Alex Cantero , individually and on behalf of all others similarly situated v. Bank of America, N.a.

21-400, 21-403
Cantero v. Bank of America, N.A., Hymes v. Bank of America, N.A.
United States Court of Appeals
For the Second Circuit
August Term 2024
Argued: March 3, 2025
Decided: May 5, 2026
Nos. 21-400, 21-403
ALEX C ANTERO ,
individually and on behalf of all others similarly situated,
Plaintiff-Appellee,
v.
B ANK OF A MERICA, N.A.,
Defendant-Appellant.
S AUL R. HYMES , ILANA H ARWAYNE -G IDANSKY ,
on behalf of themselves and all others similarly situated,
Plaintiffs-Appellees,
v.
B ANK OF A MERICA, N.A.,
Defendant-Appellant.
On Appeal from the United States District Court
for the Eastern District of New York

-- 1 of 65 --

2
Before: L IVINGSTON, Chief Judge, and PARK and PÉREZ , Circuit Judges.
Plaintiffs deposited money in mortgage-escrow accounts with
Bank of America (“BOA”), which refused to pay the two-percent
interest rate required by New York General Obligations Law (“GOL”)
§ 5-601. BOA argued that New York’s interest-on-escrow
requirement is preempted by federal banking law, which authorizes
federally chartered national banks to offer mortgage-escrow accounts
without requiring them to pay interest.
We previously concluded that GOL § 5-601 is preempted
because it exercises “control over a banking power granted by the
federal government.” Cantero v. Bank of Am., N.A., 49 F.4th 121, 125
(2d Cir. 2022). But the Supreme Court vacated that decision and
instructed us to conduct a “nuanced comparative analysis” of New
York’s interest-on-escrow law and the Court’s banking preemption
precedents. Cantero v. Bank of Am., N.A., 602 U.S. 205, 220 (2024).
Having done so, we again conclude that GOL § 5-601 is
preempted. First, New York’s interest-on-escrow requirement affects
a national banking power: the power to offer mortgages. Second, it
targets banks and limits their broad power to set the terms of
mortgage-escrow accounts, so it is similar in nature to the preempted
laws in Fidelity Federal Savings and Loan Ass’n v. de la Cuesta, 458 U.S.
141 (1982), and Barnett Bank of Marion County v. Nelson, 517 U.S. 25
(1996). Finally, its degree of interference on banks’ ability to make
real-estate loans efficiently is similarly severe to the interference

-- 2 of 65 --

3
created by the preempted advertising law in Franklin National Bank of
Franklin Square v. New York, 347 U.S. 373 (1954).
The nature and degree of GOL § 5-601’s interference with
federal law is “more akin” to the cases in which state laws were
preempted than those in which state laws were not. Cantero, 602 U.S.
at 220. We thus REVERSE the orders of the district court denying
BOA’s motions to dismiss and REMAND the cases for further
proceedings consistent with this opinion.
Judge Pérez dissents in a separate opinion.
JONATHAN E. TAYLOR (Deepak Gupta on the brief), Gupta
Wessler LLP, Washington, DC; with Hassan Zavareei &
Anna C. Haac, Tycko & Zavareei LLP, Washington, DC;
Jonathan M. Streisfeld, Kopelowitz Ostrow Ferguson
Weiselberg Gilbert, Ft. Lauderdale, FL; Todd S. Garber,
Finkelstein, Blankinship, Frei-Pearson & Garber, LLP,
White Plains, NY for Plaintiff-Appellee Alex Cantero.
JONATHAN E. TAYLOR (Deepak Gupta on the brief), Gupta
Wessler LLP, Washington, DC; with Mark C. Rifkin &
Matthew M. Guiney, Wolf Haldenstein Adler Freeman &
Herz LLP, New York, NY for Plaintiffs-Appellees Saul R.
Hymes and Ilana Harwayne-Gidansky.
L ISA S. BLATT (Enu Mainigi, Sarah M. Harris, Aaron Z.
Roper, Erin M. Sielaff on the brief), Williams & Connolly
LLP, Washington, DC; with Mark W. Mosier & Andrew
Soukup, Covington & Burling LLP, Washington, DC;

-- 3 of 65 --

4
Thomas M. Hefferon, Goodwin Procter LLP,
Washington, DC for Defendant-Appellant.
Matthew Lambert, Conference of State Bank
Supervisors, Washington, DC; Stefan L. Jouret, Jouret
LLC, Boston, MA; Arthur E. Wilmarth, Jr., George
Washington University Law School, Washington, DC for
Amici Curiae Conference of State Bank Supervisors &
American Association of Residential Mortgage Regulators in
Support of Plaintiff-Appellee in Cantero v. Bank of
America.
Brenna Bird, Attorney General, Eric H. Wessan, Solicitor
General, State of Iowa, Des Moines, IA; Letitia James,
Attorney General, Barbara D. Underwood, Solicitor
General, Andrea Oser, Deputy Solicitor General, Sarah L.
Rosenbluth, Assistant Solicitor General, State of New
York, Buffalo, NY; Rob Bonta, Attorney General, State of
California, Sacramento, CA; Philip J. Weiser, Attorney
General, State of Colorado, Denver, CO; William Tong,
Attorney General, State of Connecticut, Hartford, CT;
Kathleen Jennings, Attorney General, State of Delaware,
Wilmington, DE; Ashley Moody, Attorney General, State
of Florida, Tallahassee, FL; Christopher M. Carr,
Attorney General, State of Georgia, Atlanta, GA; Raúl R.
Labrador, Attorney General, State of Idaho, Boise, ID;
Kwame Raoul, Attorney General, State of Illinois,
Chicago, IL; Theodore E. Rokita, Attorney General, State
of Indiana, Indianapolis, IN; Aaron M. Frey, Attorney
General, State of Maine, Augusta, ME; Anthony G.

-- 4 of 65 --

5
Brown, Attorney General, State of Maryland, Baltimore,
MD; Dana Nessel, Attorney General, State of Michigan,
Lansing, MI; Keith Ellison, Attorney General, State of
Minnesota, St. Paul, MN; Austin Knudsen, Attorney
General, State of Montana, Helena, MT; Michael T.
Hilgers, Attorney General, State of Nebraska, Lincoln,
NE; Aaron D. Ford, Attorney General, State of Nevada,
Carson City, NV; Matthew J. Platkin, Attorney General,
State of New Jersey, Trenton, NJ; Gentner Drummond,
Attorney General, State of Oklahoma, Oklahoma City,
OK; Ellen F. Rosenblum, Attorney General, State of
Oregon, Salem, OR; Michelle A. Henry, Attorney
General, Commonwealth of Pennsylvania, Harrisburg,
PA; Peter F. Neronha, Attorney General, State of Rhode
Island, Providence, RI; Marty J. Jackley, Attorney
General, State of South Dakota, Pierre, SD; Sean D. Reyes,
Attorney General, State of Utah, Salt Lake City, UT;
Charity R. Clark, Attorney General, State of Vermont,
Montpelier, VT; Jason S. Miyares, Attorney General,
Commonwealth of Virginia, Richmond, VA; Robert W.
Ferguson, Attorney General, State of Washington,
Olympia, WA; Bridget Hill, Attorney General, State of
Wyoming, Cheyenne, WY; Brian L. Schwalb, Attorney
General, District of Columbia, Washington, DC, for Amici
Curiae the States of New York, Iowa, California, Colorado,
Connecticut, Delaware, Florida, Georgia, Idaho, Illinois,
Indiana, Maine, Maryland, Michigan, Minnesota, Montana,
Nebraska, Nevada, New Jersey, Oklahoma, Oregon,

-- 5 of 65 --

6
Pennsylvania, Rhode Island, South Dakota, Utah, Vermont,
Virginia, Washington, and Wyoming, and the District of
Columbia, in support of Plaintiffs-Appellees in Hymes v.
Bank of America.
H. Rodgin Cohen, Matthew A. Schwartz, Shane M.
Palmer, Sullivan & Cromwell LLP, New York, NY; Gregg
L. Rozansky & Tabitha Edgens, The Bank Policy
Institute, Washington, DC; Jonathan D. Urick & Tyler S.
Badgley, U.S. Chamber Litigation Center, Washington,
DC; David Pommerehn, Consumer Bankers Association,
Washington, DC; Thomas Pinder & Andrew Doersam,
The American Bankers Association, Washington, DC;
Justin Wiseman & Alisha Sears, Mortgage Bankers
Association, Washington, DC for Amici Curiae The Bank
Policy Institute, American Bankers Association, Consumer
Bankers Association, Mortgage Bankers Association, Chamber
of Commerce of the United States of America in Support of
Defendant-Appellant in Cantero v. Bank of America and
Hymes v. Bank of America.
PARK, Circuit Judge:
Plaintiffs deposited money in mortgage-escrow accounts with
Bank of America (“BOA”), which refused to pay the two-percent
interest rate required by New York General Obligations Law (“GOL”)
§ 5-601. BOA argued that New York’s interest-on-escrow
requirement is preempted by federal banking law, which authorizes

-- 6 of 65 --

7
federally chartered national banks to offer mortgage-escrow accounts
without requiring them to pay interest.
We previously concluded that GOL § 5-601 is preempted
because it exercises “control over a banking power granted by the
federal government.” Cantero v. Bank of Am., N.A., 49 F.4th 121, 125
(2d Cir. 2022). But the Supreme Court vacated that decision and
instructed us to conduct a “nuanced comparative analysis” of New
York’s interest-on-escrow law and the Court’s banking preemption
precedents. Cantero v. Bank of Am., N.A., 602 U.S. 205, 220 (2024).
Having done so, we again conclude that GOL § 5-601 is
preempted. First, New York’s interest-on-escrow requirement affects
a national banking power: the power to offer mortgages. Second, it
targets banks and limits their broad power to set the terms of
mortgage-escrow accounts, so it is similar in nature to the preempted
laws in Fidelity Federal Savings and Loan Ass’n v. de la Cuesta, 458 U.S.
141 (1982), and Barnett Bank of Marion County v. Nelson, 517 U.S. 25
(1996). Finally, its degree of interference on banks’ ability to make
real-estate loans efficiently is similarly severe to the interference
created by the preempted advertising law in Franklin National Bank of
Franklin Square v. New York, 347 U.S. 373 (1954).
The nature and degree of GOL § 5-601’s interference with
federal law is “more akin” to the cases in which state laws were
preempted than those in which state laws were not. Cantero, 602 U.S.
at 220. We thus reverse the orders of the district court denying BOA’s
motions to dismiss and remand the cases for further proceedings
consistent with this opinion.

-- 7 of 65 --

8
I. BACKGROUND1
A. Legal Context
1. National Banking System
Federal banking law authorizes national banks to exercise
express and incidental powers. Express powers are specifically
granted to national banks by statute. They include the powers to
“make contracts,” to “sue and be sued,” to “loan[] money on personal
security,” and, as relevant here, to “make, arrange, purchase or sell
loans or extensions of credit secured by liens on interests in real
estate.” 12 U.S.C. §§ 24, 371(a). National banks also have “all such
incidental powers as shall be necessary to carry on the business of
banking.” Id. § 24 (Seventh).
Together, the powers to make “loans . . . secured by liens on
interests in real estate” and to exercise necessary “incidental powers”
allow national banks to offer and set the terms of mortgage-escrow
accounts. Id. §§ 24 (Seventh), 371(a). A mortgage-escrow account
requires a borrower to set aside money for tax and insurance
payments on a home. The bank then uses that money to make
payments on the borrower’s behalf, reducing the possibility of default
and protecting the property from uninsured damage. Mortgage-
escrow accounts are a “crucial risk mitigation tool that supports safe
and sound mortgage lending,” Real Estate Lending Escrow Accounts,
90 Fed. Reg. 61099, 61100 (proposed Dec. 30, 2025), without which
there is a “higher probability of foreclosure,” Escrow Requirements
1 The background is set out more fully in Cantero, 49 F.4th at 125-29.
We summarize it here as necessary to explain our decision.

-- 8 of 65 --

9
Under the Truth in Lending Act, 78 Fed. Reg. 4726, 4735 (Jan. 22,
2013). A national bank’s ability to decide “how to structure its escrow
operations and whether and what extent to offer any compensation
to customers” is thus “a clear logical outgrowth of national banks’
other powers to manage and protect collateral.” Real Estate Lending
Escrow Accounts, 90 Fed. Reg. at 61102.
“In the 1970s, Congress found that some national banks were
engaging in ‘certain abusive practices’ and that ‘significant reforms’
were necessary to protect borrowers.” Cantero, 602 U.S. at 211
(quoting 12 U.S.C. § 2601(a)). Congress thus passed the Real Estate
Settlement Procedures Act of 1974 (“RESPA”), which “extensively
regulates national banks’ operation of escrow accounts.” Id. RESPA
requires national banks to “promptly return[] to the borrower” any
“balance” left over in escrow accounts after a loan is paid, 12 U.S.C.
§ 2605(g), and limits the amounts banks can require borrowers to
deposit in mortgage-escrow accounts, id. § 2609(a). But RESPA “does
not mandate that national banks pay interest to borrowers on the
balances of their escrow accounts.” Cantero, 602 U.S. at 211.2
2 Another federal statute, the Truth in Lending Act (“TILA”),
“requires national banks to operate escrow accounts for certain mortgages,”
but it “does not apply to the mortgages in this case.” Cantero, 602 U.S. at
211 n.1. TILA expressly incorporates state interest-on-escrow requirements
for those mandatory accounts: national banks “must pay interest to the
borrower in the manner as prescribed by an applicable State or Federal
law.” Id. (cleaned up).

-- 9 of 65 --

10
2. Preemption Framework
The Dodd-Frank Wall Street Reform and Consumer Protection
Act establishes the “controlling legal standard for when a ‘State
consumer financial law,’ like New York’s interest-on-escrow law, is
preempted with respect to national banks.” Cantero, 602 U.S. at 213
(quoting 12 U.S.C. § 25b(b)(1)). It states that state laws are preempted
“only if . . . in accordance with the legal standard for preemption in
the decision of the Supreme Court of the United States in Barnett Bank
. . . the State consumer financial law prevents or significantly
interferes with the exercise by the national bank of its powers.” 12
U.S.C. § 25b(b)(1)(B).
Barnett Bank requires us to survey prior Supreme Court cases
“to demarcate when a state law significantly interferes with [a]
national bank’s exercise of its powers.” Cantero, 602 U.S. at 215
(cleaned up). Those cases establish that “some (but not all) non-
discriminatory state laws that regulate national banks are
preempted.” Id. at 221.
On one hand, National Bank v. Commonwealth, 76 U.S. 353 (1869);
McClellan v. Chipman, 164 U.S. 347 (1896); and Anderson National Bank
v. Luckett, 321 U.S. 233 (1944), upheld state laws regulating national
banks. National Bank concerned a Kentucky tax on bank stock.
McClellan involved a generally applicable Massachusetts law that
prohibited banks from receiving preferential transfers. And Anderson
dealt with a Kentucky escheat law that presumed inactive deposits
were abandoned. The Court concluded that these laws were not
preempted because they did not “prevent or significantly interfere

-- 10 of 65 --

11
with the national bank’s exercise of its powers.” Barnett Bank, 517 U.S.
at 33.
On the other hand, First National Bank of San Jose v. California,
262 U.S. 366 (1923); Franklin National Bank of Franklin Square v. New
York, 347 U.S. 373 (1954); Fidelity Federal Savings and Loan Ass’n v. de la
Cuesta, 458 U.S. 141 (1982); and Barnett Bank of Marion County v.
Nelson, 517 U.S. 25 (1996), struck down state laws. In First National
Bank of San Jose, California escheated all deposits that were inactive
for over twenty years. In Franklin, New York barred bank
advertisements from using the word “savings.” In Fidelity, California
restricted the use of due-on-sale clauses. And in Barnett Bank, Florida
prevented banks from selling insurance. The Supreme Court
concluded that these laws were preempted.
3. New York’s Interest-on-Escrow Requirement
GOL § 5-601 is a consumer financial law requiring “mortgage
investing institution[s]” to credit escrow accounts for certain
residences in New York “with dividends or interest at a rate of not
less than two per centum per year.” GOL § 5-601. It allows New
York’s Superintendent of Financial Services to prescribe a higher
minimum interest rate after “consider[ing] pertinent economic and
cost factors.” N.Y. Banking L. § 14-b(2).
GOL § 5-601 was enacted in 1974, but no court has enforced
compliance with its interest-on-escrow requirement. As early as 2004,
the Office of the Comptroller of the Currency (“OCC”) advised that
banks may “make real estate loans . . . without regard to state law
limitations.” Bank Activities and Operations; Real Estate Lending
and Appraisals, 69 Fed. Reg. 1904, 1917 (Jan. 13, 2004).

-- 11 of 65 --

12
In December 2025, the OCC issued a notice proposing a rule
that would “expressly codify [national] banks’ power to establish and
maintain escrow accounts” and “clarify that the terms and conditions
of escrow accounts, including the extent of any compensation paid to
customers, are business decisions to be made by each bank.” Real
Estate Lending Escrow Accounts, 90 Fed. Reg. at 61103. It
simultaneously proposed a “preemption determination” concluding
that “the National Bank Act preempts New York’s Gen. Oblig. Law
section 5-601” and eleven other similar state laws. Preemption
Determination: State Interest-on-Escrow Laws, 90 Fed. Reg. 61093,
61094 (proposed Dec. 30, 2025).
B. Facts and Procedural History
Plaintiffs Alex Cantero, Saul Hymes, and Ilana Harwayne-
Gidansky took out mortgage loans from BOA, which required
Plaintiffs to make deposits in mortgage-escrow accounts. But BOA
refused to pay the two-percent interest required by GOL § 5-601. So
Plaintiffs filed two putative class actions, alleging breach of contract
and other claims.
BOA moved to dismiss the complaints, arguing that it did not
need to pay interest because GOL § 5-601 is preempted by federal
banking law. The district court disagreed. It denied BOA’s motions
to dismiss, see Hymes v. Bank of Am., N.A., 408 F. Supp. 3d 171
(E.D.N.Y. 2019), and certified the preemption question for
interlocutory appeal, see Hymes v. Bank of Am., N.A., No. 18-cv-2352,
2020 WL 9174972 (E.D.N.Y. Sept. 29, 2020).
On appeal, we reversed the district court’s denial of BOA’s
motions. We concluded that federal banking law preempts GOL § 5-

-- 12 of 65 --

13
601 because an interest-on-escrow requirement “would exert control
over a banking power—and thus, if taken to its extreme, threaten to
‘destroy’ the grant made by the federal government.” Cantero, 49
F.4th at 132 (quoting McCulloch v. Maryland, 17 U.S. (4 Wheat.) 316,
431 (1819)). But we did “not endeavor to assess whether the degree
of the state law’s impact on national banks would be sufficient to
undermine that power.” Id.
The Supreme Court vacated our decision and remanded for us
to conduct the “kind of nuanced comparative analysis” that the
Barnett Bank standard requires. Cantero, 602 U.S. at 220. In “Barnett
Bank and each of the earlier [banking preemption] precedents, the
Court [assessed] the nature and degree of the state laws’ alleged
interference.” Id. at 220 n.3. So rather than focus only on whether a
state law “controls” a national bank’s powers, we should make a
“practical assessment of the nature and degree of the interference”
based on “the text and structure of the laws, comparison to other
precedents, and common sense.” Id. at 219, 220 n.3.
With this instruction in mind, we revisit whether, under the
Barnett Bank standard, federal banking law preempts New York’s
interest-on-escrow requirement.
II. DISCUSSION
A. Legal Standard
To determine whether a state law is preempted, we first ask
whether the state law affects “the exercise by the national bank of its
powers.” 12 U.S.C. § 25b(b)(1)(B). If it does, we then consider
whether the state law “prevents or significantly interferes” with those

-- 13 of 65 --

14
powers by analyzing two factors: (1) the “nature” of the state law’s
interference, and (2) the “degree” of interference. Cantero, 602 U.S. at
220 & n.3.
We assess these factors in a “nuanced comparative analysis” in
light of the Supreme Court’s preemption cases. Id. at 220. If the
nature and degree of a state law’s interference with national bank
powers “is more akin to the interference in cases like Franklin, Fidelity,
First National Bank of San Jose, and Barnett Bank itself, then the state law
is preempted.” Id. Conversely, if the nature and degree of the state
law’s interference with national bank powers “is more akin to the
interference in cases like Anderson, National Bank v. Commonwealth,
and McClellan, then the state law is not preempted.” Id.
1. The Banking Power at Issue
The National Bank Act authorizes national banks to exercise
express and incidental powers. Historically, “grants of both
enumerated and incidental ‘powers’ to national banks [have been
interpreted] as grants of authority not normally limited by, but rather
ordinarily pre-empting, contrary state law.” Barnett Bank, 517 U.S. at
32.
But state laws that do not affect a national bank’s “exercise . . .
of its powers” are not preempted. 12 U.S.C. § 25b(b)(1)(B). Consider
Kentucky’s tax on “bank stock” in National Bank. 76 U.S. at 354. That
law affected—indeed, targeted—banks. But the Court held the tax
was not preempted because it was “no great[] interference with the
functions of the bank” and “in no manner hinder[ed] it from
performing all the duties of financial agent of the government.” Id. at

-- 14 of 65 --

15
362-63. State laws that merely affect banks, rather than banking
powers, are not typically preempted.
2. Nature of the Interference
If a state law affects a banking power, we must identify the
“nature” of the interference by analyzing the “text and structure” of
the relevant state and federal laws. Cantero, 602 U.S. at 220 n.3. Two
principles emerge from this analysis. First, generally applicable state
laws are unlikely to be preempted. Second, state laws that prohibit a
bank from exercising express or broad powers are likely to be
preempted.
a. Generally Applicable Laws
“[N]ational banks . . . remain subject to state law governing
‘their daily course of business’ such as generally applicable state
contract, property, and debt-collection laws.” Cantero, 602 U.S. at 219
(quoting National Bank, 76 U.S. at 361-62). So even when a generally
applicable state law has some effect on a bank’s exercise of its powers,
it is ordinarily not preempted. That was the case in McClellan, in
which the Court upheld Massachusetts’s generally applicable
fraudulent transfer law because it subjected banks to “the same
conditions and restrictions to which all the other citizens of the state
are subjected.” 164 U.S. at 358. The Court explained that “in the
broadest sense, any limitation by a state on the making of contracts is
a restraint upon the power of a national bank within the state” but
that “the purpose and object of congress in enacting the national bank
law was to leave such banks, as to their contracts in general, under
the operation of the state law.” Id. at 358-59. So it concluded that “the
general and undiscriminating law of the state of Massachusetts

-- 15 of 65 --

16
subjecting the taking of real estate to certain restrictions” was not
preempted by federal law “permit[ting] national banks to take real
estate for given purposes.” Id. at 358, 361.
Even state laws that target banks or banking powers are
unlikely to be preempted if they are part of a generally applicable
scheme. This principle helps to explain the diverging results in First
National Bank of San Jose and Anderson, which both addressed state
laws requiring banks to escheat abandoned deposits to the state.
California’s law created a non-rebuttable presumption that deposits
unclaimed for 20 years were abandoned, but Kentucky’s law created
a rebuttable presumption that deposits unclaimed for 10 years were
abandoned. See First Nat’l Bank of San Jose, 262 U.S. at 366-67;
Anderson, 321 U.S. at 236-37. The Court struck down California’s law
but upheld Kentucky’s. It emphasized that it could not “discern any
greater or different effect [of the Kentucky law] . . . from the
application of the ancient law of escheat or forfeiture of goods,” which
allowed the state to appropriate “abandoned personal property.”
Anderson, 321 U.S. at 240, 252. But California’s non-rebuttable
presumption of “seizure and escheat . . . for mere dormancy” was an
“unusual” variation on the common-law doctrine. Id. at 251. Even
though both laws targeted banks, only Kentucky’s was consistent
with the generally applicable common-law escheat rule, so it was not
preempted.3
3 This principle explains why fair-lending and false-advertising laws
are unlikely to be preempted. See, e.g., Cuomo v. Clearing House Ass'n, LLC,
557 U.S. 519, 522-23, 536 (2009) (permitting enforcement of New York fair-

-- 16 of 65 --

17
b. Prohibitions on Express and Broad Powers
The text and structure of the relevant laws also bear on the
nature of interference by illuminating how directly they conflict.
Of course, when state and federal laws “impose directly
conflicting duties on national banks,” the state law is always
preempted. Barnett Bank, 517 U.S. at 31. But when federal statutes
grant permissive powers, as opposed to imposing mandatory duties,
“compliance with both” state and federal law is not “a physical
impossibility,” because a bank can choose not to exercise its power in
the manner prohibited by state law. Fidelity, 458 U.S. at 155 (citation
omitted). In those cases, the Court has found state laws preempted
when federal law grants a “broad, not a limited, permission.” Barnett
Bank, 517 U.S. at 32.
That was the case in Fidelity and Barnett Bank, where the text
and structure of federal law indicated that Congress granted banks
broad authority to exercise express powers, such that state limits on
those powers were preempted. In Fidelity, the Court emphasized that
a federal regulation granted a federal savings-and-loan association
the power to include due-on-sale clauses in contracts “at its option,”
so California’s law prohibiting some due-on-sale clauses
impermissibly deprived banks of the “flexibility” the regulation
granted. 517 U.S. at 155. Likewise, in Barnett Bank, a federal statute
providing, “without relevant qualification, that national banks may
act as the agent for insurance sales” suggested “a broad, not a limited,
lending law against national banks). Such laws typically apply generally
applicable rules to banks, so they are similar to the escheat rule in Anderson.

-- 17 of 65 --

18
permission.” Barnett Bank, 517 U.S. at 32 (cleaned up). So a Florida
law prohibiting banks from acting as insurance agents was
preempted. Id.4
3. Degree of Interference
The Court has relied on “common sense” to assess how severely
a state law interferes with a banking power. Cantero, 602 U.S. at 220
n.3. It has suggested that a state law’s interference is severe when it
prevents a bank from efficiently exercising its powers or makes a
bank’s product undesirable to consumers.
The first consideration led the Court in Franklin to conclude that
New York’s law prohibiting banks from using the word “savings” in
advertisements was preempted by federal law permitting banks to
receive savings accounts. See Cantero, 602 U.S. at 216. There, “the
Court determined that the New York law significantly interfered with
the banks’ power because the banks could not advertise effectively”
their power to receive savings deposits without using the word
“savings.” Id. (citing Franklin, 347 U.S. at 377-78). “[S]tate law could
not interfere with the national bank’s ability to [advertise savings
accounts] efficiently,” so it was preempted. Id. (emphasis added).
4 To be sure, not all state laws that limit broad grants of power are
preempted. For example, the fraudulent transfer law in McClellan
interfered with national banks’ broad express power to take real estate, but
the law was not preempted because it was generally applicable and
restricted banks in only “particular and exceptional circumstances.” 164
U.S. at 358. A “nuanced comparative analysis” must balance the different
attributes of a state law that favor preemption against those that do not.
Cantero, 602 U.S. at 220.

-- 18 of 65 --

19
Similarly, the absence of an impact on a bank’s ability to
exercise its powers efficiently supported the Court’s conclusion in
McClellan. There, Massachusetts’s fraudulent conveyance law was
not preempted because it affected banks in only “particular and
exceptional circumstances,” and thus did not “impair[] the efficiency
of national banks.” McClellan, 164 U.S. at 358. These cases highlight
that a state law’s impact on a national bank’s efficiency informs the
degree of the state law’s interference with federal law.
The second consideration—that a state law’s interference is
more severe if it makes a bank’s services less desirable to consumers—
undergirded the Court’s conclusions in First National Bank of San Jose
and Anderson. In First National Bank of San Jose, California’s law
imposing a non-rebuttable presumption that unclaimed deposits
were abandoned, was preempted because consumers “might well
hesitate to subject their funds to possible confiscation.” 262 U.S. at
370. But in Anderson, Kentucky’s similar law imposing a rebuttable
presumption was not preempted because it “may operate for the
benefit and security of depositors” and would not “deter them from
placing their funds in national banks.” 321 U.S. at 252. The Court’s
expectations about whether California’s and Kentucky’s escheat laws
would deter customers thus supported the different outcomes in
these cases.5
5 Plaintiffs argue that BOA must prove that a “state law posed a
significant practical impediment to the exercise of an express power” with
“evidence.” Appellees’ Supp. Br. at 8. In support, they argue that the
“Court in Franklin . . . had the benefit of a ‘large record’ documenting the
[New York] law’s real-world ‘consequences upon banks.’” Id. at 7 (quoting

-- 19 of 65 --

20
B. Application
New York’s interest-on-escrow law is preempted because it
affects a broad federal grant of power to set the terms of mortgage-
escrow accounts and it impedes national banks’ ability to offer those
accounts efficiently. These characteristics make GOL § 5-601 more
like the laws in Franklin, Fidelity, First National Bank of San Jose, and
Barnett Bank, in which the Supreme Court found preemption, than
those in Anderson, National Bank, and McClellan, in which the Court
did not.
1. The Banking Power at Issue
New York’s interest-on-escrow requirement does not merely
affect BOA, it affects BOA’s “exercise of its powers.” Barnett Bank, 517
U.S. at 33 (emphasis added). Federal law gives national banks the
power to “make, arrange, purchase or sell loans or extensions of credit
secured by liens on interests in real estate,” 12 U.S.C. § 371(a), which,
in conjunction with their “incidental powers,” id. § 24 (Seventh),
authorizes them to offer mortgage-escrow accounts, see Real Estate
Lending Escrow Accounts, 90 Fed. Reg. at 61102. New York’s law
interferes with that power by limiting the terms on which banks may
offer these accounts—specifically, by requiring them to pay at least
two percent interest to customers. In short, banks may offer
Franklin, 347 U.S. at 376). But Franklin never discussed that record, nor has
the Court in any other preemption case considered record evidence of a
state law’s “real-world consequences upon banks.” Following the Court’s
instruction to ground our analysis in “the text and structure of the laws,
comparison to other precedents, and common sense,” we do not require
record evidence of a law’s real-world effects. Cantero, 602 U.S. at 220 n.3.

-- 20 of 65 --

21
mortgage-escrow accounts without interest under federal law, but
under New York law, they may not.
2. Nature of the Interference
We analyze the nature of New York’s interference with federal
law by considering the text and structure of the state and federal laws.
New York’s law is not generally applicable, like the fraudulent
conveyance law in McClellan, nor does it apply a generally applicable
rule to banks, like the escheat law in Anderson. Rather, GOL § 5-601
targets banks. That characteristic differentiates it from the non-
preempted laws in McClellan and Anderson.
We also consider whether New York’s law involves a
prohibition on an expressly- or broadly-granted power, as in Fidelity
and Barnett Bank. GOL § 5-601 does not limit an express power
because national banks’ power to set interest rates on mortgage-
escrow accounts is incidental to the power to extend “credit secured
by liens on interests in real estate.” 12 U.S.C. § 371(a). But other
federal laws governing mortgage accounts suggest that Congress
granted this incidental power broadly.
First, the Real Estate Settlement Procedures Act “extensively
regulates national banks’ operation of escrow accounts.” Cantero, 602
U.S. at 211. “But as relevant to this case, RESPA . . . does not mandate
that national banks pay interest to borrowers on the balances of their
escrow accounts.” Id. The omission of an interest rate requirement
from RESPA, a statute regulating many other aspects of escrow
accounts, suggests that national banks have a broad power to set
those rates. As the OCC explained, “RESPA, in legislating a system
of escrow account disclosures and amount limits, implicitly

-- 21 of 65 --

22
recognizes the flexibility banks have in deciding . . . whether and to
what extent to pay interest on escrowed funds.” Real Estate Lending
Escrow Accounts, 90 Fed. Reg. at 61102.
Second, the Truth in Lending Act requires national banks to
“pay interest to the borrower in the manner as prescribed by an
applicable State or Federal law” for certain mortgage-escrow
accounts, but not those at issue here. Cantero, 602 U.S. at 211 n.1
(cleaned up). The Court has relied on statutes that, like TILA,
expressly incorporated some state laws to conclude that federal law
preempted other state laws regulating the same or similar activities.
In Barnett Bank, for example, the Court inferred that federal law
granted national banks a “broad, not a limited, permission” to sell
insurance because federal law “specifically refer[red] to state
regulation, while limiting that reference to licensing—not of banks or
insurance agents, but of the insurance companies whose policies the
bank, as insurance agent, will sell.” 517 U.S. at 32. Similarly, in
Franklin, the Court found “no indication that Congress intended to
make this phase of national banking [i.e., advertising savings
accounts] subject to local restrictions, as it has done by express
language in several other instances,” thus inferring preemption in
part from statutory silence. 347 U.S. at 378. Finally, in Fidelity, the
Court concluded that “provisions incorporating specific aspects of
state law [would be] needlessly repetitive” if similar state laws were
not preempted and thus “decline[d] to construe the Act so as to render
these provisions nugatory.” 458 U.S. at 163. Congress’s decision to
adopt specific state interest-on-escrow requirements in TILA thus
also suggests that similar state laws—like GOL § 5-601—are
preempted.

-- 22 of 65 --

23
To be sure, the breadth of the federal grant of power here is not
express, as in Fidelity and Barnett Bank, where federal statutes
“explicitly” granted an express power. Barnett Bank, 517 U.S. at 34.
But the clear implication from RESPA and TILA—that national banks
have broad power to set interest rates for most mortgage-escrow
accounts—makes the nature of interference in this case “akin” to that
in Fidelity and Barnett Bank. Cantero, 602 U.S. at 220.
3. Degree of Interference
Next, we turn to the “degree” of New York’s interference by
considering whether GOL § 5-601 interferes with banks’ ability to
offer mortgage-escrow accounts efficiently and to attract customers.
The “vast majority of home mortgages come with escrow
accounts.” Cantero, 602 U.S. at 211. Banks administering these
accounts incur operational and compliance costs, which they may
recover by investing escrow funds or through other means,
depending on their “business strategy, costs, market demand, [and]
competition.” Real Estate Lending Escrow Accounts, 90 Fed. Reg. at
61100. But on each account associated with certain residences in New
York, GOL § 5-601 requires banks to pay at least two percent interest.
This requirement “raises the cost to national banks to use escrow
accounts” for residences in New York. Kivett v. Flagstar Bank, FSB, 154
F.4th 640, 648 (9th Cir. 2025). New York’s mandated interest rate may
cause national banks to “offer escrow accounts on fewer real estate
loans; attempt to recoup costs in other ways; or even reduce lending.”

-- 23 of 65 --

24
Preemption Determination: State Interest-on-Escrow Laws, 90 Fed.
Reg. at 61097.
This interference is similar in degree to that of New York’s
advertising law in Franklin. Both laws restrict an incidental power
that is essential to banks’ exercise of an express power. The law in
Franklin limited national banks’ ability to advertise savings accounts,
and advertising was “one of the most usual and useful of weapons”
to solicit and receive savings deposits. 347 U.S. at 377. Likewise, GOL
§ 5-601 limits banks’ ability to set the terms of mortgage-escrow
accounts when the ability “to effectively and efficiently set the terms
and conditions of their escrow accounts . . . is a core component of
banks’ mortgage lending powers.” Real Estate Lending Escrow
Accounts, 90 Fed. Reg. at 61100. Both laws also restrict banks’ ability
to exercise that incidental power—the Franklin law, by prohibiting all
advertisements with the word “savings,” and GOL § 5-601, by
prohibiting all mortgage-escrow accounts with interest rates below
two percent. 6
If anything, GOL § 5-601 involves a more severe limit than the
Franklin law. A “state law that alters a national bank’s pricing almost
6 The two percent floor is much higher than the prevailing interest
rate on certificates of deposit, which ranged from 0.16% to 0.91% between
2010 and 2020. Fed. Deposit Ins. Corp., National Rate on Non-Jumbo Deposits
(less than $100,000): 12 Month CD, https://perma.cc/Y4GA-97E8 (last visited
Apr. 29, 2026); see also Fed. Deposit Ins. Corp., National Rate on Non-Jumbo
Deposits (greater or equal to $100,000): 12 Month CD, https://perma.cc/7W8X-
J36L (last visited Apr. 29, 2026) (similar range for higher-value deposits).
And GOL § 5-601 authorizes the Superintendent of Financial Services to
raise the minimum interest rate even higher. See supra at 11.

-- 24 of 65 --

25
by definition interferes more with the bank’s powers than a simple
advertising restriction.” Kivett, 154 F.4th at 660 (Nelson, J.,
dissenting); see also In re Cap. One 360 Sav. Acct. Int. Rate Litig., 779 F.
Supp. 3d 666, 691 (E.D. Va. 2024) (A “requirement to impose a specific
interest rate . . . would constitute an even more severe interference
with national banks’ fundamental power to receive deposits” than the
law in Franklin.); cf. Ill. Bankers Ass’n v. Raoul, 760 F. Supp. 3d 636, 656
(N.D. Ill. 2024) (Law regulating interchange fees was “facially more
extreme than the sort of state laws that the Supreme Court intended
for national banks to be subject to.”).7
Finally, the state law’s likely effect on the attractiveness of a
bank’s products and services is inconclusive here. It is conceivable
that GOL § 5-601 could make mortgage-escrow accounts more
attractive if consumers are influenced by higher interest rates on their
escrow accounts. But forcing banks to pay customers interest may
cause them to “desist from using escrow accounts, implement fees,
otherwise increase borrower costs to offset [their] losses, or reduce
their overall mortgage lending due to decreased profitability,”
ultimately deterring consumers. Real Estate Lending Escrow
Accounts, 90 Fed. Reg. at 61103; see also U.S. G EN. A CCT. O FF., B-
114860, STUDY OF THE F EASIBILITY OF E SCROW A CCOUNTS ON
R ESIDENTIAL MORTGAGES B ECOMING INTEREST B EARING 14 (1973)
(“Most lending institutions reported that maintaining escrow
7 The interference with banks’ efficiency here is also unlike that in
McClellan. Massachusetts’s fraudulent-transfer law affected banks in only
“particular and exceptional circumstances,” so it did not “in any way
impair[] the efficiency of national banks.” McClellan, 164 U.S. at 358. In
contrast, GOL § 5-601 affects mortgage-escrow accounts for a broad swath
of properties in New York.

-- 25 of 65 --

26
accounts was costly and that it was not feasible to pay interest on
them.”). We cannot predict with confidence whether GOL § 5-601
would attract or deter customers, so we do not view New York’s law
as comparable to the state laws at issue in either First National Bank of
San Jose or Anderson in this respect.
* * *
In summary, New York’s law affects a banking power, so we
consider the nature and degree of its interference. As to the nature of
its interference, we conclude that it is more like that of the laws in
Barnett Bank and Fidelity, which affected broad grants of federal
power—here, national banks’ power to set interest rates for
mortgage-escrow accounts—than the interference caused by the
generally applicable law in McClellan or the common-law escheat rule
in Anderson. As to the degree of its interference, we conclude that the
impact on national banks’ ability to offer mortgage-escrow accounts
is at least as severe as the interference that New York’s advertising
law had in Franklin. GOL § 5-601’s interference with federal law thus
resembles other preempted state laws’ interference in both “nature
and degree,” so it is preempted too. Cantero, 602 U.S. at 220 n.3.
The dissent reaches the opposite conclusion by divining four
preemption fact patterns from the Court’s precedents and then
concluding that New York’s law fits none of them. For instance, it
writes off Franklin because New York’s law does not “drive a wedge
into national banks’ and consumers’ abilities to transact,” like the
advertising law in that case. Infra at 13. That approach offers little
guidance for evaluating state laws that affect banks in ways that differ
from the limited preemption caselaw. Indeed, it effectively
guarantees the dissent’s conclusion by reasoning that New York’s law
is preempted only upon proof that it “will materially disincentivize

-- 26 of 65 --

27
banks from offering escrow accounts—or from offering mortgages
altogether—or if banks will materially distort other mortgage terms.”
Infra at 14.
The First Circuit also reached the opposite conclusion and held
that a Rhode Island law requiring “all banks operating within the
state to pay mortgage borrowers interest on the funds they deposit
into mortgage-escrow accounts” was not preempted by the National
Bank Act. Conti v. Citizens Bank, N.A., 157 F.4th 10, 12 (1st Cir. 2025).
We disagree with that conclusion for two main reasons. First, Conti
disregarded RESPA and TILA in summarily concluding that Barnett
Bank and Fidelity were “generally inapposite” to its analysis. Id. at 20
& n.7. But federal statutes like RESPA and TILA were relevant in the
Court’s preemption precedents, including Barnett Bank, Franklin, and
Fidelity. Second, Conti failed to acknowledge the practical reality that
a state law restricting the pricing of a bank’s product would have a
“material impact . . . on banking operations.” 157 F.4th at 23-24. We
thus decline to follow the First Circuit’s analysis here.
IV. CONCLUSION
New York’s interest-on-escrow law is preempted because it is
“akin to” the laws the Supreme Court has struck down for
significantly interfering with a national bank’s powers. Cantero, 602
U.S. at 220. As a result, BOA was not required to pay interest on
Plaintiffs’ escrowed funds. We thus reverse the orders of the district
court denying BOA’s motions to dismiss and remand the cases for
further proceedings consistent with this opinion.

-- 27 of 65 --

21-400, 21-403
Cantero v. Bank of America
MYRNA PÉREZ , Circuit Judge, dissenting:
It is worth repeating that this case involves a state law which merely
requires national banks to pay a consumer 2% interest on funds that the consumer
has placed in an account with the bank. The majority opinion holds that such a
law significantly interferes with national banking powers, and thus, is preempted.
I respectfully disagree.
The Supreme Court unanimously rejected the “control” test that this panel
had articulated for evaluating National Bank Act preemption. See Cantero v. Bank
of America, N.A. (Cantero II), 602 U.S. 205, 220–21 (2024). Sharing the Supreme
Court’s concerns, I nonetheless joined that opinion because I believed the
articulated control test did not have to be understood to “imply that every state
law that impacts national banks’ business interests is preempted.” See Cantero v.
Bank of America, N.A. (Cantero I), 49 F.4th 121, 142 (2d Cir. 2022) (Pérez, J.,
concurring), vacated, 602 U.S. 205 (2024). I wrote separately to emphasize that an
overbroad application of the control test would be in tension with the Supreme
Court’s seminal precedent in Barnett Bank. Id. at 141–42. Our error, we were told,
was that we did not conduct a “nuanced comparative analysis” of the law at issue
relative to precedent, but instead “distill[ed] a categorical test that would preempt

-- 28 of 65 --

2
virtually all state laws that regulate national banks.” Cantero II, 602 U.S. at 220–21.
The control test preempted too much and left states too little room to exercise their
“power to regulate national banks.” Id. at 215 (quoting Barnett Bank of Marion
Cnty., N.A. v. Nelson, 517 U.S. 25, 33 (1996)).
While this new majority opinion nominally sets aside the categorical test
previously articulated, the majority opinion, in my view, nevertheless trudges
through a strained analysis of the Supreme Court’s precedents to reach an
approach that is just as capacious. Moreover, the majority opinion ignores the
nature of the federal banking power at issue and recharacterizes the relevant
power as broadly as possible to manufacture a direct conflict with state interest-
on-escrow laws. Having been warned of the dangers of such an expansive view
of preemption once, I must respectfully dissent here.
I.
Undoubtedly, the Supreme Court’s instructions in Cantero II left many
questions unanswered. Therefore, I believe it is most helpful to first describe the
framework for approaching National Bank Act preemption that, in my view, best
adheres to the Supreme Court’s instructions.

-- 29 of 65 --

3
A non-discriminatory state law is preempted if it “prevents or significantly
interferes with [a] national bank’s exercise of its powers.” See Cantero II, 602 U.S.
at 220. To ascertain significant interference, the Supreme Court directed us to
“make a practical assessment of the nature and degree of the interference caused
by a state law.” Id. at 219–20. That assessment requires a “nuanced comparative
analysis” of the “interference with national bank powers” caused by the law before
us relative to a specified universe of prior cases:
If the state law’s interference with national bank powers is more akin
to the interference in cases like Franklin, Fidelity, First National Bank of
San Jose, and Barnett Bank itself, then the state law is preempted. If the
state law’s interference with national bank powers is more akin to the
interference in cases like Anderson, National Bank v. Commonwealth,
and McClellan, then the state law is not preempted.
Id. “[I]n accordance with [Barnett Bank],” 12 U.S.C. § 25b(b)(1)(B), we look at “the
nature and degree of the state laws’ alleged interference with the national banks’
exercise of their powers based on the text and structure of the laws, comparison to
other precedents, and common sense,” Cantero II, 602 U.S. at 220 n.3.1
1 Because the Supreme Court held that “Dodd-Frank adopted Barnett Bank, and because Barnett Bank
was also the governing preemption standard before Dodd-Frank,” both the majority opinion and I apply
the Barnett Bank test throughout, as clarified by the Supreme Court, without leaning on the specifics of the
statutory text that incorporates Barnett Bank into 12 U.S.C. § 25b(b). See Cantero II, 602 U.S. at 214 n.2. For
similar reasons, it does not matter whether we analyze the case under § 25b(b)(1)(B) or (C), the latter of
which applies to laws “preempted by a provision of Federal law other than title 62 of the Revised Statutes.”
12 U.S.C. § 25b(b)(1)(C); see id. at 221 n.4 (reserving for remand “the relevance here (if any)” of
§ 25b(b)(1)(C)).

-- 30 of 65 --

4
At the start of this analysis, courts must identify and describe the kind of
power at issue.2 The relevant national banking power in this appeal is the power
to make real estate loans—to offer mortgages—some of which may come with
escrow accounts. See 12 U.S.C. § 371(a). One could think of regulating escrow
accounts either as regulating mortgage lending itself, or as regulating a separate
national banking power that is “incidental” to the enumerated mortgage power.
See 12 U.S.C. § 24 (Seventh). In any event, I believe we must ask how that power
is exercised, or the nature of that power, to assess the nature of the claimed
interference. Regardless of how one conceptualizes the relevant power here, it is
at base a power to offer a product and engage in a transaction with consumers,
and so our analysis should focus on how those transactions are impacted, if at all.3
The Supreme Court’s preemption cases do not “draw a bright line,” but they do
2 This first step is merely a starting point, providing the context through which a national banking
power is to be understood. It is not meant to be dispositive. See infra Section III.A. I recognize that defining
the relevant banking power is susceptible to levels of abstraction—defined too generally, and virtually all
regulation would be preempted; defined too narrowly, and there would never be preemption. But that is
precisely why defining the relevant national banking power does little analytical work. Cf. Cantero II, 602
U.S. at 220 (explaining the relevant determination as whether “the state law’s interference with national
bank powers is more akin” to the line of cases finding preemption or the line that did not). That there is
interference between state law and national banking power is all but presumed—instead, we must analyze
the nature of that interference.
3 Many national banking powers are consumer-facing, such as the power to sell insurance (at issue
in Barnett Bank) and to take deposits (at issue in San Jose and others). But others are not, including many
related to corporate organization and investments, and others can be both, including the power to make
contracts. See, e.g., 12 U.S.C. § 24.

-- 31 of 65 --

5
offer guideposts for discerning when states “prevent[] or significantly interfere[]”
with national banking powers to make consumer transactions. See Cantero II, 602
U.S. at 221. And the cases establish that where a consumer-facing national banking
power is at issue, state laws significantly interfere and are preempted when they
will materially reduce the likelihood that banks or consumers will make a rational
choice to make a mutually beneficial transaction authorized by federal law.4 The
cases illustrate at least four “types” of significant interference with these
transactions, but categorizing the cases does not yield easy answers; it is only a
framework for the “nuanced comparative analysis” we must undertake. Id. at 220.
In other words, the four types of interference identified below are useful entry
points for understanding when a state regulation might significantly interfere with
consumer-facing national banking powers.
First, a state law might improperly “interfere” with the exercise of national
banking powers to engage with consumers by outright banning certain
transactions. In Barnett Bank, the relevant “Federal Statute authorize[d] national
4 The Office of the Comptroller of the Currency (“OCC”) appears to agree with this distillation of
the relevant authority, though it would apply the resulting standard differently. Amicus Br. of OCC at 9
(“[T]he Court should conclude that a state law that requires a national bank to pay even a nominal rate of
interest on a particular category of account impermissibly conflicts with a national bank’s power by
disincentivizing the bank from continuing to offer the product.”).

-- 32 of 65 --

6
banks to engage in activities,” i.e., selling insurance in small towns, “that the State
Statute expressly forb[ade].” 517 U.S. at 31. New York’s interest-on-escrow law is
plainly not a ban on the exercise of national banks’ power to make mortgage loans,
and it is not a ban on mortgage-escrow accounts, so this type of interference
requires no further discussion.5
Second, a state law might improperly interfere with a national banking
power to engage with consumers by making a national bank product materially
less attractive, such that a meaningful number of consumers are deterred from
engaging in those transactions. In San Jose, the Court addressed California’s law
requiring escheat to the state of inactive deposit accounts after twenty years. See
First Nat’l Bank of San Jose v. California, 262 U.S. 366 (1923). The Supreme Court
reasoned that, if states were free to impose such automatic-escheat laws,
5 Of course, as Bank of America’s counsel acknowledged at oral argument, almost any regulation
can be described as a ban on something by describing every feature of a banking product as an “incidental”
national banking power and adjusting the level of generality. New York’s law is, in some sense, a ban on
“offering mortgages with escrow accounts that do not earn interest.” But one could equally say fair-lending
laws are bans on “offering mortgages using underwriting processes that include protected characteristics.”
And the law at issue in McClellan was a ban on certain kinds of loan modifications. While rhetorically
powerful, it is not helpful for engaging in the required Barnett analysis to conceptualize mere regulations
of national banking powers—e.g., whether they limit allowable interest rates on escrow accounts or factors
in underwriting processes—as bans. Under the statute, the relevant question is whether those regulations
“significantly interfere” with the exercise of the power. As discussed throughout, this is the same problem
as focusing on defining the national banking power at stake: the majority opinion indulges in this kind of
blurring-of-the-lines abstraction by conceptualizing the power at issue as a broad power, all but ensuring
a direct conflict with any attempt by the states to regulate. Instead, this type of interference occurs where
a state regulation on its face purports to prohibit or ban certain activities.

-- 33 of 65 --

7
consumers “might well hesitate to” make deposits with national banks and
thereby “subject their funds to possible confiscation.” Id. at 370.
The Supreme Court has also recognized, however, that a similar risk of
lesser magnitude would not have a material deterrent effect. The Court faced a
Kentucky escheat law in Anderson that was triggered by a shorter period of
dormancy than California’s—ten years rather than twenty—but which provided
depositors an avenue to recover their funds that California’s law may not have
offered. See Anderson Nat’l Bank v. Luckett, 321 U.S. 233, 250 (1944) (noting that the
California Supreme Court “had declined . . . to express an opinion” on that
question). The Supreme Court recognized that Kentucky’s law could cause
depositors hardship—certainly more than if they could leave their money
indefinitely without jumping through procedural hoops—even if “in many
circumstances [it] may operate for the benefit and security of depositors.” Id. at
252 (emphasis added). But the Court concluded that the hardship would be no
greater than the effect of “the tax laws, the attachment laws, or the laws for the
administration of estates of decedents or of missing or unknown persons,” which
seemed to be beyond question. Id. at 252. Ultimately, the Anderson Court
distinguished San Jose on the grounds that California’s imposition of automatic

-- 34 of 65 --

8
forfeiture, without proof of abandonment, was “so unusual and so harsh in its
application to depositors as to deter them from placing or keeping their funds in
national banks,” while the Court was simply not so persuaded that Kentucky’s
law would deter depositors “from placing their funds in national banks in that
state.” Id. at 250–52.6
Third, a state law might improperly interfere with national banking powers
directed towards consumers by making a product materially less attractive for the
bank to offer, prompting banks either to pull the product from the market or offer
it to consumers only on materially worse terms. In Fidelity, the Court addressed a
California law limiting lenders’ ability to invoke “due-on-sale” clauses in
mortgage loan agreements. See Fid. Fed. Savs. & Loan Ass’n v. de la Cuesta, 458 U.S.
141, 145–51 (1982). Due-on-sale clauses, the Court tells us, “permit[] the lender to
declare the entire balance of a loan immediately due and payable if the property
securing the loan is sold or otherwise transferred,” and they are an important risk-
6 One could argue that National Bank v. Commonwealth deals with this type of interference as well,
though it did not involve consumer transactions per se, 76 U.S. (9 Wall.) 353, 359–60 (1869), and according
to the majority opinion, did not concern a national banking power at all, Maj. Op. at 14–15. In that case,
Kentucky imposed a tax of $0.50 on each share of national bank stock worth $100. Presumably, that tax
made it marginally less attractive for people to engage in the transaction of purchasing national bank stock,
which hindered the ability of national banks to raise capital to some extent. But the Supreme Court held
that Kentucky’s small tax was “no greater interference with the functions of the bank than any other legal
proceeding to which its business operations may subject it.” Commonwealth, 76 U.S. at 362–63.

-- 35 of 65 --

9
and cash-flow-management tool for lenders. See id. at 145–46. The Federal Home
Loan Bank Board (“Board”) had even issued a regulation clarifying lenders’ power
to employ due-on-sale clauses. The Board found that restricting their use would
threaten “the financial security and stability of Federal associations,” “restrict and
impair the ability of Federal associations to sell their home loans in the secondary
mortgage market, . . . thereby reducing the flow of new funds for residential
loans,” and “cause a substantial reduction of the cash flow and net income of
Federal associations, and that to offset such losses it is likely that the associations
will be forced to charge higher interest rates and loan charges on home loans
generally.” Id. at 146–47 (quoting Late Charges and Due on Sale Clauses, 41 Fed.
Reg. 6283, 6285 (proposed Feb. 6, 1976) (codified at 12 C.F.R. §§ 545.6–11(f) (1980)).
The Court emphasized that California prohibited lenders from engaging in
practices the Board “view[ed] as critical to ‘the financial stability of the
association,’” “limit[ed] the availability of an option the Board consider[ed]
essential to the economic soundness of the . . . industry,” and impaired “[t]he
marketability of a mortgage in the secondary market,” which “is critical to a
savings and loan” for generating cash flow. Id. at 155–56 & n.10.

-- 36 of 65 --

10
While a state law that increases banks’ risk can make a product materially
less attractive for the bank to offer and thus be preempted, the Supreme Court has
recognized that if the impact is less significant, then the state law is not preempted.
In McClellan v. Chipman, Massachusetts barred banks from receiving preferential
transfers of real estate from insolvent debtors to secure existing debt. 164 U.S. 347,
357–58 (1896). Like the ability to demand repayment upon the sale of a property
(via a due-on-sale clause), the ability of a bank to take additional collateral to
secure an existing debt could be a useful tool for mitigating risk, especially if a
debtor were insolvent. But the law at issue in McClellan did not “impair[] the
efficiency of the banks to discharge the duties imposed upon them,” because it
applied only “under particular and exceptional circumstances.” Id. at 358–59. This
was a matter of degree, however, because the Court also recognized that
permitting the state to apply a narrowly targeted law such as that did not “impl[y]
the existence of a power in the State to forbid such taking in all cases.” See id. In
other words, some loans would be riskier if banks could not take additional
collateral as security in certain circumstances, but that increase in risk was small
enough that banks could manage or offset it without materially altering their
products or operations.

-- 37 of 65 --

11
Fourth, a state law might interfere without affecting the attractiveness or
viability of the product itself, but rather by making it materially harder for banks
and consumers to transact. This was the fact pattern at issue in Franklin, which
addressed a New York law that had “forbidden use of the word ‘savings,’ or its
variants, by any [national] banks.” Franklin Nat’l Bank of Franklin Square v. New
York, 347 U.S. 373, 374 (1954). The Supreme Court recognized that to exercise their
deposit-taking power, national banks had to compete for consumers’ business, and
“[m]odern competition for business finds advertising one of the most usual and
useful of weapons.” Id. at 377. As discussed above, federal law does not preempt
all state regulation of national banks’ ability to advertise their products. But in
Franklin, New York’s law impermissibly interfered with national banks’ power to
engage in a particular kind of transaction specifically authorized by federal law,
because it would “permit a national bank to engage in a business but g[i]ve no
right to let the public know about it.” Id. at 377–78. In other words, it would drive
a wedge between banks and their prospective customers. And again, state laws
that have effects similar in kind but lesser in degree are not necessarily preempted.
For example, the Supreme Court has upheld state laws “prohibiting branches.”
First Nat’l Bank in St. Louis v. Missouri, 263 U.S. 640, 659 (1924).

-- 38 of 65 --

12
II.
Applying this approach as I believe the Supreme Court has instructed, I
would hold that a 2% interest-on-escrow floor is not preempted on this record.
Bank of America and its amici are well positioned to explain to this Court just how
being required to pay 2% interest on escrow accounts would affect national banks’
power to make mortgage loans and to mitigate risk with escrow accounts. And
Bank of America bears the burden of persuasion on its affirmative defense.7 See In
re Methyl Tertiary Butyl Ether Prods. Liab. Litig., 725 F.3d 65, 96 (2d Cir. 2013) (“[T]he
party asserting that federal law preempts a state law bears the burden of
establishing preemption.”). I believe it has not satisfied that burden.
A.
To be sure, Bank of America and its industry amici make a persuasive case
that escrow accounts are an important part of the powers of national banks.
Escrow accounts are a useful tool “for national banks to protect their security
7 It remains unclear what quantum of proof, if any, is required of a party seeking preemption under
Dodd-Frank or Barnett Bank, including whether and when such an argument must be supported by
affidavits or, if necessary, live testimony. I would note that Congress requires the OCC to make a “case-
by-case” determination, 12 U.S.C. § 25b(b)(3)(A), of whether “substantial evidence, made on the record of
the proceeding” supports preemption before making its own preemption determinations, id. at § 25b(c). In
my view, those requirements suggest some kind of factual showing is required. However, I need not
address the precise substance of Bank of America’s burden because at a minimum, some sort of argument
derived from “common sense” would be required. Cantero II, 602 U.S. at 220 n.3. As explained below, Bank
of America has offered none.

-- 39 of 65 --

13
interest when they make real-estate loans,” and “mitigate the risk of loss that
would arise if the borrower failed to pay taxes or to have the property properly
insured.” Bank of America Cantero Opening Br. at 22–23. The industry amici tell
us “[n]ational banks rely on [escrow] accounts to help manage their credit risk on
multiple millions of mortgages across the United States,” they are “crucial to the
success of the modern home mortgage system,” and a large majority of new
mortgage originations come with escrow accounts. Amicus Br. of Bank Pol’y Inst.
et al. at 5–6, 9. As a report put it, “through the use of an escrow account, a lender
is able to protect the priority of its mortgage lien and ensure the protection of its
collateral.” Bruce E. Foote, Cong. Rsch. Serv., 98-979 E, Mortgage Escrow Accounts:
An Analysis of the Issues 2 (1998).
But the importance of escrow accounts is not directly at issue in this case.
The question we face is whether a state law requiring 2% interest to be paid on
those accounts will significantly interfere with national banks’ exercise of powers.
To answer that question, I turn to the four types of interference identified above.
Interest-on-escrow laws are not an outright ban like the law at issue in Barnett Bank,
so the first type is inapplicable. Nor do they drive a wedge into national banks’
and consumers’ abilities to transact like the advertising law in Franklin, so the

-- 40 of 65 --

14
fourth type of interference identified is not at issue. Instead, this case most
implicates the third type, in which a state law might affect the banks’ incentives to
offer those products in the first place, and to a lesser extent the second type, in
which a state law might operate to make a banking product less attractive for
consumers. See supra at 6–10. Thus, the interference will be significant “akin to
the interference in cases like Franklin, Fidelity, First National Bank of San Jose, and
Barnett Bank,” if New York’s law will materially disincentivize banks from offering
escrow accounts—or from offering mortgages altogether—or if banks will
materially distort other mortgage terms to address the interference which might
make the accounts less attractive to consumers. See Cantero II, 602 U.S. at 220. For
the reasons explained below, New York’s interest-on-escrow law does not pose
such a threat, and thus, does not significantly interfere with national banking
powers.
Bank of America and its amici offer platitudes and conclusions without
explanation. Bank of America has asserted several times that New York’s interest-
on-escrow law puts it to a choice between “(i) using other means (such as higher
mortgage interest rates, higher loan origination fees, or reduced loan amounts) to
mitigate its risk, (ii) doing nothing and assuming greater risk, or (iii) refraining

-- 41 of 65 --

15
from making the loan at all.” Bank of America Cantero Reply Br. at 12; Bank of
America Cantero Opening Br. at 23. But Bank of America largely omits the fourth
option: continue using escrow accounts to “mitigate its risk” to precisely the same
degree as it does right now, but pay 2% interest on them. Neither Bank of America
nor amici have put forward anything prompting concern that any bank will
change the availability of escrow accounts—a critical tool in banks’ risk-mitigation
toolkit—if required to pay 2% interest.
B.
A “nuanced comparative analysis of” the “text and structure” of the state
and federal laws at issue here, relative to the Supreme Court’s preemption cases,
confirms that New York’s interest-on-escrow law imposes—at most—an
insignificant interference with the national banking powers relevant here. See
Cantero II, 602 U.S. at 220.
The subtly different facts of San Jose and Anderson provide a useful starting
point for assessing the degree of interference caused by an incentive-shifting state
law. Cf. Roderick M. Hills, Jr., Exorcising McCulloch: The Conflict-Ridden History of
American Banking Nationalism and Dodd-Frank Preemption, 161 U. Pa. L. Rev. 1235,
1267–68 (2013) (asserting that Anderson “virtually overruled” San Jose by adopting

-- 42 of 65 --

16
“a new and narrower reading of banks’ immunity that permitted states to impose
regulations on lending and deposit-taking, so long as the states did not thereby
discriminate against any nationally chartered banks or contradict any policies of
the federal government”). As discussed, there is no suggestion in the record that
Bank of America—or any national bank—“might well hesitate to” offer mortgages,
or mortgages with escrow accounts, if required to offer 2% interest. Cf. San Jose,
262 U.S. at 370. The same was true in Anderson, where the Supreme Court did not
find the state law preempted. The law gave the state of Kentucky “the right to
demand payment of [certain deposit] accounts in the place of the depositors” after
ten years of inactivity. Anderson, 321 U.S. at 248. That meant that—instead of the
banks continuing to invest and profit from those funds indefinitely, and the
depositors always having ready access to their money—the deposits (and future
interest they might earn) went to the state, and depositors would have “to demand
from the state payment of the deposits” and “resort to the courts if payment [was]
refused.” Id. at 242; see Fed. Nat’l Mortg. Ass’n v. Lefkowitz, 390 F. Supp. 1364, 1368
(S.D.N.Y. 1975) (“In the absence of such a statute [as was at issue in Anderson], the
banks presumably would have had full use of the funds until—if ever—they were
claimed.”). But crucially, the Court recognized that “[s]omething more” was

-- 43 of 65 --

17
needed to justify preemption. Anderson, 321 U.S. at 248–49. The banks’ lost ability
to profit from the dormant accounts, and the potential burden placed on
consumers to reclaim their funds, did not amount to substantial interference.
Comparison to Fidelity and McClellan is also instructive. Both concerned
important risk-mitigation tools available to lenders under federal law: in Fidelity,
it was due-on-sale clauses, 458 U.S. at 145–47, and in McClellan, it was the power
to take real estate “conveyed to it in satisfaction of debts previously contracted in
the course of its dealings,” 164 U.S. at 357–58. In Fidelity, California law prohibited
the exercise of due-on-sale clauses “unless the lender c[ould] demonstrate that
enforcement is reasonably necessary to protect against impairment to its security
or the risk of default.” 458 U.S. at 149 (quoting Wellenkamp v. Bank of America, 582
P.2d 970, 977 (Cal. 1978)). Though the statute contained an exception, the Supreme
Court nevertheless deemed the law preempted because “further limiting the
availability of an option the Board considers essential to the economic soundness
of the thrift industry” was “an obstacle to the accomplishment and execution of
the full purposes and objectives” of express federal regulation. Id. at 156 (quoting
Hines v. Davidowitz, 312 U.S. 52, 67 (1941)). Fidelity would be more analogous here
if New York prohibited lenders from requiring escrow accounts unless they could

-- 44 of 65 --

18
show the payments were “reasonably necessary to protect against impairment to
its security or the risk of default,” or otherwise limited the availability of escrow
accounts.
But New York’s law does nothing like that. It is much more like the
Massachusetts law at issue in McClellan, which impacted a risk mitigation tool but
only on the margins. 164 U.S. at 353–54, 358. In McClellan, the state disarmed
banks from exercising a risk-mitigation tool in just the circumstance where they
would need it most—when their borrower was insolvent—and still, the law was
not preempted because such insolvency constituted “particular and exceptional
circumstances.” Id. at 358–59. The burden of paying a bit of interest, in
comparison to the total loss of a risk-mitigation tool at a crucial albeit exceptional
moment, is negligible.
C.
Amici in favor of Bank of America contend that if you add up all the money
in all the escrow accounts for all the millions of mortgages across the United States,
you get “billions of dollars.” Amicus Br. of Bank Pol’y Inst. et al. at 6. But for the
inquiry demanded by the Supreme Court’s preemption precedents, it does not
matter how much money the industry as a whole might have to pay in interest.

-- 45 of 65 --

19
The question, as explained above, is whether the interest requirement will
materially distort the bank’s incentives to offer mortgages, or prompt changes in
other mortgage terms that might deter consumers. As discussed, Bank of America
has not met its burden of persuasion on this question. In fact, there are several
reasons to think interest on escrow laws do not materially distort incentives.
Congress’s decisions in Dodd-Frank are informative. Congress required
escrow accounts, and required interest be paid “[i]f prescribed by applicable State
or Federal law,” on certain mortgages. 15 U.S.C. § 1639d(b)(1), (g)(3). All agree
that the mortgages at issue here are not covered by § 1639d. But the fact that
Congress required interest be paid on those mandatory escrow accounts indicates
that Congress at least did not think interest requirements would seriously distort
the availability or the terms of the very mortgages it sought to protect by statute.
See Lusnak v. Bank of America, N.A., 883 F.3d 1185, 1194–95 (9th Cir. 2018) (noting
the same).
Common sense also tells us interest-on-escrow laws are unlikely to provoke
material changes in the banks’ offerings and the overall incentive structure.8 The
8 Bank of America and amici’s argument that New York’s 2% interest-on-escrow floor is sometimes
higher than the national average savings rate, and that such a gap is subject to fluctuation, is not relevant
to the analysis. See, e.g., Amicus Br. of Bank Pol’y Inst. et al. at 6 & n.4. Instead, what is relevant is how
much a national bank must pay a typical individual mortgagor in interest, and whether that amount is
likely to cause the national banks to change their products in response if they seek to recoup costs, whether

-- 46 of 65 --

20
amounts of consumer funds that sit in mortgage-escrow accounts and collect
interest are far lower than the amounts of the mortgage loans themselves on which
banks collect interest. An escrow account will rarely hold much more than the
amount of one’s annual property tax and insurance bill, which is typically orders
of magnitude lower than the entire value of one’s mortgage.9
Available evidence supports the intuition that the amounts held in escrow
accounts, while important to individual consumers, are relatively insignificant to
the economics of most mortgages, whether interest is paid or not. In 1991, the
Senate Committee on Governmental Affairs held a hearing called, “Mortgage
Escrow Accounts: Loopholes in Federal Consumer Protections.” 102d Cong. 1
(1991). In a statement submitted by the Mortgage Bankers Association of America,
the average escrow balance was estimated at $674, on which a 2% annual interest
rate would require less than $15 in interest per year. See id. at 215 (statement of
by altering the interest rate on the mortgage itself, the origination fee, or some other feature. The key to
that inquiry is comparing the amount a bank will be required to pay in 2% interest on escrow with the
income it receives from a given mortgage. Though that ratio may sometimes fluctuate, such fluctuation
would be immaterial given the width of the gap.
9 Under the Real Estate Settlement Procedures Act and its implementing regulation, monthly escrow
payments are limited to the amounts needed to cover certain payments made over the course of a year—
typically property taxes and insurance—plus a modest “cushion.” See 12 C.F.R. § 1024.17(c)(1)(ii). Those
payments might be made quarterly, twice yearly, or yearly, which is why monthly escrow payments help
homeowners “budget the property taxes and hazard insurance on a monthly basis.” See Bruce E. Foote,
Cong. Rsch. Serv., 98-979 E, Mortgage Escrow Accounts: An Analysis of the Issues 2 (1998). But an escrow
account should never hold much more than a year’s worth of insurance and tax payments.

-- 47 of 65 --

21
Stephen B. Ashley, Chairman & C.E.O., Sibley Mortg. Corp.). Senator Herbert
Kohl of Wisconsin discussed the phenomenon of state-chartered lenders in his
state switching to federal charters to avoid paying Wisconsin’s required 5.25%
interest on escrow, which he explained was “worth about $40 a year to the average
mortgage holder.” Id. at 5–6. Alan B. Morrison of Public Citizen testified that the
difference between New York’s minimum 2% interest on escrow and the much
higher market rate at the time of 8% amounted to “$60 or $70 a year” for one
homeowner. Id. at 23. And John C. Weicher, an official from the Department of
Housing and Urban Development, testified that “the difference in the mortgage
payment for a borrower whose escrow account bears interest at a market rate and
one whose account bears no interest works out to a difference of a bit less than $4
a month on a mortgage for $100,000.” Id. at 33.
The Supreme Court has recognized that the National Bank Act preempts
state laws that impose significant burdens on either national banks or consumers
(or both) that materially distort their behaviors. It is “common sense,” as the
Supreme Court put it in this case, Cantero II, 602 U.S. at 220 n.3, that a consumer
might not want a deposit account that comes with a potentially catastrophic risk
of inadvertent, irretrievable loss of their savings, see San Jose, 262 U.S. at 369–70. It

-- 48 of 65 --

22
is common sense that a bank will require serious additional risk-mitigation and
potentially much higher fees, if it has doubts about its ability to enforce a due-on-
sale clause, allowing its loan to be transferred to someone whose creditworthiness
the bank did not underwrite and depriving the bank of cash flow. See Fidelity, 458
U.S. at 149, 167–70. And it is common sense that if banks are prohibited from
advertising their savings account using the word “savings,” then consumers who
want “savings” accounts will be less likely to seek out their products. See Franklin,
347 U.S. at 377–78. On the other hand, neither logic, nor anything in the record,
tells us that banks paying each consumer interest on the amount the consumer has
in their escrow account each month will cause banks to restructure their mortgage
products in ways that either banks or consumers will shy away from. Instead,
common sense indicates that it is a marginal adjustment which does not impact
the overall viability of escrow accounts, and Bank of America has not suggested
business as usual will be disrupted in any meaningful way.
III.
Because I do not think Bank of America has satisfied its burden to show the
interest-on-escrow law is preempted, I will move on to what I see as errors

-- 49 of 65 --

23
inherent in Bank of America’s arguments and my disagreement with the majority
opinion’s reasoning.
A.
Most notably, Bank of America argues that interest-on-escrow laws are
preempted because they represent a kind of ban on national banks’ flexibility.
They urge that federal law “extensively regulates national banks’ operation of
escrow accounts,” but it does not require payment of interest. Bank of America
Suppl. Br. at 2 (quoting Cantero II, 602 U.S. at 210–11). Bank of America claims that
gap means that “federal law grants flexibility,” and that “[l]aws limiting flexibility
within federal regulatory schemes are preempted.” Id. at 2, 7; see also id. at 14–17.
The majority opinion adopts this argument and insists that federal law grants a
“broad” power for national banks to set interest rates on mortgage-escrow
accounts. See Maj. Op. at 23.
But this is just a relabeling of the rejected control test. Now that the Supreme
Court has made clear that not every state law which “exercise[s] control over a
federally granted banking power” is necessarily preempted, see Cantero II, 602 U.S.
at 213, Bank of America and the majority opinion have opted to simply move their
focus away from the effect of the state law and towards the shape of the federal

-- 50 of 65 --

24
power at issue. By reframing the federal grant of power as enabling national banks
to exercise discretion and flexibility, suddenly almost every state law that imposes
any restriction on national banks at all necessarily conflicts with the federal grant
of power so conceived, risking preemption. In short, the majority opinion and
Bank of America manufacture a direct “prohibition” or “ban” where there is none,
see supra Section I, which effectively reimposes the control test. The state laws that
“limit flexibility within federal regulatory schemes” on the one hand, and the state
laws that “exercise control over a federally granted banking power” in a manner
that federal law does not on the other hand, form a perfectly overlapping Venn
diagram. Perhaps for that reason, Bank of America admitted at oral argument that
a “flexibility test” probably would not “fly.” See Oral Arg. at 13:00–10.
Putting aside the Supreme Court’s rejection of the control test, if this Court’s
preemption analysis hinged on whether a state law interfered with a national
bank’s flexibility or exercise of discretion, such a “test would obviate the need for
an inquiry into whether a state law’s interference with federal-banking powers
was significant.” See Conti v. Citizens Bank, N.A., 157 F.4th 10, 25 (1st Cir. 2025)
(concluding an analogous interest-on-escrow law was not preempted). Instead of
looking to the material effect of a regulation, such as its effect on the incentives of

-- 51 of 65 --

25
transacting parties, Bank of America and the majority opinion would simply have
us look to whether the regulation constrains a national bank at all. Such an
approach cannot possibly be squared with the Supreme Court’s instruction to
engage in a “practical assessment of the nature and degree of the interference.”
See Cantero II, 602 U.S. at 219–20.
The majority opinion relies on what, in my view, is a misreading of the
Supreme Court’s preemption precedents. Principally, the majority opinion relies
upon Fidelity and Barnett Bank, where the Supreme Court did indeed reason that
the grants of national banking power at issue were broad, and thus, state laws
which imposed certain restrictions on the banks were preempted. See Fidelity, 458
U.S. at 155; Barnett Bank, 517 U.S. at 32. But crucially, both of those cases dealt with
express powers. The majority opinion concedes that the broad power of flexibility
that it constructs here “is not express, as in Fidelity and Barnett Bank, where federal
statutes ‘explicitly’ granted an express power.” Maj. Op. at 23 (quoting Barnett
Bank, 517 U.S. at 34). To me, this is a “nuance” that distinguishes Fidelity and
Barnett Bank; they merely represent the obvious proposition that state laws are
preempted when they are in direct conflict with an express federal grant of power.

-- 52 of 65 --

26
Recognizing the need to conjure a “broad” power where there is none, the
majority opinion and Bank of America draw lessons from federal statutes. They
insist that those statutes imply that national banks are afforded flexibility to set the
terms of escrow accounts under federal law. But I believe the conclusions they
draw go too far.
Bank of America and the majority opinion first hang their hat on the fact that
the Real Estate Settlement Procedures Act of 1974 (“RESPA”), a statute which
governs banks’ abuse of escrow accounts, does not include a mandatory interest
provision. Bank of America argues that here, “federal and state laws target the
same concern with different, mutually incompatible solutions.” Bank of America
Suppl. Br. at 13. But there is no incompatibility between the safeguards prescribed
by RESPA—including prompt return of leftover funds at the end of the loan term,
limits on required escrow payments, and disclosures—and the modest amount of
interest prescribed by New York. New York’s law complements RESPA and
creates no impossibility, no conflict, and apparently not even an obstacle to
accomplishing Congress’s purposes.
The majority opinion similarly cites the “omission” of an interest provision
in RESPA and insists it shows that national banks instead “have a broad power to

-- 53 of 65 --

27
set those rates.” Maj. Op. at 21–22. But if both Congress’s expressed intent and
silence on an issue can be read to grant a broad national banking power, then states
are left between a rock and a hard place. They certainly cannot regulate in a way
that contradicts an express power, but in the majority opinion’s view, they also
cannot not regulate to fill in the gaps. Such a dilemma lacks any grounding in the
Supreme Court’s precedents; indeed, “[n]one of the cases identified by [Cantero II]
held a state law preempted based on congressional silence.” Conti, 157 F.4th at 22.
Not only that, the majority opinion’s attempt at discerning a “broad” right out of
the void is at least in tension with the federal statute governing preemption, which
establishes a presumption that state laws are not preempted. See 12 U.S.C.
§ 25b(b)(1) (providing that “State consumer financial laws are preempted, only if ”
certain conditions are met (emphasis added)). At bottom, such an approach does
not reflect a nuanced comparative analysis focused on the nature and degree of
interference, thus leaving states with little guidance on what regulation is
permissible, let alone when silence equates to broad federal powers of flexibility.
Nor does the Truth in Lending Act (“TILA”) support preemption. Congress,
through TILA, requires national banks to pay interest on escrow accounts “in the
manner as prescribed by an applicable State or Federal law” for some mortgage-

-- 54 of 65 --

28
escrow accounts, but not those at issue here. See 15 U.S.C. § 1639d(g)(3). Bank of
America and the majority opinion believe that this shows Congress wanted to
immunize all mortgages that are not governed by TILA from state interest-on-
escrow laws. But this mode of construction lacks foundation; again, the majority
opinion’s position is that a void of Congressional action is more valuable to
ascertaining Congress’s intent than its expressed preferences. The more
reasonable interpretation of TILA’s role, as noted above, is that it illuminates
“Congress’s view that such laws would not necessarily prevent or significantly
interfere with a national bank’s operations.” Lusnak, 883 F.3d at 1194–95.
B.
For good measure, my view is that Bank of America’s other arguments are
similarly unpersuasive.
First, the suggestion that New York’s banking superintendent can “set any
interest rate (up to infinity)” is irrelevant to the question before us. Bank of
America Suppl. Br. at 23. If New York were to raise its minimum interest on
escrow by orders of magnitude, then yes the analysis above would be different,
and perhaps then the law would be preempted. But the Supreme Court has
already explained that in ordinary national bank preemption cases, slippery-slope

-- 55 of 65 --

29
reasoning is inappropriate, since the power to impose an insignificant burden does
not “impl[y] the existence of a power” to impose a significant one. See McClellan,
164 U.S. at 359. Otherwise, every insignificant interference could be rendered
preempted with just a tweak to the hypothetical.
Second, Bank of America and several amici suggest that permitting states to
require a minimum interest rate on escrow accounts implies that states can also set
a minimum interest rate on savings accounts, certificates of deposit, etc. Bank of
America Suppl. Br. at 27–28; Suppl. Amicus Br. of Bank Pol’y Inst. et al. at 13–14.
But this is just more of the same slippery-slope reasoning that, if credited, would
require preempting everything. Rather than speculating about the next case, courts
“must make a practical assessment of the nature and degree of the interference”
caused by the state law before it on its own terms. Cantero II, 602 U.S. at 219–20.10
And that analysis might cash out differently for a minimum interest rate on
savings accounts, because the interest rate on a savings account (relative to other
features, like balance minimums and fees) is the key feature driving consumer
behavior. On the other hand, nobody seeks out a mortgage because it will have an
10 The other items listed in Bank of America’s parade of horribles are irrelevant for the same reason.
Bank of America Suppl. Br. at 27–28. This case does not require us to speculate about how any of those
cases might come out.

-- 56 of 65 --

30
escrow account that exists largely to protect the bank from risk, or because they
might receive a small amount of interest on the amounts they are compelled to
keep in the escrow account. The character of the interference is simply not
comparable.
Third, the suggestion that state interest-on-escrow laws threaten
“disuniformity,” and that every state interest-on-escrow law must be preempted
because there could theoretically be fifty different ones, is unpersuasive. Bank of
America Suppl. Br. at 23–24. For one thing, this rule would also prove far too
much, because if any state can regulate, every state can, and ordinarily states can
choose to regulate in different ways. A “uniformity” test would also amount to “a
categorical test that would preempt virtually all state laws that regulate national
banks.” Cantero II, 602 U.S. at 220–21. Further, the Supreme Court’s concern for
the risk of “varying limitations” state-by-state in San Jose is inapposite here. 262
U.S. at 370. In San Jose, the law required banks in California to turn over to the
state any deposits left untouched for twenty years. The Supreme Court was
concerned that the “depositors of a national bank” who “often live in many
different States and countries,” would be at risk of forfeiting their savings based
on the law of “the State where the bank happened to be located,” to which they

-- 57 of 65 --

31
might have little connection. Id. Here, consumers face no similar burden. Their
mortgages will simply be subject to the law of the state where they live (or own
property). At stake for them is merely whether their escrow account would
contain a little extra cash depending on where that account happened to be
located, as opposed to having their deposits be subject to seizure. And banks
cannot claim to be meaningfully burdened by paying different interest-on-escrow
rates in different states. Every lender and servicer already has to keep track of the
interest rates they charge on every mortgage every month—which vary not state-
to-state but loan-to-loan. And Bank of America’s suggestion that New York’s
interest-on-escrow floor could be “ever-fluctuating” is unpersuasive for similar
reasons. Bank of America Suppl. Br. at 23. Lenders already manage to administer
“ever-fluctuating” rates on variable-rate loans. There is simply no way to take a
modest interest-on-escrow law like New York’s and paint it convincingly as a
threat to the efficient functioning of national banks.
IV.
Finally, I briefly address the role that the Office of the Comptroller of the
Currency (“OCC”) and its regulations play in the preemption analysis.

-- 58 of 65 --

32
The OCC, which also supports Bank of America as amicus, charters,
supervises, and regulates national banks. It carries real weight as the expert
regulator that not only enforces the law against national banks but also supervises
them through the exercise of its “visitorial powers . . . largely to the exclusion of
other governmental entities.” See Watters v. Wachovia Bank, N.A., 550 U.S. 1, 6–7
(2007).
The OCC’s regulations purport to preempt all “state law limitations
concerning . . . [e]scrow accounts.” 12 C.F.R. § 34.4(a)(6). And in December 2025,
the OCC issued a notice of proposed rulemaking that may “codify” the national
banking power “to establish and maintain escrow accounts” and “clarify that the
terms and conditions of escrow accounts, including the extent of any
compensation paid to customers, are business decisions to be made by each bank.”
Real Estate Lending Escrow Accounts, 90 Fed. Reg. 61099, 61103 (proposed Dec.
30, 2025). Simultaneously, the OCC proposed a “preemption determination”
declaring that the law at issue here, and eleven similar state laws across the
country, are preempted. See Preemption Determination: State Interest-on-Escrow
Laws, 90 Fed. Reg. 61093, 61094 (proposed Dec. 30, 2025). If these regulations (and

-- 59 of 65 --

33
proposed11 rules) explained the OCC’s position persuasively, the OCC would be
statutorily and doctrinally entitled to meaningful deference. See 12 U.S.C.
§ 25b(b)(5)(A) (providing that the OCC’s preemption determinations are assessed
for persuasiveness); Skidmore v. Swift & Co., 323 U.S. 134 (1944).12 But the OCC is
not persuasive.
In 2004, the OCC issued a rule adding to a list of categories of state laws
related to mortgage lending that could not be enforced against national banks,
including laws concerning escrow accounts. Bank Activities and Operations; Real
Estate Lending and Appraisals, 69 Fed. Reg. 1904, 1905 (Jan. 13, 2004) (to be
codified at 12 C.F.R. pts. 7, 34). In its final rule, the OCC alluded vaguely to its
“experience supervising national banks” and its “experience with types of state
laws that can materially affect and confine—and are thus inconsistent with—the
11 For purposes of this discussion, I assume the proposed 2025 rules will be finalized in substantially
identical form.
12 In Wachovia Bank, N.A. v. Burke, 414 F.3d 305 (2d Cir. 2005), this Court deferred to different
regulations promulgated by the OCC related to different aspects of national bank preemption, under the
now-defunct framework of Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837 (1984). Because
the Court in Burke did not consider the regulations at issue here, we need not decide whether Burke remains
good law for any purpose in this Circuit, in light of Dodd-Frank’s clarification that Skidmore applies, 12
U.S.C. § 25b(b)(5)(A); the OCC’s concession that Dodd-Frank merely confirmed existing law on that issue,
Amicus Br. of OCC at 10 & n.5; the Supreme Court’s overruling of Chevron, Loper Bright Enters. v. Raimondo,
603 U.S. 369 (2024); and the Supreme Court’s decision not to defer to the OCC’s regulation in Cantero II, see
602 U.S. at 221 n.4 (leaving for us to “address as appropriate on remand . . . the significance here (if any)”
of OCC’s rules). Bank of America, in its post-Cantero II supplemental brief, appears to concede that neither
Burke nor the OCC’s regulations are binding on this Court. Bank of America Suppl. Br. at 19.

-- 60 of 65 --

34
exercise of national banks’ real estate lending powers.” Id. at 1908, 1911. But it did
not explain what experience led it to include laws concerning escrow accounts,
much less interest-on-escrow laws specifically.
In 2011, after the enactment of Dodd-Frank, the OCC essentially reenacted
the 2004 regulation in full. Office of Thrift Supervision Integration; Dodd-Frank
Act Implementation, 76 Fed. Reg. 43549, 43557 (July 21, 2011); Amicus Br. of OCC
at 13; Hills, supra, at 1238–39, 1275–96. The OCC reaffirmed the preemption
decisions it had made in 2004 “based on the OCC’s experience with the potential
impact of such laws on national bank powers and operations.” Dodd-Frank Act
Implementation, 76 Fed. Reg. at 43557. The OCC specifically reiterated that,
“based upon [its] assessment as the primary Federal supervisor of national banks,
state laws that would affect the ability of national banks to underwrite and
mitigate credit risk, manage credit risk exposures, and manage loan-related assets,
such as laws concerning the protection of collateral value, . . . risk mitigation, . . .
escrow standards,” and other matters, “would meaningfully interfere with
fundamental and substantial elements of the business of national banks and with
their responsibilities to manage that business and those risks.” Id. But again, the

-- 61 of 65 --

35
OCC offered nothing beyond that bare conclusion and vague reference to its
authority and experience.
That takes us to the two 2025 proposed rules, in which the OCC insists that
national banks may set “the terms and conditions of escrow accounts, including
the extent of any compensation paid to customers” with flexibility, see Real Estate
Lending Escrow Accounts, 90 Fed. Reg. at 61103, and that relatedly, a slate of state
interest-on-escrow laws are preempted, see Preemption Determination: State
Interest-on-Escrow Laws, 90 Fed. Reg. at 61096. Its position is unpersuasive for
many of the reasons already discussed.
First and foremost, in recognizing a purported right to flexibility in setting
the terms of escrow accounts, the OCC attempts to manufacture a direct conflict
that would short-circuit courts’ preemption inquiry, just like the majority
opinion.13 See supra at Section III.A. The OCC points out that “the flexibility to
make business judgments concerning the investment and use of escrowed funds
13 I recognize that in Fidelity, the Supreme Court noted that “[f]ederal regulations have no less pre-
emptive effect than federal statutes,” and treated a power recognized by federal regulation as the relevant
banking power for purposes of preemption. See 458 U.S. at 153–54. But Fidelity dealt with a regulation
promulgated by the Federal Home Loan Bank Board, which the Supreme Court confirmed had statutory
authority to issue the pre-emptive regulation at issue. Id. at 159. In contrast, Congress expressly limited
the OCC’s authority to preempt state regulations and instructed courts to review those determinations for
persuasiveness. See 12 U.S.C. §§ 25b(b), (c). I think it is plain, then, that Congress did not bestow the OCC
with authority to simultaneously “clarify” national banks’ unlimited discretion with one regulation and
declare state law preempted for conflicting with that discretion with another regulation.

-- 62 of 65 --

36
has long since been inherent to the business of banking codified in the National
Bank Act.” Real Estate Lending Escrow Accounts, 90 Fed. Reg. at 61101. Thus, the
OCC represents, discretion in setting the terms of escrow accounts “is a core
component of banks’ mortgage lending powers.” Id. at 61100. But if that
description were enough, one can hardly imagine a component of bank decision
making that, under the OCC’s reasoning, would not be deemed a broad and
flexible national banking power. And if that were the case, “that would preempt
virtually all state laws that regulate national banks,” contrary to the express
directive of the Supreme Court that a more exacting preemption analysis is
required. See Cantero II, 602 U.S. at 220–21. Moreover, the OCC explains that
flexibility in setting the terms of escrow accounts allows national banks “to
appropriately balance the costs and benefits . . . and the risks and rewards” of its
products and activities, and that the terms of escrow accounts “are ultimately a
business judgment made by each bank in accordance with safe and sound banking
principles.” See Real Estate Lending Escrow Accounts, 90 Fed. Reg. at 61100.
Again, that reasoning is little more than a platitude that businesses should not be
regulated—it is hard to imagine any banking activity that would not be subject to
cost-benefit analysis pursuant to business judgment. The question is not whether

-- 63 of 65 --

37
business judgment is implicated at all, but rather how much it is constrained by
the regulation in question. And on that latter front, the OCC’s reasoning is bare.
The OCC’s simultaneous preemption determination otherwise relies on the
same flawed reasoning, conclusory assertions, and forced analogies pressed in this
case by Bank of America and adopted by the majority opinion. See Preemption
Determination: State Interest-on-Escrow Laws, 90 Fed. Reg. at 61096. Most
glaringly, the OCC insists that Fidelity supports a broad reading of the relevant
grant of national banking power while glossing over the fact that Fidelity dealt with
an express grant of power. See id.
It is worth comparing the OCC’s regulations concerning escrow accounts
with the federal regulations concerning due-on-sale clauses at issue in Fidelity, 458
U.S. at 145–47. As the Supreme Court recited in that case, the agency explained in
detail how the exact type of restrictions on lenders that California imposed “would
have a number of adverse effects” in the real world, and it explained not only what
those effects were but how they would come about. Id. at 146. “[T]he financial
security and stability” of the lenders “would be endangered” if their collateral
were sold to people who could not repay the loan and protect the collateral; the
inability to call mortgages due early would “cause a substantial reduction of the

-- 64 of 65 --

38
cash flow” of federal associations; impairment of the federal associations’ ability
“to sell their home loans in the secondary mortgage market” would “reduc[e] the
flow of new funds for residential loans”; and overall, the restrictions would
“benefit only a limited number of home sellers, but generally will cause economic
hardship.” Id. (quoting Late Charges and Due on Sale Clauses, 41 Fed. Reg. at
6285). For those specific reasons, the agency concluded that banks had the express
power to include the disputed due-on-sale clauses in contracts with borrowers.
The OCC offers no comparable analysis of the impact of interest-on-escrow laws
here, and instead relies upon generalizations, which I find unpersuasive.
* * *
In sum, Bank of America has not shown that New York’s interest-on-escrow
law significantly interferes with the exercise of any national banking power, and
common sense indicates that it does not. I respectfully dissent.

-- 65 of 65 --