22-3094-cv Jean Robert Saint-Jean et al. v. Emigrant Mortg. Co., Inc. et al.

UNITED STATES COURT OF APPEALS FOR THE SECOND CIRCUIT

August Term 2023

(Argued: February 14, 2024 Decided: February 19, 2025)

Docket No. 22-3094-cv

JEAN ROBERT SAINT-JEAN, EDITH SAINT-JEAN, FELEX SAINTIL, LINDA COMMODORE, BEVERLEY SMALL, YANICK SAINTIL, JEANETTE SMALL, FELIPE HOWELL JR., AS ADMINISTRATOR OF THE ESTATE OF FELIPE R. HOWELL, Plaintiffs-Appellees,

FELIPE HOWELL, Plaintiff,

v.

EMIGRANT MORTGAGE COMPANY, EMIGRANT BANK, Defendants-Appellants,

EMIGRANT SAVINGS BANK-MANHATTAN, EMIGRANT BANCORP, INC., Defendants.

ON APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF NEW YORK

Before: CHIN, PARK, and ROBINSON, Circuit Judges. In this "reverse redlining" case, eight Black homeowners in New

York City sued the defendant-appellant lending institution and affiliated entities,

alleging that the lender violated federal, state, and city antidiscrimination laws

by making mortgage refinancing loans at extraordinarily high default interest

rates to Black and Latino individuals in poor neighborhoods who had no income,

no assets, and low credit scores, but high equity in their homes, and then

foreclosing on the loans when the individuals defaulted. Following a jury

verdict in favor of the homeowners, the United States District Court for the

Eastern District of New York (Brodie, C.J.) entered final judgment awarding four

of the homeowners $722,044 in compensatory damages and four other

homeowners nominal damages.

On appeal, the lender argues that the district court erred in three

ways: first, by finding the homeowners' claims timely under the doctrine of

equitable tolling and the discovery rule of accrual; second, in its instructions to

the jury on disparate impact and disparate treatment theories of discrimination;

and third, in holding, contrary to the first jury's verdict, that a release-of-claims

provision in a loan modification agreement signed by two homeowners was

unenforceable as a matter of law.

2 AFFIRMED.

Judge Park dissents in a separate opinion.

LILA R. MILLER (Tara K. Ramchandani, Yiyang Wu, Edward K. Olds, on the brief), Relman Colfax PLLC, Washington, DC; and Rachel Geballe, Brooklyn Legal Services, Brooklyn, NY, on the brief, for Plaintiffs-Appellees.

MATTHEW A. SCHWARTZ (Richard H. Klapper, on the brief), Sullivan & Cromwell, LLP; and Evandro C. Gigante, Proskauer Rose LLP, New York, NY, on the brief, for Defendants-Appellants.

LAUREN GORODETSKY, Counsel (Steven Y. Bressler, Deputy General Counsel, Christopher Deal, Assistant General Counsel, on the brief), for Seth Frotman, General Counsel, Consumer Financial Protection Bureau, Washington, DC, for Amicus Curiae Consumer Financial Protection Bureau, in support of Plaintiffs-Appellees.

Barbara D. Underwood, Solicitor General, Ester Murdukhayeva, Deputy Solicitor General, Anthony R. Raduazo, Assistant Solicitor General, New York, NY, for Amicus Curiae Letitia James, Attorney General for the State of New York, in support of Plaintiffs-Appellees.

Jeffrey Gentes, Stuart Rossman, Anika Singh Lemar, Pamela Heller, Theresa Dudek, Attorneys; Sydney Brun-Ozuna, Alexander Emmons, Student Attorneys, for Amicus Curiae Connecticut

3 Fair Housing Center, National Consumer Law Center, and the Housing Clinic of Jerome N. Frank Legal Services Organization at Yale Law School, in support of Plaintiffs-Appellees.

CHIN, Circuit Judge:

In 1999, defendants-appellants Emigrant Mortgage Company, Inc.

and Emigrant Bank (together, "Emigrant") introduced the STAR NINA ("no

income, no asset") loan program. The STAR NINA program did not require

potential borrowers to disclose their income or assets to receive a loan.

Plaintiffs-appellees ("Plaintiffs") are eight Black homeowners in New

York City who applied for and received STAR NINA loans between 2004 and

2009. All Plaintiffs eventually defaulted on their loans and were the subject of

foreclosure proceedings. Plaintiff Felipe Howell, for example, lost his home after

Emigrant foreclosed on his STAR NINA loan. Although Howell had $424,000 in

equity in his home when he took out the loan, he received nothing when the

home was sold in foreclosure. Three other Plaintiffs lost their homes, and four

remained in foreclosure as of oral argument on February 14, 2024. Plaintiffs sued

Emigrant under various federal, state, and city laws, including, as relevant to this

appeal, the Fair Housing Act (the "FHA"),

, 3605, the Equal

Credit Opportunity Act (the "ECOA"),

et seq., and the New York

4 City Human Rights Law (the "NYCHRL"), N.Y.C. Admin Code § 8-101 et seq.,

alleging that Emigrant's lending practices in connection with the STAR NINA

loans discriminated against Black and Latino borrowers.

After a six-week trial in 2016, the jury found that Emigrant's STAR

NINA loan program discriminated on the basis of race and awarded Plaintiffs

$950,000 in compensatory damages. In resolving the parties' post-trial motions

under Rules 50 and 59 of the Federal Rules of Civil Procedure, the district court

(Johnson, J.) sua sponte ordered a new trial as to all Plaintiffs limited to the issue

of damages after finding that "the source of the damages assessed is not clear."

Saint-Jean v. Emigrant Mortg. Co.,

(E.D.N.Y. 2018). The

second jury awarded four Plaintiffs compensatory damages totaling $722,044

and nominal damages of $1 each to the four Plaintiffs whose foreclosure

proceedings were ongoing. The district court (Brodie, C.J.) entered final

judgment on November 16, 2022. 1 Emigrant timely appealed on December 6,

2022.

1 Judge Sterling Johnson presided over the case through both jury trials. The case was reassigned to Chief Judge Margo K. Brodie on October 13, 2022, after Judge Johnson's passing.

5 On appeal, Emigrant argues that the district court erred in three

ways: first, by finding Plaintiffs' claims timely under the doctrine of equitable

tolling and the discovery rule of accrual; second, in its instructions to the jury on

disparate impact and disparate treatment theories of discrimination; and third, in

holding, contrary to the first jury's verdict, that a release-of-claims provision in a

loan modification agreement signed by two Plaintiffs was unenforceable as a

matter of law.

We conclude that the district court did not abuse its discretion in

holding that Plaintiffs' claims were timely under the doctrine of equitable tolling.

We also conclude that there was no error in the district court's instructions to the

jury. Finally, we agree that the release-of-claims provision contained in a loan

modification agreement between Emigrant and two Plaintiffs is unenforceable as

a matter of law. Accordingly, we AFFIRM the judgment of the district court.

BACKGROUND

On appeal following a jury verdict, we construe the evidence at trial

in favor of the prevailing party -- here, Plaintiffs. See Triolo v. Nassau Cnty.,

(2d Cir. 2022). During the first trial, Plaintiffs presented evidence,

through their own testimony, expert witnesses, and documents, that showed not

6 only that Emigrant's STAR NINA loan was designed to fail, but also that

Emigrant targeted Black and Latino borrowers.

I. The Facts

A. The Plaintiffs and Their Loans

In 1979, Howell and his wife moved into a new home on 158th Street

in Jamaica, Queens with their young children. 2 Over the next thirty years, the

Howells paid down their mortgage until they owned their home outright.

Howell even burned his loan paperwork with friends to celebrate being debt-

free.

By February 2008, Howell's home was worth $430,000. With only a

$6,000 lien on the property, Howell had $424,000 in equity in his home. One day,

a contractor knocked on Howell's door and told him that he could make some

extra money by constructing a rental property in his yard. Howell, who was

retired by then, expressed his interest in generating extra income and the

contractor introduced Howell to a broker who helped him apply for loans. The

broker introduced Howell to Emigrant.

2 Howell died intestate on May 1, 2020. On March 22, 2022, the district court (Johnson, J.) granted Plaintiffs' motion, pursuant to Rule 25 of the Federal Rules of Civil Procedure, to substitute Felipe R. Howell, Jr., the administrator of Howell's estate, for Howell.

7 Emigrant approved Howell for a $200,750 STAR NINA loan with an

initial interest rate of 10.375%. Howell and Emigrant closed on the loan on

February 6, 2008. Although Howell did not realize it at signing, his loan

contained a provision providing for an 18% default interest rate if he missed a

single payment. The default provision appeared in a separate rider to the main

loan papers, and no one otherwise informed him of the default rate.

Howell could not afford the $1,817.61 monthly payments on his

STAR NINA loan. Indeed, his inability to make a single payment triggered the

18% default rate, which quickly caused him to fall deeper into the debt he owed

Emigrant. In March 2009, Emigrant obtained a judgment of foreclosure on

Howell's home and purchased it for $1,000 in August 2009. Despite having had

over $400,000 in equity in his home before he took out the STAR NINA loan,

Howell lost his home -- what he had called his "castle" -- and received no part of

the foreclosure proceeds. J. App'x at 807.

The other Plaintiffs presented similar stories at trial. All are Black

homeowners who, having poor credit and low incomes but high equity in their

homes, were candidates for Emigrant's STAR NINA loan program. Plaintiffs had

largely similar experiences with Emigrant. First, seeking to refinance existing

8 loans or mortgages, Plaintiffs went to Emigrant because the STAR NINA loan,

unlike other loans, did not require income or asset verification. Then, at closing,

when Plaintiffs felt rushed, Emigrant would produce a stack of documents for

the borrowers to sign. Buried in those documents was a "default interest rate

rider" separate from the main mortgage document giving Emigrant the power to

impose an 18% default interest rate if a payment was overdue by thirty days. Ex.

App'x at 662.

Eventually, and generally soon after closing, each Plaintiff defaulted

on their loan and was subject to the 18% interest rate. Emigrant then initiated

foreclosure proceedings against Plaintiffs' homes. Like Howell, Plaintiff Linda

Commodore lost her home to Emigrant; most of the foreclosure proceeds went to

paying Emigrant. Plaintiffs Jeanette and Beverely Small, a mother and daughter

who lived with Beverley's young son, sold their apartment to avoid foreclosure,

though most of the proceeds went to paying Emigrant. Two married

couples -- Jean Robert and Edith Saint-Jean and Felex and Yanick Saintil -- were

still in foreclosure proceedings at the time of oral argument in this case.

The Saintils delayed foreclosure by requesting two loan

modifications from Emigrant. In March 2010, the couple entered into their

9 second modification agreement with Emigrant whereby Emigrant reduced their

monthly payments to $2,804.38 and reduced the interest rate on the loan to 6%

per year for five years, after which the contract provided that "the interest rate

shall be reset to the original fixed rate term of the mortgage." Ex. App'x at 631.

In exchange, the Saintils submitted an initial payment of $5,632.98 to Emigrant.

As part of the modification agreement, the Saintils agreed to "release and forever

discharge Emigrant . . . from any and all claims" they might have against

Emigrant. Id. at 632-33.

The Saintils still could not afford the reduced monthly payment, but

they entered the agreement out of fear that if they did not modify their loan

Emigrant would "take [their] house." J. App'x at 1788. Indeed, Emigrant

eventually stopped accepting payments from them and initiated foreclosure

proceedings.

B. The STAR NINA Loan Program 3

Emigrant offered STAR NINA loans in response to an apparent

"marketplace demand for a no-income verification loan product for people with

low credit scores." J. App'x at 2396. Because Emigrant did not ask potential

3 The facts in this section and the next are drawn primarily from the testimony of expert witnesses at trial.

10 borrowers to report their income or assets on the loan application, STAR NINA

loans were "based solely on the strength of collateral values with no

consideration given to the borrowers' capacity to repay." Ex. App'x at 672.

STAR NINA loans had four unusual characteristics. First, Emigrant

did not verify borrowers' income or assets; second, a credit score below 600 was

required; third, each loan came with a default interest rate of 18%; and fourth,

borrowers had to have high equity -- in the ballpark of 50% -- in their homes.

When considered together, these characteristics made it likely that

STAR NINA loans would fail from the outset: "[b]ecause these loans were

underwritten based on the collateral value of the underlying property and not on

the borrower's ability to repay, there was little assurance that borrowers who

defaulted on the loan would be able to make any payments at the default interest

rate." Ex. App'x at 662 (findings of 2009 New York State Banking Department

report).

No income, no asset loans are not uncommon, but a typical no-

income, no-asset borrower is an individual with a "[great] deal of equity" and

"extremely high credit," who has "difficulty identifying their income." J. App'x at

713. Emigrant's STAR NINA loan -- which looked to credit scores only to require

11 a low one -- was therefore "absolutely" different from a typical no-income, no-

asset loan. Id. Rebecca Walzak, a mortgage lending consultant and expert in

underwriting and mortgage industry practices, explained that STAR NINA loans

were "targeted to some of the most vulnerable individuals in the community," id.

at 771, and described the terms as "the worst [she] had ever seen in a mortgage

loan." Id. at 701.

Despite the high risk of default, the loans were highly profitable for

Emigrant; indeed, it is apparent that Emigrant made these loans because it was

likely the borrowers would default. See Ex. App'x at 662 ("[Emigrant's] sole

consideration in granting these mortgages seems to be whether there is sufficient

equity cushion to cover arrears, default costs, and legal fees in a foreclosure

proceeding.") (findings of New York Banking Department). The typical STAR

NINA borrower had high equity in their home, generally because of years of

paying down an original mortgage, and so there was a "sufficient equity

cushion" to more than cover Emigrant's original loan as well as costs and legal

fees in a foreclosure proceeding. Id. at 662. Indeed, STAR NINA loans were

"highly profitable" for Emigrant, id. at 695, and generated $50 million in interest

revenue in 2008 alone.

12 While conducting an audit of Emigrant, the New York Banking

Department determined that Emigrant did not "disclose[] [the default interest

rate] in the underlying mortgage note but only in the rider," Ex. App'x at 662,

and "there is no evidence . . . that brokers from whom [Emigrant] received

applications . . . disclosed the 18% default interest rate to the borrower at the time

of application or thereafter." Id. After the New York Banking Department

criticized the use of the 18% default interest rate in 2008, Emigrant discontinued

the STAR NINA loan program. 4

C. The Discriminatory Nature of the STAR NINA Program

There was another dimension to the STAR NINA program beyond

the unusual features of the loan that often led to default -- Emigrant also made

the loans to Black and Latino borrowers more frequently than to borrowers of

other races. Statistical evidence at trial showed that STAR NINA loans given

between 2004 and 2010 were disproportionately sold to Black borrowers in

neighborhoods that were majority Black and Latino. In communities with 10%

or less Black or Latino residents, 23% of Emigrant's refinancing loans were STAR

4 At a December 2008 board meeting, Emigrant's CEO Howard Milstein "noted that the Bank has withdrawn from the STAR mortgage lending program, resulting from a criticism . . . regarding the Bank's default interest rider." Ex. App'x at 719 (Emigrant board meeting minutes dated December 9, 2008).

13 NINA loans. But as the percentage of Black and Latino residents increased to

upwards of 80% of a given census tract, Emigrant's STAR NINA loans increased

too, almost doubling to comprise up to 45% of its refinancing loans issued in the

same area.

A different statistical analysis compared neighborhoods that were

similar in terms of creditworthiness, median household income, homeownership

rate, and level of educational attainment. As the proportion of Black and Latino

residents in a neighborhood increased, the number of STAR NINA loans also

increased by a statistically significant percentage, even when controlling for

credit score and other non-race-related factors.

Emigrant also targeted Black and Hispanic communities through

their STAR NINA advertising campaigns. Between 2005 and 2008, Emigrant

spent $102,734 -- 76% of its overall advertising budget -- on running ads in four

newspapers catering to Black and Hispanic readerships: Caribbean Life, Black Star

News, Oui (a Spanish-language newspaper), and Mi Zona Hispana.

After 2008, once the STAR NINA loan program ended, Emigrant

stopped advertising in those newspapers. From 2009 to 2010, Emigrant's

newspaper spending dropped from $102,734 to $200. Emigrant devoted 82% of

14 its total advertising spending to advertising in newspapers that circulated in

areas with a combined Black and Hispanic population of over 80%. And 96% of

the images in Emigrant's ads featured families who were either Black or

Hispanic. According to a statistical analysis, this constituted a "positive and

strong correlation" between the Black and Hispanic population and Emigrant's

advertising spending.

Emigrant's internal communications and policies showed that

Emigrant was aware of race-based disparities in its STAR NINA loan-writing,

and that it encouraged race-based targeting of the loan. For instance, Vice

Chairman James Woolsey acknowledged the racial disparity in a July 2006 email,

writing to a colleague that there is a "[v]ery simple answer to why the pricing of

loans to blacks is higher than to whites." Ex. App'x at 711. He went on to explain

that in 2004, 2005, and 2006, the percentage of loans made to Black borrowers

that were STAR NINA loans was 50%, 75%, and 70%, respectively, compared to

22%, 40%, and 48% of the loans to White borrowers. Id.

Other documents showed that Emigrant targeted "ethnic parts of the

community," J. App'x at 993, and tracked advertising expenditures in "African-

American," "Caribbean," "Haitian," and "Hispanic" newspapers. Id. at 481.

15 Emigrant did not similarly track advertising targeted at White neighborhoods or

readerships. Moreover, STAR NINA loans were sold "almost exclusively by

brokers." Id. at 989 (testimony of CEO Howard Milstein).

D. Plaintiffs Learn of Emigrant's Discriminatory Lending Practices

Even though all Plaintiffs experienced similar issues with their

STAR NINA loans -- rushed closings, hidden terms, unaffordable monthly

payments, a punitive interest rate, foreclosure or threats of foreclosure, and

mental and emotional turmoil over their debt and losing their homes -- none of

the Plaintiffs was aware of the possibility that the loans were discriminatory until

May 2009, when the Saint-Jeans arrived at foreclosure court for a proceeding

with Emigrant.

As the Saint-Jeans waited in the courthouse lobby, they realized that

there were other individuals also waiting for proceedings involving Emigrant.

Jean Robert Saint-Jean observed that the other borrowers were also "people of

color." Id. at 1601-02. He saw that when cases involving Emigrant were called,

"[e]verybody [] standing in the room" were "[a]ll these other people of color," and

that this observation "push[ed]" him to ask his lawyer about what he had

observed. Id. at 1601.

16 Felex and Yanick Saintil joined the case in May 2012. Prior to joining

the case, Felex Saintil did not know anyone else who received a STAR NINA

loan, nor did he know to whom Emigrant was targeting its advertising. Beverley

Small also did not know of anyone else who had received a STAR NINA loan

when she closed on hers. In August 2013, however, she and Jeanette learned

about other Black borrowers who had lost their homes as a result of defaulting

on STAR NINA loans, and after learning these facts they began to suspect

discrimination. Linda Commodore explained: "I met with my now attorneys and

after discussing my situation, I realized that there was information that I wasn't

aware of and I believe that it was, it was discriminatory." Id. at 1736. The Smalls

joined the case in 2014. The remaining Plaintiffs learned of Emigrant's

discrimination in 2013 and likewise joined the case in 2014, also after learning

facts of their discrimination and speaking with counsel.

II. Proceedings Below

On April 29, 2011, Jean Robert and Edith Saint-Jean filed this action

against Emigrant, alleging, inter alia, racial discrimination in violation of the

FHA,

and 2605, the ECOA,

et seq., and the

NYCHRL, N.Y.C. Administrative Code § 8-101 et seq.

17 On June 8, 2011, Emigrant moved to dismiss the complaint, arguing,

inter alia, that the claims were time-barred under the applicable statutes of

limitation. On May 4, 2012, while Emigrant's motion to dismiss was pending,

Plaintiffs filed a first amended complaint adding Felex and Yanick Saintil as

parties.

On September 25, 2014, the district court denied Emigrant's motion

to dismiss, adopting the report and recommendation of the magistrate judge

(Orenstein, M.J.) to conclude that the discovery rule and the doctrine of equitable

tolling applied to Plaintiffs' claims. On October 2, 2014, Plaintiffs filed a second

amended complaint adding the remaining Plaintiffs. On August 3, 2015,

Emigrant moved for summary judgment, arguing, inter alia, that Plaintiffs' claims

were time-barred and neither the discovery rule nor the doctrine of equitable

tolling could save the claims. On February 26, 2016, the district court heard oral

argument on the motion and denied it, ruling from the bench.

The first trial took place on issues of liability and damages from May

23 to June 27, 2016. The jury found Emigrant liable as to all Plaintiffs except the

Saintils and awarded compensatory damages of $950,000. The jury determined

that the Saintils knowingly and voluntarily agreed to release their claims against

18 Emigrant in their March 2010 modification agreement. After trial, Emigrant

moved, pursuant to Rules 50 and 59 of the Federal Rules of Civil Procedure, for

judgment as a matter of law or, in the alternative, a new trial. Plaintiffs also

moved for a new trial as to the Saintils, arguing that contrary to the jury's verdict,

the release provision in Saintils' loan modification agreement was unenforceable.

Two years later, the district court denied Emigrant's post-trial

motions in their entirety and granted Plaintiffs' Rule 50 motion as to the Saintils,

ruling that the Saintils' waiver was unenforceable as a matter of law. The district

court also sua sponte ordered a new trial as to all Plaintiffs, solely on damages,

reasoning that the original jury award was "against the weight of the evidence,"

and that "the damages . . . [did not] succeed at restoring Plaintiffs to their pre-

STAR NINA loan positions," Saint-Jean,

.

Following a second trial on damages that took place from May 7 to

May 22, 2019, a jury awarded $722,044 in compensatory damages to the four

Plaintiffs who had lost their homes, and $1 each in nominal damages to the four

Plaintiffs whose foreclosure proceedings remained ongoing. The district court

entered final judgment in favor of Plaintiffs on November 16, 2022.

This appeal followed.

19 DISCUSSION

On appeal, Emigrant argues principally that the district court: (1)

erred in holding that Plaintiffs' claims were timely pursuant to the doctrine of

equitable tolling and the discovery rule; (2) incorrectly instructed the jury on

disparate impact and intentional discrimination claims; and (3) erroneously

overturned the jury's verdict as to the Saintils by finding that the release-of-

claims provision is unenforceable as a matter of law. We address each issue in

turn after first providing an overview of the FHA.

I. The Fair Housing Act

Section 3604 of the FHA, also known as Title VIII of the Civil Rights

Act of 1968, makes it unlawful to "make unavailable . . . a dwelling to any person

because of race, color, . . . or national origin."

(a); see

LeBlanc-Sternberg v. Fletcher,

(2d Cir. 1995). Section 3605 of the

FHA makes it unlawful for "any person or other entity whose business includes

engaging in residential real estate-related transactions to discriminate . . . in

making available such a transaction, or in the terms or conditions of such a

transaction, because of race, color, . . . or national origin."

(a).

As the Supreme Court has held, disparate impact discrimination

20 claims are cognizable under the FHA. Texas Dep't of Hous. & Cmty. Affs. v.

Inclusive Cmtys. Project, Inc. ("Inclusive Communities"),

, 545-46 (2015)

("The Court holds that disparate-impact claims are cognizable under the Fair

Housing Act upon considering its results-oriented language, the Court's

interpretation of similar language in Title VII and the ADEA, Congress'

ratification of disparate-impact claims in 1988 against the backdrop of the

unanimous view of nine Courts of Appeals, and the statutory purpose.").

Courts have held that "redlining" violates the FHA. See, e.g.,

Nationwide Mut. Ins. Co. v. Cisneros,

(6th Cir. 1995); NAACP v.

American Fam. Mut. Ins. Co.,

(7th Cir. 1992); Ring v. First

Interstate Mortg., Inc.,

(8th Cir. 1993). Redlining is "the

practice of denying the extension of credit to specific geographic areas due to the

income, race, or ethnicity of its residents." United Companies Lending Corp. v.

Sargeant,

, 203 n.5 (D. Mass. 1998); see also, e.g., Crawford v.

Signet Bank,

, 928 n.4 (D.C. Cir. 1999) ("[R]edlining is the practice of

financial institutions intentionally not lending to certain neighborhoods or parts

of a community." (internal quotation marks omitted)); Raymond H. Brescia,

Subprime Communities: Reverse Redlining, The Fair Housing Act and Emerging Issues

21 in Litigation Regarding the Subprime Mortgage Crisis, 2 Albany Gov't L. Rev. 164,

178 (2009) ("The term 'redlining' derives its origins from lending practices where

bankers would literally draw a red line on maps, identifying the communities --

typically communities of color -- where the bank would not extend credit."

(emphasis added)).

Courts have also recognized "reverse redlining" claims under the

FHA. Reverse redlining is "'the practice of extending credit on unfair terms'

because of the plaintiff's race and geographic area." Steed v. EverHome Mortg. Co.,

(11th Cir. 2009) (quoting Hargraves v. Cap. City Mortg. Corp.,

(D.D.C. 2000)) (emphasis added). In reverse redlining cases,

lenders engage in "predatory" lending practices, including imposing "exorbitant

interest rates, lending based on the value of the asset securing the loan rather

than a borrower's ability to repay . . . , repeated foreclosures, and loan servicing

procedures in which excessive fees are charged" in certain neighborhoods.

Hargraves,

. 5

5 In 1993, the Senate Committee on Banking, Housing, and Urban Affairs conducted a hearing on reverse redlining. Senator Alfonse M. D'Amato called "reverse redlining [] among the most pernicious forms of racial and ethnic discrimination and consumer fraud." Reverse Redlining; Problems in Home Equity Lending: Hearings Before the Senate Comm. On Banking, Hous., and Urban Affairs, 103d Cong. 243, 246 (Feb. 17, 1993).

22 Reverse redlining is a product of redlining, because the latter creates

"a credit-vacuum filled by predatory lenders" who know that borrowers in

historically redlined areas "are a captive market with no access to reasonably-

priced credit." Equity Predators: Stripping, Flipping and Packing Their Way to Profits,

Hearing Before Special Comm. On Aging, 105th Cong. 1, 86 (Mar. 16, 1998)

(statement of William J. Brennan, Jr., Atlanta Legal Aid Society, Inc.). Instead of

not making loans, as in a traditional redlining case, a lender engages in reverse

redlining when it makes loans in a community based on race and does so in an

unfair and predatory way.

In Hargraves, to proceed on a reverse-redlining theory, the plaintiff

was required to show that the defendant's lending practices were "unfair" and

"predatory" and that the defendants either intentionally targeted borrowers on

the basis of race or that there was a disparate impact on the basis of race.

Hargraves,

. Hence, the court considered two prongs of a

reverse redlining claim: first, whether there were "unfair" and "predatory" loan

terms, and second, whether there was discrimination based on the familiar

theories of disparate treatment or disparate impact discrimination.

23 While we have not addressed a reverse redlining case in our Circuit,

the Supreme Court has instructed that the FHA is to be given a "generous

construction," City of Edmonds v. Oxford House, Inc.,

(1995), and

district courts (including in our Circuit) have permitted reverse redlining claims

to proceed under the FHA, the ECOA, and the TILA. See, e.g., M&T Mortg. Corp.

v. White,

(E.D.N.Y. 2010) (FHA and ECOA); Barkley v.

Olympia Mortg. Co., Nos. 04-CV-875 (RJD), 05-CV-187 (RJD), 05-CV-4386 (RJD),

05-CV-5302 (RJD), 05-CV-5632 (RJD), 05-CV-5679 (RJD),

, at *13-

15 (E.D.N.Y. Aug. 22, 2007) (FHA); Hargraves,

; Henderson

v. Vision Prop. Mgmt., LLP,

, at *4-5 (E.D. Mich. Aug. 23, 2021)

(FHA and ECOA). Moreover, Emigrant has not contested the cognizability of a

reverse redlining claim under the FHA. Accordingly, we hold that where a

defendant lender engages in "unfair" or "predatory" lending practices targeting

borrowers based on race, or where those lending practices have a disparate

impact based on race, this is reverse redlining in violation of the FHA.

II. Timeliness of Plaintiffs' Claims

Emigrant contends the district court erred in determining that

Plaintiffs' claims, though brought after the expiration of the otherwise applicable

24 limitations periods, were timely under the discovery rule of accrual and the

doctrine of equitable tolling.

To bring a claim under the FHA, "[a]n aggrieved person may

commence a civil action . . . not later than 2 years after the occurrence or the

termination of an alleged discriminatory housing practice."

(a)(1)(A). The applicable statutes of limitation are two years under the FHA

and the ECOA, and three years under the NYCHRL.

(a)(1)(A)

(FHA); 15 U.S.C. § 1691e(f) (eff. Oct. 28, 1974) (ECOA);

(NYCHRL).

The timeliness dispute stems from the parties' different

conceptualizations of Plaintiffs' injury. Claims under the FHA, like other federal

causes of action, accrue when a "plaintiff knows or has reason to know of the

injury that is the basis of the action." Leon v. Murphy,

(2d Cir.

1993) (internal quotation marks omitted). Accordingly, we must first identify the

injury that animates Plaintiffs' claims. Emigrant's position is that Plaintiffs'

injury is the unfavorable loan itself, meaning that any claims accrued at closing,

or, at the latest, upon default. On the other hand, Plaintiffs contend that they are

not suing because they received financially unfavorable loans; rather, their

25 claims for relief arise from the fact that, as the jury found, they were

"discriminated against in a systemic fashion." Pl. Br. at 17, 26. Alternatively,

Plaintiffs argue that even if their claims accrued at closing or upon default,

equitable tolling should apply.

We agree that the relevant injury is discrimination and not the loan

itself. The "discriminatory housing practice" is not the making of an onerous

loan but it is, instead, Emigrant's targeting of Black and Latino borrowers and

pattern of writing STAR NINA loans to a disproportionate number of Black

borrowers knowing that the loans were likely to fail. While Plaintiffs were no

doubt harmed by the loans themselves, their claims against Emigrant sound in

discrimination, and not, for instance, fraud related to the loans . As the district

court in Phillips v. Better Homes Depot, Inc. put it, "[t]here is a difference between

being aware that you got a bad deal and being aware that you were

discriminated against in a systematic fashion." No. 02-CV-1168 (ERK),

, at *25 (E.D.N.Y. Nov. 12, 2003); see also

(discussing Barrett v. United

States,

(2d Cir. 1982)) (holding that, "in a case of

discrimination, a victim may not know he or she has been the target of

26 discrimination until meeting other victims or learning more about lending

practices in minority communities").

Accordingly, and as we explain further below, we hold that even if

Plaintiffs' claims accrued at closing or upon default when they first learned of the

onerous default interest rate, the district court did not abuse its discretion in

exercising its equitable power to toll the statute of limitations until the date when

Plaintiffs knew or had reason to know of their injury -- that they were victims of

Emigrant's sophisticated and systemic pattern of discriminatory lending.

Because we conclude that the claims are timely under the doctrine of equitable

tolling, we do not reach the issue of whether they are also timely under the

related, but distinct, doctrine of the federal rule of discovery. 6

6 While "[t]he distinction between equitable tolling and the diligence-discovery rule has not always been clear in our caselaw," we see the difference as follows: the discovery rule "delays the date of accrual" whereas "the doctrine of equitable tolling applies after the claim has already accrued, suspending the statute of limitations to prevent unfairness to a diligent plaintiff." Ellul v. Congregation of Christian Bros.,

(2d Cir. 2014) (internal citations and quotation marks omitted). We decline to consider whether Plaintiffs may have an argument that the discovery rule applies to their claims, and instead analyze Plaintiffs' claims under the equitable tolling doctrine because even assuming an accrual date of signing (which is Emigrant's position), the statutes of limitation should still be tolled to when Plaintiffs began or had reason to suspect discrimination, as a matter of equity. Dillman v. Combustion Eng'g, Inc.,

(2d Cir. 1986) (holding that in the context of a Title VII claim, "when an employer's misleading conduct is responsible for the employee's unawareness of his cause of action" and is "extraordinary" enough, "equitable tolling will defer the start of

27 A. Standard of Review

"A decision whether to equitably toll a statute of limitations is left to

the sound discretion of the district court, and should not be disturbed absent an

abuse of that discretion." Kregos v. Associated Press,

(2d Cir. 1993);

accord A.Q.C. ex rel. Castillo v. United States,

(2d Cir. 2011)

(district court's grant of equitable tolling is generally reviewed for "abuse of

discretion"). We have also explained, however, that "the operative review

standard" for equitable tolling depends on "what aspect of the lower court's

decision is challenged." Belot v. Burge,

(2d Cir. 2007). "[W]e

review the legal premises for [the district court's] conclusion de novo, the factual

bases for clear error, and the ultimate decision for abuse of discretion." DeSuze v.

Ammon,

(2d Cir. 2021). Only where the court "has understood

the governing law correctly, and has based its decision on findings of fact which

were supported by the evidence, but the challenge is addressed to whether the

court's decision is one of those within the range of possible permissible

decisions," will we review for abuse of discretion in both "name" and "operation."

Belot,

.

the . . . filing period from the time of the discriminatory action to the time the employee should have discovered the action's discriminatory nature").

28 The district court addressed the issue of timeliness four times

throughout the life of the case: in denying Emigrant's motion to dismiss, its

motion for summary judgment, its motion at the close of trial, and its posttrial

motion. The magistrate judge also addressed the issue of timeliness. We have

emphasized that, because "equitable tolling often raises fact-specific issues

premature for resolution on a Rule 12(b)(6) motion, before a plaintiff can develop

the factual record . . . it is generally improper to dismiss a complaint as

untimely." Clark v. Hanley,

(2d Cir. 2023). On summary judgment,

"where a reasonable factfinder could conclude" that a plaintiff's claims are timely,

"courts routinely . . . deny summary judgment."

. Judge Johnson's

denials of Emigrant's motions to dismiss and for summary judgment based on

timeliness were proper.

At the close of trial and again in subsequent motion practice, with

the benefit of six weeks of evidence and a jury verdict for Plaintiffs, Judge

Johnson denied Emigrant's requests for judgment as a matter of law on the

timeliness issue. The facts as established by the evidence at trial show that the

29 district court did not abuse its discretion in equitably tolling the statutes of

limitation.

B. Applicable Law

"Statutes of limitations are generally subject to equitable tolling

where necessary to prevent unfairness to a plaintiff who is not at fault" for

lateness in filing. Veltri v. Bldg. Serv. 32B-J Pension Fund,

(2d Cir.

2004). "The taxonomy of tolling, in the context of avoiding a statute of

limitations, includes at least three phrases: equitable tolling, fraudulent

concealment of a cause of action, and equitable estoppel." Pearl v. City of Long

Beach,

(2d Cir. 2002). We conclude here that the doctrine of

equitable tolling applies to render Plaintiffs' claims timely in this case.

A district court may exercise its discretion to equitably toll the

statute of limitations once a litigant has demonstrated "'that some extraordinary

circumstance stood in her way' and . . . 'that she has been pursuing her rights

diligently.'" Doe v. United States,

(2d Cir. 2023) (alterations

adopted) (quoting A.Q.C. ex rel. Castillo v. United States,

(2d Cir.

2011)). "This standard calls for reasonable diligence, not maximum feasible

diligence, which a [plaintiff] may satisfy by showing that he acted as diligently as

30 reasonably could have been expected under the circumstances." Harper v. Ercole,

(2d Cir. 2011) (emphases, internal quotation marks, and

internal citations omitted). Importantly, equitable tolling is appropriate where,

"despite all due diligence," a plaintiff "is unable to obtain vital information

bearing on the existence of his claim" within the statute of limitations period.

Valdez ex rel. Donely v. United States,

(2d Cir. 2008) (quoting Cada

v. Baxter Healthcare Corp.,

(7th Cir. 1990)). Equitable tolling may

be appropriate even if there are lengthy delays in filing. See, e.g., Baskin v.

Hawley,

(2d Cir. 1986) (tolling the statute of limitations

for eight years).

"While equitable tolling extends to circumstances outside both

parties' control, the related doctrines of equitable estoppel and fraudulent

concealment may bar a defendant from enforcing a statute of limitation when its

own deception prevented a reasonably diligent plaintiff from bringing a timely

claim." Sebelius v. Auburn Reg'l Med. Ctr.,

(2013) (Sotomayor, J.,

concurring). This Circuit has explicitly clarified that "fraudulent concealment is

not essential to equitable tolling." Valdez,

; Veltri,

("The relevant question is not the intention underlying defendants' conduct, but

31 rather whether a reasonable plaintiff in the circumstances would have been

aware of the existence of a cause of action.").

To show fraudulent concealment, a plaintiff must establish that: (1)

the defendant concealed the existence of the cause of action from the plaintiff; (2)

the concealment prevented plaintiff's discovery of the claim within the

limitations period; and (3) the plaintiff's ignorance of the claim did not result

from a lack of diligence. New York v. Hendrickson Bros., Inc.,

(2d Cir. 1988). A plaintiff can prove concealment by showing "either that the

defendant took affirmative steps to prevent the plaintiff's discovery of his claim

or injury or that the wrong itself was of such a nature as to be self-concealing."

C. Application

In the circumstances here, we agree with the district court that

equitable tolling was appropriate because, through no fault of their own,

Plaintiffs did not learn of their cause of action, and could not reasonably be

expected to do so with the exercise of due diligence, within the limitations

period. Plaintiffs demonstrated that their inability to discover the discriminatory

practice was an extraordinary circumstance standing in their way of bringing

32 suit, and that they were inhibited from pursuing their rights diligently until they

were on notice of their claims. Though that alone would be enough to support

the district court's discretionary equitable tolling determination, the court's

determination is further supported by the fact that Emigrant took steps to

conceal the discriminatory nature of the STAR NINA loan such that Plaintiffs

were reasonably prevented from learning about their discrimination cause of

action within the statutory period.

The core inquiry of our equitable tolling analysis is whether there is

"unfairness to a plaintiff who is not at fault" such that a district court may

exercise its discretion and equitable power to toll the statute of limitations. See

Veltri,

. While the dissent criticizes our decision as creating a

"fairness-based tolling rule," Diss. Op. at 8, equitable tolling has always been

based on principles of fairness and equity. See Cerbone v. Int'l Ladies' Garment

Workers' Union,

, 48-49 (2d Cir. 1985) ("Though ordinarily the

applicable period starts to run on the employer's commission of the unlawful act

and is not tolled pending the employee's realization that the conduct was

discriminatory, a court might in some cases permit tolling as a matter of

fairness." (internal quotation marks omitted)). Of course, we do not rely solely

33 on notions of fairness to conclude that equitable tolling is appropriate in this

case.

It is the egregious nature of Emigrant's discriminatory lending

practice that makes this case extraordinary. Emigrant targeted Black and

Hispanic borrowers who owned substantial equity in their homes and imposed

exorbitant default interest rates designed to lead them to default on their loans.

Although Plaintiffs had owned their homes from a minimum of five years to as

many as thirty years, foreclosure proceedings were initiated against them within

two years of the origination of their Emigrant origination loans. And

unbeknownst to Plaintiffs until years later, they were targeted for the loans

because of their race.

We therefore agree with the district court that the extraordinary

facts here support equitable tolling because Plaintiffs, through no lack of

diligence of their own, were unaware of the facts of discrimination within the

statutory period. Each Plaintiff signed his or her loan with Emigrant without

knowledge of other borrowers or Emigrant's marketing and business strategies

surrounding the STAR NINA loan. It was not until 2009, when the Saint-Jeans

34 observed other Black borrowers in foreclosure court, that those Plaintiffs learned

that there were others like them who were subjected to the predatory loans.

Moreover, Emigrant took steps to conceal the predatory nature of

the loans and Plaintiffs had no reason to suspect the discriminatory nature of the

lending practices. Indeed, Emigrant almost exclusively targeted Black and

Hispanic media outlets. Plaintiffs had no way of knowing that these loans were

not similarly targeted towards White communities. And, because of the

individualized nature of each Plaintiff's transaction with Emigrant, there was

nothing at closing to indicate to Plaintiffs that the STAR NINA loans were being

given disproportionately to Black borrowers. All Plaintiffs testified that

Emigrant rushed them to sign stacks of documents at closing, and a New York

Banking Department review of Emigrant's policy revealed that the 18% default

interest rate was left out of the main documents and included only in a separate

rider to the loan, further obfuscating the true financial impact of the STAR NINA

loan. Emigrant also dissuaded Plaintiffs from bringing lawyers to closing,

representing to Plaintiffs that a lawyer at closing would represent their interests.

On this record, and construing the facts in Plaintiffs' favor, there is sufficient

evidence that Emigrant took steps to conceal the discriminatory nature of STAR

35 NINA loans. 7 Jones v. Treubig,

(2d Cir. 2020) ("[W]e must

'consider the evidence in the light most favorable to the party against whom the

motion was made . . . .'" (quoting Black v. Finantra Cap., Inc.,

, 209 (2d

Cir. 2005))).

Our conclusion flows from well-settled principles of equitable

tolling and fraudulent concealment. Lozano v. Montoya Alvarez,

(2014) ("Congress is presumed to incorporate equitable tolling into federal

statutes of limitations because equitable tolling is part of the established

backdrop of American law.") (citing Rotella v. Wood,

(2000)). The

dissent paints equitable tolling as a rigid, stepwise doctrine, but that is not

correct. Indeed, "[a]s the courts in this country recognized early on, 'the essence

of the doctrine of equitable tolling is that a statute of limitations does not run

against a plaintiff who is unaware of his cause of action.'" Joseph A. Seiner, Time,

Equity, and Sexual Harassment, 12 U.C. Irvine L. Rev. 573, 595 (2022) (alterations

adopted) (quoting Serdarevic v. Advanced Med. Optics, Inc.,

(Fed. Cir. 2008)). Additionally, despite the dissent's assertion that "Title VII [and

7 Though we agree with the district court that there was concealment here, we emphasize that concealment is not a requirement for equitable tolling -- avoiding unfairness to the plaintiff is reason enough to equitably toll a statute of limitations. See Veltri,

.

36 ADEA] claims accrue at the time a plaintiff learns of the discriminatory act," Diss.

Op. at 17, we have held that the time period is tolled or delayed pending the

employee's realization that the conduct was discriminatory when "the employee

was actively misled by his employer, he was prevented in some extraordinary

way from exercising his rights, or he asserted his rights in the wrong forum, in

which event tolling of the time bar might be permitted as a matter of fairness,"

Miller v. Int'l Tel. & Tel. Corp.,

, 24 (2d Cir. 1985).

And there are district court cases in the Second Circuit that apply

both the doctrines of equitable tolling and fraudulent concealment in similar

discriminatory lending contexts where, as here, the parties did not have any

reason to suspect discrimination until they learned facts that caused them to

realize that their experience was part of a larger pattern of discrimination.

For instance, Barkley v. Olympia Mortgage Co. involved an allegedly

discriminatory and fraudulent property-flipping scheme, where plaintiffs alleged

that defendants bought damaged properties at foreclosure sales and, working

with appraisers, lenders, and lawyers, targeted persons of limited financial

means and savvy in low-income Black and Latino communities and sold them

the properties at inflated appraisal values while saddling the buyers with

37 "unconscionable loans."

, at *1, 9. In denying defendants'

motion to dismiss on timeliness grounds, the court noted that, "[w]ithout

meeting [the] other [defendant's] clients or explaining their circumstances to an

attorney who responsibly represented their interests, plaintiffs had no way of

knowing exactly how and why they had been victimized," even though the

plaintiffs "quickly realized they were on the receiving end of a defective

product."

.

In Council v. Better Homes Depot, Inc., the district court concluded it

was "reasonable that the Plaintiffs were not aware that they were the victims of

wrongdoing," because a "victim of discrimination may not know that he or she

has been the target of discrimination until meeting other victims or becoming

familiar with lending practices in minority communities."

, at

*9.

The principles articulated by the district courts in these cases are

readily applicable here. The structure of the STAR NINA lending program

included features by which Emigrant sought to conceal the discriminatory nature

of its predatory practices. Plaintiffs may have been injured individually by harsh

interest rates and hidden terms, but it took knowledge of Emigrant's treatment of

38 Plaintiffs in the aggregate to alert Plaintiffs to the probability of a claim for

discrimination. See Gabelli v. SEC,

(2013) ("Usually when a

private party is injured, he is immediately aware of that injury and put on notice

that his time to sue is running. But when the injury is self-concealing, private

parties may be unaware that they have been harmed. Most of us do not live in a

state of constant investigation; absent any reason to think we have been injured,

we do not typically spend our days looking for evidence that we were lied to or

defrauded. And the law does not require that we do so."). Indeed, had Plaintiffs

individually sued Emigrant at the time of closing (or default) for discrimination,

"they would not have survived a motion to dismiss the claims they now bring."

Pl. Br. at 18.

For instance, to make out a claim of disparate impact, Plaintiffs

would have needed to know facts showing a substantial adverse impact on Black

borrowers at the time they took out their loans. To show disparate treatment,

they would have needed to show that Emigrant was motivated at least in part by

race to target them and other Black and Latino borrowers. Without any

knowledge or facts as to other borrowers, and without knowledge of the steps

Emigrant took to target Black and Latino neighborhoods, none of the Plaintiffs

39 would have been able to state a claim based on the information known to them at

signing. And nothing in the record indicates that any Plaintiff had reason to

inquire more into the possibility of discrimination at the time of signing. The

district court thus did not abuse its discretion in tolling the statutes of limitation

to the date when the Saint-Jeans first began to suspect discrimination. We affirm

that decision; "[t]o hold otherwise would reward [Emigrant] for [its] evasiveness"

stemming from its sophisticated, multifaceted pattern of targeting the STAR

NINA loan to Black and Latino borrowers. Phillips,

, at *25.

We are not persuaded by Emigrant's arguments to the contrary. For

instance, Emigrant cites Pantoja v. Scott as an example of a case where, in an FHA

claim, the "alleged discriminatory act concluded with the closing on the

apartment," No. 96-CV-8593 (AJP),

, at *10 (S.D.N.Y. Oct. 26,

2001). The facts of Pantoja are readily distinguishable. Pantoja involved an

individual Puerto Rican resident who, pursuant to a lease providing the right to

purchase the apartment he rented, applied to the owner of his condominium

building for financing. The plaintiff alleged "that it was [the owner's] practice to

provide secondary financing to purchasers but that [the owner] refused to do so

for him, based on [racially] discriminatory reasons."

.

40 In granting defendants' motion to dismiss on timeliness grounds, the

district court held that the owner's "refusal to provide secondary financing

(which Pantoja claims was due to a discriminatory motive) 'occurred or

terminated,' at the latest, on [] the date of the real estate closing, when Pantoja

purchased the apartment without secondary financing from [the owner]."

(internal citations omitted). Thus, unlike our case, Pantoja involved one

plaintiff and one alleged discrete act of discrimination -- an instance that plaintiff

himself alleged he was aware of even prior to closing. See

. The plaintiff

in Pantoja had all the information he needed to assert a claim under the FHA by

his closing date, if not earlier. Not so here. At their respective STAR NINA

closings, Plaintiffs were, at best, arguably advised of the loan terms -- but they

had no basis to infer racial discrimination. 8 While Pantoja is an example of an

injury occurring, and ending, at closing, it does not help Emigrant's argument

here because the evidence at trial supports a conclusion that Plaintiffs did not

have any information about a possible discrimination claim at their respective

closings.

8 As the extensive record in this case shows, whether Plaintiffs were properly apprised of the actual loan terms is also not clear. See Ex. App'x at 662 (New York Banking Department report noting conflicting information about the cost of Emigrant loans in its disclosures). We resolve this ambiguity on appeal in favor of Plaintiffs.

41 By arguing that Plaintiffs were given all the terms of the predatory

loans upfront at signing, Emigrant obfuscates the injury that forms the basis of

Plaintiffs' discrimination claims. Emigrant is wrong that the "default interest rate

. . . forms the core of Plaintiffs' case." Def. Br. at 31. Plaintiffs are not suing for

the bad loans or the predatory terms; instead, they are suing for discrimination --

Emigrant subjected them to predatory lending practices at least in part because

of their race. Thus, even if Plaintiffs were apprised of the loan terms at closing,

that would be of no moment in analyzing whether Plaintiffs knew or should

have known about the discrimination at the time of closing. For the same reason,

Emigrant's argument that "Plaintiffs also failed to offer any evidence that

Emigrant representatives concealed material terms of the loans through untrue

representations" is also irrelevant to the issue of whether Emigrant's conduct at

the closings concealed their discrimination.

.

In short, reasonable borrowers in Plaintiffs' position could not have

known that they were victims of discrimination at signing, closing, default, or

even foreclosure. Moreover, by misleading Plaintiffs into thinking that counsel

at closing represented their interests, Emigrant took steps to conceal its

discriminatory scheme. Accordingly, we hold that, on these facts, the district

42 court did not abuse its discretion in applying the doctrine of equitable tolling,

and the statute of limitations was thus tolled to the date on which Plaintiffs

became aware of the discriminatory lending scheme.

III. The Jury Instructions

Emigrant argues that the district court's jury instructions on

disparate impact and disparate treatment theories of discrimination were

erroneous and require reversal. We find no reversible error in the challenged

portions of the district court's instructions.

A. Standard of Review

We review challenges to a district court's jury instructions de novo.

Mirlis v. Greer,

(2d Cir. 2020) (citing Uzoukwu v. City of New York,

(2d Cir. 2015)); Boyce v. Soundview Tech. Grp., Inc.,

(2d Cir. 2006) ("When a party challenges a court's jury charge, [the Second

Circuit] reviews the jury instructions de novo and as a whole."). We will overturn

a verdict on a challenge to jury instructions only if (1) the instructions were

43 erroneous, and (2) the error was prejudicial. See Turley v. ISG Lackawanna, Inc.,

(2d Cir. 2014).

"Jury instructions are erroneous if they mislead the jury or do not

adequately inform the jury of the law," Uzoukwu,

, and error is

prejudicial where the appellant can show that the error, "in light of the charge as

a whole," Turley,

, improperly "influence[d] the jury's verdict,"

Gordon v. N.Y.C. Bd. of Educ.,

(2d Cir. 2000).

Moreover, in determining whether a jury instruction was so

prejudicial as to warrant overturning the verdict, we must examine the jury

charge "in its entirety," rather than "scrutinize[] [it] strand-by-strand." Warren v.

Dwyer,

(2d Cir. 1990); accord Coquina Invs. v. TD Bank, N.A.,

, 1309 n.8 (11th Cir. 2014) ("[W]hen the instructions, taken together,

properly express the law applicable to the case, there is no error even though an

isolated clause may be inaccurate, ambiguous, incomplete or otherwise subject to

criticism." (internal quotation marks omitted)). And, though our review is de

novo, we emphasize that "a trial court has discretion in the style and wording of

jury instructions, so long as the instructions . . . do not mislead the jury as to the

44 proper legal standard." Parker v. Sony Pictures Ent., Inc.,

(2d

Cir. 2001) (internal citation and quotation marks omitted).

B. Disparate Impact

We conclude that the district court's disparate impact jury

instruction was neither erroneous nor unfairly prejudicial. The disparate impact

instruction read, in relevant part:

Plaintiffs are alleging that Defendants' practice of making STAR NINA loans has a discriminatory effect. For you to assess Plaintiffs' claim, you will consider the following.

First, Plaintiffs must establish by a preponderance of the evidence that Defendants' practice of making STAR NINA loans actually or predictably had a substantial adverse impact on African-American[] or Hispanic[] borrowers.

Second, if you find that Plaintiffs have proven the first factor, then you must decide whether Defendants have established by a preponderance of the evidence that the practice of making STAR NINA loans was necessary to achieve one or more substantial, legitimate, nondiscriminatory interests of Defendants. If you find that Defendants failed to establish that the practice was actually necessary to achieve their substantial, legitimate and nondiscriminatory interests, you must find for Plaintiffs on their discriminatory effect claim.

Third, if you find that the STAR NINA loan program was necessary to achieve one or more substantial, legitimate, nondiscriminatory interests of Defendants, then you must decide whether Plaintiffs have established by a preponderance of the evidence that Defendants' interests could have been served by another practice

45 that had a less discriminatory effect. If Plaintiffs make this showing, then you must find for Plaintiffs on their discriminatory effect claim.

I instruct you that Plaintiffs are not required to show that Defendants intended to discriminate in order to establish their claim of discriminatory effect.

J. App'x at 2461-63 (emphases added).

Emigrant asserts three challenges to the disparate impact

instruction. First, Emigrant argues that the district court's articulation of the

disparate impact burden -- that Plaintiffs had to show a "substantial adverse

impact on African-American[] or Hispanic[] borrowers" -- was erroneous and

prejudicial because it "materially altered the [disparate impact] standard" by

"fail[ing] to instruct the jury that disparate impact discrimination requires proof

that African-American borrowers . . . suffered a disproportionately adverse effect

from the STAR NINA loan program, as compared to non-African-American

borrowers." Def. Br. at 24-25 (emphasis in original). Second, Emigrant argues

that the instruction inadequately conveyed the requirement of a causal

relationship between STAR NINA loans and the adverse impact on Black and

Latino communities required to make a finding of disparate impact. Finally,

Emigrant contends that the district court erred by not instructing the jury on

Plaintiffs' "burden of proving an available alternative practice" that serves

46 Emigrant's legitimate nondiscriminatory interests. Def. Br. at 46 (emphasis in

original) (quoting MHANY Mgmt., Inc. v. Cnty. of Nassau,

(2d

Cir. 2016)).

1. Applicable Law

In the FHA context, "[a] disparate impact analysis examines a

facially-neutral policy or practice . . . for its differential impact or effect on a

particular group." Huntington Branch, NAACP v. Town of Huntington,

, 933 (2d Cir. 1988). To make out a claim of disparate impact under the FHA,

a plaintiff must first establish a prima facie case by showing "(1) the occurrence of

certain outwardly neutral practices, and (2) a significantly adverse or

disproportionate impact on persons of a particular type produced by the

defendant's facially neutral acts or practices." Regional Econ. Cmty. Action

Program, Inc. v. City of Middletown,

, 52-53 (2d Cir. 2002). This is a

"modest" burden. Francis v. Kings Park Manor, Inc.,

(2d Cir. 2021).

If the plaintiff makes this showing, the burden shifts to the defendant to "prove

that its actions furthered, in theory and in practice, a legitimate, bona fide

[business] interest and that no alternative would serve that interest with less

discriminatory effect." Huntington Branch, NAACP, 844 F.2d at 936 (citing

47 Resident Advisory Bd. v. Rizzo,

, 148-49 (3d Cir. 1977)).

Thus, "[t]he basis for a successful disparate impact claim involves a

comparison between two groups -- those affected and those unaffected by the

facially neutral policy." Tsombanidis v. West Haven Fire Dep't,

(2d Cir. 2003). This comparison, which lies at the core of disparate impact

liability, "must reveal that although neutral, the policy in question imposes a

significantly adverse or disproportionate impact on a protected group of

individuals."

(internal quotation marks omitted).

As for the causation requirement, the Supreme Court in Inclusive

Communities noted that "a disparate-impact claim that relies on a statistical

disparity must fail if the plaintiff cannot point to a defendant's policy or policies

causing that disparity." 576 U.S. at 542. Requiring a link between a defendant's

policy and disparate impact therefore ensures that "racial imbalance does not,

without more, establish a prima facie case of disparate impact." Id. (alterations

adopted and internal quotation marks omitted).

2. Application

Examining the jury charge in its entirety, we conclude that there was

no error, much less reversible error, in the district court's disparate impact

48 instructions. Despite Emigrant's arguments to the contrary, the instructions

properly apprised the jury of the elements of a disparate impact claim under

settled Supreme Court and Second Circuit precedent. 9

First, Emigrant argues that the district court should have instructed

the jury that to succeed on their disparate impact claims, Plaintiffs had to show

that Emigrant's practice of making STAR NINA loans "actually or predictably"

had a "significantly adverse or disproportionate impact" on Black or Latino

borrowers. J. App'x at 571. Emigrant contends that by omitting the word

"disproportionate" in this phrase, the district court committed reversible error.

We disagree.

We have consistently used the "significantly adverse or

disproportionate" language to describe the comparative aspect of a disparate

9 As a threshold matter, Plaintiffs argue that Emigrant has waived its challenge to two of its objections to the jury instructions -- the comparative element and robust causation. See Pl. Br. at 33. Emigrant responds that it properly preserved all challenges to the jury instructions it now raises on appeal. Based on our review of the record, we find that Emigrant properly preserved its objection to the lack of "disproportionate" language. At the charge conference, counsel for Emigrant requested that "significantly disproportionate" be added to the charge. J. App'x at 2246. After the district court instructed the jury, Emigrant stated that it had "no additional [objections] besides the ones counsel mentioned" before. Id. at 2475. With respect to Emigrant’s objections on appeal to the lack of "robust causation" and a disparity "over and above" a certain baseline with respect to causation, for the reasons set forth in the discussion, we conclude that even if Emigrant had preserved these objections, its arguments have no merit.

49 impact claim under the FHA. See, e.g., Fair Hous. in Huntington Comm. Inc. v.

Town of Huntington, N.Y.,

, 366 (2d Cir. 2003) ("[A] plaintiff must

demonstrate that an outwardly neutral practice actually or predictably has a

discriminatory effect; that is, has a significantly adverse or disproportionate impact

on minorities.") (emphasis added); Tsombanidis,

("To establish a

prima facie case under [a disparate impact] theory, the plaintiff must show[] . . . a

significantly adverse or disproportionate impact on persons of a particular type

produced by the defendant's facially neutral acts or practices.") (emphasis in

original) (internal quotation marks omitted); MHANY,

(setting

forth the standard as "a significantly adverse or disproportionate impact on

persons of a particular type").

Here, the district court instructed the jury that, to prevail on their

disparate impact claim, Plaintiffs had to prove that Emigrant's STAR NINA

lending practices "actually or predictably had a substantial adverse impact on

African-American[] or Hispanic[] borrowers," J. App'x at 2462. The district

court's "substantial adverse impact" language is not significantly different from

our Circuit's settled "significantly adverse or disproportionate impact" language.

It also mirrors, nearly word-for-word, the model instruction from the leading set

50 of pattern federal jury instructions. Compare Sand et al., Modern Federal Jury

Instructions, Civil Instruction 87-33 (a defendant violates Title VIII under a

disparate impact theory where the conduct "actually or predictably had a

substantial discriminatory impact on the protected group," where "substantial

discriminatory impact" means "a manner plainly disproportionate to how it

affects other people") with J. App'x at 2462 ("Plaintiffs must establish by a

preponderance of the evidence that Defendants' practice of making STAR NINA

loans actually or predictably had a substantial adverse impact on African

Americans or Hispanic[] borrowers.").

Emigrant contends that by leaving out the word "disproportionate,"

the district court improperly relieved Plaintiffs of their burden of proof and

permitted the jury to find for Plaintiffs simply if the jury found that Plaintiffs

suffered an adverse impact from the loans, without regard to the effect on Black

borrowers as compared to non-Black borrowers. While we agree that disparate

impact claims must apprise the jury of the requirement of a disproportionate or

disparate effect on a protected class, we disagree that the district court's failure to

use the word "disproportionate" at that particular point in the charge was

prejudicial.

51 The district court's failure to include the word "disproportionate" in

this part of the charge does not render the instruction as a whole an inaccurate

statement of law requiring reversal. Indeed, the word appears elsewhere, and

read as a whole, the charge made clear to the jury that, to find for Plaintiffs, the

jury had to compare the impact on Plaintiffs to similarly situated non-Black or

Hispanic borrowers. For instance, in describing Plaintiffs' claims at the outset of

the charge, the district court stated to the jury that

the plaintiffs . . . claim that the defendants, Emigrant Mortgage Company and Emigrant Bank violated their rights under the Fair Housing Act, Equal Credit Opportunity Act and New York City Human Rights Law by offering loans on terms that were grossly unfavorable to the borrowers and by allegedly making those loans disproportionately in African-American and Hispanic communities.

J. App'x at 2438-39 (emphasis added). 10

Accordingly, we are satisfied, based on our review of the charge in

its entirety, that the instruction sufficiently described the requirement that the

10 We also note that Emigrant included the "significantly adverse or disproportionate impact" language in its own proposed jury instructions submitted to the district court. See J. App'x at 571. The proposed jury instruction reads verbatim, "Plaintiffs must prove by a preponderance of the evidence that Defendants (1) maintained a race-neutral practice or policy, that (2) had a significantly adverse or disproportionate impact on Black borrowers. Plaintiffs must also establish that the race- neutral policy or practice is the cause of the significantly adverse or disproportionate impact." J. App'x at 571 (footnotes omitted).

52 adverse impact be disproportionate on a protected class, in this case, on Black

borrowers. We conclude that "the entire charge delivered a correct interpretation

of the law." United States v. Quattrone,

(2d Cir. 2006). By

focusing the jury on a "substantial adverse impact on African-American or

Hispanic borrowers," J. App'x at 2462, the charge captured the necessary

comparison of White borrowers to Black and Hispanic borrowers. 11 Moreover,

there was substantial evidence adduced at trial from which a jury could (and

did) find that Black borrowers were disproportionately affected by the predatory

STAR NINA loans.

Second, Emigrant argues that the district court erred in its

instruction on the causation element. According to Emigrant, the district court

should have instructed the jury on the so-called "robust" causation requirement

for disparate impact claims. For starters, nothing in Emigrant's proposed

11 While Emigrant may have wanted additional language explicitly advising the jury that it must base its finding by comparing the experience of Black borrowers to similarly situated White borrowers, there is no requirement that jury instructions be favorable to a party, see Riverwoods Chappaqua Corp. v. Marine Midland Bank, N.A.,

(2d Cir. 1994) ("While a more specific instruction might have been helpful, there is no basis for concluding that the jury was given a misleading or inaccurate impression of the law.").

53 instruction references "robust" causation, and Emigrant does not provide its

desired language in its brief before our Court. 12

Emigrant argues that the Supreme Court "has emphasized that 'a

robust causality requirement ensures that racial imbalance does not, without

more, establish a prima facie case of disparate impact and thus protects

defendants from being held liable for racial disparities they did not create.'" Def.

Br. at 47-48 (quoting Inclusive Communities, 576 U.S. at 542 (alteration adopted)).

Relying on this language in Inclusive Communities, and language in a Second

Circuit case requiring plaintiffs "to show that there existed some demonstrable

prejudicial treatment of minorities over and above that which is the inevitable

result of disparity in income," Emigrant argues that the district court failed to

instruct the jury that Emigrant could not be held liable for racial disparities

outside of its control. Boyd v. Lefrak Org.,

(2d Cir. 1975).

We disagree that the cases require "robust" causation, and we agree

with Plaintiffs that the requirement of a "causal link" between the STAR NINA

lending practices and the adverse impact on Plaintiffs "is apparent on the face of

12 As for its proposed jury instructions submitted to the district court, Emigrant requested the court instruct the jury that "Plaintiffs must also establish that the race- neutral policy or practice is the cause of the significantly adverse or disproportionate impact." J. App'x at 571.

54 the instruction." Pl. Br. at 39. Here, the district court instructed the jury to

consider whether STAR NINA loans "had a substantial adverse impact" on Black

and Hispanic communities. J. App'x at 2462 (emphasis added). A plain reading

of the instruction makes clear that Plaintiffs were required to demonstrate a

causal link between the STAR NINA loans and the discriminatory effect on

Plaintiffs. We are unconvinced by Emigrant's argument that the district court

should have instructed on a requirement of "robust causality" -- that is not the

law of disparate impact in our Circuit, or under the applicable regulations.

While we agree that a defendant may not be held liable for racial

disparities it did not cause, we think that the standard causation language in the

charge is sufficient to convey this principle, and the "robust causality" language

is not necessary. In MHANY, for example, we affirmed the district court's

finding of race-based disparate impact even though that disparity flowed from

underlying socioeconomic disparities across races, which the defendant in

MHANY did not create.

. Moreover, we read

Inclusive Communities to require, in terms of causation, that "a defendant's policy

or policies caus[ed] [a] disparity." 576 U.S. at 521. There was ample evidence

presented at trial that Emigrant's lending policies and practices caused a

55 disproportionate number of STAR NINA loans to be written to Black borrowers.

We see no reason to overturn the jury's verdict based on Emigrant's reliance on

non-binding language from the above-referenced cases. 13

Finally, Emigrant's issue with the district court's instruction on the

requirement of a "less discriminatory alternative" boils down to the district

court's omission of the word "available." See Def. Br. at 46 (arguing the

"instructions omitted entirely the requirement that any alternative actually be

available"). According to Emigrant, this failure permitted Plaintiffs to suggest

alternatives that were "mere conjecture" and not practically possible. Id. at 47.

We are not persuaded that this was error. As Plaintiffs point out,

the jury instructions track HUD regulations nearly word-for-word. Compare J.

App'x at 2462 (instructing that jury must consider whether Plaintiffs established

that Emigrant's interests "could have been served by another practice that had a

13 The HUD regulations support our conclusion as well. The relevant regulation simply requires "a plaintiff to link a specific practice to a current or predictable disparity" to prove causation. Reinstatement of HUD's Discriminatory Effects Standard,

, 19461 (Mar. 31, 2023). Thus, though we acknowledge that Inclusive Communities states that a "robust causality requirement ensures that racial imbalance does not, without more, establish a prima facie case of disparate impact," 576 U.S. at 542 (alterations adopted and internal quotation marks omitted), we do not read Inclusive Communities to set forth a new rule requiring use of the words "robust causality," and, in any event, that language is not at odds with the instructions in this case.

56 less discriminatory effect") with

(c)(3) (a plaintiff may prevail

by proving that the interests "could be served by another practice that has a less

discriminatory effect"). And implicit in the language the district court used is the

suggestion that there had to have been a viable alternative: the requirement that

Plaintiffs establish that Emigrant's interests could have been served by another

practice implies that the practice must be a viable alternative rather than mere

speculation.

Accordingly, there was no error in the disparate impact jury

instructions.

C. Disparate Treatment

The district court gave the following instruction on disparate

treatment:

In order to prevail on their claim that the Defendants intentionally discriminated in lending practices that violated the Fair Housing Act, the Equal Credit Opportunity Act, and the New York City Human Rights Law, Plaintiffs must establish that:

(1) the STAR NINA loan product was grossly unfair to the borrower; and

(2) Defendants' effort to make STAR NINA loans in certain communities was motivated [] at least in part[] by race, color, or national origin.

57 If you find that Plaintiffs have established these elements by a preponderance of the evidence, then you must find that Defendants violated the Fair Housing Act, the Equal Credit Opportunity Act, and the New York City Human Rights Law.

Plaintiffs are not required to show that Defendants acted with racial animus, which means hatred or dislike for a particular racial or ethnic group. Nor do they need to prove that race, color, or national origin was the only reason for Defendants' conduct. Rather, they are only required to show that race, color, or national origin was one motivating factor. This means that in order for Defendants to be found liable for violating the Fair Housing Act, Equal Credit Opportunity Act, and New York City Human Rights Law, race, color, or national origin need only have played some role in Defendants' conduct.

J. App'x at 2460-61.

Emigrant contends that the district court erred by instructing that a

finding of animus, defined as "hatred or dislike," was not a predicate to finding

discrimination. Emigrant also argues that the district court erred by not stating

that race had to be a "significant" factor in the decision. In our view, the

disparate treatment instruction was also legally sufficient and thus we affirm.

Disparate treatment discrimination "involves differential treatment

of similarly situated persons or groups" on account of race. Huntington Branch,

NAACP, 844 F.2d at 933. An FHA plaintiff can show disparate treatment or

intentional discrimination, however, without having to establish that the

58 defendant was motivated by hatred, dislike, or bias. See Cmty. Servs., Inc. v. Wind

Gap Mun. Auth.,

(3d Cir. 2005); see also Horizon House

Developmental Servs., Inc. v. Township of Upper Southampton,

(E.D. Pa. 1992) ("In order to prove intentional discrimination it is not necessary to

show an evil or hostile motive. It is a violation of the FHAA to discriminate even

if the motive was benign or paternalistic.").

The disparate treatment instruction here accurately stated the law

and was sufficient to instruct the jury. 14 The district court correctly stated that a

finding of ill will, hatred, or bias is not required to find intentional

discrimination. Read as a whole, the district court's instruction on animus reads

as a caution to the jury that animus, defined as hatred or ill-will, is not necessary

for a finding of disparate treatment. This is correct; a party may intentionally

discriminate without harboring hatred or ill-will toward a particular group. See

14 Like the court's instructions on disparate impact, the disparate treatment instruction also mirrored Judge Sand's model jury instruction. Compare Sand et al., Modern Federal Jury Instructions, Civil Instruction 87-35 ("To establish a claim of disparate treatment under Title VIII, a plaintiff must prove by a preponderance of the evidence that the defendant intentionally deprived the plaintiff of a right protected by Title VIII, . . . and that the defendant did so, at least in part, because he intended to discriminate against plaintiff on account of plaintiff's [race].") with J. App'x at 2460. ("Defendants' effort to make STAR NINA loans in certain communities was motivated[,] at least in part, by race, color, or national origin.").

59 Boykin v. KeyCorp,

(2d Cir. 2008) (A plaintiff does "not need to

allege discriminatory animus for her disparate treatment claim to be sufficiently

pleaded" under the FHA.).

Our sister circuits have come to similar conclusions. See, e.g.,

Bangerter v. Orem City Corp.,

(10th Cir. 1995) (A "plaintiff need

not prove the malice or discriminatory animus of a defendant to make out a case of

intentional discrimination.") (emphasis added); Cmty. Servs., Inc.,

(A defendant's "discriminatory purpose need not be malicious or invidious" to

constitute intentional discrimination.); Ferrill v. Parker Grp., Inc.,

,

472-73 & n.7 (11th Cir. 1999) ("[R]acial animus and intent to discriminate are not

synonymous. . . . In other words, ill will, enmity, [and] hostility are not

prerequisites of intentional discrimination."); accord Weiss v. La Suisse,

(2d Cir. 2005) (summary order) (same).

Emigrant's next argument -- that race must be a significant

motivating factor in a disparate treatment case -- is also unavailing. In MHANY,

we held that "if one of the motivating factors for an act was unlawful, the act

violated the FHA."

(emphasis added). Although we also said in

MHANY that "[a] plaintiff can establish a prima facie case of disparate treatment

60 'by showing that animus against the protected group was a significant factor in

the position taken by the municipal decision-makers themselves or by those to

whom the decision-makers were knowingly responsive,'"

at 606 (quoting

LeBlanc-Sternberg,

), we did not hold that this was a requirement of

a successful disparate treatment claim.

Plaintiffs were entitled to, and did, prove disparate treatment at trial

by showing that one of Emigrant's motives was to intentionally target

communities of color for the STAR NINA loan. And it is not required, as part of

a disparate treatment theory, that Emigrant harbored malicious or invidious

biases against those communities while doing so. Accordingly, because the

district court's instructions accurately captured governing law on disparate

treatment, the jury was properly instructed.

IV. The Saintils' Release

We now turn to the enforceability of the release-of-claims provision

contained within the Saintils' 2010 loan modification agreement with Emigrant.

The district court determined that the provision is unenforceable as a matter of

law because of legal prohibitions against waiver in residential mortgages and

61 because it is void as against public policy. Emigrant contends on appeal that this

was error. The release provision reads:

Release in Favor of Emigrant. Except for the obligations and rights expressly set forth and reserved in this Agreement and those portions of the Loan Documents not modified herein, Borrower . . . hereby unconditionally and irrevocably remise[s], release[s], and forever discharge[s] Emigrant, and any entities related to it . . . from any and all claims, counterclaims, actions, causes of action, suits, set- offs, costs, losses, expenses, sums of money, accounts, reckonings, debts, charges, complaints, controversies, disputes, damages, judgements, executions, promises, omissions, duties, agreements, rights, and any and all demands, obligations and liabilities, . . . arising at law or in equity, . . . which Borrower may have against [Emigrant], including . . . any and all claims that were or that could have been asserted in any legal proceeding or action; . . . and any and all claims that are relating to, concerning, or underlying the Loan.

Ex. App'x at 632-33.

We agree with the district court that the release provision of the loan

modification agreement is unenforceable because it contravenes public policy

against broad waivers of claims in mortgage transactions.

A. Standard of Review

We review "de novo [] legal issues as to . . . contract enforceability."

United States v. Twenty Miljam-350 IED Jammers,

(2d Cir. 2011).

62 B. Applicable Law

Federal law determines the validity of release of federal statutory

claims. Olin Corp. v. Consol. Aluminum Corp.,

, 15 (2d Cir. 1993). "[A]

release that is clear and unambiguous on its face and which is knowingly and

voluntarily entered into will be enforced." Pampillonia v. RJR Nabisco, Inc.,

(2d Cir. 1998). This principle extends to the release of claims under

the federal antidiscrimination laws. For instance, we have held that employees

may prospectively waive a claim of discrimination under Title VII and the ADEA

so long as it is done knowingly and voluntarily. Bormann v. AT&T Commc'ns,

Inc.,

, 402 (2d Cir. 1989). In Bormann, we adopted the Third Circuit's

"totality of the circumstances" standard to determine whether an employee in the

ADEA context waived discrimination claims "knowingly, willingly and free from

coercion." Id. at 403; see Coventry v. U.S. Steel Corp.,

, 523 (3d Cir.

1988) (citing a six-factor totality of the circumstances test).

When faced with the question of whether a contractual provision

violates federal public policy, we recognize that "[t]he term public policy is

obviously a broad one; it embraces a multitude of virtues and sins. It is clear that

public policy circumscribes agreements between private parties in order to

63 prevent the courts from becoming vehicles of discrimination." Stamford Bd. of

Educ. v. Stamford Educ. Ass'n,

(2d Cir. 1982) (internal quotation

marks and citation omitted). "Federal 'public policy' is typically found in the

Constitution, treaties, federal statutes and regulations, and court cases." Thomas

James Assocs., Inc. v. Jameson,

(2d Cir. 1996). And:

while violations of public policy must be determined through definite indications in the law of the sovereignty, courts must not be timid in voiding agreements which tend to injure the public good or contravene some established interest of society.

Stamford Bd. of Educ.,

(internal quotation marks and citation

omitted).

"It is the policy of the United States to provide, within constitutional

limitations, for fair housing throughout the United States."

. It is

also "unlawful for any creditor to discriminate against any applicant, with

respect to any credit transaction . . . on the basis of race, color, religion, national

origin, sex or marital status or age."

(a)(1).

Section 1639c(e)(3) of the Truth in Lending Act ("TILA") provides, in

pertinent part:

No waiver of statutory cause of action. No provision of any residential mortgage loan . . . and no other agreement between the consumer and the creditor relating to the residential mortgage loan

64 . . . shall be applied or interpreted so as to bar a consumer from bringing an action in an appropriate district court of the United States . . . for damages or other relief in connection with . . . any . . . Federal law.

15 U.S.C. § 1639c(e)(3).

New York contract law governs the enforceability of provisions

purporting to waive state claims. Albany Sav. Bank, FSB v. Halpin,

(2d Cir. 1997). The New York Court of Appeals "deem[s] a contractual

provision to be unenforceable where the public policy in favor of freedom of

contract is overridden by another weighty and countervailing public policy." 159

MP Corp. v. Redbridge Bedford, LLC,

, 133 (N.Y. 2019). "[B]ecause

freedom of contract is itself a strong public policy interest in New York," New

York courts "may void an agreement only after balancing the public interests

favoring invalidation of a term chosen by the parties against those served by

enforcement of the clause and concluding that the interests favoring invalidation

are stronger."

(internal citation and quotation marks omitted). New York

courts conducting this balancing have found unenforceable agreements that

contravene the protections of New York's housing laws and regulations. See

Riverside Syndicate, Inc. v. Munroe,

(N.Y. 2008); Thornton v.

Baron,

(N.Y. 2005).

65 C. Application

We agree with the district court that the release provision contained

within the Saintils' 2010 loan modification agreement violated public policy and

is therefore unenforceable as a matter of law. In reaching our conclusion, we

find support in both New York state and federal public policy against the waiver

of claims in mortgage transactions and modifications.

The federal housing laws articulate a broad and clear public policy

against discrimination in housing and housing-related transactions, like the loan

modification at issue here. See

;

(a)(1). Section

1639c(e)(3) of TILA also sets forth public policy against broad waivers of the

right to bring a cause of action in agreements concerning residential mortgage

loans. See 15 U.S.C. § 1639c(e)(3).

Emigrant argued in its opposition to Plaintiffs' Rule 50 motion that

Section 1639c(e)(3) is inapplicable to the public policy analysis for two reasons:

first, because the section deals with arbitration, and second, because the section

took effect in June 2013 and is not retroactive to the Saintils' 2010 modification

agreement. While we do not agree with either proposition, the section provides

useful guidance even if they are correct.

66 We begin with the text of the statute. Dobrova v. Holder,

(2d Cir. 2010) ("Statutory analysis necessarily begins with the plain meaning

of a law's text and, absent ambiguity, will generally end there.") (internal

quotation marks omitted and alterations adopted). Nothing in the plain text of

Section 1639c(e)(3) limits its application to arbitration. Emigrant reasons that the

title of Section 1639c(e) is "Arbitration," but the text of Section 1639c(e)(3) uses the

broad language of "any residential mortgage loan" and "no other agreement" in

setting forth its prohibition on a broad release of claims pursuant to a mortgage

agreement. See 15 U.S.C. § 1639c(e)(3); see also Pennsylvania Dep't of Corrs. v.

Yeskey,

(1998) ("The title of a statute cannot limit the plain

meaning of the text. For interpretive purposes, it is of use only when it sheds

light on some ambiguous word or phrase.") (alterations adopted). Sections

1639c(e)(1) and (2), by contrast, explicitly reference "arbitration or any other

nonjudicial procedure." Accordingly, we do not read Section 1639c(e)(3) to

govern only arbitration agreements. See also Attix v. Carrington Mortg. Servs.,

LLC,

(11th Cir. 2022) (noting that "Section 1639c(e)(3) governs

several types of agreements, including an agreement between a consumer and a

67 creditor relating to a residential mortgage loan") (internal quotation marks

omitted and alterations adopted). 15

Emigrant also contends (and it argued below) that the prohibition

on releases in mortgage contracts did not become effective until 2013 and that it

does not have retroactive effect. But as Plaintiffs point out, however, the statute

was passed in July 2010. While the statute's interpreting regulation did not

become effective until June 2013, that does not mean the statute did not take

effect when it was passed. Indeed, federal courts have disagreed about the

effective date of Section 1639c(e). Compare Weller v. HSBC Mortg. Servs., Inc.,

(D. Colo. Sept. 11, 2013) (statute became effective July

2010), with Richards v. Gibson, No. 15-CV7-LG-RHW,

, at *2 (S.D.

Miss. Mar. 4, 2015) (statute became effective June 2013).

15 Our reading is further supported by the CFPB Regulation, which states:

The statute further provides in section [1639c(e)(3)] that no covered transaction secured by a dwelling, and no related agreement between the consumer and creditor, may be applied or interpreted to bar a consumer from bringing a claim in court in connection with any alleged violation of Federal law.

, at 11387 (emphasis added).

68 We agree with Plaintiffs that Section 1639c took effect on the date of

enactment, i.e., in July 2010. There is simply nothing in the statute that suggests

otherwise. The statute does not specify a date when it took effect, nor does it

contain language to the effect that the statute becomes effective upon the

adoption of an implementing regulation. See Johnson v. United States,

(2000) (The omission of an express effective date "simply remits us to the

general rule that when a statute has no effective date, 'absent a clear direction by

Congress to the contrary, [it] takes effect on the date of its enactment.'") (quoting

Gozlon-Peretz v. United States,

(1991)).

Moreover, Emigrant did not even raise the defense until November

2014, after the date the regulation was adopted. The statutory text provides that

"no provision . . . shall be applied or interpreted" to bar borrowers from seeking

judicial resolution of their claims related to the mortgage. 15 U.S.C. § 1639c(e)(3)

(emphasis added). This language brings the focus of our inquiry to the time of

enforcement or application, and not the time the agreement was entered.

Compare id. (applying to the "appli[cation] or interpret[ation]" of waiver clauses)

with id. § 1639c(e)(1) (applying to "inclu[sion]" of waiver clauses).

69 Even assuming Section 1639c(e)(3) is limited to arbitrations or that it

does not apply retroactively, it still serves as persuasive authority. Against the

backdrop of clear federal public policy against racial discrimination in housing

and lending transactions and given Section 1639c(e)(3)'s clear expression of a

policy against waiving those claims pursuant to mortgage agreements, we agree

that the release provision in the Saintils' modification agreement undermined

federal public policy.

We conclude that New York state public policy supports this result

as well. A state banking regulation enacted prior to the Saintils' modification

agreement discourages waiver of claims by providing that "[a] servicer shall not

require a homeowner to waive legal claims and defenses as a condition of a loan

modification, reinstatement, forbearance or repayment plan." N.Y. Comp. Codes

R. & Regs, tit. 3, § 419.7(j). 16 This clear articulation of a state policy against

waiver of legal claims as a condition of loan modification supports a finding that

enforcing the loan modification agreement here would undermine New York

state public policy.

16 This provision was initially codified at 3 N.Y.C.R.R. § 419.11(h).

70 In urging reversal, Emigrant correctly notes that we have stressed

the important public policy in favor of settling disputes. See Anita Founds., Inc. v.

ILGWU Nat. Ret. Fund,

, 190 (2d Cir. 1990). But, upon our own

review of the extensive trial record in this case, we agree with Plaintiffs and the

district court that the strong public policy against coercive agreements and

against the broad waiver of discrimination claims specifically in residential

housing outweighs the policy in favor of settlement. Accordingly, we affirm the

district court's conclusion that Paragraph 19 of the Saintils' loan modification

agreement is unenforceable as against federal and state public policy.

CONCLUSION

For the reasons set forth above, we AFFIRM the district court's

judgment.

71 22-3094 Saint-Jean v. Emigrant Mortg. Co.

PARK, Circuit Judge, dissenting:

In this case, the district court sent untimely claims to an im- properly instructed jury two separate times. To affirm the resulting judgment, the majority blesses significant legal errors and upends es- tablished doctrine.

First, this suit is time-barred. The statute of limitations for claims under both the Fair Housing Act and the Equal Credit Oppor- tunity Act is two years. But Plaintiffs did not file suit until 2011 and 2014, even though their claims accrued between 2003 and 2008. Equi- table tolling can extend statutory deadlines only if Plaintiffs exercise diligence and show extraordinary circumstances. The district court found neither, and the majority’s decision to affirm based solely on fairness defies binding precedent, breaks with other circuits, and un- settles doctrine surrounding analogous statutes.

Second, even if the claims were timely, the district court’s charge allowed the jury to find disparate-impact discrimination based only on a “substantial adverse impact,” with no showing of dispro- portionate effects on minority borrowers. The majority’s approval of the erroneous jury instructions based on commentary outside the in- struction itself also stretches our precedent.

Third, the district court erred in presenting the Saintil Plaintiffs’ claims to the jury even though a provision in their loan agreement explicitly released such claims. The majority reads into federal law novel grounds for invalidating such knowing and voluntary releases based on an unprecedented public-policy analysis. I would vacate the judgment of the district court with instruc- tions to dismiss Plaintiffs’ claims as untimely. Short of that, I would vacate and remand for a new trial with a proper disparate-impact in- struction after dismissing the released claims. I respectfully dissent.

I. BACKGROUND

A. Facts

Plaintiffs claim that Emigrant discriminated against them by targeting minority homeowners with unfavorable loans. Emigrant offered STAR NINA (“no income, no asset”) loans to subprime bor- rowers with low credit scores, no proof of income, and substantial home equity between 1999 and 2008. All such loans included an 18% penalty interest rate after 30 days in default. Most borrowers were white.

Plaintiffs received and defaulted on STAR NINA loans be- tween 2004 and 2008. Plaintiffs found these loans through their self- selected, third-party brokers. Although Emigrant’s advertising alleg- edly targeted minority communities, Plaintiffs do not claim to have seen these ads. Like many borrowers, Plaintiffs chose the STAR NINA program because they wanted to borrow against their home equity but lacked the credit or verifiable income required for other loans. Plaintiffs received and signed the same set of documents as other STAR NINA borrowers. These documents included standalone sheets explaining their payment obligations and the penalty default rate. Plaintiffs, like other borrowers, either had counsel at closing or chose to disregard Emigrant’s suggestion that they retain counsel.

2 Despite receiving the same loans on the same terms as thou- sands of other borrowers, Plaintiffs sued Emigrant for “targeting them” with unfavorable loans based on their race. As relevant here, Plaintiffs alleged disparate-impact and intentional discrimination in violation of the Fair Housing Act (“FHA”), the Equal Credit Oppor- tunity Act (“ECOA”), and the New York City Human Rights Law (“NYCHRL”). The FHA and ECOA limit suits to two years from the alleged violation, and the NYCHRL has a three-year limitation.

Plaintiffs filed their claims three to ten years after closing on their loans. The original Plaintiffs, the Saint-Jeans, closed on a $370,000 loan in January 2008 and defaulted in September of that year. They claim to have noticed discrimination because of the number of black Emigrant debtors during their foreclosure proceedings in the spring of 2009. They met with a lawyer about their potential claims in July 2009. But for reasons unexplained in the record, the Saint- Jeans did not file this suit until April 2011—more than three years af- ter closing on their loan.

The other Plaintiffs joined the Saint-Jeans’ amended complaint in October 2014—three and a half years later. Ms. Commodore had closed and defaulted on her loan back in 2004—a decade before suing. The Saintils had closed and defaulted in 2006—eight years before su- ing. The Smalls had closed and defaulted in 2007—seven years before suing. Finally, Mr. Howell had closed and defaulted in 2008—six years before suing. None of these Plaintiffs offered any evidence of their efforts to investigate potential claims relating to their loans be- fore joining this suit. Howell, Commodore, and the Smalls joined only after the Saint-Jeans’ lawyer approached them in 2013.

3 B. Proceedings Below

Emigrant argued that Plaintiffs’ suit was untimely, but the dis- trict court denied its motions to dismiss, for summary judgment, and for judgment as a matter of law. Although all claims were filed more than three years after Plaintiffs closed on their loans, the district court held that the discovery rule and equitable tolling rendered the claims timely. First, the district court concluded that the FHA and ECOA “are silent as to the discovery rule,” so it applied the rule to Plaintiffs’ claims because Emigrant’s scheme of discriminatory lending “would be invisible to individual borrowers” without conferring with coun- sel.

The district court also concluded that equitable tolling applied to Plaintiffs’ claims because Emigrant’s discriminatory lending scheme was “self-concealing.” See Special App’x at 175 (“[D]iscrimi- natory mortgage lending is inherently self-concealing.”). As part of its concealment analysis, the district court pointed to Emigrant’s “overt and intentional acts,” including allegedly rushing borrowers during closing and dissuading them from bringing counsel. Id. at 70. The district court held that Plaintiffs were entitled to equitable tolling because they “had little or no basis to fully understand their claims and the nature of a systemic” scheme. Id. But the district court made no findings about Plaintiffs’ diligence in pursuing their claims or whether extraordinary circumstances prevented them from filing sooner.

Plaintiffs tried their case twice. The first jury returned a verdict for Plaintiffs, awarding $950,000 in compensatory damages, but

4 found that the Saintils’ waiver precluded their recovery. The district court rejected Emigrant’s motions for judgment as a matter of law and for a new trial. But it ordered a new trial sua sponte because the award did not “succeed at restoring Plaintiffs to their pre-STAR NINA loan positions.” The second jury awarded Plaintiffs $722,048— that is, less—in compensatory damages, including nominal damages to the Saint-Jeans and Saintils.

The district court instructed both juries that they could find dis- parate-impact liability if the loans “actually or predictably had a sub- stantial adverse impact on African-American or Hispanic borrowers.” Emigrant repeatedly objected to this instruction for failing to require a comparison to non-minority borrowers. Plaintiffs’ own proposed instructions had also included that element. The district court never- theless omitted the comparative instruction from its disparate-impact charge.

Finally, Emigrant argued that the Saintils’ 2010 loan-modifica- tion agreement barred their recovery in this suit. The Saintils had agreed to “release and forever discharge Emigrant . . . from any and all claims” arising from its prior loan. In exchange, Emigrant reduced the Saintils’ interest rate and waived some of their interest and late charges. The district court, however, denied Emigrant’s motion for summary judgment as to the Saintils’ claims without explanation. Af- ter the first jury found that the Saintils had knowingly and voluntarily released their claims, the district court declared the provision void as against public policy. Acknowledging that no particular federal law supported invalidation, the district court considered the release in the context of its earlier “historical overview” of housing discrimination.

5 Based on this history and the Saintils’ continuing struggle with pay- ments, the district court allowed the Saintils’ claims to proceed, and the second jury awarded them nominal damages.

II. TIMELINESS

It is undisputed that Plaintiffs failed to file their claims within the applicable statutes of limitations. The question is whether equita- ble tolling or the discovery rule excuses their delay. Neither does.

A. Legal Standards

The FHA provides that “[a]n aggrieved person may commence a civil action . . . not later than 2 years after the occurrence or the ter- mination of an alleged discriminatory housing practice.”

(a)(1)(A). Until 2010, the ECOA stated that no action “shall be brought later than 2 years after the date of the occurrence.” 15 U.S.C. § 1691e(f) (2009). And New York law limits claims under the NYCHRL to those “commenced within three years.”

(2).

Plaintiffs filed all of their claims outside the applicable statutes of limitations, whether they accrued at closing or upon default. 1 A

1 The parties dispute when Plaintiffs were injured. Plaintiffs say it was when Emigrant discriminated against them, while Emigrant says any injury would have occurred when it “imposed a higher default interest rate.” Even accepting that “the relevant injury is discrimination and not the loan itself,” ante at 26, that means Plaintiffs’ claims accrued at closing, when they suffered the dignitary harm of a discriminatory loan. But regardless whether the limitations period began at closing or default, Plaintiffs had a complete and present cause of action more than two years before they sued.

6 statute of limitations begins to run when a claim accrues. See Clark v. Iowa City,

(1875). Accrual is the date on which a plaintiff is first able to sue—usually the date on which the plaintiff is injured. See Green v. Brennan,

(2016); see also Bay Area Laundry & Dry Cleaning Pension Tr. Fund v. Ferbar Corp. of Cal.,

(1997) (accrual occurs “when the plaintiff has a complete and present cause of action” (quotation marks omitted)).

The district court excused the untimeliness of Plaintiff’s law- suits based on the discovery rule and equitable tolling. The discovery rule delays accrual until a plaintiff knows that he has been injured or learns of the cause of the injury. 2 See, e.g., United States v. Kubrick,

(1979) (noting that lower courts apply a discovery rule for medical-malpractice claims); Kronisch v. United States,

(2d Cir. 1998). But this rule typically applies only when a statute’s text provides for it. Rotkiske v. Klemm,

(2019) (“It is not our role to second-guess Congress’ decision to include a ‘violation oc- curs’ provision, rather than a discovery provision.”).

While the discovery rule delays the start of a limitations period, equitable tolling “pauses the running of, or ‘tolls,’ a statute of limita- tions when a litigant has pursued his rights diligently but some ex- traordinary circumstance prevents him from bringing a timely ac- tion.” Arellano v. McDonough,

(2023) (quotation marks omitted); see Menominee Indian Tribe of Wis. v. United States,

Accrual never depends, however, on a plaintiff’s knowledge of his legal rights—only of the facts that would allow him to sue. See Valez ex rel. Donely v. United States,

(2d Cir. 2008).

7 250, 255 (2016) (“[A] litigant is entitled to equitable tolling of a statute of limitations only if the litigant establishes two elements: (1) that he has been pursuing his rights diligently, and (2) that some extraordi- nary circumstance stood in his way and prevented timely filing.” (quotation marks omitted)).

“Generally, a litigant seeking equitable tolling bears the burden of establishing [these] two elements.” Pace v. DiGuglielmo,

(2005). First, a plaintiff must show a “level of diligence which could reasonably be expected in the circumstances,” Gonzalez v. Hasty,

(2d Cir. 2011) (quotation marks omitted), and that he “acted with reasonable diligence throughout the period he seeks to toll,” Smith v. McGinnis,

(2d Cir. 2000) (emphasis added). “[T]he second prong of the equitable tolling test is met only where the circumstances that caused a litigant’s delay are both ex- traordinary and beyond its control.” Menominee,

. “Ex- traordinary circumstances” thus refers “not to the uniqueness of a party’s circumstances, but rather to the severity of the obstacle im- peding compliance with a limitations period.” Harper v. Ercole,

(2d Cir. 2011).

B. Equitable Tolling

The majority misapplies equitable tolling by excusing Plaintiffs of their burden to prove the threshold elements of diligence and ex- traordinary circumstances. In so doing, it creates a new fairness- based tolling rule for discrimination claims.

To begin, the majority mischaracterizes the district court’s eq- uitable-tolling decision as factual and discretionary. Ante at 28-31.

8 But the district court committed an error of law by concluding that Plaintiffs did not need to prove diligence and extraordinary circum- stances because “[t]he law prohibits a judge from exercising her dis- cretion where these two elements are missing.” Doe v. United States,

(2d Cir. 2023).

The majority’s affirmance of the district court’s equitable-toll- ing decision disregards binding precedent. The majority says that Plaintiffs meet the requirements for equitable tolling because they “could not reasonably be expected” to learn of their cause of action through due diligence within the limitations period. Ante at 34. And “that alone would be enough” because “unfairness to a plaintiff who is not at fault” is the “core inquiry of our equitable tolling analysis.” Id. at 33. (quotation marks omitted). That analysis is novel and wrong. Plaintiffs here have shown neither diligence nor extraordinary cir- cumstances, even though these are requisite “elements, not merely factors of indeterminate or commensurable weight.” Menominee,

(quotation marks omitted). The majority’s decision to ex- cuse these deficiencies is untenable.

First, the majority simply assumes that all Plaintiffs were dili- gent. It concludes that Plaintiffs “through no lack of diligence of their own, were unaware of the facts of discrimination within the statutory period.” Ante at 34. But the district court made no findings at all about Plaintiffs’ diligence. Instead, the district court disregarded that requirement because the “nature of this type of discrimination” shows that Plaintiffs’ “ignorance of their claim was not for lack of due diligence.” Special App’x at 71. But a plaintiff’s actual “ignorance” is

9 irrelevant and a court may not assume that diligence would have been futile.

Plaintiffs did not satisfy their burden of showing diligence. The Saint-Jeans, the original Plaintiffs, closed on their loan in 2008. They claim to have noticed discrimination during their foreclosure pro- ceedings in the spring of 2009. But they offered no evidence of their diligence between contacting an attorney in 2009 and filing this suit in April 2011. In other words, the Saint-Jeans were aware of potential claims in 2009. The other Plaintiffs fare no better. The record contains no evidence of diligence by Commodore, Howell, the Smalls, or the Saintils, all of whom joined the amended complaint in October 2014— six to ten years after closing on their loans and three and a half years after the Saint-Jeans filed suit. Of these Plaintiffs, Howell, Commo- dore, and the Smalls admitted that the Saint-Jeans’ attorney contacted them in 2013 about potential claims, but they did not join this suit until October 2014. These Plaintiffs failed to pursue their claims diligently under any standard. Absent a showing of diligence, the district court had no discretion to afford equitable tolling.

Second, the majority also overlooks Plaintiffs’ failure to prove extraordinary circumstances. It simply offers the conclusory state- ment that “Plaintiffs demonstrated that their inability to discover the discriminatory practice was an extraordinary circumstance.” Ante at

10 32. 3 To the majority, Plaintiffs’ failure to recognize their claims “would be enough to support the district court’s discretionary equi- table tolling determination.”

. This is wrong.

Ignorance is not an extraordinary circumstance. Such circum- stances must be “both extraordinary and beyond [a plaintiff’s] con- trol.” Menominee,

. “Nor can equitable tolling be prem- ised on . . . ignorance of the right to bring a claim.” Watson v. United States,

(2d Cir. 2017). The district court found only that the Saint-Jeans “alleged that they did not discover the discrimi- natory scheme underlying their claims until they met with counsel within the limitations period.” Special App’x at 23. But that says nothing about the mortgage scheme being “extraordinary” or whether any external obstacle prevented the Saint-Jeans—or any of the other dilatory Plaintiffs—from exercising their rights within the limitations period.

In the employment-discrimination context, a plaintiff alleging discriminatory discharge can satisfy the extraordinary-circumstances requirement only if “it would have been impossible for a reasonably prudent person to learn that his discharge was discriminatory.” Mil- ler v. Int’l Tel. & Tel. Corp.,

, 24 (2d Cir. 1985) (emphasis added). Employment-discrimination claims thus “are not tolled or

3 The majority also offers that “the egregious nature of Emigrant’s

discriminatory lending practice . . . makes this case extraordinary.” Ante at 35. But the nature of the claim has no bearing on whether an “extraordinary circumstance stood in [Plaintiffs’] way and prevented timely filing.” Me- nominee,

(cleaned up).

11 delayed pending the employee’s realization that [his employer’s] con- duct was discriminatory unless the employee was actively misled by his employer, [or] he was prevented in some extraordinary way from exercising his rights.”

. 4 The same principle applies here.

The majority contends that its finding of extraordinary circum- stances is “further supported by the fact that Emigrant took steps to conceal the discriminatory nature of the STAR NINA loan.” Ante at 34. But its concealment analysis is unfounded. 5 “Concealment by

4 The majority asserts that “equitable tolling has always been based

on principles of fairness and equity.” Ante at 33 (citing Cerbone v. Int’l La- dies' Garment Workers' Union,

, 48-49 (2d Cir. 1985)). But Cerbone affirmed a timeliness dismissal precisely because “fairness” applied only to “relieve a plaintiff who was ‘actively misled by his employer’ or who was ‘prevented in some extraordinary way from exercising his rights.’” 768 F.2d at 49 (quoting Miller, 755 F.2d at 24). 5 The majority purports to ground its decision in “well-settled prin-

ciples of equitable tolling and fraudulent concealment.” See ante at 36. But these are distinct doctrines. See Pearl v. City of Long Beach,

(2d Cir. 2002) (“The taxonomy of tolling, in the context of avoiding a statute of limitations, includes at least three phrases: equitable tolling, fraudulent concealment of a cause of action, and equitable estoppel.”). To be sure, “[o]ur Court has used ‘equitable tolling’ to mean fraudulent concealment of a cause of action.”

. But we have been clear that fraudulent con- cealment requires the plaintiff to establish “(1) that the defendant concealed from him the existence of his cause of action, (2) that he remained in igno- rance of that cause of action until some point within four years of the com- mencement of his action, and (3) that his continuing ignorance was not at- tributable to lack of diligence on his part.” New York v. Hendrickson Bros.,

(2d Cir. 1988); see Phhhoto Inc. v. Meta Platforms, Inc., 123

12 mere silence is not enough. There must be some trick or contrivance intended to exclude suspicion and prevent inquiry.” Wood v. Carpen- ter, 101 U.S. (11 Otto) 135, 143 (1879). Plaintiffs never alleged—let alone pleaded with the specificity Rule 9 requires—that Emigrant lied to them or otherwise attempted to cover up its alleged discrimination after execution of the loan documents.

, 601-04 (2d Cir. 2024) (discussing “equitable tolling based on fraudulent concealment” under Hendrickson Bros.). In contrast, standard eq- uitable tolling applies “only if the litigant establishes two elements: (1) that he has been pursuing his rights diligently, and (2) that some extraordinary circumstance stood in his way and prevented timely filing.” Menominee,

(quotation marks omitted). The majority errs by muddling the two doctrines to conclude that any concealment entitles Plaintiffs to toll- ing—regardless of their ability to prove each element of fraudulent conceal- ment or equitable tolling. For example, the majority cites Baskin v. Hawley,

(2d Cir. 1986), for the proposition that “[e]quitable tolling may be appropriate even if there are lengthy delays in filing.” Ante at 31. But Baskin concerns a jury’s finding of fraudulent concealment, not eq- uitable tolling. See Baskin,

. 6 The majority nevertheless concludes that “there is sufficient evi-

dence that Emigrant took steps to conceal the discriminatory nature of STAR NINA loans.” Ante at 35-36. This active-concealment analysis finds no support in the record. First, the majority says that “Plaintiffs testified that Emigrant rushed them to sign stacks of documents at closing.” Id. at 35. But simply feeling “rushed” at closing cannot constitute active conceal- ment when Plaintiffs received and signed these documents before closing. Second, it claims that the default rate was “included only in a separate rider to the loan, further obfuscating the true financial impact.” Id. at 35. But this ignores the fact that all Plaintiffs signed this standalone sheet acknowledg- ing their payment obligations and the default interest rate. See, e.g., Exhibit

13 The majority’s misguided reasoning threatens to unsettle equi- table tolling doctrine beyond this case. The majority discards the pre- requisites of diligence and extraordinary circumstances for equitable tolling, replacing them with a fairness-based discovery rule for dis- crimination claims. See ante at 27 (“[T]he district court did not abuse its discretion in exercising its equitable power to toll the statute of limitations until the date when Plaintiffs knew or had reason to know of their injury—that they were victims of Emigrant’s sophisticated and systemic pattern of discriminatory lending.”). It insists that con- cealment—which was not proved here—is unnecessary for its hold- ing because “avoiding unfairness to the plaintiff is reason enough to equitably toll a statute of limitations.” Id. at 36 n.7. But that is wrong—unfairness is not sufficient for equitable tolling; courts must find both diligence and extraordinary circumstances before weighing the equities.

C. Discovery Rule

The majority also leaves open the district court’s expansive mis- understanding of the discovery rule. Id. at 27 n.6. It incorrectly sug- gests that the discovery rule applies to all federal causes of action. See id. at 25 (“Claims under the FHA, like other federal causes of action, accrue when a plaintiff knows or has reason to know of the injury that

App’x at 585. Finally, we are told that Emigrant “dissuaded Plaintiffs from bringing lawyers to closing.” Ante at 35. But the record again disproves this falsity: Emigrant encouraged the Saint-Jeans to retain counsel, see Ex- hibit App’x at 623 (“The hiring of an attorney is not required but is defi- nitely recommended.”), and Howell was, in fact, represented by outside counsel, see Joint App’x at 840-42.

14 serves as the basis for the action.” (quotation marks omitted)). But “[t]his expansive approach to the discovery rule is a bad wine of re- cent vintage.” Rotkiske,

(quotation marks omitted).

The discovery rule requires a statute-by-statute textual inquiry that neither the FHA nor ECOA can support. The applicable limita- tions provisions are clear in their focus on the occurrence—rather than discovery—of discrimination. See

(a)(1)(A) (FHA) (“not later than 2 years after the occurrence”); 15 U.S.C. § 1691e(f) (2009) (ECOA) (no “later than 2 years after the date of the occur- rence”). Whether these statutes implicitly incorporate the discovery rule is a matter of first impression in this Circuit. The Supreme Court has held that a similar statute of limitations in the Fair Debt Collection Practices Act (“FDCPA”) does not incorporate the discovery rule. See Rotkiske,

; 15 U.S.C. § 1692k(d) (“within one year from the date on which the violation occurs”). As with the FDCPA, the statutory language in the FHA and ECOA “unambiguously sets the date of the violation as the event that starts the . . . limitations pe- riod.” Rotkiske,

. Such language represents Congress’s decision to reject the discovery rule for these statutes, and we are not free to decide otherwise. 7

7 Some district courts have concluded that the FHA’s status as a tort analogue, see Curtis v. Loether,

(1974), means that the statute of limitations does not begin to run until “the wrongful act or omission re- sults in damages,” see, e.g., Fair Hous. Just. Ctr., Inc. v. JDS Dev. LLC,

, 501 (S.D.N.Y. 2020) (citing Wallace v. Kato,

, 391

15 The two other courts of appeals that have considered the ques- tion agree. The Ninth Circuit has held that the FHA is not subject to the discovery rule. See Garcia v. Brockaway,

(9th Cir. 2008) (en banc) (“Holding that each individual plaintiff has a claim until two years after he discovers the [violation] would contradict the text of the FHA, as the statute of limitations for private civil actions begins to run when the discriminatory act occurs—not when it’s en- countered or discovered.”). And the Fifth Circuit has held the same in the ECOA context. See Archer v. Nissan Motor Acceptance Corp.,

(5th Cir. 2008). In short, the FHA and ECOA make clear that the discovery rule does not apply to Plaintiffs’ claims. The dis- trict court erred by relying on the rule and the majority’s tacit en- dorsement of that error is misguided.

D. Title VII and the ADEA

The majority’s opinion also departs from precedent involving analogous antidiscrimination statutes. The Supreme Court instructs that “cases interpreting Title VII and the [Age Discrimination in Em- ployment Act (“ADEA”)] provide essential background and instruc- tion” for interpreting the FHA. Texas Dep’t of Hous. and Cmty. Affs. v. Inclusive Cmtys. Project, Inc.,

, 533 (2015). The same is true of the ECOA. See Smith v. City of Jackson,

(2005) (“[When Congress uses the same language in two statutes having similar purposes, . . . it is appropriate to presume that Congress

(2007)). These courts mistake the injury: discrimination results in damages immediately. That is why dignitary harm alone is recoverable. Here, for example, the jury awarded nominal damages to the Saint-Jeans and Saintils.

16 intended that text to have the same meaning in both statutes.”). Like the FHA and ECOA, Title VII and the ADEA provide that the limita- tions period begins with the occurrence—not the discovery—of an unlawful act. See

(e)(1). Timeliness precedents un- der these analogous statutes thus guide this Court’s reading of the FHA and ECOA.

Title VII claims accrue at the time a plaintiff learns of the dis- criminatory act, not when he realizes that the act was discriminatory. See Flaherty v. Metromail Corp.,

(2d Cir. 2000) (“It has long been settled that a claim of employment discrimination accrues for statute of limitations purposes on the date the employee learns of the employer’s discriminatory conduct.”); Miller, 755 F.2d at 24 (hold- ing the same under the ADEA). This Court applies this principle broadly: “As with all discrimination claims, plaintiffs’ claims accrued when they knew or should have known of the discriminatory ac- tion”—not the discriminatory nature of that action. Washington v. County of Rockland,

(2d Cir. 2004) (Sotomayor, J.) (em- phases added). Other circuits agree. 8

8 See, e.g., Ayala v. Shinseki,

(1st Cir. 2015); Hamilton v. 1st Source Bank,

(4th Cir. 1990) (“To the extent that notice enters the analysis, it is notice of the employer’s actions, not the notice of a discriminatory effect or motivation, that establishes the commencement of the pertinent filing period.”); Pacheco v. Rice,

(5th Cir. 1992) (“To allow plaintiffs to raise employment discrimination claims whenever they begin to suspect that their employers had illicit motives would effec- tively eviscerate the time limits prescribed for filing such complaints.”);

17 The majority barely acknowledges our Title VII and ADEA ju- risprudence. Not only does this raise doubts about the soundness of the majority’s legal reasoning, it leaves open the possibility that its new fairness-based approach to equitable tolling will reach far be- yond this case.

III. JURY INSTRUCTIONS

The majority also overlooks the district court’s erroneous jury instruction on disparate-impact liability, which required no finding that Emigrant harmed Plaintiffs more than borrowers of other races. The district court’s instruction allowed the jury to find disparate im- pact based on only a “substantial adverse impact” on minority bor- rowers, not a “disproportionately” adverse impact. This contradicted the plain meaning of “disparate impact” and binding precedent. The legal error is obvious: for an impact to be “disparate,” a comparison is necessary. The district court’s general-verdict form thus left open

Amini v. Oberlin Coll.,

(6th Cir. 2001) (“[T]he starting date for the . . . limitations period is when the plaintiff learns of the employ- ment decision itself, not when the plaintiff learns that the employment de- cision may have been discriminatorily motivated.”); Thelen v. Marc’s Big Boy Corp.,

(7th Cir. 1995) (“A plaintiff’s action accrues when he discovers that he has been injured, not when he determines that the injury was unlawful.”); Humphrey v. Eureka Gardens Pub. Facility Bd.,

(8th Cir. 2018); Lukovsky v. City & County of San Francisco,

(9th Cir. 2008) (“Plaintiffs argue that their claims did not ac- crue until they knew both that they were not being hired and of the Defend- ants’ alleged discriminatory intent. . . . [But] the claim accrued when the plaintiffs received notice they would not be hired.”); Hulsey v. Kmart, Inc.,

(10th Cir. 1994).

18 the possibility that the jury imposed liability without considering, much less finding, any disproportionate effects on minority borrow- ers. This error requires vacatur.

A. Error

We review a district court’s jury instructions de novo. “An in- struction is erroneous if it misleads the jury as to the correct legal standard or does not adequately inform the jury of the law.” Ashley v. City of New York,

(2d Cir. 2021) (quotation marks omitted).

The district court’s instruction here misstated the law. The Su- preme Court has made clear that a disparate-impact claim requires a disproportionate effect on minorities. See Inclusive Cmtys., 576 U.S. at 524 (“[A] plaintiff bringing a disparate-impact claim challenges prac- tices that have a ‘disproportionately adverse effect on minorities’ and are otherwise unjustified by a legitimate rationale.” (quoting Ricci v. DeStefano,

(2009)). 9 The district court thus erred by

9 Although Inclusive Communities, 576 U.S. at 545, held “that dispar-

ate-impact claims are cognizable under the Fair Housing Act,” we have yet to decide whether the ECOA provides for such liability. Appellants do not raise this issue on appeal, so we should “express no opinion about whether a disparate impact claim can be pursued under ECOA.” Garcia v. Johanns,

, 633 n.9 (D.C. Cir. 2006); see

(“Both Title VII and the Age Discrimination in Employment Act (ADEA) prohibit actions that ‘otherwise adversely affect’ a protected individual. The Supreme Court has held that this language gives rise to a cause of action for disparate impact discrimi- nation under Title VII and the ADEA. ECOA contains no such language.”

19 instructing the jury that it could find liability based on a “significantly adverse impact” even if it found that non-minority borrowers suf- fered the same impact.

The majority, however, concludes that the district court’s in- struction was close enough. That’s because a “substantial adverse im- pact” is “not significantly different” from the correct instruction, which would require a “significantly adverse or disproportionate” impact. Ante at 50. But even that instruction would not be proper.10 The disjunctive phrasing would permit liability for conduct that has only a “significantly adverse” impact, but not a disproportionate one. Such an instruction strays from Inclusive Communities. The majority ignores that precedent and misreads others.

For example, the majority notes that this Court used the phrase “significantly adverse or disproportionate” effect to describe

(citations omitted)); cf. Golden v. City of Columbus,

, 963 n.11 (6th Cir. 2005) (“Neither the Supreme Court nor this Court have previously de- cided whether disparate impact claims are permissible under ECOA. How- ever, it appears that they are.”). 10 The majority maintains that “Emigrant included the ‘significantly adverse or disproportionate impact’ language in its own proposed jury in- structions submitted to the district court.” Ante at 52 n.10. But Emigrant’s proposed instructions included the critical element of disproportionate im- pact: “In order to show a ‘significantly advance or disproportionate im- pact,’ Plaintiffs must show that a race-neutral practice or policy actually or predictably resulted in discrimination against Black borrowers. If Plaintiffs only raise an inference of discrimination, they have not proven a ‘signifi- cantly adverse or disproportionate impact.’” Joint App’x at 571 (footnote omitted) (emphasis added).

20 disparate-impact liability in Regional Economic Community Action Pro- gram, Inc. v. City of Middletown (“RECAP”),

, 53 (2d Cir. 2002). But RECAP did not concern the proper wording of a jury in- struction. We have “cautioned that trial judges should not import un- critically language . . . developed by appellate courts for use by judges” into jury charges. Gordon v. New York City Bd. of Educ.,

(2d Cir. 2000) (cleaned up). Even if we could lift jury instructions from language in appellate opinions, RECAP stated that a “disparate impact analysis examines a facially-neutral policy or practice . . . for its differential impact or effect on a particular group.” RECAP, 294 F.3d at 52 (quotation marks omitted) (emphasis added). Indeed, we concluded that the plaintiff had failed to allege a dispar- ate-impact claim because “[n]o comparison of the act’s disparate im- pact on different groups of people” in that case was “possible.” Id. at 53. The majority similarly overlooks comparisons among groups in our other disparate-impact precedents. 11 These cases make clear that disparate impact requires a disproportionate effect on a particular

11 See Tsombanidis v. West Haven Fire Dept.,

(2d Cir. 2003) (“The basis for a successful disparate impact claim involves a com- parison between two groups.”); MHANY Mgmt., Inc. v. County of Nassau,

(2d Cir. 2016) (“The district court concluded that Plaintiffs had established a prima facie case of disparate impact, finding that Garden City’s [zoning decision] had a significant disparate impact on minorities be- cause it largely eliminated the potential for the type of housing that minor- ities were disproportionately likely to need.” (quotation marks omitted)). See also Huntington Branch, NAACP v. Town of Huntington,

, 933 (2d Cir. 1988) (“A disparate impact analysis examines a facially-neutral pol- icy or practice . . . for its differential impact or effect on a particular group.” (emphasis added)).

21 group of people. 12 A jury instruction that omits this requirement fails to inform the jury of the law. 13

B. Prejudice

The jury here returned only a general-verdict form, so the error in the jury instruction means that “a new trial will be required, for there is no way to know that the invalid claim . . . was not the sole basis for the verdict.” United N.Y. & N.J. Sandy Hook Pilots Ass’n v.

12 Our sister circuits take a similar approach. The Ninth Circuit has required a “significant, adverse, and disproportionate effect on a protected class” to show disparate impact under the FHA. Sw. Fair Hous. Council, Inc. v. Maricopa Domestic Water Improvement Dist.,

(9th Cir. 2021) (emphasis added). The Fourth and Fifth Circuits agree that “disproportion- ate”—not merely “substantial”—adverse effects are necessary. See Reyes v. Waples Mobile Home Park Ltd.,

(4th Cir. 2018); Inclusive Cmtys. Project, Inc. v. Lincoln Prop. Co.,

(5th Cir. 2019); see also

(Davis, J., concurring in part and dissenting in part). 13The majority says that the district court’s instruction “mirrors, nearly word-for-word, the model instruction” from Sand’s Modern Federal Jury Instructions. See ante at 50-51. But those model instructions plainly state that disparate-impact liability requires a “substantial discriminatory impact”—meaning an impact “plainly disproportionate to how it affects other people.” Id. at 51. The district court’s phrase of choice—“substantial adverse impact”—does not include the comparative element of “discrimi- natory” or “disproportionate.” The ordinary meaning of “adverse” is merely “unfavorable,” “harmful,” or “opposed to one’s interests.” See Ad- verse, Merriam-Webster’s Collegiate Dictionary 19 (11th ed. 2020). Absent any clarifying instruction from the district court, we must presume that the jury gave the term its ordinary meaning. See, e.g., Victor v. Nebraska,

(1994) (evaluating jury instructions according to their ordinary meaning).

22 Halecki,

(1959); see Hathaway v. Coughlin,

(2d Cir. 1996) (“[W]here jury instructions create an erroneous impression regarding the standard of liability, it is not harmless error because it goes directly to plaintiff’s claim, and a new trial is war- ranted.” (citing Hendricks v. Coughlin,

(2d Cir. 1991)). The jury may have returned a verdict for Plaintiffs because it found that Emigrant’s conduct had an adverse—but not dispropor- tionate—impact on minority borrowers. As a result, the error re- quires remand for a new trial.

The majority asserts that Emigrant was not prejudiced by any error in the disparate-impact instruction because “read as a whole, the charge made clear to the jury that, to find for Plaintiffs, the jury had to compare the impact on Plaintiffs to similarly situated non-Black or Hispanic borrowers.” Ante at 52 (emphasis added). In support, the majority points to a sentence near the beginning of the jury colloquy in which the district court said that “the plaintiffs . . . claim that the defendants . . . violated their rights . . . by offering loans on terms that were grossly unfavorable to the borrowers and by allegedly making those loans disproportionately to African-American and Hispanic communities.” Joint App’x at 2438-39; ante at 52. But that was merely the judge’s comment on Plaintiffs’ claims, not a jury instruction. See United States v. Tracy,

(2d Cir. 1993) (distinguishing “the court’s own comments” from its charge in a jury colloquy).

“Disparate impact” requires a jury to find an impact that is dis- parate. But the district court here instructed the jury that a “substan- tial adverse impact” is enough. That instruction was wrong, contrary

23 to precedent, and prejudicial. The error cannot be saved by a separate comment describing Plaintiffs’ claims preceding the jury instruction.

IV. SAINTILS’ RELEASE

Finally, a claims-release provision in the Saintil Plaintiffs’ rene- gotiated loan agreement barred their recovery in this suit. The first jury found as much in denying damages to the Saintils because they had “knowingly and voluntarily agreed to release their claims against Emigrant.” The district court nevertheless voided the Saintils’ release as against public policy based on its “historical overview of mortgage lending in the United States.” Special App’x at 177-78. But there is no legal basis for invalidating the Saintils’ waiver.

The majority points to the federal policy of ending housing dis- crimination and a bar on waiving claims under the Truth in Lending Act, a statute not at issue here. See ante at 64-65. But none of the dis- trict court, Plaintiffs, or the panel majority identifies a single case in which federal fair-housing policy trumped the longstanding policy of respecting settlements. See, e.g., ABKCO Music, Inc. v. Harrisongs Mu- sic, Ltd.,

, 997 (2d Cir. 1983) (collecting cases). The district court thus erred by disregarding the Saintils’ knowing and voluntary release of their federal claims. It then presented those waived claims to the second jury, which in turn awarded only nominal damages. However small, that award was based on yet another error of law.

* * *

In conclusion, the claims underlying this suit were time-barred when the Saint-Jean Plaintiffs filed in 2011 and when the other

24 Plaintiffs joined in 2014. The majority excuses this untimeliness by turning equitable tolling into a fairness-based discovery rule in dis- crimination cases. In doing so, the majority defies Supreme Court and Second Circuit precedent and raises troubling questions about the reach of its holding to Title VII and ADEA cases. On top of that, the majority affirms a jury instruction for disparate-impact liability that lacked any comparative component. Finally, the majority takes the unprecedented step of voiding a contractual release provision on fed- eral housing policy grounds. All three of these conclusions are wrong. I respectfully dissent.

25