20-1666 (en banc) Farhane v. United States

United States Court of Appeals For the Second Circuit

August Term, 2023

Argued en banc: May 22, 2024 Decided: October 31, 2024

No. 20-1666

ABDERRAHMANE FARHANE,

Petitioner-Appellant,

v.

UNITED STATES OF AMERICA,

Respondent-Appellee.

Appeal from the United States District Court for the Southern District of New York No. 18-cv-11973, Loretta A. Preska, Judge.

Before: WALKER, WESLEY, CARNEY, SULLIVAN, PARK, NARDINI, MENASHI, LEE, ROBINSON, PÉREZ, NATHAN, MERRIAM, and KAHN, Circuit Judges. *

CARNEY, J., filed the majority opinion in which WESLEY, LEE, ROBINSON, PÉREZ, NATHAN, MERRIAM, and KAHN, JJ., joined.

∗ Judge Walker, Judge Wesley, and Judge Carney, who are senior judges, participated in this rehearing en banc pursuant to

(c)(1) and

(c). WESLEY, J., filed a concurring opinion in which LEE, ROBINSON, NATHAN, and MERRIAM, JJ., joined.

PÉREZ, J., filed a concurring opinion in which LEE, ROBINSON, NATHAN, and MERRIAM, JJ., joined.

WALKER, J., filed a dissenting opinion in which SULLIVAN, PARK, and MENASHI, JJ., joined in full, and NARDINI, J., joined as to Part I.

PARK, J., filed a dissenting opinion in which SULLIVAN, NARDINI, and MENASHI, JJ., joined.

NARDINI, J., filed a dissenting opinion.

Over a decade ago, the Supreme Court ruled that the Sixth Amendment requires criminal defense counsel to advise her client whether a guilty plea carries a risk of deportation. Today we hold that the Sixth Amendment entitles a naturalized U.S. citizen facing the risk of deportation following denaturalization to no less protection than a noncitizen facing the risk of deportation. A risk of denaturalization cannot be decoupled from a risk of deportation. A naturalized U.S. citizen considering whether to enter a guilty plea has a constitutional right to be advised by counsel that he may lose his citizenship and be banished from the country as a result.

Petitioner-Appellant Abderrahmane Farhane came to the United States almost thirty years ago, settling with his family in Brooklyn. He became a naturalized U.S. citizen in 2002. In 2006, he pleaded guilty on advice of counsel to serious crimes and served over eleven years in federal prison as a result. The government filed a complaint for denaturalization against him in 2018, over a year after his release from prison, based on conduct admitted to in his plea. Upon learning of the government’s intent to denaturalize him, Farhane moved to vacate his plea, conviction, and sentence under

. He asserted an ineffective assistance of counsel claim, alleging that his trial counsel never advised him of the risk of denaturalization and thus, removal, and that he would not have agreed to plead guilty had he known of this risk. The District Court (Preska, J.) denied his motion. On appeal, a divided panel affirmed the denial.

In these en banc proceedings, we VACATE the decision of the prior panel Majority; VACATE the judgment of the District Court denying habeas relief to Farhane; and REMAND the case to allow the District Court to reevaluate Farhane’s Strickland claim consistent with this opinion.

2 RAMZI KASSEM (Naz Ahmad, Mudassar Toppa, CLEAR Project, Main Street Legal Services, Inc., CUNY School of Law, Long Island City, NY; Alan E. Schoenfeld, Emily Barnet, Sandra Redivo, Trena M. Riley, Dylan Reichman, Wilmer Cutler Pickering Hale and Dorr LLP, New York, NY; Thad Eagles, Jeremy W. Brinster, Wilmer Cutler Pickering Hale and Dorr LLP, Washington, DC; Asma S. Jaber, Wilmer Cutler Pickering Hale and Dorr LLP, Boston, MA, on the brief), for Petitioner-Appellant Abderrahmane Farhane.

KARL METZNER (Jun Xiang, Hagan Scotten, on the brief), for Damian Williams, United States Attorney for the Southern District of New York, New York, NY, for Respondent-Appellee United States of America.

John C. Yang, Niyati Shah, Marita Etcubañez, Asian Americans Advancing Justice, Washington, DC; Andrew Z. Michaelson, Mateo de la Torre, King & Spalding LLP, New York, NY, for Amicus Curiae Asian Americans Advancing Justice in Support of Petitioner- Appellant Abderrahmane Farhane.

Andrew Wachtenheim, Nabilah Siddiquee, Immigrant Defense Project, New York, NY, for Amicus Curiae Immigrant Defense Project in Support of Petitioner- Appellant Abderrahmane Farhane.

Joel B. Rudin, Matthew A. Wasserman, National Association of Criminal Defense Lawyers, New York, NY; Richard D. Willstatter, New York State Association of Criminal Defense Lawyers, White Plains, NY; S. Isaac Wheeler, Federal Defenders of New York, Inc., New York, NY, for Amici Curiae National Association of Criminal Defense Lawyers, New York State Association of Criminal Defense Lawyers, Federal Defenders of New York, Inc., Federal Public Defender’s Office for the Western District of New

3 York, Office of the Federal Public Defender for the District of Connecticut, Office of the Federal Public Defender for the District of Vermont, and Office of the Public Defender for the Northern District of New York in Support of Petitioner-Appellant Abderrahmane Farhane.

Alyssa Barnard-Yanni, Andrew D. Silverman, Daniel A. Rubens, Orrick, Herrington & Sutcliffe LLP, New York, NY; Lauren A. Weber, Orrick, Herrington & Sutcliffe LLP, Seattle, WA, for Amici Curiae Professors of Criminal Law, Criminal Procedure, and Immigration Law in Support of Petitioner-Appellant Abderrahmane Farhane.

CARNEY, Circuit Judge:

In these en banc proceedings, we consider whether a naturalized United States

citizen has a Sixth Amendment right to be advised by counsel that he may be

denaturalized and deported as a result of his entry of a guilty plea. In its 2010 decision

in Padilla v. Kentucky,

, the Supreme Court ruled that the Sixth Amendment

requires criminal defense counsel to advise her client of a risk of deportation associated

with such a plea. Today we hold that the Sixth Amendment entitles a naturalized U.S.

citizen facing the risk of deportation following denaturalization to no less protection

than a noncitizen facing the risk of deportation. Deportation following denaturalization

proceedings is a severe, adverse immigration consequence that is covered by Padilla. To

provide constitutionally effective advice, counsel must address the risk of this

consequence with her naturalized citizen client before he decides to enter a guilty plea.

Like most courts, ours has generally drawn a distinction between “direct” and

“collateral” consequences of a conviction to provide a useful boundary between what

the Sixth Amendment requires counsel to address (the possible sentence of

incarceration, for example, is a “direct” consequence) and what counsel need not

address (the loss of an occupational license, on the other hand, is likely “collateral”). In

4 the past, we have treated adverse immigration consequences including deportation as

collateral, being beyond both the power of the sentencing court itself to impose and

perhaps beyond a criminal defense counsel’s presumptive area of expertise. See Michel

v. United States,

, 465–66 (2d Cir. 1974).

In Padilla, however, the Supreme Court disavowed that binary framework as a

tool for assessing counsel’s obligations when permanent removal from the country was

a possible consequence. That immigration result was too severe, its imposition too

closely connected with the criminal process, and its effect upon families too drastic, the

Court reasoned, to be lumped categorically together with other “collateral”

consequences of a guilty plea. Accordingly, the Court determined that counsel has a

constitutional duty to advise her client on this subject.

Accepting this premise in Padilla’s case, and then applying the principles

established in Strickland v. Washington,

(1984), to govern ineffective

assistance claims, the Court found that counsel fell below objective standards of

competence by neglecting to alert Padilla that his plea agreement with the government

would open the door to his removal from the country. That risk, the Court said, could

not constitutionally be ignored by counsel. Nor was affirmative misadvice necessary to

raise the constitutional concern, it stressed: silence on the subject breached the Sixth

Amendment duty to advise. The Court explained that the prospect of removal from the

country may be more important to a defendant than time served behind bars, and

counsel bears a duty at least to call to her client’s attention the risk of such serious

adverse immigration consequences.

5 In the government’s 2023 fiscal year, 878,500 persons became U.S. citizens. In the

decade from 2012 to 2023, the United States naturalized more than 7.7 million persons. 1

Those individuals earned that cherished status and swore allegiance to the Constitution

after a lengthy, multi-layered process that calls for determination and commitment.

More will do so this year. If it is later determined that a naturalized citizen obtained his

citizenship by fraud or false statements, the law provides a mechanism by which that

citizenship may be revoked. See

. But the Constitution has long been

understood to protect naturalized citizens—even those who run afoul of our criminal

laws—just as it does U.S.-born citizens, affording each the effective advice of counsel

when they are caught up in the criminal process. As the Court long ago observed in

Schneider v. Rusk, aside from the Constitution’s requirement that only natural-born

citizens are eligible to be President, a naturalized citizen “possess[es] all the rights of a

native citizen, and stand[s], in the view of the constitution, on the footing of a native.”

(1964) (internal quotation marks omitted).

We now reaffirm that understanding and decide that, applying Padilla and its

clarifying companion case, Chaidez v. United States,

(2013), criminal defense

attorneys have a Sixth Amendment obligation to inquire into and advise a naturalized

citizen client of any risk of deportation following denaturalization proceedings that

accompany the client’s guilty plea, just as they do for a deportation risk facing a

noncitizen client. If the plea carries such a risk and counsel has failed to address it, then

the two-pronged Strickland test applies: we ask whether counsel’s performance fell

below an objective standard of reasonableness at the time, and whether the client has

shown that he was prejudiced by counsel’s deficient performance. The outcome of a

1U.S. Citizenship & Immigr. Servs., Naturalization Statistics, https://www.uscis.gov/citizenship- resource-center/naturalization-statistics (last updated May 9, 2024).

6 Sixth Amendment challenge to the conviction will hinge on the answers to those

questions.

Petitioner-Appellant Abderrahmane Farhane came to the United States from

Morocco almost thirty years ago, in 1995. He was naturalized in 2002. He has six

children, two of whom were naturalized through him and two of whom are U.S.

citizens through their birth here. In 2006, Farhane pleaded guilty on advice of counsel to

serious crimes charged in the Southern District of New York. He served more than

eleven years in federal prison as a result.

Within a year of his release from prison, however, the government began

proceedings to denaturalize him based on conduct admitted to in his plea agreement—

proceedings acknowledged to be a precursor to deportation. Farhane then brought a

challenge by motion for habeas relief under

. He asserted that, contrary

to his Sixth Amendment rights, his counsel never advised him of the risk of

denaturalization and deportation, and that he would not have agreed to enter the plea

had he known that he might be banished from this country and separated from his

family as a result. The District Court denied his motion. On appeal, a divided prior

panel affirmed the denial.

On review, we VACATE the decision of the prior panel Majority; VACATE the

judgment of the District Court denying habeas relief to Farhane; and REMAND the case

to allow the District Court to address the Strickland ineffective assistance questions in

the first instance, in the context of the Sixth Amendment understandings set forth here.

7 BACKGROUND

I. Factual and procedural background

A. Farhane’s immigration history

Abderrahmane Farhane immigrated to this country from Morocco in 1995, at

about age 40, after visiting the country on three occasions beginning in the late 1980s.

He and his wife and four Moroccan-born children settled in Brooklyn, where he ran a

small bookstore.

In March 2001, he applied to become a naturalized U.S. citizen by completing

and submitting the form provided for that purpose by the Immigration and

Naturalization Service (“INS”). Among the many questions that form requires the

applicant to answer is one that is particularly broad: “Have you ever . . . knowingly

committed any crime for which you have not been arrested?” Farhane checked the box

marked “No” in response. About one year later, in March 2002, he participated in an in-

person interview with an INS official as his next step toward naturalization. There, he

was asked and repeated under oath his negative answer to that question. Then, on April

19, 2002, after swearing to the veracity of that and other answers given on the form a

third time, he took the oath of citizenship and became a naturalized U.S. citizen.

In time, his answers in March 2001 and in March and April 2002 to that sweeping

question became one of two bases for the government’s denaturalization complaint

against him: that by answering “No,” he purportedly concealed material facts related to

his application; and that the material facts that the “No” concealed further

demonstrated, the government alleged, that he was not a person of good moral

character. On these two bases, the government later charged, a federal district court was

bound to revoke his naturalization.

8 B. Farhane’s criminal conduct, guilty plea, and sentence

In 2005, the U.S. Attorney for the Southern District of New York charged Farhane

by complaint with providing false statements in violation of

. In 2006, a

superseding indictment that now named other codefendants (Tarik Shah, Rafiq Sabir,

and Mahmud Faruq Brent) charged Farhane with two criminal counts: in Count 5,

conspiracy to provide material support for terrorism, see 18 U.S.C. § 2339A; and in

Count 6, making false statements to law enforcement officers in connection with a

terrorism investigation, see id. § 1001(a)(2). 2 The superseding indictment further alleged

that from about November 2001 through June 2005, Farhane and Shah “agreed with

each other to assist another individual to transfer money from the United States to

locations overseas to purchase weapons and communications equipment for jihadists in

Afghanistan and Chechnya, and attempted to hide the nature of that assistance from

United States authorities.” App’x at 75–76. The “[]other individual” referred to in this

count was a confidential informant for the FBI. The record contains no evidence that

any money was transferred overseas in connection with the charged conspiracy. Count

5 carried a statutory maximum sentence of fifteen years.

The allegations of Count 6 rested primarily on a conversation that took place in

2005 between Farhane and an FBI agent. In that conversation, Farhane denied having a

conversation with the confidential informant about sending money overseas to jihadists

or meeting the informant in person. He falsely explained that he had provided a phone

number to the informant merely to help that person send money overseas by wire

transfer at lower cost. (The conversation in question was one referred to in the

2The superseding indictment appears to have been the first indictment obtained against Farhane. But because it was entered on the District Court’s docket as a “superseding indictment,” App’x at 46, we (like the parties) refer to it in this way as well.

9 conspiracy count, Count 5.) Count 6 carried a statutory maximum sentence of eight

years.

On November 9, 2006, on advice of counsel, Farhane entered a plea agreement in

which he pleaded guilty to one count of conspiracy to commit money laundering, see

, and one count of making materially false statements involving

international terrorism, see

§ 1001(a)(2). The agreement stipulated a maximum

sentence of 13 years—the combined statutory maximum for the counts of conviction. In

pleading guilty, he also waived his right to appeal or attack collaterally any sentence at

or below the stipulated maximum. He had no known prior criminal record.

At his change-of-plea hearing, Farhane stated that he was guilty of the

conspiracy charge because he had “agreed with others in the month[s] of November

and December of 2001 to transfer money for mujahideen in Afghanistan and

Chechnya.” App’x at 194. As to the false statements charge, he acknowledged that in

2005 he gave false answers when questioned as part of an FBI terrorism investigation.

In April 2007, the District Court (Preska, J.) entered judgment against Farhane

and sentenced him to the statutory maximum term of 13 years’ incarceration and a two-

year term of supervised release. Farhane timely filed a notice of appeal and expressed to

appellate counsel his desire to raise ineffective assistance and other claims related to the

District Court proceedings. But in 2008, his appointed appellate counsel moved under

Anders v. California,

(1967), to withdraw the appeal, citing the waiver in

Farhane’s plea agreement and other considerations. Three years later, in February 2011,

after the trial and appeal of Farhane’s codefendant Sabir, we granted counsel’s Anders

10 motion and dismissed Farhane’s appeal in a brief order. 3 See United States v. Farhane

(Sabir),

, 132 n.2 (2d Cir. 2011).

Meanwhile, Farhane was incarcerated for eleven years. In May 2017, after his

release from prison, he began his two-year term of supervised release.

C. The 2018 denaturalization proceedings

In July 2018, just over a year after his release from prison and into his term of

supervised release, the government advised Farhane that it had begun a civil

denaturalization action against him. Proceeding under

(a), the

government alleged in its civil complaint that Farhane had obtained his naturalization

unlawfully. The complaint contained two pivotal assertions, both based on Farhane’s

2006 guilty plea. First, it asserted that his guilty plea conclusively showed that Farhane

knowingly committed crimes pre-dating his April 2002 naturalization. Farhane’s

admission in 2006 to those pre-April 2002 crimes, in turn, established that he gave false

testimony under oath in his March 2002 INS interview. Thus, he had procured his

naturalization by concealing a material fact: that he had committed a crime for which he

had not been arrested.

Second, it asserted that this false testimony and his pre-April 2002 criminal acts

themselves conclusively established that he could not show in April 2002 that he was a

“person of good moral character,” as he had to, to naturalize then. App’x at 315–17. No

more was needed to require the court to enter the requested order of denaturalization:

the government reminded the District Court that it enjoyed no equitable discretion to

second-guess these facts and deny entry of a judgment of denaturalization against

3Counsel’s Anders motion in 2008 pre-dated the 2010 Padilla decision, and so did not address its import; our 2011 order made no mention of Padilla.

11 Farhane as a naturalized citizen whose “citizenship was procured illegally or by willful

misrepresentation of material facts.”

.

D. Farhane’s § 2255 motion

Four months later, in December 2018, Farhane moved under

to

vacate his plea and sentence. In his accompanying affidavit, he averred that his trial

counsel did not advise him that his plea carried a risk of denaturalization and

deportation. Farhane asserted:

I would not have entered a guilty plea if [counsel] had told me that I could lose my U.S. citizenship as a result. I would not have entered a guilty plea if [counsel] had told me that I could face deportation as a result. I would not have entered a guilty plea if [counsel] had told me that my children could lose their U.S. citizenship as a result.

App’x at 299 (para. numbering omitted). He alleged that counsel knew he was a

naturalized citizen and that counsel had so advised the court at Farhane’s first detention

hearing, in 2005. Yet counsel never alerted him to the possibility of denaturalization and

deportation raised by his plea or advised him with respect to that risk. On this basis,

Farhane submitted, his plea was invalid.

The government opposed Farhane’s motion. It maintained that the Sixth

Amendment does not require counsel to advise about “collateral consequences” such as

denaturalization. As to his risk of deportation, it responded, “That argument is

meritless because, at the time of the plea, Farhane was a citizen and therefore could not

be deported.” S.D.N.Y. No. 1:18-cv-11973 Doc. 22, at 14 n.9. It further suggested that the

criminal case outcome would have been the same because, even had he been advised of

the risk of denaturalization and deportation, he would have chosen to take the 13-year

term of imprisonment over the 23-year term to which he was exposed absent the plea

agreement. Farhane countered that the stakes to him were such that he would have

chosen to go to trial and take the chance of receiving a longer sentence.

12 The District Court denied Farhane’s § 2255 motion. It explained that it saw

nothing in the record to suggest that his lawyer “should have known” about his

exposure to denaturalization. United States v. Farhane, No. 18 Civ. 11973,

, at *2 (S.D.N.Y. Mar. 31, 2020). Padilla had been handed down in the interim

between Farhane’s sentencing and his § 2255 motion (and before the judgment in the

criminal case against him became final). But Padilla had no bearing on Farhane’s case,

the Court reasoned, because Farhane’s “conviction itself did not give rise to an

imminent risk of deportation, as was the case in Padilla.” Id. Instead, Farhane’s risk of

denaturalization “stemmed from his misrepresentations . . . and his having illegally

procured naturalization.” Id. And in any event, although Farhane’s counsel knew

Farhane was naturalized, he had “no basis for suspecting that the guilty plea could have

immigration consequences[.]” Id.

In August 2023, a divided panel of this Court affirmed the District Court’s denial

of relief, concluding that civil denaturalization is a collateral and not a direct

consequence of a conviction, and that the Sixth Amendment therefore imposes no

obligation on attorneys to warn of that risk. See Farhane v. United States,

, 126

(2d Cir. 2023). In 2024, we ordered this rehearing en banc.

The denaturalization proceedings brought against Farhane have been stayed

pending resolution of his appeal in our Court.

DISCUSSION

We first consider whether, under Padilla, the Sixth Amendment requires a

criminal defense attorney to inform a naturalized citizen client whether his guilty plea

carries a risk of denaturalization and deportation. Concluding that it does, we then turn

to Farhane’s Strickland claim and decide to remand his case to the District Court. On

remand, the District Court should allow the parties to develop the record further and

13 conduct an evidentiary hearing as needed. It should then reevaluate, in light of the Sixth

Amendment understandings set forth here, whether Farhane has established two

propositions: that his lawyer’s conduct was objectively unreasonable; and that he was

prejudiced as a result.

Whether defense counsel rendered ineffective assistance presents a mixed

question of law and fact that we review de novo. Doe v. United States,

(2d Cir. 2019).

I. The Sixth Amendment requires an attorney for a naturalized U.S. citizen to provide advice about the risk of denaturalization and deportation flowing from the citizen’s guilty plea.

The question we confront is whether Farhane’s Sixth Amendment right to

counsel was violated when his attorney failed to inform him that his guilty plea could

lead to his denaturalization and deportation. The Supreme Court has described the

analysis undertaken in Padilla as follows: “[P]rior to asking how the Strickland test

applied (‘Did this attorney act unreasonably?’), Padilla asked whether the Strickland test

applied (‘Should we even evaluate if this attorney acted unreasonably?’).” Chaidez,

(emphasis in original). We take the same approach here, considering first

whether the Sixth Amendment requires counsel to advise a naturalized citizen client of

the risk of denaturalization and related deportation that may result from a guilty plea.

We conclude that, under Padilla, a naturalized U.S. citizen has a Sixth

Amendment right to be advised by counsel that he may be denaturalized and deported

as a result of his entry of a guilty plea. 4 Because a guilty plea to a conviction that

4The Sixth Amendment provides in relevant part that “[i]n all criminal prosecutions, the accused shall enjoy the right . . . to have the Assistance of Counsel for his defence.” U.S. Const. amend. VI. The Sixth Amendment right has long been recognized to encompass the right to effective counsel in deciding whether to enter a guilty plea. See Lafler v. Cooper,

, 162

14 exposes a criminal defendant to the risk of denaturalization necessarily exposes him to

the risk of deportation, a straightforward application of Padilla—holding that “counsel

must inform her client whether his plea carries a risk of deportation,” 559 U.S. at 374—

resolves the question. Alternatively, because denaturalization is a “particularly severe”

consequence, and “nearly an automatic result” of a guilty plea to certain offenses,

advice regarding denaturalization, like advice regarding deportation, “is not

categorically removed from the ambit of the Sixth Amendment right to counsel.”

at

365–66. To provide effective representation, counsel must advise her client as to that

paired set of serious adverse immigration consequences by at least flagging the risk. As

many have observed, the risk of banishment may weigh more heavily in a criminal

defendant’s assessment of his options than any particular sentence of incarceration. See,

e.g., Lee v. United States,

, 370 (2017) (“[W]e have recognized that preserving

the client’s right to remain in the United States may be more important to the client than

any potential jail sentence.” (internal quotation marks omitted)). This is so whether the

defendant is a noncitizen weighing the risk of deportation or a naturalized citizen

weighing the risk of denaturalization and deportation.

In sum, whether viewed as a direct application of Padilla’s holding or a necessary

result of its core reasoning, the constitutional bottom line is clear: criminal defense

counsel has a Sixth Amendment duty to advise a naturalized citizen client that entering

a guilty plea exposes him to a risk of denaturalization and deportation. 5

(2012); Hill v. Lockhart,

, 58–59 (1985); Kovacs v. United States,

(2d Cir. 2014).

5Congress’s 1996 amendments to our immigration laws “involved a change in nomenclature; the statutory text now uses the term ‘removal’ rather than ‘deportation.’” Padilla,

n.6. We use these terms interchangeably throughout.

15 The government argues that a summary application of the familiar distinction

between direct and collateral consequences of a conviction places the risk of

denaturalization and deportation outside the scope of the Sixth Amendment. Civil

denaturalization falls into the “collateral” pile, the government asserts, urging on this

analysis that counsel simply has no constitutional obligation to advise clients of such a

risk. Further, the government maintains, Padilla addressed the Sixth Amendment rights

to effective counsel of noncitizens only and said nothing about the rights of naturalized

citizens. We reject these contentions. As previewed above and as detailed below, the

government’s arguments cannot be reconciled with Padilla.

A. Farhane’s guilty plea carried a risk of deportation.

Applying the Supreme Court’s holding in Padilla answers the question before us.

A risk of denaturalization simply is a risk of deportation. 6 And, as Padilla instructs, to

satisfy the Sixth Amendment’s requirement of effective assistance in a criminal

prosecution, “counsel must inform her client whether his plea carries a risk of

6See Anthony D. Bianco et al., Civil Denaturalization: Safeguarding the Integrity of U.S. Citizenship, 65 U.S. Attys’ Bull. 5, 17 (July 2017) (stating that the government typically “does not expend resources on civil denaturalization actions unless the ultimate goal is the removal of the defendant from the United States”). Judge Walker’s Dissent takes issue with our reliance on the government’s representation in the noted publication that it typically does not pursue civil denaturalization unless it also intends to seek deportation. It argues that the representation deserves little weight, reasoning that the government’s process is “subject to change depending on the presidential administration.” Dissent (Walker, J.) at 8 n.7. We agree that, as that Dissent observes, “[o]ur Sixth Amendment analysis should not depend on the varying prosecutorial priorities of the Department of Justice.”

The Majority’s Sixth Amendment analysis does not depend, however, on potentially variable prosecutorial priorities. Rather, the risk that triggers the Sixth Amendment duty here derives from the statutory schemes that render certain individuals subject to deportation. See infra nn.7, 9, 11. In contrast, no statutory scheme compels the result reached by Judge Walker’s Dissent. The Sixth Amendment interpretation and the direct/collateral framework that it invokes represent doctrines adopted by the lower courts over the years. Applying those doctrines alone would leave naturalized citizens uniquely, and paradoxically, without the Sixth Amendment protection that Padilla affords.

16 deportation.”

. Thus, applying Padilla to the case at bar, it is clear that the

Sixth Amendment requires criminal defense counsel to advise his client of the risk of

denaturalization and the associated heightened risk of deportation that flow from his

guilty plea.

Farhane’s case illustrates why a guilty plea that carries a risk of denaturalization

necessarily carries a risk of deportation. The government initiated civil denaturalization

proceedings against him in 2018, soon after his release from prison, using the

admissions in his 2006 guilty plea as the foundation of its civil complaint. In that

complaint, the government charges that Farhane’s naturalization was “illegally

procured” or “procured by concealment of a material fact or by willful

misrepresentation,”

(a), referring to the statements Farhane made about

his 2001 conduct in his guilty plea. 7 Indeed, in its denaturalization complaint, the

7“[F]ailure to comply with the statutory prerequisites for naturalization renders [a] certificate of citizenship revocable as ‘illegally procured’ under

(a).” Fedorenko v. United States,

(1981). One such prerequisite is that the applicant be “a person of good moral character” for the five years (the “statutory period”) preceding his application for citizenship.

(a)(3). The current standard citizenship application form asks all applicants, “Have you EVER committed . . . a crime or offense for which you were NOT arrested?” Maslenjak v. United States,

, 346 (2017) (bold in original), and the form submitted by Farhane used almost identical language, see App’x at 334 (“Have you ever . . . knowingly committed any crime for which you have not been arrested?”). By regulation, an admission to having engaged in many types of criminal conduct within the statutory period automatically precludes an applicant from demonstrating the “good moral character” required for naturalization. See

(b)(2)(vi) (lack of good moral character if, during the statutory period, the applicant has given false testimony to obtain an immigration benefit), (b)(3)(iii) (lack of good moral character if, during the statutory period, the applicant committed unlawful acts that adversely reflect upon his moral character, or was convicted or imprisoned for such acts).

A naturalized citizen who willfully fails to disclose such conduct (occurring within the statutory period) when applying for citizenship will also have procured his naturalization “by concealment of a material fact or by willful misrepresentation.”

(a); see also Kungys v. United States,

(1988). Thus, a naturalized citizen who pleads guilty in any context to any such pre-naturalization conduct that occurred within the statutory period—

17 government argues that Farhane, by way of his guilty plea, “is collaterally estopped

from contesting those matters determined by the judgment in the criminal case.” App’x

at 314. As the government conceded at oral argument in these en banc proceedings, the

collateral estoppel effect of Farhane’s plea therefore makes it “provably easier” for the

government to establish the facts necessary to denaturalize him. See En Banc Oral Arg.

Tr. at 67:10–13. The court adjudicating the government’s denaturalization petition will

have “no discretion to excuse the conduct.” Fedorenko v. United States,

(1981). The government admonished the District Court accordingly in its

denaturalization complaint against Farhane. See App’x at 314. And once his citizenship

is revoked, Farhane will be subject to removal as a noncitizen convicted of an

aggravated felony. See

(a)(43)(D), (U), 1227(a)(2)(A)(iii); Sessions v.

Dimaya,

, 153 (2018) (“[R]emoval is a virtual certainty for an alien found to

have an aggravated felony conviction, no matter how long he has previously resided

here.”). 8

including, for example, a guilty plea to driving while under the influence, see U.S. Citizenship & Immigr. Servs., Instructions for Application for Naturalization, at 21 (Apr. 1, 2024)—will necessarily have illegally procured his citizenship and failed to disclose that conduct when applying for citizenship. See

(a), 1451(a). That guilty plea makes the citizen vulnerable to a civil complaint under the denaturalization statute, which provides that “[i]t shall be the duty of the United States” government “to institute [denaturalization] proceedings” in these circumstances.

§ 1451(a). 8 As Judge Wesley notes in concurrence, two Circuits have held that a noncitizen is not

removable under the aggravated felony provision,

(a)(2)(A)(iii), if the person was a U.S. citizen at the time of conviction. Concurrence (Wesley, J.) at 1 n.1. See Singh v. Att’y Gen. of U.S.,

(3d Cir. 2021); Hylton v. U.S. Att’y Gen.,

(11th Cir. 2021). Even if this Circuit were to adopt our sister Circuits’ construction of the aggravated felony provision— an issue not presented to the en banc court—Farhane would still face a risk of deportation. As Farhane’s counsel conceded at oral argument, if now denaturalized, Farhane would still be deportable on grounds other than the instant conviction. En Banc Oral Arg. Tr. at 77:24–78:3. Moreover, the unsettled nature of the question of the aggravated felony provision’s applicability to naturalized citizens further demonstrates the risk facing those like Farhane, who must assess whether to enter a guilty plea. Padilla instructs that even such an “unclear or

18 The government’s request that we apply Padilla only where a “direct linkage”

exists between a conviction resulting from a guilty plea and deportation cannot be

reconciled with Padilla. En Banc Oral Arg. Tr. at 43:7–15. To be sure, for a naturalized

citizen defendant, the risk of deportation is one step further removed from entry of a

guilty plea than it is for a noncitizen: the government must obtain his denaturalization

before it can initiate removal proceedings. The Padilla Court, however, expressly ruled

that the Sixth Amendment requires counsel to advise her client regarding the

deportation consequences of his plea, whether or not those consequences are clear or

certain. In situations where “the law is not succinct and straightforward,” the Court

clarified, “a criminal defense attorney need do no more than advise a noncitizen client

that pending criminal charges may carry a risk of adverse immigration consequences.”

Padilla,

; see also

(Alito, J., concurring) (expressing view that

counsel “must . . . advise the defendant that a criminal conviction may have adverse

immigration consequences and that, if the alien wants advice on this issue, the alien

should consult an immigration attorney,” but that counsel need not “explain what those

consequences may be”). Where entry of an order of deportation is a step removed from

the guilty plea, as it is for naturalized citizen defendants, the path from plea to eventual

deportation may indeed be less straightforward than that faced by Padilla. But that

alone does not remove advice regarding the risk from the ambit of the Sixth

uncertain” risk of deportation is sufficient to trigger the Sixth Amendment duty to advise. Padilla,

. The Dissent attempts to downplay the connection between Farhane’s plea and his denaturalization by emphasizing the defenses that Farhane might still theoretically raise against the three denaturalization counts in the government’s petition. Dissent (Walker, J.) at 6–7. Farhane’s counsel stated at oral argument in our en banc proceedings that he is not aware of any actually viable defenses to those counts based on current caselaw. En Banc Oral Arg. Tr. at 78:4–79:5. The merits of those defenses are not before us, of course. It suffices to note that the mere possibility that the government might be unsuccessful in its attempts to denaturalize or deport Farhane does nothing to negate the fact that Farhane faced a cognizable, serious risk when he decided to enter his guilty plea.

19 Amendment: “Lack of clarity . . . does not obviate the need for counsel to say something

about the possibility of deportation, even though it will affect the scope and nature of

counsel’s advice.”

at 369 n.10 (emphasis added).

Farhane’s guilty plea carried not only the risk of denaturalization but also, and

inevitably, the risk of deportation. Under Padilla, counsel advising a client in Farhane’s

circumstances is constitutionally obligated to advise that his plea “may carry a risk of

adverse immigration consequences.”

.

B. The direct/collateral distinction is also “ill suited” to evaluating the risk of denaturalization, and so counsel has a constitutional obligation to advise of such a risk.

A faithful application of Padilla’s underlying reasoning similarly compels the

conclusion that advice about the risk of denaturalization falls within the scope of the

Sixth Amendment. In finding deportation “uniquely difficult to classify as either . . .

direct or . . . collateral” and thus “ill suited” to categorization under the direct/collateral

framework,

, the Padilla Court highlighted two aspects of the consequence.

First, it stressed that deportation is a “particularly severe penalty.”

(internal

quotation marks omitted). Second, it observed that deportation was “nearly an

automatic result for a broad class of noncitizen offenders,” emphasizing “its close

connection to the criminal process” and noting how “[o]ur law has enmeshed criminal

convictions and the penalty of deportation for nearly a century.”

at 365–66. The same

considerations apply fully to the consequence of denaturalization.

It is not disputed here that denaturalization is a “particularly severe”

consequence.

. The Supreme Court “has long recognized the plain fact that to

deprive a person of his American citizenship is an extraordinarily severe penalty.”

Klapprott v. United States,

(1949); see also

at 616–17 (Rutledge, J.,

concurring) (stating that “[t]o take away a man’s citizenship deprives him of a right no

20 less precious than life or liberty” and that “in its wake may follow the most cruel

penalty of banishment”). If Farhane loses his citizenship, he loses his home of over

thirty years, his business, his life with his family; in addition, two of his children, who

derived citizenship through him, stand to lose their citizenship here as well. See

(d). 9

For naturalized citizens who plead guilty to criminal conduct pre-dating their

naturalization—and within the statutory period—their convictions create a “nearly . . .

automatic” risk of denaturalization. See Padilla,

. Farhane’s circumstances

exemplify this fact. As outlined above, he admitted through his guilty plea that,

beginning in November 2001, he conspired to commit money laundering in violation of

. In the denaturalization proceedings that the government initiated

against him soon after his release from his thirteen-year sentence, he will be unable to

contest or explain in any way the operative facts: that he illegally procured his

citizenship based on criminal conduct occurring shortly before his 2002 naturalization,

and that he concealed a material fact or made a willful misrepresentation regarding that

criminal conduct by failing to disclose it in his March 2002 interview. See

(a). And, as already noted, the court adjudicating the government’s

denaturalization petition will have no discretion to excuse the conduct. 10 Thus,

9When an individual is denaturalized based on concealment of a material fact or willful misrepresentation in procuring naturalization—as would be the case with Farhane—any person claiming citizenship through such denaturalized person “shall be deemed to have lost . . . citizenship.” 8 U.S.C. 1451(d); see also Bianco et al., at 16 (“If the defendant's spouse or children obtained citizenship based on the defendant’s naturalization, the denaturalization judgment revokes the spouse’s and children’s naturalization as a matter of law.”).

10The lack of discretion afforded the courts “to ameliorate unjust results on a case-by-case basis” also informed the Padilla Court’s analysis with respect to deportation.

. For the better part of the twentieth century, the Court stated, sentencing judges had “conclusive authority to decide whether a particular conviction should be disregarded as a basis for deportation,”

(quoting Janvier v. United States,

(2d Cir. 1986)), through a

21 Farhane’s guilty plea and conviction, on one hand, and the consequence of

denaturalization, on the other, are “enmeshed”; and denaturalization has a “close

connection to the criminal process.” Padilla, 559 U.S. at 365–66.

The government attempts to limit Padilla’s application, contending that

denaturalization is not as “nearly . . . automatic” a consequence of a conviction as is

deportation. Id. at 366. But its arguments in this vein fail to persuade.

First, we reject the government’s suggestion that “nearly . . . automatic” means

“definite” or “immediate.” Id.; see also Gov’t Br. at 18 (maintaining that civil

denaturalization is neither a “definite” or “immediate” consequence of a criminal

conviction). As discussed above, the Court in Padilla did not limit its holding to cases in

which deportation would certainly or immediately follow a criminal conviction. See

Padilla,

(clarifying that counsel’s Sixth Amendment duty to advise

attaches even to cases where “the deportation consequences of a particular plea are

unclear or uncertain”). Instead, it repeatedly stressed that constitutionally effective

counsel has an obligation to warn of a “risk of deportation.”

(emphasis added);

see also

(“The weight of prevailing professional norms supports the view that

counsel must advise her client regarding the risk of deportation.” (emphasis added));

(“[W]e now hold that counsel must inform her client whether his plea carries a

risk of deportation.” (emphasis added)). Put simply, Padilla was focused on cases

procedure known as a judicial recommendation against deportation, or JRAD.

at 361–62. But by 1990, Congress had entirely eliminated the JRAD procedure.

. Courts thus went from having ample discretion to grant relief to deportable noncitizens to having none at all.

at 363–64. “These changes to our immigration law,” the Padilla Court observed, “have dramatically raised the stakes of a noncitizen’s criminal conviction,” and “accurate legal advice for noncitizens accused of crimes has never been more important.”

. These concerns carry equal weight in the denaturalization context, in which courts, too, have no discretion to grant relief to naturalized U.S. citizens eligible for denaturalization, even when granting such relief would “ameliorate unjust results.”

.

22 involving advice as to whether a guilty plea would make a criminal defendant

deportable, not just on cases where deportation would inescapably follow. See, e.g.,

(“We agree with Padilla that constitutionally competent counsel would have

advised him that his conviction for drug distribution made him subject to automatic

deportation.” (emphasis added));

(“When attorneys know that their clients

face possible exile from this country and separation from their families, they should not

be encouraged to say nothing at all.” (emphasis added));

(observing that

“informed consideration of possible deportation can only benefit both the [government]

and noncitizen defendants during the plea-bargaining process” (emphasis added)).

Farhane’s guilty plea and conviction created a “nearly . . . automatic” risk of

denaturalization,

, because his admissions to criminal conduct in his plea

established his ineligibility for naturalization, in accordance with the applicable

statutory and regulatory provisions—just as Padilla’s guilty plea made him subject to

deportation under the relevant immigration laws. That Farhane’s plea makes it

“provably easier,” in the government’s words, for the government to denaturalize him

renders the risk of denaturalization a sufficiently automatic consequence of his guilty

plea, at least under Padilla’s reasoning. See En Banc Oral Arg. Tr. at 67:10–13.

Second, and relatedly, even deportation in Padilla’s circumstances is not as

“automatic” as the government now makes it out to be. For instance, the government

suggests the comparison is invalid because “[c]ivil denaturalization cannot occur unless

[it] initiates a separate civil proceeding under

(a).” Gov’t Br. at 18. The

same is true, however, of deportation proceedings that follow entry of a guilty plea: an

additional judicial or agency proceeding is required before deportation will occur. See 8

U.S.C. § 1229a; Padilla,

(recognizing that removal is “civil in nature” and

“not, in a strict sense, a criminal sanction”). With respect to both denaturalization and

23 deportation, the consequence is not imposed by the sentencing judge in the predicate

criminal case.

Additionally, the government highlights the role that prosecutorial discretion

plays in denaturalization proceedings, advising that it “rarely seeks denaturalization,

even following criminal convictions.” Gov’t Br. at 18. But deportation proceedings, too,

are initiated by the government through an “exercise of prosecutorial discretion . . . on a

case-by-case basis . . . .” Mem. from Sec’y John Kelly, Dep’t of Homeland Sec.,

Enforcement of the Immigration Laws to Serve the National Interest 4 (Feb. 20, 2017). Even

once the government has placed a noncitizen in removal proceedings, removal is not an

entirely foregone conclusion: the Attorney General retains “limited remnants of

equitable discretion” to cancel removal, Padilla,

, and certain individuals

may avoid deportation if they can, for instance, convince their immigration judge that

their conviction is not for a qualifying crime of moral turpitude, or establish eligibility

for cancellation of removal. And deportation is not necessarily “immediate” in the

temporal sense: it can occur long after a conviction renders the noncitizen eligible. See

Restrepo v. Att’y Gen. of U.S.,

, 800–02 (3d Cir. 2010) (concluding that no

statute of limitations applies to removal proceedings predicated on a conviction).

The government also contends that a conviction is not a necessary condition for

denaturalization: it may obtain an order of denaturalization even without a criminal

conviction, so long as it can prove that the defendant committed the underlying

conduct that rendered his naturalization “illegally procured,” or that the defendant lied

about his conduct during the naturalization process,

(a). Again, the same

is true of deportation. See, e.g.,

§ 1227(a)(1), (3)–(6) (providing numerous grounds for

deportation not predicated on a criminal conviction). At oral argument in these en banc

proceedings, in fact, the government conceded that the Sixth Amendment would apply

with respect to counsel’s advice to a noncitizen about entering a plea to a nonremovable

24 offense, where in the course of pleading guilty the defendant would admit to conduct

that would make him removable. See En Banc Oral Arg. Tr. at 42:5–17. If that is so, we

see no reason why, in cases involving the risk of denaturalization, the fact that the

conviction itself would not serve as the predicate for denaturalization should remove

advice about that consequence from the Sixth Amendment’s ambit. In short, the

government’s asserted distinctions between deportation and denaturalization do not

withstand scrutiny.

Third, the government’s concession at oral argument regarding the applicability

of Padilla to asylees underscores the weakness of its urged reading of Padilla. When

asked whether Padilla’s holding applies to a noncitizen asylee who is exposed to the risk

of deportation because of a criminal conviction, the government answered in the

affirmative. See En Banc Oral Arg. Tr. at 52:7–8 (“Absolutely, your Honor. They’re a

noncitizen. Padilla applies to any noncitizen.”). But the deportation risk faced by asylees

who are deciding whether to enter a guilty plea is virtually indistinguishable from that

faced by naturalized citizens confronted with the same choice. The path from conviction

to removal for asylees, like that for naturalized citizens, comprises two steps: in the case

of asylees, the government must first terminate their asylum status before it can remove

them. 11

The similarities between denaturalization and termination of asylum status do

not end there. Just as a conviction is not a necessary condition for denaturalization (in

the sense that an individual can be denaturalized based on conduct admitted to in a

11Asylees are subject to removal only after their asylum status has been terminated. See

(c)(1)(A) (stating that the Attorney General “shall not remove or return” a noncitizen granted asylum);

§ 1158(c)(2) (providing a mechanism for terminating asylum); id. § 1158(c)(3) (providing that a noncitizen whose asylum status is terminated is subject to any applicable grounds of inadmissibility or removal under

(a) and § 1227(a)).

25 guilty plea or on a lie about that conduct, even if the conviction itself is not a predicate

for denaturalization), a conviction is also not a necessary condition for termination of

asylum. The government can terminate asylum on grounds unrelated to a conviction. 12

And as for the government’s contention that a conviction is not a sufficient

condition for denaturalization and removal, the same is also true for termination of

asylum and removal. For example, the government posits that establishing at a

denaturalization proceeding that nondisclosure or misrepresentation of a material fact

during the naturalization process was “willful” might require an additional showing

not already established by the criminal conviction. 13 In the case of asylees, however, to

terminate an individual’s asylum, the government must likewise prove more than the

fact of a conviction of a qualifying crime: it must show that the asylee “constitutes a

danger to the community of the United States.”

(b)(2)(A)(ii). 14

Accordingly, the government’s concessions with respect to the Sixth Amendment duties

owed asylees expose the same flaws in its contention that denaturalization is not a

sufficiently automatic consequence of a guilty plea and conviction.

12 See

§ 1158(b)(2)(A)(i), (iii)–(vi), (c)(2)(A)–(E).

13See id. § 1451(a); Kungys,

; Gov’t Br. at 19–20 (“[A]n inadvertent oversight or sincere failure of recollection in omitting the criminal conduct on a naturalization form would not have been a ‘willful misrepresentation.’”).

14In more detail: An individual’s asylum-based status may be terminated if the noncitizen asylee, “having been convicted by a final judgment of a particularly serious crime, constitutes a danger to the community of the United States.”

(b)(2)(A)(ii); see also

§ 1158(c)(2)(B) (identifying the conditions in § 1158(b)(2) as a basis for terminating asylum); id. § 1158(b)(2)(B)(i) (defining “particularly serious crime” to include an aggravated felony). Accordingly, in addition to establishing that the noncitizen has been convicted of a particularly serious crime, the government must also prove that the noncitizen “constitutes a danger to the community of the United States.” Id. § 1158(b)(2)(A)(ii).

26 C. The government’s remaining arguments against applying Padilla to the risk of denaturalization and related deportation fail.

The government offers a host of other arguments in support of its position that

Padilla should not apply to a client in Farhane’s circumstances. None is convincing.

As an initial matter, we easily disavow any suggestion that Padilla protects only

noncitizens. That the Padilla Court referred at points to “noncitizen defendants” and

“noncitizen offenders,”

, is unsurprising, since Padilla himself was a

noncitizen. Notably, however, the Court in Padilla stated its holding in broad terms:

“[W]e now hold that counsel must inform her client whether his plea carries a risk of

deportation.”

(emphasis added). And it stressed that “[i]t is our responsibility

under the Constitution to ensure that no criminal defendant—whether a citizen or not—is

left to the mercies of incompetent counsel.”

(emphasis added) (internal quotation

marks omitted). It seems to us perplexing and paradoxical, both, to construe Padilla as

providing stronger protection to a noncitizen at risk of deportation (like Padilla) than to

a U.S. citizen at risk of denaturalization followed by deportation (like Farhane). 15 Under

the government’s view, Farhane would have been better off had he never pursued

naturalization: had he been merely a lawful permanent resident when he pleaded

guilty, rather than a naturalized citizen, Padilla would settle that Strickland governed his

ineffective assistance claim. Cf. United States v. Winter,

, 989 n.45 (5th Cir.

15Judge Walker’s Dissent fails to address the apparent paradox that, under the government’s and the prior panel majority’s views, the Sixth Amendment would provide greater protection to a noncitizen faced with a risk of deportation than to a naturalized citizen also facing a risk of deportation. Indeed, it seems to us telling that it devotes much space to analyzing the deficient performance and prejudice prongs of Strickland, see Dissent (Walker, J.) at 9–21, rather than the merits of the Sixth Amendment claim. It does not face the puzzling situation its Sixth Amendment interpretation creates for criminal defense attorneys who would, under its view, be constitutionally required to advise only some of their immigrant clients about the immigration risks of a possible plea.

27 1975) (“We are unaware of any context . . . in which [a noncitizen] has greater rights

under law in our Courts than an American citizen. Indeed, such a circumstance might

in itself raise serious questions of equal protection.”).

The government also maintains that the Sixth Amendment applies to affirmative

misadvice regarding the risk of denaturalization but not to a wholesale failure to advise

on that risk. But the Court roundly rejected an argument based on this very distinction

in Padilla in terms equally apt here:

A holding limited to affirmative misadvice would invite two absurd results. First, it would give counsel an incentive to remain silent on matters of great importance, even when answers are readily available. Silence under these circumstances would be fundamentally at odds with the critical obligation of counsel to advise the client of the advantages and disadvantages of a plea agreement. . . . Second, it would deny a class of clients least able to represent themselves the most rudimentary advice on deportation even when it is readily available.

559 U.S. at 370–71 (internal quotation marks omitted). We decline the government’s

invitation to adopt a rule that would lead to “absurd results,” id. at 370, and that would

incentivize counsel to remain silent when advice is most needed.

Nor do we agree with the government’s suggestion that the rule we set forth

today will call into question the applicability of the direct/collateral framework to

ineffective assistance claims more broadly. As explained above, our conclusion here is

dictated by Padilla’s holding that “advice regarding deportation is not categorically

removed from the ambit of the Sixth Amendment right to counsel,” even when “the

deportation consequences of a particular plea are unclear or uncertain.” Id. at 366, 369.

The government recites a litany of other collateral consequences that it claims will by

28 virtue of our decision now fall in the Sixth Amendment’s scope. 16 Virtually none of

them, however, “carries a risk of deportation,” the risk that the Padilla court held to be

“unique.” Id. at 365, 374. The government’s argument therefore misses the mark. 17

Next, like the Court in Padilla, we see no reason to expect that our decision today

will open the floodgates to meritless challenges to long-final convictions. See id. at 371–

72 (describing the “serious consideration” given to the concerns expressed “regarding

the importance of protecting the finality of convictions obtained through guilty pleas”

and nonetheless finding it “unlikely that our decision today will have a significant effect

on those convictions already obtained as the result of plea bargains”). Any floodgates

concerns here do not rest on solid ground. To start, the government has represented

16See Gov’t Br. at 32 (listing as examples of other collateral consequences that would be called into question by virtue of a decision in Farhane’s favor: “civil or criminal forfeiture, mandatory restitution, court martial or disqualification from the armed services, loss of . . . licenses granted by the state, loss of civil rights, loss of federal benefits, denial of certain types of employment, mandatory HIV testing, registration of sex offenders, use of the conviction in a subsequent civil or criminal case, and, for non-citizens, immigration consequences . . .” (internal quotation marks omitted)).

17With respect to the potential application of Padilla’s underlying reasoning to these other consequences, we stress the Padilla Court’s characterization of deportation as “particularly severe,” id. at 365, and the role that severity played in our analysis of denaturalization. Take civil forfeiture, one of the government’s proffered examples. Although we do not decide this question today, it seems to us unlikely that civil forfeiture would come close to denaturalization and deportation in severity as well as irreversibility. As the Supreme Court has recognized, “denaturalization, like deportation, may result in the loss of all that makes life worth living.” Knauer v. United States,

(1946) (internal quotation marks omitted); see also Klapprott,

(characterizing the loss of citizenship as “an extraordinarily severe penalty,” one that cannot compare to “a mere money fine or a short imprisonment”); Delgadillo v. Carmichael,

(1947) (describing deportation as “the equivalent of banishment or exile”); Schneiderman v. United States,

(1943) (describing the benefits that derive from citizenship as “priceless,” and characterizing the loss of citizenship as “more serious than a taking of one’s property, or the imposition of a fine or other penalty”). The government has presented no instance in which civil forfeiture of assets meaningfully compares in permanence or familial consequences.

29 that it rarely seeks denaturalization, even following criminal convictions. See Gov’t Br.

at 18 (reporting that, based on its own review of dockets, just 228 denaturalization cases

were filed nationwide between 2008 and 2020). In any event, the record offers no basis

for estimating how many naturalized citizen defendants would even be eligible for

denaturalization and deportation as a result of their guilty pleas. In addition, statutes of

limitations and other constraints applicable to petitions for state and federal habeas and

coram nobis relief can further be expected to reduce the likelihood that a flood will

follow in this decision’s wake. See, e.g.,

(f) (setting forth one-year statute

of limitations, which runs from the latest of four dates, for § 2255 motions); Jobe v.

Comm’r of Corr.,

, 159 (Conn. 2020) (concluding that “a petitioner whose

conviction has expired fully prior to the filing of a habeas petition is not in ‘custody’ on

that conviction within the meaning of [Connecticut’s habeas statute]” (internal

quotation marks omitted)); State v. Grisgraber,

(Conn. 1981) (stating

that, in Connecticut, a writ of coram nobis “authorize[s] the trial judge, within three

years, to vacate the judgment”); People v. Friedgood,

(N.Y. 1983)

(finding no abuse of discretion in the trial court’s denial of the defendant’s motion to

vacate his conviction under

where the defendant “waited

for over three years to bring” his challenge). But see In re D.C.,

, 470 n.2 (Vt.

2016) (explaining that petitions for post-conviction relief “are not subject to a statute of

limitations”); Thompson v. Comm’r of Corr.,

(Conn. App. 2017)

(clarifying that there is no statute of limitations for a writ of habeas corpus in

Connecticut). And the constitutional ruling on the Sixth Amendment’s applicability to

an ineffective assistance claim is only the threshold inquiry. It leads to a demanding

analysis under Strickland, one that—the Court observed—rarely leads to overturning a

conviction. See Padilla, 559 U.S. at 371–72 (“Surmounting Strickland’s high bar is never

an easy task.”).

30 Finally, the government’s focus on the collateral estoppel effect of Farhane’s plea

and conviction is a red herring. See, e.g., Gov’t Br. at 17 (“The possibility of collateral

estoppel in a civil denaturalization proceeding is a collateral consequence of a criminal

conviction[.]”); id. at 28 (asserting that, even if Farhane’s “criminal conviction would

work to his detriment by virtue of collateral estoppel, . . . this Court has repeatedly held

that collateral estoppel alone does not transform a collateral consequence into a direct

one” (internal quotation marks omitted)). In reaching our conclusion today, we follow

the Padilla Court’s guidance by focusing instead on how Farhane’s guilty plea carried

the risk of adverse immigration consequences, specifically denaturalization and

deportation. As we have already observed, the Padilla Court instructed that counsel

must advise a client “that pending criminal charges may carry a risk of adverse

immigration consequences.”

. Whether a conviction carries such a risk by

virtue of collateral estoppel consequences or by serving as a predicate for the later

immigration result is beside the point. Here, the government’s denaturalization

complaint expressly relied on Farhane’s guilty plea, asserting that he “is collaterally

estopped from contesting those matters determined by the judgment in the criminal

case,” and reminding the court that it was not free to take any related circumstances

into account before entering a denaturalization order. App’x at 314. By making it

substantially easier for the government to obtain Farhane’s denaturalization, his guilty

plea materially increased his risk of denaturalization and eventual deportation. Under

Padilla, that is enough to trigger counsel’s Sixth Amendment obligation to say something

about the risk.

D. The government forfeited its Teague argument.

The government also now contends that Teague v. Lane,

(1989),

precludes us from addressing Farhane’s Sixth Amendment arguments. But the

government forfeited their argument. As the original panel Majority observed, the

31 government “did not make this argument in the district court” and still has “offered no

justification for the omission.” Farhane, 77 F.4th at 126 n.4.

What is more, the government failed to raise Teague at the first opportunity it had

in this Court, in its opposition to Farhane’s request for a certificate of appealability. It

was not until the appeal merits briefing stage—after a three-judge panel had already

heard argument on Farhane’s motion for a certificate of appealability and granted it—

that the government mentioned Teague for the first time. The government has therefore

forfeited its Teague argument, and we are under no obligation to consider it. See, e.g.,

Collins v. Youngblood,

(1990) (instructing that Teague “is not

‘jurisdictional’ in the sense that [courts] . . . must raise and decide the issue sua sponte”

(emphasis in original)); Young v. Conway,

(2d Cir. 2012) (noting that

courts “have the discretion, but are by no means required, to address [Teague] defenses

for the first time on appeal”).

Although it is true that we may nonetheless exercise our discretion to consider

the application of Teague to our holding today, the government does not offer any

compelling reason why we should do so. In fact, the government’s still-unexplained

delay in raising the issue weighs heavily against the notion. If we were to consider and

find persuasive the government’s “eleventh-hour Teague argument” now, Farhane’s

Sixth Amendment contentions—the issues that the panel that granted Farhane a

certificate of appealability, the panel that originally decided his appeal, and a majority

of the en banc Court all “thought worthy of review”—“would be insulated from our

consideration.” Buck v. Davis,

, 127–28 (2017). We therefore decline to

consider the government’s forfeited Teague argument.

* * *

32 In sum, we conclude that criminal defense counsel has a Sixth Amendment

obligation to inform a client that his plea carries a risk of serious adverse immigration

consequences: here, of denaturalization and related deportation. Padilla dictates this

result.

II. We remand Farhane’s case to permit the District Court to conduct a Strickland analysis in accordance with the Sixth Amendment understandings we reach.

Having established that the Sixth Amendment requires counsel to advise

naturalized citizen clients of the risk of denaturalization and deportation arising from a

guilty plea, the next step is consideration of the merits of Farhane’s claim that his

counsel’s failure to do so in 2006 constituted ineffective assistance. To establish that his

counsel was constitutionally ineffective, Farhane must show that (1) “counsel’s

representation fell below an objective standard of reasonableness,” and (2) “any

deficiencies in counsel’s performance [were] prejudicial to the defense.” Strickland,

.

A. The District Court should reevaluate whether Farhane’s counsel’s performance was objectively unreasonable under professional norms prevailing in 2006.

Under Strickland’s first prong, the defendant must show that his lawyer’s

performance “falls outside the ‘wide range of professionally competent assistance.’”

Kovacs v. United States,

(2d Cir. 2014) (quoting Strickland,

).

To assess objective reasonableness, we evaluate “prevailing professional norms” at the

time of representation, using “American Bar Association [“ABA”] standards and the

like [as] guides to determining what is reasonable.” Padilla,

(internal

quotation marks and alterations omitted).

Here, in concluding that Farhane failed to establish that his counsel’s

performance was objectively unreasonable, the District Court’s reasoning cannot be

33 squared with Padilla. In its analysis, the District Court relied principally on the fact that,

because Farhane “was a U.S. citizen at the time of his plea,” his conviction did not

create an “imminent risk of deportation.” Farhane,

, at *2. But by

requiring Farhane to show that his conviction gave rise to an imminent risk of

deportation, the District Court impermissibly narrowed Padilla. As explained above, the

Supreme Court held in Padilla that counsel must inform her client “whether his plea

carries a risk of deportation.”

(emphasis added). It stressed that counsel’s

duty to advise her client about the risk of deportation applies whenever that risk exists,

not solely when that risk is imminent or crystal clear. See

at 369 & n.10 (explaining

that lack of clarity in the law affects only the scope and nature of the advice counsel has

a duty to provide, not whether counsel has such a duty at all). The District Court, in

focusing only on an “imminent risk of deportation,” Farhane,

, at *2,

therefore assessed the objective reasonableness of counsel’s performance in the first

instance using an incorrect understanding of Padilla.

Accordingly, we remand for the District Court to reevaluate whether Farhane’s

counsel’s failure to advise was objectively reasonable. The correct question—and the

one that the District Court should answer on remand—is whether prevailing

professional norms in 2006, when Farhane pleaded guilty, required defense counsel to

advise naturalized citizen clients of the risk of denaturalization and deportation

stemming from entry of a guilty plea. 18 In determining what was reasonable in 2006, it

18We note that the defendant whose guilty plea was at issue in Padilla was counseled and entered his guilty plea in 2002—four years before Farhane pleaded guilty. See Padilla v. Commonwealth,

(Ky. Ct. App. 2012). The Supreme Court in Padilla reviewed “authorities of every stripe” before concluding that “[t]he weight of prevailing professional norms supports the view that counsel must advise her client regarding the risk of deportation.”

(internal quotation marks and alterations omitted). It emphasized that professional norms had imposed such a duty on defense counsel “[f]or at least the past 15 years”—that is, since at least 1995.

.

34 should consult ABA standards and similar materials, which “may be valuable measures

of the prevailing professional norms of effective representation”—but “are only guides,

. . . and not inexorable commands.” Padilla, 559 U.S. at 366–67 (internal quotation marks

omitted).

Farhane’s trial counsel should also have an opportunity to be heard. The District

Court denied Farhane’s § 2255 petition without an evidentiary hearing, and so counsel

had no opportunity either to describe or to explain his conduct: we have Farhane’s

account, through his affidavit, but not counsel’s. Our “usual practice” is “to remand

Strickland cases to the district court to permit the attorney in question to testify and

explain h[is] actions.” Jackson v. Leonardo,

(2d Cir. 1998); see also Sparman

v. Edwards,

(2d Cir. 1998) (“[A] district court facing the question of

constitutional ineffectiveness of counsel should, except in highly unusual

circumstances, offer the assertedly ineffective attorney an opportunity to be heard and

to present evidence, in the form of live testimony, affidavits, or briefs.”). We see no

reason to depart from that practice here. And Farhane may be heard in person, too,

instead of solely by affidavit.

To clarify: on remand, the District Court should determine whether Farhane’s

counsel’s performance in 2006 was objectively unreasonable by analyzing relevant

authorities on professional norms and providing Farhane’s trial counsel with an

opportunity to speak.

B. The District Court should address in the first instance whether Farhane established prejudice.

Under Strickland’s second prong—prejudice—the defendant must establish “a

reasonable probability that, but for counsel’s unprofessional errors, the result of the

proceeding would have been different.” Padilla,

(quoting Strickland,

). In the plea context, “we do not ask whether, had he gone to trial, the result

35 of the trial would have been different than the result of the plea bargain.” Lee, 582 U.S.

at 364 (internal quotation marks omitted). Instead, we consider whether the defendant

has demonstrated a “reasonable probability that, but for counsel’s errors, he would not

have pleaded guilty and would have insisted on going to trial.” Id. at 364–65 (quoting

Hill v. Lockhart,

(1985)); see also Kovacs,

(concluding that

prejudice also exists if the defendant can show that “there was a reasonable probability

that [he] could have negotiated a plea that did not impact [his] immigration status”).

We decline to determine the prejudice question as to Farhane now. 19 The

Supreme Court has emphasized that determining “what an individual defendant would

have done” as part of the prejudice inquiry “demands a case-by-case examination of the

totality of the evidence.” Lee, 582 U.S. at 367–68 (internal quotation marks omitted). In

the District Court, Farhane did not have a robust opportunity to develop a relevant

record nor did the government have a chance to rebut it. On remand, both will have

that opportunity, and further, the District Court will be positioned to assess the

credibility of Farhane’s assertion that, had he been properly advised, he would have

insisted on proceeding to trial. 20 See, e.g., Padilla,

(“Whether Padilla is

entitled to relief on his claim will depend on whether he can satisfy Strickland’s second

19The District Court should consider, among other things, whether Farhane might have “placed particular emphasis on immigration consequences in deciding whether or not to plead guilty,” Doe v. United States,

(2d Cir. 2019) (quoting Kovacs,

) (alterations adopted)—enough to forgo the ten-year reduction in his maximum possible sentence that resulted from his entry of a guilty plea.

20Farhane additionally argues that he could have presented viable defenses at trial, and that his trial counsel could have “leveraged” these defenses “to negotiate an alternative plea deal that foreclosed the risk of denaturalization and deportation.” Farhane Br. at 51; see also

at 46–49 (contending that he could have challenged the credibility of the government’s “key witness” at trial);

at 49–51 (asserting that he could have mounted an entrapment defense). Unsurprisingly, the government disputes these contentions. We express no view on these issues here; on remand, the District Court may consider these arguments as it sees fit.

36 prong, prejudice, a matter we leave to the Kentucky courts to consider in the first

instance.”); Rodriguez v. United States,

, 44 (2d Cir. 2018) (finding it “not

clear on this record whether Rodriguez would have proceeded to trial had she known

she could face denaturalization” and remanding “for the district court to develop a

fuller record concerning the issue of prejudice”).

* * *

For the reasons stated above, we remand to the District Court to permit it to

make fresh determinations on both prongs of the Strickland inquiry after allowing the

parties to develop the relevant records and conducting an evidentiary hearing as

appropriate. 21

CONCLUSION

For the foregoing reasons, we VACATE the decision of the prior panel Majority;

VACATE the judgment of the District Court; and REMAND the case for further

proceedings consistent with this opinion.

21Of course, “there is no reason for a court deciding an ineffective assistance claim . . . to address both components of the inquiry if the defendant makes an insufficient showing on one.” Strickland,

.

37 20-1666 (en banc) Farhane v. United States

WESLEY, Circuit Judge, joined by LEE, ROBINSON, NATHAN, and MERRIAM, Circuit Judges, concurring:

A number of my colleagues endorse what was my initial view of this case.

I respect their position in this difficult matter, but I have come to disagree with that

view. I join the majority opinion to vacate and remand.

Farhane’s guilty plea admitted to criminal activity that preceded his U.S.

citizenship and was not revealed in his naturalization process. His plea, entered

18 years ago, now endangers his U.S. citizenship and eligibility to remain in this

country. Only after Farhane was released from prison did the Government initiate

civil denaturalization, invoking the estoppel of Farhane’s long ago plea

admissions. “[O]nce his citizenship is revoked, Farhane will be subject to removal

as a noncitizen convicted of an aggravated felony.” 1 Maj. Op. at 18. Farhane’s

2006 guilty plea therefore carried a risk of “adverse immigration consequences,”

1 Farhane and the Government agree that, if denaturalized, Farhane is removable under the Immigration and Nationality Act (“INA”)’s aggravated felony provision. Appellant’s Br. at 23; En Banc Oral Arg. Tr. at 51:5–7. Like the majority, I therefore assume the same. At argument, the Government remarked that the Third and Eleventh Circuits, reviewing decisions of the Board of Immigration Appeals, have held that a noncitizen is not removable under the INA’s aggravated felony provision if they were a U.S. citizen at the time of conviction. En Banc Oral Arg. Tr. at 50:4–15 (citing Singh v. Att’y Gen. of U.S.,

(3d Cir. 2021); Hylton v. U.S. Att’y Gen.,

(11th Cir. 2021)). This Court has yet to reach the issue and cannot here. However, Farhane’s counsel confirmed that, even if the aggravated felony provision did not apply, once denaturalized, Farhane would still be deportable on other grounds.

at 77:24–78:3. In any event, Padilla applies even when a plea’s deportation risk is “unclear or uncertain.” Padilla v. Kentucky,

(2010). including “deportation,” and Padilla teaches that, under the Sixth Amendment,

counsel must advise her client of that risk. Padilla v. Kentucky,

(2010).

The Government contends that Padilla does not impose a Sixth Amendment

concern with regard to that risk here because, unlike the noncitizen in Padilla,

Farhane is a naturalized citizen; his denaturalization and subsequent deportation

are dependent on a “rare, contingent risk that a separate civil denaturalization

proceeding could be initiated.” Gov’t Br. at 30. The dissent agrees, concluding

that “Padilla’s core reasoning does not apply in Farhane’s case, in which the

preexisting denaturalization risk was independent of and at most collateral to his

conviction and any deportation risk was entirely dependent on Farhane first

becoming denaturalized.” Dissent (Walker, J.) at 9.

The snag in that conclusion, as I see it now, is that asking whether Farhane,

regardless of his conviction, could be denaturalized, is the wrong way to assess

the immigration implications of his conviction. No doubt, the Government could

have sought to denaturalize Farhane with or without his conviction. However,

Farhane’s conviction, for pre-naturalization conduct, weakens his defense to

denaturalization; and, his conviction, to the type of crimes here, separately

2 guarantees that, upon his denaturalization, he will be deported. 2 See Maj. Op. at

16–18. There is no doubt a significant connection between his criminal activity, his

naturalization status, and his continued presence in the United States.

Consider the following: what if the Government had initiated

denaturalization proceedings before (instead of after) Farhane’s plea? Would the

dissent say, even then, with the Government having revealed its intent to

denaturalize Farhane, that his counsel need not have warned him of the

denaturalization and deportation risks of pleading guilty in his criminal case?

Nothing about Padilla mandates that we leave naturalized citizens with Sixth

Amendment rights uniquely dependent on the Government’s whim, at best, or

manipulation, at worst. 3 Nor does Padilla suggest that we should.

For the reasons thoroughly discussed by the majority opinion, the

differences between the deportation risk connected to a noncitizen’s guilty plea, as

in Padilla, and the coupled denaturalization and deportation risk resulting from a

naturalized citizen’s guilty plea, as that here, fail to convince that the Sixth

2See supra at 1 n.1. 3At least one Circuit has concluded that there is no statute of limitations for removal proceedings predicated on a conviction, or for civil denaturalization. See United States v. Phattey,

, 1279 (9th Cir. 2019) (concluding that the five-year statute of limitations generally applicable to civil penalties does not apply to civil denaturalization); Restrepo v. Att’y Gen. of U.S.,

, 800–02 (3d Cir. 2010) (concluding that the federal “catch all” statute of limitations does not apply to removal proceedings predicated on a conviction). 3 Amendment requires counsel to warn of the former, but not the latter. See Maj.

Op. at 16–20, 27–28.

The majority recognizes that the precise likelihood of deportation following

a naturalized citizen’s guilty plea may be more difficult to assess than that of most

noncitizens, but Padilla explains that the Sixth Amendment requires counsel to

advise of a plea’s “risk of adverse immigration consequences,” even where “the

deportation consequences of a particular plea are unclear or uncertain.” Padilla,

; see Maj. Op. at 19–20. Padilla does not suggest that the Supreme

Court had canvassed the entire federal code and every immigration hypothetical

to confirm that Padilla arose out of the one and only set of factual circumstances in

which a guilty plea could carry an immigration risk giving rise to Sixth

Amendment concerns.

Instead, the Supreme Court made clear that, when it comes to deportation,

the direct/collateral framework is not the measuring stick for the Sixth

Amendment’s reach. Under the direct/collateral framework, “[d]eportation as a

consequence of a criminal conviction is, because of its close connection to the

criminal process, uniquely difficult to classify as either a direct or a collateral

consequence.” Padilla,

. Indeed, “deportation, because the

4 consequence of a distinct civil proceeding, could well be viewed as [a collateral]

matter” of no Sixth Amendment concern. Chaidez v. United States,

(2013). The Supreme Court concluded that “[t]he collateral versus direct

distinction is thus ill suited to . . . the specific risk of deportation.” Padilla,

. It is because deportation “is a ‘particularly severe’ penalty, and one

‘intimately related to the criminal process,’” not because it is a so-called direct

consequence, that the Sixth Amendment requires counsel to warn of it. Chaidez,

(quoting Padilla,

). After all, if deportation were

completely, not “nearly” automatic, Padilla,

, the direct/collateral

framework would have easily deemed it a direct consequence; the Supreme Court

would have had no need to decide Padilla.

Here, the test for determining whether Padilla applies should not

functionally reimpose the direct/collateral framework that Padilla explicitly set

aside. The Government’s argument for why Padilla does not apply to Farhane’s

case merely establishes that, if the direct/collateral framework were to apply,

Farhane’s denaturalization and subsequent deportation could be viewed collateral

to his plea. That is not enough to distinguish Padilla.

5 Although the denaturalization and deportation risk posed by Farhane’s plea

raises a Sixth Amendment concern under Padilla, the degree of risk remains highly

relevant to the Sixth Amendment analysis under Strickland v. Washington,

(1984). In other words, if a naturalized citizen’s guilty plea poses a risk of

denaturalization and deportation like Farhane’s here, arguments about the

likelihood or certainty that his conviction will result in the realization of that risk

will be considered at a different step of the Sixth Amendment framework: during

the court’s Strickland analysis. 4

For this reason, I agree with the majority that, as in Padilla, Strickland

prevents opening the “floodgates to meritless challenges” that some of my

colleagues fear. 5 Maj. Op. at 29–30; see Dissent (Park, J.) at 6–7; see also Padilla, 559

U.S. at 371–72. For example, if a naturalized citizen pleaded guilty to an

undisclosed pre-naturalization crime, as here, but denaturalization and

deportation were unlikely—perhaps the crime was not of a type for which the

Government was generally seeking denaturalization or was a tenuous basis for

4 In my view, unlike Padilla’s more amorphous analysis, Strickland’s clear, two-part test is much better equipped to tackle the intensely factual and particularized evaluation of the exact level of immigration risk posed by a plea. 5 As the majority observes, if there were, as the Government says, only 228 denaturalization cases

filed between 2008 and 2020, how deep are the flood waters? See Maj. Op. at 29–30 (citing Gov’t Br. at 18). I concede that changes to executive branch policy could perhaps raise the tide, but Strickland should prove an adequate levee. 6 deportation—then the court might conclude that the immigration risk carried by

the conviction was so low that it was outside the norms for counsel to consider, or,

that even if counsel had warned of it, there was no reasonable probability that he

would have rejected the plea.

This case presents a convergence of facts that make the Strickland analysis

more challenging: Farhane pleaded guilty to crimes that would clearly render a

noncitizen deportable; some of the criminal activity occurred before he became a

U.S. citizen and was not disclosed before his naturalization; and the criminal

activity was of a type that may have made it more likely that the Government

would seek his denaturalization. The majority wisely exercises restraint in

remanding for the district court to conduct the Strickland analysis, now with the

guidance of our decision, and to develop the factual record if needed. See Maj. Op.

at 33–37; see also Concurrence (Pérez, J.) at 12 (“It is not the job of this Court to

conduct this analysis without giving the parties the opportunity to develop the

record.”); Dissent (Nardini, J.) at 1 (reasoning that “the district court should

undertake that analysis in the first instance”).

Lastly, a word or two on Teague—the one issue on which the original panel

was unanimous. See Farhane v. United States,

, 126 n.4 (2d Cir. 2023)

7 (citing Teague v. Lane,

(1989)). We agreed, as the majority does

now, that the Government had forfeited a Teague defense, and that we should

decline to resurrect it. 6

Some of my dissenting colleagues now suggest that we should excuse the

Government’s forfeiture and embrace Teague here, presenting their view as an

exercise of judicial restraint and constitutional avoidance. Dissent (Park, J.) at 1,

4–5. Unfortunately, it is not.

In Teague, the Supreme Court announced that “new constitutional rules of

criminal procedure will not be applicable to those cases which have become final

before the new rules are announced.” Teague,

. In other words,

“new constitutional rules of criminal procedure will not be announced or applied

on collateral review.” Lockhart v. Fretwell,

(1993). However,

Teague is not “jurisdictional.” Collins v. Youngblood,

(1990). The

Supreme Court “ha[s] held that States can waive a Teague defense, during the

course of litigation, by expressly choosing not to rely on it, or by failing to raise it

6 My colleagues note that “forfeited,” not “waived,” is the appropriate term here. Dissent (Park, J.) at 4 n.1. In adopting “forfeited,” I do not suggest that the Government’s previous silence on Teague was not knowing and intelligent. As the majority correctly observes, the Government has never explained why it did not earlier raise Teague. See Maj. Op. at 31–32 (citing Farhane v. United States,

, 126 n.4 (2d Cir. 2023)). 8 in a timely manner.” Danforth v. Minnesota,

(2008) (internal

citations omitted).

To be clear, because the Government undisputedly never raised a Teague

defense in the district court—first raising it in this Court, and only after we granted

a certificate of appealability—the Government forfeited the defense, and the

Supreme Court’s retroactivity doctrine bars no aspect of the majority’s opinion.

The finality interests served by Teague are insufficient to render its commands

absolute. See id.; Collins,

. Forfeited, as here, Teague bars nothing.

True, we have discretion to excuse the Government’s forfeiture of Teague.

See Buck v. Davis,

, 127 (2017). But there is no persuasive reason to do

so here. At the outset, this is not a case in which the finality interests weigh

strongly in favor of excusing forfeiture. My colleagues claim that the “interest in

finality is particularly salient here” because of the age of Farhane’s conviction and

the potential evidentiary issues that make re-trying old offenses difficult. Dissent

(Park, J.) at 6. However, this case does not involve a defendant challenging an old

conviction while he has years of imprisonment left to serve, a circumstance in

which the Government’s ability to re-prosecute is of heightened concern. Farhane

has completed his term of imprisonment and been released. Even if Farhane’s

9 conviction were vacated, the most severe direct consequence of his conviction—

his term of imprisonment—will remain irreversible.

I add that, in this case, any threat to finality is a problem of the

Government’s own making. It was the Government’s decision, and its alone, to

wait 18 years after Farhane’s plea to initiate denaturalization proceedings. If the

Government had earlier revealed its intent to denaturalize and deport Farhane, he

would have long ago learned of those consequences of his plea. Instead, the

Government sat idle. And then, when Farhane filed his

motion,

does the Government raise Teague? No.

Regardless, even if we were to excuse the Government’s forfeiture, Teague’s

retroactivity rule, in this case, would pose no simple bar. As we all recognize,

because Padilla was decided before Farhane’s conviction became final, Padilla, in

this case, is not retroactive. See Beard v. Banks,

(2004) (defining

finality for retroactivity purposes). The only way Teague could bar Farhane’s

ineffective assistance claim is if we not only excused the Government’s forfeiture,

but also decided that applying Padilla here, to conclude that the Sixth Amendment

required counsel to advise of the risk in this case, would itself announce a new

constitutional rule of criminal procedure.

10 Deciding that question—whether Farhane, by invoking Padilla, asks us to

announce a new rule—presents no opportunity for judicial restraint or

constitutional avoidance. The question may be framed in procedural terms, but

the answer turns on substance. “‘A case announces a new rule . . . when it breaks

new ground or imposes a new obligation’ on the government.” Chaidez,

(alterations accepted) (quoting Teague,

). We “announce[] a

new rule if the result was not dictated by precedent existing at the time the

defendant’s conviction became final,” meaning it would not “have been apparent

to all reasonable jurists.”

(internal quotation marks and citations omitted).

Importantly, however, “a case does not ‘announce a new rule, when it is merely an

application of the principle that governed’ a prior decision to a different set of

facts.”

at 347–48 (alterations accepted) (quoting Teague,

). This

is why “garden-variety applications of the test in Strickland, for assessing claims of

ineffective assistance of counsel do not produce new rules.”

. “[W]hen

all we do is apply a general standard to the kind of factual circumstances it was

meant to address, we will rarely state a new rule for Teague purposes.”

These intricate standards offer no chance to avoid the substance of Padilla

and its Sixth Amendment implications here. If Padilla had been decided after

11 Farhane’s conviction became final, then excusing the Government’s forfeiture of

Teague would have offered the opportunity for constitutional avoidance that some

of my colleagues seek. See

(holding that Padilla “announced a new rule”).

In those circumstances, at least post-Chaidez, the Teague analysis would have

amounted to a quick confirmation of dates. Here, though, where excusing the

Government’s forfeiture leads us only to questions of what a constitutional

decision like Padilla dictates, or to what kinds of facts it was meant to address—

questions that ultimately determine the Sixth Amendment’s present reach—we do

not have that opportunity. 7

There is no avoiding the constitutional question here. “The right of one

charged with crime to counsel may not be deemed fundamental and essential to

fair trials in some countries, but it is in ours.” Gideon v. Wainwright,

,

7My colleagues contend that the majority’s application of Padilla announces a new constitutional rule of criminal procedure for purposes of Teague. Dissent (Park, J.) at 1. I note that their conclusion may affect our general preference that defendants raise ineffective assistance claims, often dependent on evidence outside the record, on collateral, not direct, review. See United States v. Doe,

, 154–56 (2d Cir. 2004); see also Massaro v. United States,

(2003). The Supreme Court has made clear that “appl[ying] Strickland’s general standard to yet another factual situation” does not announce a new rule. Chaidez v. United States,

, 348–49 (2013). Therefore, we have generally not had to worry that if we decline to hear an ineffective assistance claim (based on Strickland) on direct appeal, we permit the Government to then bar that claim by raising Teague on collateral review. Given that some may conclude that applying Padilla in certain circumstances amounts to announcing a new rule for purposes of Teague, we may wish to consider that factor when deciding whether to hear, on direct appeal, ineffective assistance claims that invoke Padilla. 12 344 (1963). Padilla tells us that, when defendants “face possible exile from this

country,” we cannot apply the direct/collateral framework without eroding that

right, a right our country extends to all, regardless of citizenship or path to

citizenship. See Padilla,

. Farhane’s case is different from Padilla, but

not in a way that asks for a different approach to understanding what the Sixth

Amendment owes the accused. I vote to vacate and remand.

13 20-1666 (en banc) Farhane v. United States

PÉREZ, Circuit Judge, joined by LEE, ROBINSON, NATHAN, and MERRIAM, Circuit Judges, concurring:

Padilla v. Kentucky stood for a simple proposition: to provide effective

assistance under the Sixth Amendment, “counsel must inform her client whether

his plea carries a risk of deportation.”

(2010). Padilla and its

progeny recognize no exceptions to this rule. Yet the government here asks us to

exclude a sweeping class of people from Padilla’s holding: naturalized citizens.

The majority opinion rightly declines to do so.

Padilla’s reasoning was powerful in its simplicity: deportation is a severe

penalty, and the risk of deportation increases when certain defendants plead

guilty to certain offenses. See

. Before assuming such a risk, those

defendants must be apprised of it by their lawyers. See

. While I concur

in full in the majority opinion, I write separately to emphasize that this case breaks

no new ground—all of the pertinent legal questions in this case were asked and

answered in Padilla, including the professional norms expected of Farhane’s

counsel at the time of Farhane’s guilty plea.

I. Padilla Controls This Case.

As amply and ably explained by the majority opinion and Judge Wesley’s

concurrence, Padilla and its progeny carve out no exceptions as to when Padilla’s protections apply or whom it protects: when a guilty plea increases a defendant’s

risk of deportation, every defendant, no matter their citizenship status, must be

informed of that risk.

Of course, it is correct that there are a number of important consequences of

a guilty plea about which lawyers are not constitutionally required to warn their

clients. See Walker Dissenting Op. at 2 n.2. But those circumstances have no

bearing on this matter because the Supreme Court already determined that

deportation is different. Noting the severity of deportation, the Padilla Court

described it as “the equivalent of banishment or exile.” Padilla,

(quoting Delgadillo v. Carmichael,

, 390–91 (1947)). This logic is in line

with long-standing Supreme Court precedent reinforcing the seriousness of

subjecting any individual to possible removal from this country. See Knauer v.

United States,

(1946) (“For denaturalization, like deportation, may

result in the loss of all that makes life worth living.”) (internal quotation marks

and citation omitted)). It is the severity of the potential outcome and the

uniqueness of deportation that “underscores how critical it is for counsel to inform

[their] . . . client that he faces a risk of deportation.” Padilla, 559 U.S. at 373–74.

2 One dissenting opinion essentially asks this Court to invent a “closely tied”

test to govern when a lawyer has a duty to inform a client about a deportation

possibility after a criminal plea. Walker Dissenting Op. at 9. But Padilla does not

call on us to do so. Rather, the Court acknowledged that deportation risks resist

neat classification as either “direct” or “collateral” consequences of a guilty plea.

That is, counsel’s obligations under Padilla are not governed by an analysis of how

many steps there are from criminal plea to deportation. Instead, the Padilla Court

set forth a binary inquiry when the client is considering a guilty plea: does the plea

increase the risk of deportation?

Certainly, the Padilla Court did explain that the connection between certain

offenses and deportation muddies the distinction between direct and collateral

consequences. See

. But it is the very muddiness of the tie that

explains why it is nonsensical to treat clients who must be denaturalized before

deportation differently from those who were never naturalized, or to focus on how

many steps exist between the guilty plea and deportation. 1

1 Further, in certain cases, a conviction requires the sentencing court to denaturalize the defendant. See

(e) (“When a person shall be convicted under section 1425 of title 18 of knowingly procuring naturalization in violation of law, the court in which such conviction is had shall thereupon revoke, set aside, and declare void the final order admitting such person to citizenship, and shall declare the certificate of naturalization of such person to be canceled. Jurisdiction is conferred on the courts having jurisdiction of the trial of such offense to make such adjudication.”). 3 In sum, the distinctions a dissenting opinion is asking this Court to draw

would mean defense counsel would never be required to warn naturalized citizens

of any risk of deportation. This even includes situations where defense counsel

knew there was little doubt that deportation would follow, like in cases where an

individual will be left without a lawful immigration status after denaturalization.

This result is unsupportable under Padilla.

II. Teague Does Not Apply Because This Case Establishes No New Rule.

As discussed both in the majority opinion and Judge Wesley’s concurring

opinion, the rule set forth in Padilla was clear. Since the risk of denaturalization is

a risk of deportation, no new rule is needed to reach a decision in this case. See

Chaidez v. United States,

(2013) (“A case announces a new rule . . .

when it breaks new ground or imposes a new obligation on the government.”

(internal quotation marks and citation omitted)). And absent a new rule, Teague v.

Lane,

(1989), is inapplicable. See

(holding that a new

rule of criminal procedure should not be applied retroactively to cases on collateral

review).

The majority opinion concluded that the government forfeited any Teague

argument. I have no qualms with that conclusion. I note, however, that even if

we were to consider the Teague argument, Farhane has the better of the 4 debate. Contrary to the assertion in one of the dissenting opinions, see Park

Dissenting Op. at 1–3, for the purposes of Teague, “a case does not announce a new

rule, when it is merely an application of the principle that governed a prior

decision to a different set of facts.” Chaidez, 568 U.S. at 347–48 (internal quotation

marks and citation omitted). That dissenting opinion overreads the “all

reasonable jurists” standard observed in Chaidez and other cases when implying

that Teague applies when there is even a suggestion that a rule may be novel. See

Park Dissenting Op. at 2. In practice, the Supreme Court has found that a rule

breaks new ground where, for example, it contradicts uniform practice, see Chaidez,

568 U.S. at 353–54, or where the courts of appeals have taken “differing positions,”

Butler v. McKellar,

(1990). These circumstances are not present

here. Rather, with respect to Farhane’s case, “all we do is apply a general standard

to the kind of factual circumstances it was meant to address.” Chaidez,

; see also Stringer v. Black,

, 228–29 (1992) (explaining that “it would

be a mistake to conclude” that a holding in a previously decided case “was

limited” to the precise facts of that prior case).

5 III. Failure to Advise Farhane of the Risk of Deportation Satisfies the Deficient Performance Prong of Strickland v. Washington.

This Court correctly remands to the district court the question of whether

Farhane’s trial counsel rendered ineffective assistance per the first prong of the

two-pronged analysis set forth in Strickland v. Washington,

(1984). But

I believe Padilla makes the district court’s task on remand easier than the majority

opinion suggests because the Padilla Court already told us what constitutes

deficient performance per Strickland: failing to advise Farhane of his deportation

risks.

A. The Padilla Court Determined the Relevant Professional Norms Applicable in Farhane’s Case.

The deficient performance prong of Strickland asks whether counsel’s

representation “fell below an objective standard of reasonableness.”

.

Reasonableness is dictated by prevailing professional norms at the time of the

representation. See

When the Padilla Court ruled on the professional norms in existence in 2002,

it helped clarify what norms already existed by the time Farhane’s counsel advised

Farhane in 2006.

6 In determining the professional norms and conduct applicable at the time

of Padilla’s 2002 guilty plea, the Padilla Court relied on authorities that studied

professional norms and conduct from as early as 1995. See

. For

example, it looked to a publication from 1997 that emphasized the need for defense

counsel to “advise the defendant of . . . [a]ll of the consequences and ramifications

of a particular plea, including possible . . . effects on . . . immigration status.” G.

Nicholas Herman, Plea Bargaining § 3.03, p. 20–21 (1997). It also relied on the 1999

American Bar Association Standards for Criminal Justice, which called on counsel

to be “familiar with the basic immigration consequences that flow from different

types of guilty pleas” and to “keep this in mind in investigating law and fact and

advising the client.” ABA Standards for Criminal Justice: Pleas of Guilty,

Commentary to Standard 14-3.2(f) (3d ed. 1999). 2

The Padilla Court pointed not only to outside authorities, but also to its own

2001 decision in INS v. St. Cyr, where it emphasized the importance of considering

immigration consequences when deciding whether to accept a guilty plea. See

2 Additionally, Farhane’s counsel at oral argument pointed to a 2003 publication titled “Criminal Defense of Immigrants” which raised that a denaturalization risk could arise “from post-naturalization pleas to pre-naturalization conduct.” May 22, 2024 Hr’g Tr. 5:14–18. This source is also cited by Padilla in its discussion of applicable professional norms. See

. 7 Padilla,

. In St. Cyr, the Supreme Court cited a 1999 publication on

Criminal Defense Techniques which indicated that “[p]reserving [a] client’s right

to remain in the United States may be more important to the client than any

potential jail sentence.” INS v. St. Cyr,

(2001) (quoting 3 Bender,

Criminal Defense Techniques §§ 60A.01, 60A.02[2] (1999)).

After this thorough survey of contemporaneous professional norms and

conduct, the Padilla Court determined that, since at least 1995, counsel had to

inform clients of the risk of deportation, and because Padilla’s lawyer had not, the

deficient performance prong of Strickland was met. By remanding only on the

second prong of Strickland, the Padilla Court made unequivocally clear that the

professional norms at the time Padilla entered his plea required counsel to advise

their client on any deportation risks that accompany a guilty plea. See 559 U.S. at

367–69.

Farhane’s guilty plea was entered four years after Padilla’s guilty plea.

Given the timing of Padilla and Farhane’s respective guilty pleas, the prevailing

professional norms determined by the Padilla Court had been in force for four

more years. To conclude that there is an open question as to what the prevailing

norms were at the time Farhane’s plea was entered in 2006 requires ignoring

8 Padilla’s holding, reasoning, and analysis. That is not an exercise this Court, or the

district court on remand, is authorized to undertake.

A dissenting opinion makes too much of the sources for professional norms

cited to in Padilla not using the specific word “denaturalization.” See Walker

Dissenting Op. at 13–15.

Denaturalization is naturally encompassed within immigration risks. More

importantly, it is often synonymous with deportation. Farhane’s case is one such

example. The government is affirmatively seeking to denaturalize Farhane. See

generally App’x at A306–A321. And the government has said it rarely pursues

denaturalization when deportation is not the end goal. See Anthony D. Bianco et

al., Civil Denaturalization: Safeguarding the Integrity of U.S. Citizenship, 65 U.S. Attys’

Bull. 5, 17 (July 2017) (“[T]he government does not expend resources on civil

denaturalization actions unless the ultimate goal is the removal of the defendant

from the United States.”). The relevant inquiry in Padilla is whether deportation

is a risk. The sources cited to are concerned about the outcome of deportation. It

is not surprising that the sources are not embroiled in the intricacies of how to get

there.

9 A dissenting opinion also makes much of the fact that denaturalization is

relatively uncommon, at least compared to post-conviction removal of non-

citizens. See Walker Dissenting Op. at 12–13. Even if true, this argument misses

the mark in two respects.

First, the frequency of denaturalization proceedings—which may well

pertain to a narrower set of criminal defendants than deportation—does not carry

the force that a dissent claims. A defendant’s right to competent counsel does not

depend on the uniqueness of his circumstances. To render effective assistance, a

lawyer must “inform[] himself or herself fully on the facts and the law” as

applicable to the defendant’s particular case. ABA Standards for Criminal Justice,

Prosecution Function and Defense Function, Standard 4-5.1(a) (3d ed. 1993). “The

starting point and continuing focal point of the plea bargaining process is, of

course, the client.” 2 Crim. Prac. Manual § 45:3 (West 2024). For Sixth Amendment

purposes, what matters is the effect of a guilty plea on a particular defendant’s risk

of deportation; unimportant is the precise means by which that risk might

materialize.

Second, assistance can be ineffective when it makes the government’s ability

to deport the defendant much easier, even if the government ultimately does not

10 exercise its prosecutorial prerogative. After all, Padilla’s deportation was only

“automatic” in a legal sense, not in a practical one. As the Supreme Court

recognized in a different case, “the Executive Branch does not possess the

resources necessary to arrest or remove all of the noncitizens covered by” the

statutes that mandated Padilla’s removal. United States v. Texas,

(2023). “That reality is not an anomaly—it is a constant.”

So, although Padilla’s

plea gave the government a nearly “automatic” right to deport him, his risk of

actual deportation was far from certain. Nonetheless, because his guilty plea

increased his deportation risk, his lawyer was required to apprise him of that fact.

Here, Farhane’s guilty plea similarly made the government’s burden to

deport him much lighter, necessarily increasing his risk that the government

would successfully seek his deportation. The historic frequency with which the

government sought to denaturalize defendants after conviction does little to

change that fact; as the dissent recognizes, “[o]ur Sixth Amendment analysis

should not depend on the varying prosecutorial priorities of the Department of

Justice.” Walker Dissenting Op. at 8 n.7.

B. Examining the Evidentiary Record Is Now Left to the District Court.

In my view, this means that with respect to the deficient performance prong

of Strickland, the district court need only determine whether Farhane’s trial counsel 11 warned him that there were potential immigration consequences. 3 If Farhane was

not so warned, that would be sufficient to establish deficient performance under

Strickland’s first prong.

This Court today correctly left this determination to the district court. It is

not the job of this Court to conduct this analysis without giving the parties the

opportunity to develop the record.

Yet, a dissenting opinion conducts its own Strickland analysis and argues

that the record lacks sufficient evidence that Farhane’s counsel knew of “any of the

circumstances” or the “timing” of Farhane’s denaturalization. See Walker

Dissenting Op. at 11. But there are several pieces of information that could be

3 Inarguably, the obligation that Padilla imposed upon attorneys like Farhane’s counsel is manageable.

Immigration law, as the Padilla Court acknowledged, is a complex field. See

.

The elegance of Padilla is that it avoids all the complication with a plain rule for attorneys to warn clients when she or he faces an increased risk of deportation from a guilty plea. In determining that the professional obligation to warn attaches where there is a risk of deportation, Padilla does not encumber defense counsel with the obligations of knowing all the ins and outs of immigration law, such as what exact proceedings are required to effectuate that deportation. See

(sufficiently effective counsel “need do no more than advise a noncitizen client that pending criminal charges may carry a risk of adverse immigration consequences”); see also

(Alito, J., concurring) (defense counsel should advise their client of potential adverse immigration consequences and encourage them to seek the assistance of immigration counsel for additional advice).

All Farhane’s counsel was required to do here under Padilla was to advise him of the immigration risks he faced pursuant to his guilty plea. 12 relevant to the district court’s Strickland analysis. For example, Farhane’s counsel’s

statements at the November 2, 2005 detention hearing, a year before the guilty

plea, make plain that he knew Farhane was a naturalized citizen. See App’x at

A128:24–25 (Farhane’s counsel stating that “[Farhane’s] an immigrant . . . [h]e’s a

naturalized United States citizen.”). At the time of the detention hearing,

Farhane’s counsel also knew the immigration status of Farhane’s kids, i.e., that two

of his six children were born in the United States. See

at A128:17–19.

Additionally, Farhane’s counsel was aware of the conduct to which Farhane was

pleading guilty and the dates of those alleged crimes. See

at A128:3–15

(Farhane’s counsel discussing the allegations in the complaint);

at A178:20–

A179:3 (Farhane confirming to the district court judge at the beginning of his guilty

plea proceedings that he and his counsel had reviewed the charges against him).

These few key pieces of information shed much light on what Farhane’s counsel

knew. 4 It is up to the district court on remand to determine whether Farhane’s

counsel advised Farhane about his risk of deportation if he pleaded guilty.

4 Farhane’s counsel reiterated many of these facts in his Memorandum of Law in Aid of Sentencing. See App’x at A203 (Farhane’s counsel stating in the Memorandum that “Mr. Farhane his [sic] a 52 year-old naturalized United States citizen who has lived in Brooklyn, New York for over 20 years.”); see also

at A203–05 (emphasizing the immigration status of Farhane’s wife and kids, i.e., that his wife was “also a naturalized United States citizen,” and that his two youngest children were “native born United States citizen[s]” while three of his four oldest children were “naturalized United States citizen[s]” (emphasis added)). 13 * * *

The Sixth Amendment right to effective assistance of counsel is an integral

part of our justice system. This right cannot be fulfilled without advising clients

when a guilty plea puts them at risk of deportation. Padilla established this, and

in doing so concluded that Padilla’s counsel’s failure to so advise Padilla was

inconsistent with the prevailing professional norms of the day. On remand, the

district court needs to look no further than those Padilla norms.

14 20-1666 (en banc) Farhane v. United States

WALKER, Circuit Judge, joined in full by SULLIVAN, PARK, and MENASHI, Circuit Judges, and joined as to Part I by NARDINI, Circuit Judge, dissenting:

The majority insists that this case should be decided by a rote application of the Supreme Court’s decision in Padilla v. Kentucky,

(2010). But this reductive approach ignores critical distinctions between Padilla and the present case—distinctions that make all the difference if we carefully follow the reasoning of Padilla.

Padilla “breach[ed] the previously chink-free wall between direct and collateral consequences” by holding that, because “[d]eportation . . . is ‘unique,’” “criminal defense attorneys must inform non-citizen clients of the risks of deportation arising from guilty pleas.” Chaidez v. United States,

, 345– 46, 352–53 (2013) (quoting Padilla,

). In concluding that this case is a direct application of Padilla, the majority overlooks important differences between a noncitizen facing “virtually mandatory” deportation as a direct consequence of his guilty plea, Padilla,

, and a naturalized citizen facing possible civil denaturalization—and then, as an even more remote and uncertain consequence, potential deportation. The majority then remands this case to the district court to conduct another Strickland analysis, even though (1) the district court has already correctly determined that Farhane’s counsel’s representation was not objectively unreasonable 1 and (2) the record before us plainly establishes that Farhane was not prejudiced by his counsel’s purported ineffectiveness.

Because this case fundamentally differs from Padilla on the threshold Sixth Amendment issue—and because, in any event, Farhane’s representation was not constitutionally deficient, nor did it cause him prejudice—I respectfully dissent.

1 See Farhane v. United States, No. 18-CV-11973,

, at *2 (S.D.N.Y. Mar. 31, 2020). I. The Threshold Sixth Amendment Question

The majority contends that the Sixth Amendment obligates defense counsel to warn naturalized citizen clients of both denaturalization and deportation risks. I address each conclusion in turn.

A. Civil Denaturalization

The Supreme Court has recognized that, in determining the scope of the Sixth Amendment’s guarantee of effective assistance of counsel, “state and lower federal courts [have] almost unanimously concluded that . . . attorneys [are not required] to inform their clients of a conviction’s collateral consequences,” meaning those consequences that are not “a component of the criminal sentence.” Chaidez, 568 U.S. at 349–50; see also Padilla,

& n.8 (noting that, although “[t]here is some disagreement among . . . courts over how to distinguish between direct and collateral consequences,” collateral consequences tend to be “those matters not within the sentencing authority of the [criminal] court”). 2

Nonetheless, the Court determined in Padilla, a case involving a noncitizen, that the direct/collateral distinction was “ill suited to evaluating” whether the “specific risk of deportation” was covered by the Sixth Amendment duty to warn. Padilla,

. It did so because of (1) deportation’s “particularly severe” nature and (2) the “uniquely” “close connection [between] the criminal process” and “[d]eportation as a consequence of a criminal conviction.”

at 365–66

2 See also Jenny Roberts, Ignorance Is Effectively Bliss: Collateral Consequences, Silence, and Misinformation in the Guilty-Plea Process, 95 IOWA L. REV. 119, 124 (2009) (explaining that, under the direct/collateral framework, defense counsel must inform pleading clients of the “‘direct,’ or penal, sanctions—such as jail or prison time, probationary period or a fine—[that may] result from [a] conviction,” but need not inform clients of the collateral consequences of a conviction, such as “sex-offender registration, . . . involuntary civil commitment . . . , the loss of voting rights, . . . the loss of housing and employment opportunities,” and, before Padilla, deportation); Padilla,

(Alito, J., concurring in the judgment) (noting that collateral consequences include “civil commitment, civil forfeiture, the loss of the right to vote, disqualification from public benefits, ineligibility to possess firearms, dishonorable discharge from the Armed Forces, and loss of business or professional licenses”).

2 (emphasis added). While civil denaturalization is undoubtedly a severe penalty, it is not “intimately related to the criminal process,”

, since, unlike most avenues to deportation, civil denaturalization is almost never dependent on a criminal conviction. Understanding this critical difference between deportation and denaturalization requires taking a closer look at the statutory scheme of each consequence.

Although noncitizens can be deported for reasons other than having been criminally convicted, 3 “[f]ederal immigration enforcement turns increasingly on . . . criminal convictions triggering deportability.” 4 Being “convicted of,” rather than simply committing, a broad swath of criminal offenses is what renders noncitizens “deportable.” 5 Thus, a criminal conviction is strictly necessary, not just sufficient, for many forms of deportation.

3 See, e.g.,

(a)(1), (a)(3), (a)(5), (a)(6).

4 Note, States’ Commandeered Convictions: Why States Should Get a Veto Over Crime-Based Deportation, 132 HARV. L. REV. 2322, 2322 (2019); see

(“[C]onviction-based deportations skyrocketed from fewer than 1000 through 1984 to almost 70,000 by 1999 and more than twice that today.” (internal citations omitted)).

5 See, e.g.,

(a)(2)(A)(i)(I), (a)(2)(A)(ii) (crimes “involving moral turpitude”);

at (a)(2)(A)(i)(II) (crimes for which a sentence of at least a year may be imposed);

at (a)(2)(A)(iii) (aggravated felonies);

at (a)(2)(A)(iv) (high speed flight from an immigration checkpoint);

at (a)(2)(A)(v) (failure to register as a sex offender);

at (a)(2)(B)(i) (most controlled substance offenses);

at (a)(2)(C) (certain firearm offenses);

at (a)(2)(D)(i) (offenses related to espionage, sabotage, and treason and sedition);

at (a)(2)(D)(ii) (offenses related to threatening the President, Vice President, and other political officials via mail);

at (a)(2)(D)(iii) (violations of the Military Selective Service Act or the Trading With the Enemy Act);

at (a)(2)(D)(iv) (violations of certain travel-control restrictions or importation of a noncitizen for an immoral purpose);

at (a)(2)(E)(i) (domestic violence, stalking, child abuse, child neglect, or child abandonment);

at (a)(3)(B)(i) (knowingly filing registration application with false statements or attempting to procure registration through fraud);

at (a)(3)(B)(ii) (violations of the Foreign Agents Registration Act of 1938);

at (a)(3)(B)(iii) (violations relating to fraud and misuse of visas, permits, and other entry documents).

3 By contrast, civil denaturalization is virtually never dependent on a criminal conviction. A citizen may be denaturalized civilly if his naturalization was either (1) procured “by concealment of a material fact or by willful misrepresentation” or (2) “illegally procured.”

(a). “Citizenship is illegally procured any time the applicant has failed to comply with any of the congressionally imposed prerequisites to the acquisition of citizenship,” United States v. Sprogis,

, 117 n.2 (2d Cir. 1985) (internal quotation marks and citation omitted), including being “a person of good moral character” for the five years immediately preceding the date of filing a citizenship application through the date the applicant is admitted to citizenship,

(a). A person cannot qualify as having good moral character if, within the statutory period, he fell within eight categories enumerated in

(f). 6 Section 1101(f) also contains a catch-all

6 Under

(f), a person cannot have the requisite good moral character if, during the period for which good moral character was required, he was:

(1) a habitual drunkard;

(2) Repealed.

(c)(1), Dec. 29, 1981,

.

(3) a member of one or more of the classes of persons, whether inadmissible or not, described in paragraphs (2)(D), (6)(E), and (10)(A) of section 1182(a) of this title; or subparagraphs (A) and (B) of section 1182(a)(2) of this title and subparagraph (C) thereof of such section (except as such paragraph relates to a single offense of simple possession of 30 grams or less of marihuana), if the offense described therein, for which such person was convicted or of which he admits the commission, was committed during such period;

(4) one whose income is derived principally from illegal gambling activities;

(5) one who has been convicted of two or more gambling offenses committed during such period;

(6) one who has given false testimony for the purpose of obtaining any benefits under this chapter;

(7) one who during such period has been confined, as a result of conviction, to a penal institution for an aggregate period of one hundred and eighty days or more, regardless of

4 provision, which states that “[t]he fact that any person is not within any of the foregoing classes shall not preclude a finding that for other reasons such person is or was not of good moral character.”

Pursuant to § 1101(f), the Department of Homeland Security promulgated regulation

, which includes additional circumstances under which an applicant “shall be found to lack good moral character.”

(b)(1)– (3). The regulation contains its own catch-all provision, which instructs that, “[u]nless the applicant establishes extenuating circumstances,” he shall be found to lack good moral character if he “[c]ommitted unlawful acts that adversely reflect upon [his] moral character, or was convicted or imprisoned for such acts.”

at § 316.10(b)(3)(iii). Therefore, in the vast majority of circumstances, a criminal conviction is not necessary for civil denaturalization, which can instead result from the offender’s underlying behavior. Moreover, criminal conduct may have no impact on a naturalized citizen’s citizenship if it occurs after naturalization.

Jose Padilla and Abderrahmane Farhane themselves exemplify this distinction between deportation and civil denaturalization. Padilla could not have been deported under

(a)(2)(B)(i) for the conduct of distributing drugs, but he automatically became “deportable” under that provision the instant he was “convicted of ” distributing drugs.

(a)(2)(B)(i) (emphasis added). By contrast, Farhane’s conviction did not render him denaturalizable.

whether the offense, or offenses, for which he has been confined were committed within or without such period;

(8) one who at any time has been convicted of an aggravated felony (as defined in subsection (a)(43)); or

(9) one who at any time has engaged in conduct described in section 1182(a)(3)(E) of this title (relating to assistance in Nazi persecution, participation in genocide, or commission of acts of torture or extrajudicial killings) or 1182(a)(2)(G) of this title (relating to severe violations of religious freedom).

(f) (internal footnote omitted).

5 Farhane was subject to denaturalization at least five years before he ever pled guilty: he supplied the grounds for his denaturalization in 2001, when he conspired to commit money laundering, and again in 2002, when he lied to a U.S. Citizenship and Immigration Services (“USCIS”) officer by denying, multiple times, that he had ever “knowingly committed any crime for which he had not been arrested.” Gov’t Br. at 5. Thus, the government sought Farhane’s denaturalization not on the basis of his conviction but for: (1) “commit[ing]” the crime of conspiracy, an act “that adversely reflected on his moral character,” App’x 315–16 (citing

(b)(3)(iii)); (2) giving false testimony under oath for the purpose of obtaining naturalization,

at 316–17 (citing

(f)(6) and

(b)(2)(vi)); and (3) procuring his naturalization by concealment of material facts and by willful misrepresentation,

at 318–19 (citing

(a)). In short, even if Farhane had gone to trial and been acquitted, the government could still have sought to denaturalize him.

Farhane’s conviction serves only to collaterally estop him from denying his participation in the conspiracy, thereby saving the government from having to prove its case for denaturalization through agent testimony and damaging recordings that captured Farhane “plotting to send money to self-proclaimed ‘mujahideen’ then waging war against the United States overseas.” Gov’t Br. at 2. The majority stresses this point, asserting that the fact that Farhane’s guilty plea makes it “easier” for the government to denaturalize him renders the risk of denaturalization a “sufficiently automatic consequence” of the plea. Maj. Op. at 23. But the fact that the government now has a somewhat easier path to prove an already strong case for denaturalization does not make this case like Padilla, where the conviction instantly rendered Padilla deportable. See Padilla,

(noting that Padilla’s guilty plea “made his deportation virtually mandatory”). Farhane can still assert defenses to each of the three bases for denaturalization, none of which is foreclosed by estoppel. He could, for example, contend that: (1) “extenuating circumstances” excuse any bad acts that would otherwise demonstrate his lack of good moral character,

(b)(3); (2) his false

6 testimony was not “made . . . with an intent to obtain an immigration benefit” but instead was made unintentionally or out of embarrassment, fear, or a desire for privacy,

(b)(2)(vi); or (3) his misrepresentations were not “willful,”

(a). See Kungys v. United States,

, 780–81 (1988) (indicating that a lie told in the naturalization process out of “embarrassment, fear, or a desire for privacy” is not a sufficient basis for denaturalization under

(f)(6)); Maslenjak v. United States,

, 345–46 (2017) (same but under

(a)). In other words, not only is Farhane’s guilty plea not necessary for the government to denaturalize him; it is not even sufficient.

The fact that a criminal conviction is frequently necessary for deportation while, in almost every case, civil denaturalization can occur without a conviction explains why Padilla’s logic does not extend to civil denaturalization. Padilla painstakingly emphasized the degree of “enmesh[ment] [between] criminal convictions and the penalty of deportation,” Padilla, 559 U.S. at 365–66, concluding that deportation was “uniquely difficult to classify as either a direct or a collateral consequence” because of the “close connection” between “[d]eportation as a consequence of a criminal conviction [and] the criminal process,”

(emphasis added). While criminal convictions now make deportation “virtually inevitable for a vast number of noncitizens,”

, who, without a conviction, often would not be eligible for deportation, civil denaturalization does not share the same reliance on—and therefore the same enmeshment with—criminal convictions. This is why the Supreme Court emphasized that “[d]eportation . . . is ‘unique,’” Chaidez,

, and that “Padilla . . . did not eschew the direct-collateral divide across the board,”

. Thus, Padilla’s narrow deportation exception to the otherwise “chink-free wall between direct and collateral consequences,” Chaidez, 568 U.S. at 352–53—a framework which this court has indicated continues to operate post-Padilla, see United States v. Youngs,

(2d Cir. 2012) (Fifth Amendment context); Santiago v. Laclair,

(2d Cir. 2014) (summary order) (Sixth Amendment context)—should not be expanded to civil denaturalization.

7 b. Deportation Following Civil Denaturalization

Not confident in the notion that denaturalization standing alone should be removed from the direct/collateral framework, the majority also asserts that “[a] risk of denaturalization simply is a risk of deportation.” Maj. Op. at 16. 7 But nothing in Padilla requires that a naturalized citizen like Farhane be advised of the potential risk of deportation that might follow if, and only if, he is first civilly denaturalized. Despite the broad language of Padilla’s holding that “counsel must inform her client whether his plea carries a risk of deportation,”

, the rest of the opinion’s language and its core reasoning do not support requiring such advice when counsel advises a naturalized citizen client. Consistent language throughout the opinion makes clear that the Court was focused only on what advice the Sixth Amendment requires defense attorneys to offer noncitizen clients. For example, the opinion described how changes in federal immigration law have made removal “practically inevitable” for “noncitizen[s],”

at 363–64 (emphasis added), discussed how being informed of deportation risk can benefit “noncitizen defendants,”

(emphasis added), and stressed “how critical it is for counsel to inform her noncitizen client [of deportation risks],”

at 373–74 (emphasis added).

More importantly, Padilla’s reasoning confirms that its Sixth Amendment holding should not apply to counsel advising a naturalized citizen client at risk of deportation only if he is first civilly denaturalized. It is because of changes in immigration law that “have made removal nearly an automatic result for a broad class of noncitizen offenders” that the Court found it “most difficult to divorce the penalty from the conviction in the deportation context.”

(emphases

7To support this proposition, the majority relies on a statement in an article in a 2017 U.S. Attorneys’ Bulletin that the government “[t]ypically . . . does not expend resources on civil denaturalization actions unless the ultimate goal is . . . removal . . . .” Maj. Op. at 16 n.6 (quoting Anthony D. Bianco et al., Civil Denaturalization: Safeguarding the Integrity of U.S. Citizenship, 65 U.S. Attys’ Bull., July (I) 2017, at 5, 17). That may indeed have been the case historically, but of course would be subject to change depending on the presidential administration. Our Sixth Amendment analysis should not depend on the varying prosecutorial priorities of the Department of Justice.

8 added) (internal quotation marks omitted); see

(noting that Padilla’s guilty plea made his “deportation virtually mandatory”). But no equivalent automaticity exists in those exceedingly rare cases where a citizen faces a risk of deportation that is first contingent upon a separate civil denaturalization proceeding. If it is true that the Sixth Amendment does not require defense attorneys to advise clients on civil denaturalization risk because civil denaturalization is not as closely tied to a criminal conviction as is deportation, then it would be nonsensical to conclude that the Sixth Amendment requires counsel to advise of a deportation risk that is entirely dependent on a client first becoming civilly denaturalized. Deportation under such circumstances simply cannot be deemed “nearly an automatic result.”

. To conclude otherwise would be to strip Padilla’s holding from its careful, contextual reasoning.

In sum, Padilla’s core reasoning does not apply in Farhane’s case, in which the preexisting denaturalization risk was independent of and at most collateral to his conviction and any deportation risk was entirely dependent on Farhane first becoming denaturalized. I therefore see no reason to forgo the direct/collateral distinction in this case, much less impose an additional Sixth Amendment burden on defense counsel.

II. Strickland’s Ineffective-Assistance Test

Even if the Sixth Amendment requires defense counsel to warn naturalized citizen clients of potential denaturalization and deportation risks, Farhane’s ineffective assistance of counsel claim nonetheless fails because he has not shown (1) that “counsel’s representation fell below an objective standard of reasonableness,” or (2) that he suffered prejudice, meaning “there is a reasonable probability that, but for counsel’s [alleged] unprofessional errors, the result of the proceeding would have been different.” Strickland v. Washington,

(1984).

9 A. The Deficient-Performance Prong

I agree with the district court (Preska, J.) that “it was not objectively unreasonable for [Farhane’s] lawyer to give no advice on [the immigration consequences]” of Farhane’s guilty plea and see no reason to remand this issue, which is fully briefed before us. Farhane v. United States, No. 18-CV-11973,

, at *2 (S.D.N.Y. Mar. 31, 2020); see also Sepulveda v. Block,

(2d Cir. 1986) (concluding that there was “no need for a remand” when an issue did not “require further fact-finding or ventilation”); United States v. Kurti,

(2d Cir. 2005) (rejecting ineffective-assistance claim where record, “[v]iewed in its entirety,” showed that attorney’s performance “fell within prevailing professional norms” (internal quotation marks omitted)).

Strickland’s deficient-performance inquiry measures “whether an attorney’s representation “amounted to incompetence under ‘prevailing professional norms.’” Harrington v. Richter,

(2011) (quoting Strickland,

). This inquiry must be: (1) fact-specific, with any conclusion drawn “in light of all the circumstances,” Strickland,

(emphasis added); (2) “highly deferential” to counsel’s conduct,

; and (3) made without “the distorting effects of hindsight,”

It is not enough that the attorney’s conduct might have “deviated from best practices or most common custom.” Harrington,

; see also Strickland,

(noting that the purpose of the right to effective counsel is not to improve the quality of legal representation). Rather, Strickland’s first step is met only if Farhane’s attorney’s failure in 2006 to provide his client advice regarding denaturalization (or the possibility of deportation following denaturalization) prior to his guilty plea was “so serious” a mistake as to fall outside “the wide range of professionally competent assistance.” Strickland,

. That standard is not met here.

In remanding the question of deficient performance to the district court, the majority conflates that question with the threshold Sixth Amendment issue discussed in Section I of this dissent. The majority concludes that the district court

10 “impermissibly narrowed Padilla” by considering the fact that Farhane’s conviction did not create “an imminent risk of deportation,” when, in the majority’s view, the Sixth Amendment requires counsel to inform naturalized citizen clients of even remote risks of deportation. Maj. Op. at 34 (quoting Farhane,

, at *2). But the district court appropriately made that point as part of its deficient performance analysis, explaining that—due to the fact that Farhane’s denaturalization risk stemmed from his conduct prior to pleading guilty, rather than from the guilty plea itself—“Farhane’s lawyer had no basis for suspecting that the guilty plea could have immigration consequences.” Farhane,

, at *2. In essence, Farhane failed to provide his counsel with the facts necessary to recognize the indirect denaturalization risk (and potential subsequent deportation risk) posed by his conviction. In Padilla, by contrast, Padilla’s deportation risk was “nearly an automatic result,”

(emphasis added), of his plea, such that his counsel needed no additional factual background to appreciate that risk.

“The reasonableness of counsel’s actions may be determined or substantially influenced by the defendant’s own statements or actions.” Strickland,

. Thus, “[a] trial counsel cannot be ineffective for failing to raise claims as to which his client has neglected to supply the essential underlying facts when those facts are within the client’s possession . . . .” Dooley v. Petsock,

, 890–91 (3d Cir. 1987). The majority does not dispute the district court’s factual conclusion that “nothing in the record suggests [Farhane’s] lawyer knew, or should have known, about the circumstances giving rise to . . . Farhane’s denaturalization exposure.” Farhane,

, at *2. Nor could it—there is simply no evidence that, prior to pleading guilty, Farhane (or any other source) informed his defense attorney of any of the circumstances surrounding his naturalization, not even its timing. Without that information, Farhane’s attorney could not have begun to appreciate a potential risk of denaturalization or deportation, however remote those possibilities might have been. Therefore, it was not objectively unreasonable for his lawyer to give no advice on those issues.

11 See, e.g., Hill v. Lockhart,

, 61–62 (1985) (White, J., concurring in the judgment) (concluding that there could be no ineffective assistance where there was no allegation that defendant told attorney about previous conviction); Jackson v. Scully,

(2d Cir. 1986) (concluding that counsel was not ineffective for failing to raise an alleged Fourth Amendment violation where petitioner did not inform counsel of the circumstances surrounding the allegedly improper arrest).

The alternative deficient-performance analysis that the majority implicitly suggests—which would rely on the Supreme Court’s conclusion that Padilla’s counsel was constitutionally ineffective in 2002 for affirmatively misadvising his noncitizen client of deportation risk to conclude that Farhane’s counsel was similarly deficient in 2006 for failing to advise his naturalized citizen client of denaturalization and deportation risk—is misguided. See Maj. Op. at 34 n.18. For the reasons outlined below, this case fundamentally differs from Padilla on the issue of deficient performance, making dependence on that case inapposite.

First, the drastic difference in prevalence between denaturalization and deportation bears on what knowledge should be expected of a reasonable defense attorney practicing at the time of Farhane’s plea in 2006. Fewer than 150 denaturalization proceedings took place nationwide between 1968 and 2012 (a 44- year period), resulting in an average of fewer than four such proceedings per year. 8 By contrast, between 1997 and 2007 (only a 10-year period), 897,099 noncitizens were deported, for an average of approximately 27,184 deportations a year—and that figure includes only “individuals deported on criminal grounds.” 9 The

8 Amber Qureshi, The Denaturalization Consequences of Guilty Pleas, 130 YALE L.J. F. 166, 170 (2020).

9 Forced Apart (By the Numbers), Human Rights Watch (April 15, 2009), https://www.hrw.org/report/2009/04/15/forced-apart-numbers/non-citizens-deported-mostly- nonviolent-offenses [https://perma.cc/VQ3X-ZMSC] (capitalization altered); see

(noting that

12 extreme historical rarity of denaturalization makes it exceedingly unlikely that an average defense attorney in 2006, when Farhane pled guilty, would have thought to explore denaturalization as a contingent effect of a naturalized citizen’s guilty plea—let alone considered the even more tenuous consequence of deportation that was first dependent on denaturalization. By contrast, given the large number of criminally based deportations occurring annually, it would not have been uncommon in 2002, when Padilla pled guilty, for defense attorneys to know that a noncitizen client might be imminently at risk of deportation due to a criminal conviction. In sum, it was less unreasonable for Farhane’s counsel to fail to warn of an extremely uncommon risk than for Padilla’s counsel to fail to warn of a much more prevalent risk.

Second, although the materials cited by Farhane and the amici as establishing prevailing professional norms in 2006 (the “Norms Materials”) discuss at length the importance of warning pleading defendants of deportation consequences, not one states that defense attorneys should also advise clients of denaturalization consequences. 10 Instead, Farhane relies on the fact that some of these materials instructed defense counsel to advise clients about the “immigration consequences” that could follow a guilty plea. Appellant’s Br. at 34. In hindsight, one might be tempted to assume—as Judge Pérez does in concurrence—that denaturalization was one of the “immigration consequences”

20% of the 897,099 deported noncitizens had, like Jose Padilla, been living legally in the U.S. prior to their deportation).

10 See, e.g., National Legal Aid and Defender Association, Performance Guidelines for Criminal Defense Representation §§ 6.2(a)(3), 8.2(b)(8) (1995) (specifically citing deportation, but not denaturalization, as an example of a collateral consequence that defense counsel should discuss with pleading clients); American Bar Association, ABA Standards for Criminal Justice: Pleas of Guilty, at xi, 116, 125–26 & n.25 (3d ed. 1999) (“ABA 1999”) (same); Gabriel J. Chin & Richard W. Holmes, Jr., Effective Assistance of Counsel and the Consequences of Guilty Pleas, 87 CORNELL L. REV. 697, 713–18 (2002) (same); New York State Bar Association, Standards for Providing Mandated Representation Standard I-7(a), at 16 (Apr. 2, 2005) (“NYSBA”) (same); see also ABA 1999, at 36, 58 & nn.95–96, 59 & n.98 (similar but regarding courts’ advice to defendants).

13 referenced by these materials. But this assumption has no basis. We must put such assumptions aside and interpret what standards of reasonableness the Norms Materials actually evinced in 2006. See Strickland,

(“[E]very effort [must] be made to eliminate the distorting effects of hindsight . . . .”). And the fact that every such source that enumerates specific “immigration consequences” does not list denaturalization 11 provides powerful evidence that, in 2006, prevailing professional norms did not require defense counsel to warn pleading clients of the possibility of denaturalization. That conclusion is reinforced by how uncommon denaturalization was at the time.

Third, virtually all of the Norms Materials that require defense counsel to advise of “immigration consequences” only require such advice for noncitizen clients like Padilla. 12 This reinforces the point that denaturalization, which can

11 See, e.g., Katherine A. Brady et al., Representing the Noncitizen Criminal Defendant, in CALIFORNIA CRIMINAL LAW PROCEDURE AND PRACTICE 1297 § 48.2, at 1301–02 (Robert Waxman, ed., 5th ed. 2000) (“Brady et al.”) (denaturalization not included in list of immigration consequences); New York State Defenders Association Immigrant Defense Project, Deportation 101, at 23 (Feb. 2005) (“Deportation 101”) (same). Deportation 101 does mention denaturalization in passing, but only to explain the obvious fact that citizens who are denaturalized “may . . . be vulnerable to deportation.” Id. at 21 (emphasis added). More to the point, it does not list denaturalization as a “potential effect[]” or “immigration consequence[]” of a criminal conviction. Id. at 21, 23 (capitalization altered). Although we must judge ineffective assistance of counsel in this case based on the professional norms that existed in 2006, when Farhane pled guilty, post-2006 materials continued the trend of excluding denaturalization from lists of “immigration consequences.” See, e.g., NORTON TOOBY, TOOBY’S GUIDE TO CRIMINAL IMMIGRATION LAW § 3.3(A), at 48 (2008) (denaturalization not included in list of immigration consequences); AMERICAN BAR ASSOCIATION, STANDARDS FOR CRIMINAL JUSTICE § 4-5.5(c) (4th ed. 2017) (“ABA 2017”) (same).

12 See, e.g., ABA 1999, supra, at 126 (implying that defense counsel must only advise “non-citizens” of “immigration consequences”); Brady et al., supra, at 1298–1303 (stating that effective counsel must advise clients on immigration consequences in chapter titled “Representing the Noncitizen Criminal Defendant” (capitalization altered)); MATTHEW BENDER & CO., PUB. NO. 202, RELEASE 95, 3 CRIMINAL DEFENSE TECHNIQUES § 60A.01 (2008) (treatise page last updated in September 2002 stating that “[w]hen representing an alien in criminal court, an attorney must consider the effect a criminal conviction may have on the client’s immigration status”); NORTON TOOBY WITH KATHERINE A. BRADY, 1 CRIMINAL DEFENSE OF IMMIGRANTS § 1.3,

14 only happen to a naturalized citizen, is not an “immigration consequence” within the meaning of these materials. But it also undercuts the majority opinion’s conclusion that Farhane’s counsel was unreasonable for failing to advise him of potential deportation consequences: if prevailing norms in 2006 only required defense counsel to discuss deportation consequences with noncitizens, then Farhane’s counsel was not ineffective for failing to discuss such consequences with his naturalized-citizen client. Padilla, decided four years later, similarly emphasized that “[a] reasonably competent lawyer will tell a non-citizen client about a guilty plea’s deportation consequences.” Chaidez, 568 U.S. at 356–57 (emphasis added); see also Padilla,

(“When the law is not succinct and straightforward . . . , a criminal defense attorney need do no more than advise a noncitizen client that pending criminal charges may carry a risk of adverse immigration consequences.” (emphasis added)).

Fourth, while Farhane’s lawyer was not prompted to and thus did not discuss denaturalization with Farhane, Padilla alleged that his lawyer affirmatively misadvised him about deportation risk. See Padilla,

(“Padilla claims that his counsel not only failed to advise him of [the deportation] consequence prior to his entering the plea, but also told him that he ‘did not have

at 12–18 (3d ed. 2003) (stating that effective counsel must advise clients on immigration consequences in chapter titled “Criminal Defense of Noncitizens”); Deportation 101, supra, at 23 (list of “immigration consequences of convictions” only lists consequences “apply[ing] to . . . noncitizens” (capitalization altered)). Although we must judge ineffective assistance of counsel in this case based on the professional norms that existed in 2006, post-2006 norms materials indicate that requiring defense counsel to warn only noncitizen clients of potential immigration consequences continues to be a widespread practice. See, e.g., Peter L. Markowitz, Protocol for the Development of a Public Defender Immigration Service Plan, app. G at 13 (2009) (reproducing a training presentation encouraging defense attorneys to “Determine if Your Client is a Noncitizen and thus at Risk of Suffering Negative Immigration Consequences”); KARA HARTZLER, SURVIVING PADILLA: A DEFENDER’S GUIDE TO ADVISING NONCITIZENS ON THE IMMIGRATION CONSEQUENCES OF CRIMINAL CONVICTIONS (2011) (implying that only noncitizens need be advised of immigration consequences); ABA 2017, supra, at § 4-5.5(b) (same); Brittany Brown & Megan Hu, Immigration, 29 ATTICUS 21, 24–26 (2017) (same); Dawn Seibert & Isaac Wheeler, Representing Immigrant Clients: Ethics and Practice 2 (Apr. 3, 2014) (instructing defense counsel to ask clients where they were born in order to protect “noncitizen defendants”).

15 to worry about immigration status since he had been in the country so long.’ . . . Padilla relied on his counsel’s erroneous advice when he pleaded guilty to the drug charges that made his deportation virtually mandatory.” (internal citation omitted)). Although Padilla did not rely on any distinction between affirmative misadvice and keeping silent for purposes of conducting the threshold Sixth Amendment analysis of whether Strickland applies at all to deportation, see id. at 369–71, it endorsed the use of affirmative misadvice as a factor in the subsequent Strickland deficiency analysis, see id. at 368–69 (“Padilla’s counsel provided him false assurance that his conviction would not result in his removal from this country. This is not a hard case in which to find deficiency: . . . his counsel’s advice was incorrect.”). The distinction between misadvice (in Padilla) and non-advice (here) is particularly important given that before Padilla we had consistently held “that an attorney’s failure to inform a defendant of the immigration consequences of a guilty plea [did] not constitute ineffective assistance of counsel . . . , but that an attorney’s affirmative misrepresentation about the [immigration] consequences of a guilty plea [could] constitute ineffective assistance.” Creary v. Mukasey,

(2d Cir. 2008) (summary order) (first citing United States v. Santelises,

(2d Cir. 1975) (per curiam); and then citing United States v. Couto,

(2d Cir. 2002)). It would be strikingly unfair to fault Farhane’s lawyer for not advising his client in 2006 about potential denaturalization consequences when, until 2010, we informed defense counsel that they were not ineffective for not providing such advice even about potential deportation. See Maj. Op. at 4-5 (acknowledging that, prior to Padilla, we “treated adverse immigration consequences . . . as . . . perhaps beyond a criminal defense counsel’s presumptive area of expertise”).

In sum, any suggestion that Padilla should control our deficient-performance analysis here misses the mark, regardless of whether Farhane’s counsel’s conduct is framed as a failure to warn of denaturalization risk, a failure to warn of deportation risk, or both. Padilla crucially differs from this case in at least five ways: (1) Padilla’s counsel had all the information necessary to know

16 of his client’s deportation risk; (2) Padilla’s counsel did not need to understand rare denaturalization risks to be cognizant of his client’s more common deportation risk; (3) the Norms Materials discuss explicitly and at length the importance of warning pleading noncitizen defendants of deportation consequences 13; (4) Padilla was a noncitizen client; and (5) Padilla’s counsel gave affirmative misadvice. These differences in circumstance must be accounted for and, when they are, it becomes evident that Farhane has failed to overcome the “strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance.” Strickland,

; see

(“[T]he performance inquiry must be whether counsel’s assistance was reasonable considering all the circumstances.”).

B. The Prejudice Prong

Farhane also fails to demonstrate prejudice, i.e., that his “counsel’s errors were so serious as to deprive [him] of a fair trial.”

. In order to show prejudice in the plea-bargaining context, Farhane must prove either that, had he known of the denaturalization risk, (1) there would have been “a reasonable probability that [he] could have negotiated a plea that did not impact [his naturalized] status,” Doe v. United States,

(2d Cir. 2019), or (2) he would have “insisted on going to trial,” Hill,

. I see no reason to remand this issue to the district court because the existing record plainly establishes that Farhane cannot carry his burden as to either possibility. See Rainbow Line, Inc. v. M/V Tequila,

(2d Cir. 1973) (noting that, even though the district court did not make “findings of fact on a number of the issues involved” in the case, remand was “unnecessary” because the record was

13See supra note 10. Pre-2006 Supreme Court case law also indicated that reasonable counsel should warn pleading noncitizen clients of deportation consequences but did not address whether denaturalization risk warnings were similarly required for naturalized citizen clients. See Chaidez v. United States,

, 356–57 (2013) (“[INS v. St. Cyr,

(2001)] stated [that] . . . [a] reasonably competent lawyer will tell a non-citizen client about a guilty plea’s deportation consequences . . . .”).

17 “sufficiently clear on the key points”); see also Rodriguez v. United States,

, 44 (2d Cir. 2018) (summary order) (remanding the issue of prejudice to the district court because it could not be determined “[o]n the present record” “whether [the appellant] would have opted to [go to trial] had she known she could be denaturalized”).

First, the record demonstrates that Farhane could not have negotiated a plea deal that foreclosed a denaturalization risk. As the government points out, Farhane had minimal leverage with which to negotiate because of the strength of the government’s case against him and the gravity of the charges. Farhane, who accepted a plea deal that resulted in a substantial prison sentence (albeit, reduced from a maximum of 23 to 13 years) rather than go to trial, now tries to make this seem like a close case, asserting he had “several pathways to defend against the government’s charges.” Appellant’s Br. at 46. In reality, as Judge Preska stated at Farhane’s bail hearing, the “the government ha[d] very strong evidence,” including inculpatory recordings of Farhane agreeing to provide material support to terrorists. App’x 108. And at the time of his plea, Farhane’s counsel—who was aware of all the issues surrounding the government’s informant cited by the majority—admitted that Farhane had no viable defense and has not been faulted by Farhane for that admission.

Second, the record does not support—and indeed undermines—Farhane’s claim that he would have insisted on going to trial, risking an additional decade in prison and still facing potential denaturalization. “The decision whether to plead guilty . . . involves assessing the respective consequences of a conviction after trial and by plea.” Lee v. United States,

, 367 (2017). Here, Farhane’s “consequences of taking a chance at trial were . . . markedly harsher than pleading,”

, and he had very little to gain: it was almost certain that Farhane, having engaged in terrorism-related conduct, would be denaturalized anyway, whether he took the plea or went to trial (even upon an acquittal), see

(concluding that it was not inherently irrational for Lee to risk “a year or two more

18 of prison time” by going to trial when his plea agreement would otherwise guarantee deportation). Farhane has failed to present “substantial and uncontroverted” “contemporaneous evidence” that remaining in the United States was of such “paramount importance” to him that he would have forgone a ten-year sentence reduction for only a slightly increased chance of staying in the country.

at 369–71; see also Doe,

(“[A] petitioner alleging ineffective assistance based on immigration misadvice must clearly demonstrate that he placed particular emphasis on [immigration consequences] in deciding whether or not to plead guilty.” (second alteration in original) (internal quotation marks omitted)).

Farhane insists that statements he made in connection with his sentencing below about his desire to avoid familial separation indicate that he would have gone to trial had he been aware of the denaturalization risk. These statements instead demonstrate that Farhane was not concerned with potential immigration consequences, but rather was focused on minimizing his potential prison time, which the plea deal did by reducing his possible sentence by ten years. If Farhane were denaturalized and deported back to Morocco (where he was educated, previously worked, and continued to visit after his naturalization), he could take his family with him and find work to support them—but he could not have fulfilled his predominant aims of helping to raise his children and serving as the family breadwinner during the additional decade he may well have spent in prison had he gone to trial. Farhane’s general statements about wanting to avoid familial separation not only fail to show that “avoiding [immigration consequences] was the determinative factor for him,” Lee, 582 U.S. at 367: they affirmatively indicate that, at the time he pled guilty, Farhane was substantially prioritizing reducing the risk of a lengthier prison sentence over negligibly minimizing his denaturalization risk. This makes sense because Farhane could be denaturalized even if he had been acquitted, so he had nothing to gain and ten years to lose by proceeding to trial.

19 Finally, there is affirmative evidence that, around the time of his plea, Farhane did not care about his future ability to remain in the U.S. In 2004, he was recorded saying: “We should not think of this as our country. This is a country where we temporarily live,” and “[t]his is not a country where we should stay. One should work hard here and then take off. The longer one stays here, the more people make you sick.” App’x 91–92. Although Farhane dismisses these statements as “acontextual asides,” Reply Br. at 27, he offers no context (nor is there any in the record) that might suggest that he was expressing anything other than, at best, a lack of commitment to his continued residence in the United States.

In sum, I see no reason to remand this issue because the ample evidence in the existing record is sufficient for us to confidently conclude that Farhane was not prejudiced by his counsel’s performance.

CONCLUSION

Padilla’s break from the longstanding and widespread direct/collateral distinction was premised on the uniquely intimate relationship between criminal convictions and the deportation of noncitizens. That same close relationship simply does not exist in the case of civil denaturalization—and thus also cannot exist in instances where deportation is first predicated on civil denaturalization. Therefore, it is misguided to eschew the direct/collateral framework in this case on the basis of Padilla. I would instead adhere to that framework and hold that the Sixth Amendment does not require defense counsel to warn naturalized citizen clients of the collateral risks of denaturalization and (if denaturalized) deportation.

Even assuming, arguendo, that the Sixth Amendment does require such a warning, Farhane’s ineffective assistance of counsel claim fails both prongs of Strickland’s test. As the district court correctly concluded, Farhane’s representation was not objectively unreasonable because Farhane failed to give his attorney the factual information necessary to understand how Farhane’s plea might pose a future risk of denaturalization or deportation. Any suggestion that

20 the deficient-performance inquiry here should be synonymous with the one in Padilla ignores the numerous factors that make Farhane’s counsel’s representation much more reasonable than that of Padilla’s counsel, such that Farhane’s representation, even if imperfect, was not so seriously flawed as to be constitutionally deficient. Finally, the record before us is entirely sufficient to demonstrate that Farhane was not prejudiced by his counsel’s purportedly inadequate performance.

I do not believe that we should attempt to shoehorn this case, which is premised on different statutes and facts, into the narrow mold of Padilla. I therefore respectfully dissent.

21 20-1666 (en banc) Farhane v. United States

PARK, Circuit Judge, joined by SULLIVAN, NARDINI, and MENASHI, Circuit Judges, dissenting:

A narrow majority of our court today announces a new consti- tutional rule of criminal procedure that is not only wrong but barred by the Supreme Court’s doctrine against retroactive application of such rules on federal collateral review. See Teague v. Lane,

(1989). Although the government did not raise Teague in the district court, the issue has been squarely presented to us, so we can and should reach it. The majority’s refusal to do so disregards well- settled principles of constitutional avoidance and undermines the fi- nality of convictions.

I.

The Supreme Court “has repeatedly stated that a decision an- nouncing a new rule of criminal procedure ordinarily does not apply retroactively on federal collateral review.” Edwards v. Vannoy,

, 258 (2021). “[A]pplying ‘constitutional rules not in exist- ence at the time a conviction became final seriously undermines the principle of finality which is essential to the operation of our criminal justice system.’”

at 263 (quoting Teague,

).

“Under Teague, a case announces a new rule when it breaks new ground or imposes a new obligation on the government; in other words, it does so if the result was not dictated by precedent existing at the time the defendant’s conviction became final.” Hall v. United States,

, 60 (2d Cir. 2023) (cleaned up). A result “is not so dictated ‘unless it would have been apparent to all reasonable ju- rists.’”

(quoting Chaidez v. United States,

(2013)). The rule the majority announces—that the Constitution re- quires counsel to advise criminal defendants about the risk of denat- uralization—is plainly new and thus barred by Teague. Our disa- greement about the scope of the Sixth Amendment and the absence of any case law adopting the majority’s rule before today show that the majority’s decision is clearly not so dictated by precedent as to be “apparent to all reasonable jurists.”

(internal quotation marks omitted); see Butler v. McKellar,

(1990) (concluding that a rule was new under Teague because it “was susceptible to de- bate among reasonable minds”).

The majority characterizes its holding as merely an application of Padilla v. Kentucky,

(2010), and asserts that “the risk of denaturalization simply is a risk of deportation.” Ante, at 16. But that oversimplification strains credulity for the reasons explained in Judge Walker’s dissent. See ante, at 8-9. It also does not avoid Teague, which bars both “the invocation of a rule that was not dictated by precedent” and “the application of an old rule in a manner that was not dictated by precedent.” Stringer v. Black,

(1992); see also Butler,

(“[T]he fact that a court says that its de- cision is within the logical compass of an earlier decision, or indeed that it is controlled by a prior decision, is not conclusive for purposes of deciding whether the current decision is a new rule under Teague.” (quotation marks omitted)). Put simply, the majority’s rule is new because “[d]eportation . . . is ‘unique.’” Chaidez v. United States,

(2013) (quoting Padilla,

). The Supreme Court explicitly limited Padilla to deportation only, so we cannot simply apply that case to encompass other things that carry a “risk of deportation.” The very exercise of explaining why Padilla covers

2 denaturalization requires the court to “develop new law[] establish- ing that the Sixth Amendment applie[s] at all.” Id. at 355.

Padilla does not dictate an outcome in this case, much less one that is apparent to all reasonable jurists. Like the merits panel and nearly half of the en banc court, the panel that granted the Anders mo- tion in Farhane’s direct appeal also did not think that Padilla dictates today’s result. See Order, United States v. Farhane, No. 07-1968(L) (2d Cir. Feb. 4, 2011). This is unsurprising. When the Judicial Confer- ence Committee on Rules of Practice and Procedure amended Rule 11 after Padilla, it said nothing about citizens or denaturalization. In- deed, it specifically addressed only “a defendant who is not a United States citizen.” Fed. R. Crim. P. 11(b)(1)(O). So the claim that Pa- dilla dictates the outcome here is untenable—until today, neither the Rules Committee nor any court in the nation had accepted the major- ity’s rule. See, e.g., United States v. Reeves,

(7th Cir. 2012) (“Padilla is rife with indications that the Supreme Court meant to limit its scope to the context of deportation only.”). For Teague purposes, that ends the inquiry, and we should conclude that Farhane is barred from seeking his new rule on collateral review. See Lockhart v. Fretwell,

(1993) (“Teague stands for the proposition that new constitutional rules of criminal procedure will not be an- nounced or applied on collateral review.”).

II.

The majority tries to avoid Teague by pointing to the govern- ment’s failure to raise it below. Ante, at 32. It is true that the gov- ernment did not raise the issue in the district court. But that does not matter. Courts have discretion to overlook a forfeiture generally and should exercise that discretion to address Teague specifically.

3 A “Teague analysis is ordinarily our first step when we review a federal habeas case.” Schiro v. Farley,

(1994). When the government “argue[s] that the defendant seeks the benefit of a new rule of constitutional law, the court must apply Teague before considering the merits of the claim.” Caspari v. Bohlen,

(1994). To be sure, a Teague argument can be waived or forfeited in certain circumstances. The Supreme Court has stated that it is not required to consider Teague when a party raises it for the first time after the Court has issued a writ of certiorari—because mandatory consideration of such an “eleventh-hour Teague argument” could “in- sulate[]” from the Court’s review the question on which the Court granted certiorari. Buck v. Davis,

, 127–28 (2017); see Schiro,

. Based on Schiro, we have held that a Teague argument raised for the first time in a petition for rehearing (i.e., after the panel has already announced a decision on the merits) may, but need not, be addressed. See Agard v. Portuondo,

(2d Cir. 1998). We have never held, however, that we may reverse a dis- trict court’s judgment—and announce a new rule of constitutional law—without addressing a Teague argument timely raised on appeal in defense of that judgment. 1 We should not do so today for three reasons.

1 The panel majority incorrectly said that the government waived (rather than forfeited) Teague, see Farhane v. United States,

, 126 n.4 (2d Cir. 2023), but not even Farhane argues waiver, see United States v. Graham,

, 79-80 (2d Cir. 2022) (distinguishing forfeiture from waiver). In any event, we have in unusual circumstances overlooked a Teague argument to reject the merits of a habeas petition, as the panel ma- jority did here. See Ayala v. Speckard,

(2d Cir. 1997).

4 First, and most important, Teague would resolve this case with- out requiring us to decide a constitutional question. The en banc majority justifies its refusal to overlook forfeiture by pointing out that a successful Teague argument would prevent the en banc Court from announcing a new rule of constitutional law. See ante, at 31-33. Precisely. “[A] longstanding principle of judicial restraint requires that courts avoid reaching constitutional questions in advance of the necessity of deciding them.” Camreta v. Greene,

(2011) (emphasis added) (internal quotation marks omitted). Given that we are free to overlook forfeiture, it cannot be said that this case requires us to announce a new constitutional rule.

Second, Teague’s applicability is a purely legal question, one that Farhane had the opportunity to address—and did in fact ad- dress—both before the panel and on rehearing en banc. The issue is thus fully briefed for our review. Moreover, as the appellee, the gov- ernment “is entitled to rely on any legal argument in support of the judgment below.” Schiro,

; see also Dandridge v. Wil- liams,

, 475 n.6 (1970) (“The prevailing party may, of course, assert in a reviewing court any ground in support of his judg- ment, whether or not that ground was relied upon or even considered by the trial court.”).

Third, Teague protects the finality of convictions. Finality en- sures stability and predictability, upholds the rights of victims, and affords defendants the closure necessary for rehabilitation. See, e.g., McCleskey v. Zant,

(1991) (“Neither innocence nor just punishment can be vindicated until the final judgment is known.”); Edwards, 593 U.S. at 263 (“When previously convicted per- petrators of violent crimes go free merely because the evidence

5 needed to conduct a retrial has become stale or is no longer available, the public suffers, as do the victims.” (citing United States v. Mechanik,

(1986))). The values that undergird finality explain why Teague applies even if a defendant never had a prior opportunity to present his claim. Cf. Breard v. Greene,

(1998) (noting that, even if the defendant could not have discovered his claims earlier, they “would be barred on habeas review under Teague”). That interest in finality is particularly salient here, given that Farhane pleaded guilty 18 years ago to conduct that began 23 years ago. Indeed, Farhane now seeks—13 years after his conviction became final—to force the government to retry him “hampered by problems of lost evidence, faulty memory, and missing witnesses.” Allen v. Hardy,

(1986) (cleaned up).

I note one final concern with the majority’s decision to ignore Teague. In treating Teague as forfeited, the majority assures us that its decision will not “open the floodgates” to additional challenges. But the logic of its decision says otherwise. The majority reasons that its rule flows inexorably from Padilla and is thus not a new rule at all. 2 Ante, at 33 (“Padilla dictates this result.”). For the reasons already stated, I think that is clearly wrong. But if the rule

2 Farhane’s arguments on this point are not only incompatible with finality; they are internally inconsistent. On one hand, he says his § 2255 motion is timely because he could not have known the facts necessary to file it until the government began denaturalization proceedings. On the other hand, he argues that the risk of deportation flows automatically from his conviction, which became final after Padilla. He cannot have it both ways. Either Padilla dictated the outcome here the day it was decided (meaning that Farhane could have discovered his claim through the exer- cise of due diligence on that date, see § 2255(f)(4)), or deportation does not flow automatically from his guilty plea.

6 announced today is “old,” the majority invites thousands of natural- ized citizens convicted in the fourteen years since Padilla to challenge long-final state and federal convictions—virtually none of which will involve any real risk of denaturalization or deportation. See Dissent of Walker, J., ante, at 2-9. Had the majority addressed Teague—and still announced its constitutional rule—it would at least have been forced to confront the consequences of its decision.

In sum, Teague applies here and guards against unnecessarily announcing a new constitutional rule, which is the only thing the ma- jority does. The majority does not say that Farhane’s counsel was objectively unreasonable or that Farhane was prejudiced by any defi- ciency. Instead, the majority reaches out to announce a new Sixth Amendment rule without deciding whether Farhane is even entitled to relief under any approach. This turns constitutional avoidance on its head.

For these reasons, I respectfully dissent. 3

3 Most of the court concludes that Teague is not an impediment to reaching the merits, so having been outvoted on that issue, I would affirm for the reasons articulated by Judge Walker. See, e.g., Hanover 3201 Realty, LLC v. Village Supermarkets, Inc.,

(3d Cir. 2015) (Ambro, J., concurring in part) (discussing voting protocols in multi-issue cases).

7 20-1666 (en banc) Farhane v. United States

NARDINI, Circuit Judge, dissenting:

I join Judge Park’s dissenting opinion explaining that Teague v.

Lane,

(1989), bars the majority’s holding; and I join Part I

of Judge Walker’s dissenting opinion, which explains why the Sixth

Amendment does not require criminal defense attorneys to advise

naturalized citizen clients of the collateral risks of denaturalization

and (if denaturalized) deportation that may arise from a guilty plea.

However, while I thoroughly agree with the application of Strickland

v. Washington,

(1984), in Part II of Judge Walker’s dissent,

I believe that the district court should undertake that analysis in the

first instance. On remand, the district court will be free to address the

performance and prejudice prongs in whichever order it thinks best

and can supplement the record as it deems necessary. And if the

district court addresses the prejudice prong, it will be able to make

factual findings—including as to the dispositive question of whether

Farhane would have proceeded to trial had his attorney informed him of the risk of denaturalization and deportation—that would be subject

to reversal only for clear error.

Accordingly, I respectfully dissent for the reasons stated by

Judge Park with respect to the Teague issue, and for the reasons stated

by Judge Walker with respect to the Padilla issue.

2