -�GMI

-Dl�a-.-r

"lll:OOP'¥1NC:lltrtl'DWM"-'/:i;tE�
Pr,uttlOEMtar\ ...tinwck
,nee for•· h•IIP•i�,

"
"

.

__

�_. .. ...,.••
i

"

"
"

.-

.

..

p,ar,1?",j�]"""k
w1111M1 �rhlb)p..::k
h<••nt "1"
.t1an(l!Ml&Cl:kahni!CDt$
Md]co!;thilll'JC.IU<IIO!'li,$

5

�w

,_

""'
1 ��-� �;;,;.
j
$\.4&fi&I;
i:v;,,.4u
1
n11>4:t
S176Z'I,
j
Ull.:s.t
520,�

"'

'*;�·::

.....

MOO

"'

wt�;,:;. ��-� �.1;�

"'

w;�;;;� ��;��

""

w;�;;;_:1 W1��::.:,.

"""
w����:J

"'

��i;:;1, "';�;::�1

,�JQ

J§,f;§l..11
3n1,,

nn.i,

S1.1U'I

"""
"'

w��;,:�,
,,,._MS_n

"''"
"'

.. ��1=�•

"""'

�
w ��-:i ��;:;i

MOO

"'

��;;;:a

:----� ------- � � - � -1
I
:

I

,�.59
$l06ll

$11.1:..n

Sffl.00

·----*.m

l'o�

M5

""

""

EV75

sun
$lllffl.
$17214
$'9$.00

"T&2.111
Ul(L•J

Sf.:IO.'H

s:,iw1
$1U
..
a
SIH.&O

S,,S.1)11

$�08
UUNI

Sl'ill��

saue
Sl.19.1'
JlU..66
s-,.m

�.-.,...
J1!1Ul
su.11.w

s.auo
,ill.Iii

�llll.J.8

1i!8'i.ll!I

---lC�<IO ___ JG,00 ---�-..__..fUao

E"•t90

F:l«J

PMFl.?0

•HF;.AO

�

s,c,ni-u
,nu,

.suu.1
,i1..1,
$1111.l'il
J2'1 U

'j;�

,�,1)1.1.11
j1all.Ui
9,lla..41

,._,i
1..11 ..w
$,119.114

PH-«!

---�}JG ___

lJl�E.V

IJI

��,1-S.uli
'SJ.�.n

�1.1u,

�;,.J,6
11uD1
$,Ult.in

t,:11§.Jll'l

s.un.n
S14'!1.U

"4.al�.�
S14'.U
$uu1

$39!i.OO

,is,.1,

,11u1

5.u.m
5llll.A1
$,211.1'4

iJQ41
S\Q,¥)
bniDO

,,.ru,c�
SU'l\\

S1li4.n
uu9
srnu,
SJT&.ZJ

Slr.iia:J

,,.1n�,

Slhl.12:1
,.u.as
,u4Jli5

S.1SJJ(J

S15UU
�.i,
U1UJ

5-415.00

f,Lll.21

s1v.,.o

$nm
SU0.69
,.:l!r2&

$4}S.00

S',2'M4'j,
�UJ'1
,1,1.lio'
s"�
jlHM
)3CN.U.

S4J',--'l(I

J6)1Un
JtUH
$r.5:11

fue1111
$11;\.4:.l
sutim

$31.UCI

ua, 1T

541.l:J
$194!19

,-.75.!ll

s:i.9.'»
$Z11l.17

i1.nu,
su..u�
su,�l

n.11s
$l�f:i.M

,:u,

$113100

i�

$ts5.lD

.JO.OU ---�!jd ___ .},!B'I ___ l11U1D ___ J£0J ---1t:IJKl ___ _JJ.CO ___ 1,.0.Cll ___ JrlJXJ ___i_O}ll _. __ J£W ___ 1.0.00

-

&,II

�

•

Figure 3, on the next page, is the same screenshot as in Figure 2, above, but the
information that has not been made available for the SAFE Vehicles Rule has been
covered up. The file has several other tabs of battery model information where the
entire worksheet would have to be covered to align with what was provided for the
SAFE Vehicles Rule.

•

E:l

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page9
Figure 3

.

.

'

-•t.-tan,,

..".
"
..
..,..

---"--

.

.''

"'
,

'

'

4

.
--.. .......
.,._..
�--·· ...
---�­

""'-lw-llnor

��

IVf!i.

�

M!>

�

[V1!>

flfr.i

MS

D'ts

tvn

M5

n

n

�

�

�

hi£O

DUIO

MU>

MOO

Mm

Min

NXID

fYJal

f\'2IIID

l.YaO

�

.fVmJ

ffltll

�

IWMI

ft'lm

ma

,v.,.

*

l'YU8

-

f'W'JID

-

rwJa

�

.....

�

ma

-

ma

tvna

n

•

�

w

"

•

�

•

a

a

�

�

....,..,;v.t.HmJY.l'KV.�
......:,...-...

.....,_ftw lc'M,

,.

"'

,.
•

llllllhiC,� ...
-

JS
JI

fiJ
m

ll

.

"

:n

.....:��., ow,

"'

FX"'"
�

"
11W

rn

aapM

f}M
n:a
U11i
fJ)llli

.... ,,..._
lltt_,_,_
hn.Q:
•
��

,..

..
-.. -�""°--�
..
.......... ''' ':'
..... .. :'
!II
u
t.l
,.,

w

JIii

m•

ea..

�·��

s

kW

D
•

M

"l
Al
,u

...

"'

•
4

,.

-•DIMfm' ....
p,rt,

,.
......... fM,lpKbPtt"blih

"'""
""'"'"'�:..n

WlffWIIWp!rlswl)lad:

hll.�raa,�s
IMtcOISl1'r'IC-...._,S

.diwdwk-•

t
S

''

'

r.Yn

ft

ft

-,!l,&t5..,.-

�

ft

8

MS

B

•

ft

ft

ft

ft

d

Wl'Ml,"I W.(11151';t #t(;liF;l! llfn(.1-· MClli,iol!li Wll�I. -�l M.CIMI.J -CIN.,;J wtC.U-6 -CIM&!I wtl:eauti _.dK&I ""'C.C.J WIOu-�l -U.U•
Mer , lll'ICi-1
....
1- �
s.usz.• .. ..,...- �
$1,,.WlUS - 'f4,.Da.lli - l,,t.JB11 - ......_.,- J!U>1u -,vun- »..-LO - ».-.u - -.-e - $&.ll,ln- ,1......,, ·,,;,uu
"15!"
: U..4".U
$.U.-.H
,VUN
"--•
s,,t.MU&
S4>'U"'1
l,.UJIU
s.,naH
.....
Jt
,..,,.z:11•
,�:n1.u
SS.,SU�
fUIIU,.
�
,.,.,..,
$UU,1J
$1....JI
P)U_JI

-

_.-

'
:

!

As can be seen in the difference between the three figures, the data supplied by the
Agencies for the SAFE Vehicles Rule does not provide the level of detail required to
comment on the appropriateness of the battery cost modeling. The data previously
provided as shown in Figure 2, and the set of files it represents, contain much more
detail of the inputs to, and outputs from, the BatPaC model that are critical to the
modeling of the battery packs for evaluating the compliance costs of vehicle
regulations.
Because the information is complex, we continue to evaluate it in the event additional
conclusions may be drawn relevant to the proposed rule. We reiterate our objection
that this information should have been made available to all interested persons in a
timely manner and with adequate time for review.
CARB's Request:
c. The proposal and PRIA provide conflicting information about which battery
chemistries the agencies considered. For instance, the proposal and PRIA
refer to NMC441-Gr chemistry for both plug-in hybrid-electric vehicles and
battery-electric vehicles, but the ANL summary refers to NMC333. See,
e.g., PRIA, pp. 372, Table 6-27, 373 ["We selected NMC441 as choice of

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 10
chemistry for PHEVs and BEVs. NMC441 more suitable for high energy
batteries capable of discharge rates."}. The Excel file titled "ANL-Summary
of Main Component Performance Assumptions NPRM" has a tab labeled
'Description - BatPac' with the same table listed as in the PRIA, except the
chemistry listed for PHEVs and BEVs is NMC333-G instead of NMC441Gr. See Docket ID Nos. EPA-HQ-OAR-2018-0283-0054 and NHTSA2018-0067-0003. The proposal and PRIA do not directly reference this file.
NHTSA's Response: NHTSA and EPA used the battery chemistries associated with
the BatPaC version 3. 0 model.
CARB's Rebuttal: GARB appeals this response, pursuant to FOIA (5 U.S.C. § 552)
and Department of Transportation FOIA regulations (see, e.g., 49 C.F.R. § 7.32(d))
because it does not appear to reflect the record, and so indicates that records in
NHTSA's possession were responsive but have not been provided. The chemistry of
NMC441 is not available in version 3.0 (or even the recently-released version 3.1) of
the BatPaC model, even though the SAFE Vehicles Rule and supporting regulatory
impact analysis refer to this chemistry. The Agencies have either incorrectly stated
which version of BatPaC was used, identified the incorrect chemistry when disclosing
what was used for modeling batteries for some of the vehicle technologies, modified
BatPaC without providing the requested documents describing the modifications, or
used inputs without identifying them - despite repeated requests for this information.
Additionally, the files that NHTSA cites in its response are Autonomie model outputs.
They contain only battery pack energy capacities, battery pack power capability, and
battery pack cost for a reference year. As U.S. EPA and NHTSA stated, they did not
disclose any information about the battery pack configurations or provide other
BatPaC input information. This precludes meaningful analysis and comment, as U.S.
EPA's interagency review made clear. 8. Several pieces of information would be
needed to analyze how BatPaC was used and if the results were appropriate and
reflective of reality. Some of that information would include, but is not limited to, the
following:
•

Basic
o
o
o
o
o

battery pack information for each unique battery pack
Total number of cells in the battery pack
Total number of battery cells in wired in series
Total number of cells in wired in parallel
Number of cells in each module
Number of modules in each battery pack

See GARB Detailed Comments, pp. 139-140, citing EO 12866 Review: NHTSA responses to
interagency comments sent to 0MB, Docket ID: EPA-HQ-OAR-2018-0283-0453.

8

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Pag e 11

•

•

•

o Battery pack nominal voltage
o Battery pack mass, in kilo-grams (kg)
Cell specifications
o Electrode thicknesses and any applied limits
o Resultant cell capacity, in kWh
o Nominal cell voltage
o Mass of constituent materials used in each unique cell, particularly for
the cell's anode and cathode
o Cell mass, in grams (g)
Material cost inputs, yields, and assembly costs
o Cost of cell constituent materials, like $/kg of nickel or cobalt
o % yield of each manufactured cell component
o Unit cell hardware costs, like the positive and negative battery terminals
o Electrode processing costs
Manufacturing volume, in number of battery packs per year

Batteries are large cost drivers for electrification technologies, and the public cannot
meaningfully comment on the battery sizing and cost development methodologies
without the requested information. 9 Effectively, it is impossible to replicate what the
Agencies did with BatPaC. NHTSA and U.S. EPA must provide sufficient responsive
records to satisfy CARB's request. The Agencies have not offered any relevant FOIA
exemptions to support their non-response. The appeals officers at the Agencies
should direct a full search for records and promptly supply them to GARB.
2.
CARB's Request: The PRIA references Polk registration data, including survival
rates aggregated by model year, calendar year, and body style. These data are
needed to verify the coefficients of the new model predictions for vehicle retirement
(scrappage), but have not been made available. See, e.g., PRIA at pp. 1008, 1014,
1023, fig. 8-23, 1025, fig. 8-24, and 1027, fig. 8-25.
NHTSA's Response: The Polk registration data is proprietary information and is being
withheld in its entirety from disclosure because it is related to trade secrets and
commercial or financial information pursuant to FOIA Exemption 4. 49 U.S.C. §
552(b)(4). To purchase the series of National Vehicle Population Profile (NVPP)
datasets, please contact /HS Markit (formerly R.L. Polk & Company) directly.
CARB's Rebuttal: GARB appeals this response, pursuant to FOIA (5 U.S.C. § 552)
and Department of Transportation FOIA regulations (see, e.g., 49 C.F.R. § 7.32(d)).

9

See CARB Detailed Comments, pp. 139-140, fn. 261.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 12
The FOIA response appears to have erred in two related regards. First, FOIA itself
directs "partial disclosure of information" where possible, and specifies that agencies
must take "reasonable steps necessary to segregate and release nonexempt
information." (5 U.S.C. 552(a)(8)(A)(ii).) NHTSA does not appear to have made any
attempt to segregate and produce responsive information. Nor does it appear to have
followed its FOIA regulations in making this trade secret determination. Those
regulations require NHTSA to notify the submitter of the data "expeditiously" and ask
for any written objections to release. Notably, "[t]he burden is on the submitter to
identify with specificity all information for which exempt treatment is sought and to
persuade the agency that the information should not be disclosed." (49 C.F.R.
7.29(a).) NHTSA has not demonstrated that all relevant information is trade secret
under FOIA in the first instance, that the submitter (here, likely, IHS Markit) objects to
their release, or even that NHTSA has sought IHS Markit's views. 10
Moreover, CARB specifically requested "aggregated" data. Such aggregated data
protects individual manufacturers' information and so obviates trade secret concerns.
We note that such data, as a critical basis for a rulemaking, would be improper to
withhold from public review. Indeed, IHS Markit does allow for publication of
aggregate data, provided that publication rights have been purchased. Given the
Agencies' choices to purchase and rely on this information, the Agencies should have
purchased those rights or should do so now.
NHTSA and U.S. EPA should provide the data aggregated to the same degree as
used for developing (or "estimating" as it is also described) the CAFE Model so that
CARB and the public have a reasonable opportunity to evaluate and comment on the
Model's scrappage coefficients. The burden is on the Agencies, not the public, to
provide the necessary factual information upon which the Agencies' proposal is based
so that the public can meaningfully participate in the rulemaking. The failure to
provide this data and information, which are critical components of the Agencies'
analysis and are necessary in order to analyze the Agencies' modeling, deprives the
public of their right to participate in the rulemaking. 11

10

To the degree 49 C.F.R. Part 512's trade secret rules for NHTSA apply (which NHTSA has not
asserted), it is unclear whether any final determinations have been made under those provisions
11 See Allina Health Services v. Sebelius, 746 F.3d 1102, 1110 (D.C. Cir. 2014) ("Still, we have held for
many years that an agency's failure to disclose critical material, on which it relies, deprives commenters
of a right under§ 553 [of the Administrative Procedure Act] to participate in rulemaking." (internal
quotations omitted)); Air Transport Ass'n of Am. v. FAA, 169 F. 3d 1, 7 (1999) ("[W]e have cautioned that
the most critical factual material that is used to support the agency's position on review must have been
made public in the proceeding and exposed to refutation."); Ass'n of Data Processing Service Orgs., Inc.
v. Bd. of Governors of Fed. Reserve System, 745 F.2d 677,684 (D.C. Cir. 1984) ("[A]t least the most
critical factual material that is used to support the agency's position on review must have been made
public in the proceeding and exposed to refutation.").

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 13
CARB's Request: New vehicle sales and price data referenced in the proposal.
3.
This includes:
a. Data provided by the National Automobile Dealers Association (NADA) and
others. See 83 Fed. Reg. at 43,095; PRIA, pp. 1017-1018.
b. Data describing historical transaction prices, and quarterly new vehicle
sales data used to develop the dynamic new sales model. See PRJA, pp.
954-961.
c. Economic data used to develop the autoregressive distributed-lag (ARDL)
model that predicts new vehicle sales and is used in the CAFE model. See
83 Fed. Reg. at 43,074.
CARB's Rebuttal to 3.a. - 3.c.: NHTSA's response is omitted for brevity in this letter,
but is under GARB review. This data is necessary to evaluate the proposal's
predictions for fleet population, sales, and fatalities. We continue to object that the
Agencies have not fully explained how they manipulated and used the NADA data. 12
Because the information is complex, we continue to evaluate it in the event additional
conclusions may be drawn relevant to the proposed rule.
Data were provided to CARB only four days prior to the close of the comment period,
which did not allow for sufficient time for review and analysis. Separately, NHTSA
also provided identical data to Professors James Stock and Kenneth Gillingham, who
found numerous errors in the estimation of the new sales model as discussed in their
comment letter. 13 Moreover, though NHTSA provided this data to GARB and
Professors Stock and Gillingham, NHTSA did not make this data public for all
stakeholders to review. The failure to provide this data and information - critical
components of the Agencies' modeling and necessary in order to analyze and refute
the Agencies' modeling - to the public generally and meaningfully during the
proceeding (not four days before the close of the comment period) deprives the public
of their right to participate in the rulemaking. 14
12 See Stock, J., Gillingham, K., and Davis, W., Comments on Notice of Proposed Rulemaking for [the
SAFE Vehicles Rule], EPA-HQ-OAR-2018-0283-6220, pp. 5-6.
13
See EPA Docket ID: EPA-HQ-OAR-2018-0283-6220.
14 See Allina Health Services v. Sebelius, 746 F.3d 1102, 1110 (D.C. Cir. 2014) ("Still, we have held for
many years that an agency's failure to disclose critical material, on which it relies, deprives commenters
of a right under§ 553 [of the Administrative Procedure Act] to participate in rulemaking." (internal
quotations omitted}}; Air Transport Ass'n of Am. v. FAA, 169 F.3d 1, 7 (1999) ("[W]e have cautioned that
the most critical factual material that is used to support the agency's position on review must have been
made public in the proceeding and exposed to refutation."}; Ass'n of Data Processing Service Orgs., Inc.
v. Bd. of Governors of Fed. Reserve System, 745 F.2d 677, 684 (D.C. Cir. 1984} ("[A]t least the most
critical factual material that is used to support the agency's position on review must have been made
public in the proceeding and exposed to refutation.").

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 14
4.
CARB's Request: Report of analysis of the standard errors and significance of
the ARDL [autoregressive distributed lag] sales model coefficients, F-statistic and R2
of the overall model, and variable stationarity and co-integration indicators. This
information is needed to verify the statistical significance and errors of the coefficients
used in the Volpe model. The coefficients for the ADRL sales model listed on p. 957,
Table 8-1 of the PRIA, are not consistent with those implemented in the model. See
CAFE Model Documentation, PRIA, p. 78, Table 17, available at:
https://www.nhtsa.gov/corporate-average-fuel-economy/compliance-and-effects­
modeling-system ["2018 NPRM for Model Years 2021-2026 Passenger Cars and Light
Trucks," Model Documentation].
NHTSA's Response: NHTSA has identified an error in Table 8-1. The agency published
a revised PR/A correcting the error in Table 8-1, which is reproduced below and
available on page 949 of the revised PR/A at
https:llwww.nhtsa.gov!sites!nhtsa.dot.gov!files!documents/ld cafe co2 nhtsa 2127a/76 epa pria 181016.pdf

I'

od1ldr11h
ln1,r.-1 't

I hllP bw.,.,h R•1•

1 l>d� ltcm,a.t1110 rn....-

l.u.,.w 1,n• r�""'""''"m

I � .......

I 1 :ui.• h-.d'Jll"'-lll.llhll' L>gl
11.11,

I \dj1n1....:I R·"'l•arrd

, "1.i:,oo1111i

' IIOH•��ti

-llflOtl "1

0

1110 ·���

',td.�rror
11 ;.!JI

, ''i:·nlfkH,"t'

lilth� "'II

• ,,l,n:ah�
u ,11•111•1

I 0

11111 ·1�
111,, 1 t!"•J4

1110111111

0
•

J

Ulli

I 11, I!

1 U "I'�

CARB's Rebuttal: GARB appeals this response, pursuant to FOIA (5 U.S.C. § 552)
and Department of Transportation FOIA regulations (see, e.g., 49 C.F.R. § 7.32(d)).
All information requested has not been provided, and NHTSA has offered no valid
exemption to justify withholding records. Although NHTSA provided a revised Table
8-1, the requested variable stationarity and co-integration indicators were not
provided. Using the corrected coefficients, the outputs from CAR B's run of the CAFE
model do not match the output values published in PRIA Table-8-2. 15 GARB
continues to request these records both pursuant to its FOIA request and pursuant to
the Agencies' responsibilities to disclose the bases for their proposed actions and to
allow comment on those bases.
5.
CARB's Request: The coefficients for the dynamic fleet share equation
described in the CAFE Model Documentation on p. 79. These are not listed
15 See CARS Detailed Comments, pp. 218-219, Table Vl-4 showing discrepancy between CARB and
Agency outputs.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 15
anywhere. Additionally, according to the PRIA on p. 955, the model was based on
EIA's National Energy Modeling System (NEMS), but no reference is provided for the
NEMS model. This information is necessary to evaluate the equation used in the
model.
NHTSA's Response: Records for the dynamic fleet share equation is provided in the
Energy Information Administration's 2016 National Energy Modeling System (NEMS)
documentation beginning on page 48 and can be found at
https:llwww. eia.govloutlooks/aeolnemsldocumentationltransportationlpdf/m070(2016).

RSJt.

A copy of the Fortran code for the NEMS implementation is detailed below. The
NEMS implementation in the CAFE model is available in the public source code and
described in the CAFE Model Documentation on page 78 at
ftp:llftp.nhtsa.dot.gov!CAFE/20212026 CAFE NPRM/CAFE Model/CAFE Model/CAFE Model Documentation NPR
M 2018.pdf [NHTSA's image of relevant code is omitted from this CARB letter for
brevity.]

CARB Rebuttal: CARB appeals this response, pursuant to FOIA (5 U.S.C. § 552) and
Department of Transportation FOIA regulations (see, e.g., 49 C.F.R. § 7.32(d)). All
information requested has not been provided, and NHTSA has offered no valid
exemption to justify withholding records.
It is unreasonable and improper to direct the public to extrinsic sources of information,
when the Agencies themselves have the information and provide no reason it should
not be disclosed. It is also unreasonable and improper for the Agencies to vaguely
reference, as they do in the Vehicles Rule and the PRIA, the "NEMS model" without
citing the version used, when there are many versions of this model available. The
public should not be required to guess which version the Agencies used. The exact
model used for this rulemaking should be provided and accurately cited. This is
especially pertinent where the Agencies do not consistently use the most recent
version of available models, such as with the BatPaC model. The Agencies' errors
remain where the instructions and information have not been made publicly available.
Further, the information supplied by NHTSA referencing the CAFE Model
Documentation does not include the dynamic fleet share coefficient values. Footnote
52 in the CAFE Model Documentation states: "Refer to Section A.3.10 of Appendix A
for more information regarding the input parameters used for the Dynamic Fleet Share
model." However, Section A.3.10 refers to ZEV Credit Values:

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 16
A.3.10ZEV Credit Values
The ZEV Credit Values worksheet contains parameters allowing the modeling system to target the
ZEV requirements of CA+S 177 states during compliance simulation. Presently, usage of ZEV
credits within the CAFE Model should be considered as experimental.

.

Table 39 ZEV Credit Values Worksheet

CatNOn'

�lodd C'llaradrmtlc

ti
.a

ZEV R"'luin,rm,nt t"•I

p�n:1:ntaI,.<e

Ma.x Credits from PIIEV ("•I

p�n:�ntage

�

t i alts

.,:;

;.,

::>
..,
"'

Drlhlltioa/Nekl

Minimum pen:entagi: ufJ!�ru cm1»ion ,cli..,le (ZEV)
cn....ti1.s lhat a manufuclun:r mu�\ gt,'1\Cfale in order lu
mcC"t the ZEV ruquinmEnt 1n 1:ach s�ified model
\(."ar.
I\ 1:uimum p,:n:<:nln)!C of ZEV .:n,Jii,, that u
rnanufa,turer may genera!,: from PIIEV, ,n ottl1:r lo
meet th<: ZEV raiuiremen1 1n �ach spccilictl model
war.

Presumably, the correct reference is Section A.3.11, but, while this table does
describe the coefficients, the values themselves are not present. Without these
coefficient values, it is not possible to understand the extent to which the different
variables (fuel price, fuel economy, curb weight, horsepower) affect the model's
estimate of production volumes and fleet shares in future model years, and whether
those effects are reasonable predictions. In turn, without a proper understanding of
the fleet share model, it is not possible to evaluate whether this model's interaction
with the other components of the CAFE Model is sound and reasonable.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 17

A.3.11 DFS Model Values
The DFS Model Values worksheet contains fine tuning parameters for utilizing the Dynamic Fleet
Share and Sales Response model (DFS/SR) within the CAFE modeling system. When enabled,
the DFS/SR model adjusts the production volumes and fleet shares in future model years as a
response to increasing fuel economies and costs of vehicle models.
Cattton'

Model Cllanc:trriltk

Sl!l!d Va/Ul!S {pl!r Modi!/
YmrJ

Table 40. DFS Model Values Worksheet

Uaib

variollS

Share of Total Fleet
ll

>

i

Cl

percentage

Fuel Economv
Horscoowcr
Curb Wciruit

mo<>
ho
lbs.

Coeff1r.la,ts

number

Constant
Rho
FP
HP

number
number
number
number
number
number
number

CW

MPG
Dummv

Ddlaidea/NoCft
Flce1-spccific seed volucs for the Dynamic Fleet Share and Sales RCSJ>OOSC model,
soccificd for LDV and LDTl/2 fleets and for model vcars 2014 and 2015.
Observed share of either LDV or LDTl/2 fleets versus the total light duty fleet.
during a soccific model vear.
Avcrnec fuel cconomv for a soccific fleet. durine a =..-ific model vcnr.
Avcrngc horscoowcr for a soccific fleet. durina a <:necific model vcar.
Avcrngc curb wciclit for a sn«ific fleet, durine a .,,,.,,,ific model vcar.
Flce1-spccific coefficients for the Dynamic Flee1 Share and Sales Response model.
...,....ificd for LDV and LDTI/2 fleets.
Soccifics the NEMS "constant" coefficient
Soccifies the NEMS "rho" coefficient.
Soccifies the NEMS "fuel oricc" coefficient.
Soccifics the NEMS "horscoowcr" coefficient.
Snecifics the NEMS "curb wcil!ht" coefficient.
Soccifics the NEMS "mpg" coefficient.
Soccifics the NEMS "dummv" coefficient.

Finally, the CAFE Model documentation lacks any justification or elaboration on how
the NEMS coefficients were "applied at a different level" and why it is appropriate to
repurpose coefficients developed for vehicle categories for body styles instead. 16
CARB continues to maintain that the information provided, both in the rulemaking
record and in response to its FOIA request, was entirely insufficient.
8. CARB's Request: The agencies' detailed explanation and derivation of their point
estimates for the increase in fatalities per hundred pounds of mass reduction over a
constant footprint based on historical crash data, for model years 2004-2011 and
calendar years 2006-2012. Previously, these details were provided in a separate
report such as the "2016 Puckett and Kindelberger report." No such report is available
this time. The PRIA only provides a summary table of the results of this analysis, yet
states an "updated analysis" exists. See PRIA, p. 1357, section 11.4.
NHTSA's Response: The "updated analysis" referenced in the PR/A at p. 1357, refers
to information available in the PR/A in Section 11.4, pps 1345-51. NHTSA intends to
publish a technical summary of the logistic regression analysis and its results in the
near future. In addition, NHTSA intends to publish a report similar to the "2016
Puckett and Kinde/berger report" that will describe the methodological process by
which the results were derived. Accordingly, I am withholding these records as
16

See CARB Detailed Comments, pp. 222-223.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 18
exempt from the statutory disclosure requirement that contains information related to
pre-decisional agency deliberation, opinions or recommendations pursuant to FOIA
Exemption 5. 5 U.S.C. § 552(b)(5).
CARB's Rebuttal: CARS appeals this response, pursuant to FOIA (5 U.S.C. § 552)
and Department of Transportation FOIA regulations (see, e.g., 49 C.F.R. § 7.32(d)).
CARB also objects that the Agencies continue to withhold information necessary for
meaningful public comment on its proposed action.
First, even if technical summaries and methodological reports are still being drafted,
NHTSA's response ignores that NHTSA already made the relevant decisions here-to
propose the SAFE Vehicles Rule, relying on certain data and analysis. NHTSA
cannot, therefore, assert that the data and analysis it relied on for the proposal are
protected from disclosure as pre-decisional. Rather, it must make those available for
public review. This includes information and analysis on how the point estimates for
the increase in fatalities from mass reduction were derived. This information is
needed in order to justify the Agencies' point estimates - even more critically here,
where all of these point estimates are not statistically significant at the 95 percent
confidence level (and three are not even statistically significant at an 85 percent
confidence level, and no explanation is provided why these estimates as reasonable in
light of such statistical uncertainty). Without the detailed regression analysis, the use
of these point estimates is not justifiable and the Agencies' analysis in the proposed
rule remains opaque and improperly insulated from meaningful public review and
critique, contrary to law.
9. CARB's Request: Data used by the agencies to derive the new statistical model that
predicts fatality rates by vehicle age. See PRIA Table 11-21, p. 1397. The
coefficients of the model are provided, but without the data it is not possible to
evaluate whether the coefficients were properly derived. Additionally, the coefficients
provided in the PRIA are different (significant digits and sign changes) than those
identified in the actual model source code (which are also commented out such that
they are non-functional[ 171), and are different from the model year based coefficients
used in the input files. This renders unclear what coefficients the analysis in the
NPRM is based upon.
NHTSA's Response: The data used to derive the new statistical model for fatality rates
was obtained from /HS Markit (formerly R.L. Polk & Company) is proprietary
information. Thus, I am withholding the data in its entirety from disclosure because it
is related to trade secrets and commercial or financial information pursuant to FOIA
17 Code that is "commented out" is a programming technique that instructs the computer to skip the
calculations or other instructions in those lines of code.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 19
Exemption 4. 49 U.S.C. § 552(b)(4). To request a copy of the data, please contact
/HS Markit directly.
GARB Rebuttal: GARB appeals this response, pursuant to FOIA (5 U.S.C. § 552) and
Department of Transportation FOIA regulations (see, e.g., 49 C.F.R. § 7.32(d)).
Again, FOIA itself directs "partial disclosure of information" where possible, and
specifies that agencies must take "reasonable steps necessary to segregate and
release nonexempt information." (5 U.S.C. 552(a)(8)(A)(ii).) Yet, NHTSA has denied
all information without making this determination, and again does not appear to have
followed its FOIA regulations in making this trade secret determination. Those
regulations require NHTSA to notify the submitter of the data "expeditiously" and ask
for any written objections to release. Notably, "[tJhe burden is on the submitter to
identify with specificity all information for which exempt treatment is sought and to
persuade the agency that the information should not be disclosed." (49 C.F.R
7.29(a).) NHTSA has not demonstrated that a// relevant information is trade secret
under FOIA in the first instance, that the submitter (here, likely, IHS Markit) objects to
their release, or even that NHTSA has sought IHS Markit's views. 18
Further, GARB requested all information used to derive the new modeling coefficients.
Even if some of the data is truly proprietary, NHTSA has not demonstrated that a//
information or relevant records based upon it are confidential. Further, aggregated
data is very unlikely to be proprietary. We again note that aggregated data may be in
the Agencies' possession and must be produced.
As described in the request for information, different sets of coefficients are referenced
in the SAFE Vehicles Rule and supporting documents. The Agencies have failed to
acknowledge or clarify which coefficients the Vehicles Rule is actually based upon;
that is, whether it was the coefficients provided in the PRIA or those identified in the
actual model source code. When using the different coefficients to attempt to
understand the Agencies' analysis, the model irrationally predicts negative fleet
populations. 19 GARB appeals the response for this reason as well. GARB further
notes, as we have discussed in more detail above, that failing to provide information
upon which proposed rules are based is contrary to core principles of administrative
law.

18 To
the degree 49 C.F.R. Part 512's trade secret rules for NHTSA apply (which NHTSA has not
asserted), it is unclear whether any final determinations have been made under those provisions
19 See GARB Detailed Comments, pp. 217-218.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 20
12. CARB's Request: 20 Modeling tools developed by U.S. EPA including:
a. All files necessary to utilize - with the Advanced Light-Duty Powertrain and
Hybrid Analysis (ALPHA) and the Optimization Model for reducing
Emissions of Greenhouse gases from Automobiles (OMEGA) - the
response surface equations developed by U.S. EPA as identified or
referenced in: "Peer Review of EPA's Response Surface Equation Report"
(Docket ID No. EPA-HQ-OAR-2018-0283-0025); and SAE paper 2018-011273 authored by U.S. EPA (Docket ID No. EPA-HQ-OAR-2018-02830028).
b. All new or modified input files, source code, and executable files for U.S.
EPA's OMEGA model developed since the release of the Proposed
Determination in late 2016.
c. All current and new input files, source code, and executable files for
ALPHA used for the Proposed Determination in late 2016 and/or modified
since then.
d. All current and new pre-processors and their inputs used for the Proposed
Determination in late 2016 or modified since then to categorize, sort, and
rank technology packages and costs for use with OMEGA.
NHTSA's Response: Records related to the EPA's modeling tools fall under that
agency's jurisdiction and must be requested from the EPA directly.
U.S. EPA's Response: EPA's last publicly available version of the ALPHA and
OMEGA model is on the EPA web site (and MTE docket) and dated November 2016
(released as part of the Proposed Determination). While EPA has draft updates to the
OMEGA and ALPHA models since November 2016, these updates have not been
made available to the public. In any event, the ALPHA and OMEGA models were not
used to develop the proposed rule.
CARB's Rebuttal: This response is improper and fails to meet U.S. EPA's FOIA
obligations. FOIA obligates the agency to promptly provide records (5 U.S.C. § 552).
U.S. EPA has failed to meet its statutory obligations. Moreover, U.S. EPA created the
OMEGA model to assist in the analysis of technology costs when developing its
greenhouse gas emissions standards.2 1 The results of the OMEGA and ALPHA
models are relevant for comparison to the CAFE model for checking its reliability.
Even if these versions of the models were not used to develop the proposed rule,
2° CAR B's Requests nos. 1 O and 11 are omitted.
See Proposed Rulemaking To Establish Light-Duty Vehicle Greenhouse Gas Emission Standards and
Corporate Average Fuel Economy Standards, 74 Fed.Reg. 49,454, 49,545 (September 29, 2009).

21

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 21
there is no justification provided for withholding them from public review. Please
provide the draft updates.

II. Reliance on the NERA-Trinity Analysis, in Whole or in Part, Would Compound
the Procedural and Substantive Errors in the
As our substantive comments outlined, in the areas for which sufficient information was
provided to allow for public review the SAFE proposal is rife with errors. The failure to
provide meaningful information upon which the Agencies' relied for public review as
outlined above and in other public comments increases the risk that a final rule will be
similarly flawed. During the Agencies' review, it is important, therefore, that they rely
upon only accurate information, and that - if they seek to rely upon new information the public be given access to the information and sufficient opportunity for notice and
comment.
The NERA-Trinity analyses provided by the Auto Alliance22 purport to attempt to resolve
some of the substantive problems with the Agencies' analyses, but it would be improper
for U.S. EPA and NHTSA to rely upon them in a final rule-because the public is unable
to fully understand and provide comment on the Alliance analyses, because critical
information and modeling tools have been withheld (compounding similar procedural
errors and informational gaps by the Agencies), because there has been inadequate
opportunity to comment upon these analyses, and because even with inadequate
information it is clear that the analyses are deeply flawed. 23 We highlight some of these
obvious and critical flaws here to demonstrate the critical importance of a further public
access to and review of this information, if it were to form the basis of any regulatory
decision.
The information provided by the Auto Alliance and NERA-Trinity in support of its
analysis is limited and inadequate to provide a reasonable opportunity for meaningful
review. Should the Agencies adopt or rely on it, at minimum the following data and
information would need to be made public for review:
•

•

The NERA-Trinity models of the "New Vehicle Market" for new sales, scrappage,
fleet composition, and vehicle miles traveled (VMT) and their underlying
equations, coefficients, parameters, data and inputs, results, sensitivity tests, and
justifications.
The explanation and justifications for NERA-Trinity's calculation of consumers'
willingness-to-pay for fuel economy, including any assumptions or data used or

NHTSA-2018-0067-12073.
Likewise, for all the reasons stated herein, it would be improper for NHTSA to rely upon the flawed
NERA-Trinity Analyses in its final environmental impact statement. Any introduction of data that lacks
scientific integrity would entirely undermine and distort the analysis of alternatives and environmental
impacts in the final environmental impact statement. thereby depriving the document of any legitimacy.
See DEIS Comments, Section II, B.
22

23

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 22
relied upon.
The explanation and justifications for NERA-Trinity's calculation of "Petroleum
Market Externalities," including any assumptions or data used or relied upon.
The NERA-Trinity models have not been subject to peer review. For the Agencies to
adopt or rely upon the NERA-Trinity analysis, the Agencies would need to subject these
models to the peer review process and make appropriate improvements based upon
that review.
•

What review can be performed shows the analysis is fundamentally flawed and does
not support relaxing the existing standards. The NERA-Trinity analyses expressly
adopts many of the same fundamentally-flawed assumptions and approaches employed
by the Agencies in the proposal, including assumptions and approaches concerning the
costs of vehicle technology, the rebound effect, and that fleet size, rather than demand
for transportation, determines total vehicle miles traveled.24
Our review of the limited information provided by NERA-Trinity and the Auto Alliance
suggests that the underlying data has been aggregated, but we are unable to determine
how or to what extent. For example, the final list of vehicle models used in the analysis,
the nesting structure, and method for developing the coefficients used in the sales
model to make its predictions are not provided. Without complete information, we have
been unable to evaluate the soundness of the model's design, reproduce the analysis to
confirm it reaches the results described, or perform any additional statistical tests. Such
analysis is necessary to be able to determine whether the results are significant,
unbiased, or reasonable and reliable.
A more troubling component of the analysis, which is equally, if not more, difficult to
disentangle without more details or interim model outputs, is the scrappage model. As
discussed in the enclosed memo from Professor Gillingham, the purportedly more
reasonable fleet sizes produced in the Auto Alliance's analysis may be the result of one
error masking another. The on-road fleet population is a combination of both new
vehicle sales and used vehicle scrappage. Simply put:
The Total On-Road Fleet (this year)=

Total On-Road Fleet (last year)+
New Sales (this year)
Scrapped Vehicles (this year).

In its analysis for the Auto Alliance, NERA-Trinity use a price elasticity for new vehicles
of -1.0, 25 which apparently was selected based on a study from nearly two decades ago
and is potentially an over-estimate of current and future elasticities. The NERA-Trinity
See, e.g., Comment by the Institute for Policy Integrity at New York University School of Law (Policy
Integrity), Docket #EPA-HQ-OAR-2018-0283-5083; NHTSA-2018-0067-12213, at 79-86.
25 A price elasticity of -1.0 means that the percent increase in new vehicle prices will result in an
equivalent percent decrease in new vehicle sales (all other attributes remaining constant), e.g. 5%
increase in prices results in a 5% decrease in sales.
24

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 23
elasticity therefore predicts a greater decline in new vehicle sales under the existing
standards (or conversely, a greater increase in new sales under the rollback standards)
than a smaller elasticity, such as -0.2 that was derived by the Agencies' new sales
model. As a result, even if the NERA-Trinity scrappage model were producing results
inconsistent with economic theory as the Agencies' model does, i.e. greater demand of
used vehicles at a higher price, then the elasticity of -1.0 would obscure that effect and
paint the overall fleet size as being more stable. The predicted new vehicle sales would
offset any economically inconsistent prediction of the number of vehicles scrapped.
However, both the underlying components (i.e., the new sales estimate and the
scrappage estimate) may be incorrect. Without additional information about the NERA­
Trinity scrappage model, we cannot assess what is actually happening within the model
to check for consistency with economic theory, but it seems possible that NERA-Trinity
have simply masked, rather than corrected, the fundamental flaws in the Agencies'
scrappage modeling.
Even without the information above, the analysis is patently flawed due to two major
decisions made by NERA-Trinity that bias the analysis to ensure that it predicts benefits
from a rollback. First, like the Agencies, the NERA-Trinity analysis assumes that
automakers will install, without any regulations, available fuel-saving technology that will
pay for itself within a specified time. As noted in CAR B's Detailed Comments, there is
no historical evidence for the assumption that automakers will systematically do so in
the absence of standards requiring this technology. 26 In addition, the Auto Alliance
provides no evidence in support of its assumption. 27 This has the subsequent effect of
significantly, but erroneously, diminishing the increase in fuel consumption caused by a
rollback. In other words, this assumption makes the rollback appear more beneficial, or
at least less costly, than it would be by minimizing the increase in fuel consumption that
would result from the rollback.
The NERA-Trinity analysis assumes manufacturers would install all technologies with a
60-month payback period, which is twice the 30-month payback period the Agencies
used in the SAFE Vehicles Rule and further reduces the harm created by rolling back
the standards. NERA-Trinity's 60-month payback period appears to be based on the
willingness-to-pay calculation for fuel economy derived from its new sales model.
However, this value represents the purported consumer's willingness-to-pay, which
does not necessarily measure or capture a manufacturer's decision-making process. It
is inappropriate to substitute, without reason, a consumer's valuation into an auto
manufacturer's decision-making process, and NERA-Trinity and the Auto Alliance fail to
26 See CARB Detailed Comments, p. 164, et seq.
21 However, as noted in CARB's Detailed Comments pp. 164-66, such "over-compliance" has historically
never occurred and average new fuel economy tracks very closely to the standards, while vehicles
improve along other dimensions such as performance or size.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 24
provide any justification for doing so. While consumer preferences may be important
considerations for automakers' vehicle design decisions, other factors may also be
relevant, related to engineering limitations, manufacturing capability, supplier
constraints, or other financial and market conditions that may necessitate some
compromises such that demand and supply for fuel efficiency will not be aligned
perfectly. Even if it were appropriate to substitute a consumer's valuation, NERA­
Trinity's 60-month willingness-to-pay estimate falls within a wide range of other
estimates in the economic literature (as discussed in the Vehicles Rule) and they
provide no justification as to why their value is superior to others, aside from circular
consistency with the New Vehicle Market Model.
The Auto Alliance's suggested analysis uses an improper methodology for calculating
the societal benefits from fuel economy improvements. The analysis relied solely on the
CAFE Model as developed by the Agencies for technology costs and effectiveness,
thereby incorporating the errors in that analysis. 28 However, NERA-Trinity considered
only some of the fuel savings produced by the existing standards or, to put it more
precisely, only some of the lost fuel savings that would be produced by the proposed
rollback.
Specifically, the analysis includes only 60-months' worth of fuel savings rather than all
of the savings that actually accrue over the life of a vehicle. While consumers may not
value all future fuel savings at the time of a new vehicle purchase, there is no
justification provided by the analysis for why society should not account for the benefit
from all the actual fuel savings that actually occur-savings that leave money in the
consumer's pocket and thus produce a real benefit, whether or not the consumer
factored those savings into the initial purchase.
As Professor Gillingham noted, this is not supported in the relevant economic
literature. 29 To not include the post-payback period fuel savings, all of which will be
realized by the consumer and by society regardless of whether or not the individual
consumer values them at the time of making a vehicle purchase, is wholly inconsistent
with proper regulatory impact analysis. The amount a consumer is "willing to pay" for
fuel savings when purchasing a vehicle-the consumer's ex ante valuation of fuel
savings-is not relevant to the question of what costs and benefits actually accrue to
society under emission and fuel economy standards. When undertaking cost benefit
analysis, it is the costs and benefits that will actually accrue-ex post-that are relevant.
NERA-Trinity provide no explanation of why consumers (or a society) would fail to fully
value the money saved by driving more fuel efficient vehicles, even if consumers did not
fully value these savings when making decisions about which vehicles to purchase.
28
29

See CARS Detailed Comments at p. 93, et seq.
See enclosed comments from Professor Ken Gillingham, p. 5.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 25

NERA-Trinity's analysis here is internally inconsistent. It accounts for all of the mobility
and refueling benefits associated with improved fuel economy over the full lifetime of
vehicles, as opposed to accounting for only some of these benefits as with the approach
taken for fuel savings. NERA-Trinity provides no explanation for why two different
methodologies are used for related benefits.
The analysis appears to calculate the increase in fuel tax revenue from more than the
first five years of vehicle usage; this is also inconsistent with the approach taken for fuel
savings. Moreover, the fuel tax revenue is not appropriate to include as a societal
benefit because, as the Agencies note, fuel taxes are transfer payments and thus
should be excluded. 30
As noted above, the Auto Alliance's analysis perpetuates some of the Agencies'
erroneous assumptions and thus cannot support a rollback. Regardless of what
payback period is selected, by relying on the Agencies' CAFE Model and inputs used to
support the proposal, the technology costs remain overstated due to the reasons
previously discussed in CARB's comment letter. These include, but are by no means
limited to, invalid input assumptions, algorithms that do not function correctly, and a
failure to maintain performance neutrality with the addition of new fuel-saving/emission­
reducing technologies.
The NERA-Trinity analysis uses the same. exaggerated rebound effect of 20 percent
based on a selective review of the literature as well as the Agencies' dramatically
undervalued domestic social cost of carbon. NERA-Trinity also adopts the Agencies'
incorrect assumptions regarding the sources of crude oil and where it will be refined dramatically limiting the upstream emissions impacts of the increased fuel consumption.
under the rollback. Like the Agencies' analysis, the NERA-Trinity analysis fails to
interact fleet size and total vehicle miles traveled, assuming that, regardless of the
number of vehicles in operation, vehicles will each be driven a fixed, age-specific
number of miles, which leads to unsupportably large VMT estimates under the original
standards and therefore falsely inflates the benefits of the rollback.
Given the limited information provided by the Auto Alliance, it is impossible to recreate
NERA-Trinity's analysis and correct all of the errors described here (or identify any
others that may exist). However, these errors alone illustrate that any reliance on the
NERA-Trinity analysis, in whole or in part, would be entirely arbitrary. In comparison,
CARB's prior critique of the SAFE Vehicles Rule analysis, the previous analyses by the
Agencies and CARB that concluded that existing standards remain appropriate, and the
30

83 Fed.Reg. at 43,088.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 26
underlying TAR demonstrate that the NERA-Trinity conclusions are likely wrong and
that the rollback would result in significant net societal costs. 31
Ill. Conclusion.

Having failed to meet their obligations under the law, NHTSA and U.S. EPA must
withdraw the SAFE Vehicles Rule. CARB continues to evaluate progress towards
reducing motor vehicle emissions and remains willing to discuss sensible, supported
adjustments to ensure the emissions are reduced while promoting a sustainable
economy, clean transportation system, and an innovative, competitive manufacturing
capability.
If you have any questions regarding the enclosed materials, please contact Mr. Pippin
C. Brehler, Senior Attorney, at [e-mail] or by phone at (916) 4458239.

Tu�-�\
Ellen M. Peter
Chief Counsel
Legal Office
Enclosure
cc:

Richard W. Corey
Executive Officer
Executive Office
Heidi King
Deputy Administrator
National Highway Traffic Safety Administration
U.S. Department of Transportation
1200 New Jersey Avenue. SE
Washington, D.C. 20590
Andrew Wheeler
Acting Administrator

31

See CARB's Detailed Comments at 330-336.

Mr. Christopher Lieske and Mr. James Tamm
December 19, 2018
Page 27
U.S. Environmental ProtectionAgency
1301 ConstitutionAvenue NW
Washington, D.C. 20460
John Shoaff
Director, Office ofAir Policy and Program Support
U.S. Environmental ProtectionAgency
Office ofAir and Radiation
1301 Constitution Avenue NW
Washington, D.C. 20460
Chief Counsel
FOIAAppeal
Attn: Andrew J. DiMarsico, Senior Attorney
National Highway Traffic SafetyAdministration
U.S. Department of Transportation
1200 New JerseyAvenue SE
West Building, W41-227
Washington, D.C. 20590
A
NHTS Executive Secretariat
FOIAAppeal
1200 New Jersey Avenue SE
W41-307
Washington, D.C. 20590