17-874-pr Dhinsa v. Krueger

17-874-pr Dhinsa v. Krueger

United States Court of Appeals FOR THE SECOND CIRCUIT ______________

August Term, 2017

(Argued: March 7, 2018 Decided: February 20, 2019)

Docket No. 17‐874‐pr ______________

GURMEET SINGH DHINSA,

Petitioner‐Appellant,

–v.–

J. E. KRUEGER, as Warden of FCI Schuylkill,

Respondent‐Appellee. ______________

B e f o r e :

CABRANES, CARNEY, Circuit Judges, and VILARDO, District Judge.*

______________

Petitioner‐appellant Gurmeet Singh Dhinsa appeals from a judgment in the United States District Court for the Eastern District of New York (Edward R. Korman, Judge) denying his petition for a writ of habeas corpus. Dhinsa brings his petition under

by virtue of the “savings clause” in

(e). Dhinsa is serving six concurrent life sentences arising out of multiple convictions concerning his

*Judge Lawrence J. Vilardo, of the United States District Court for the Western District of New York, sitting by designation.

leadership role in a racketeering enterprise. Dhinsa’s petition challenges only two of those convictions. Although the habeas relief he seeks would not affect the length of his incarceration, we conclude that Dhinsa has standing to bring the petition in light of the $100 special assessment that attaches to each of his convictions. If he successfully challenges any of his convictions, he may be entitled to vacatur of the corresponding sentence, including the special assessment. The special assessment, while a small sum, nevertheless is a concrete, redressable injury sufficient to establish Dhinsa’s Article III standing. We therefore next consider whether Dhinsa has satisfied the jurisdictional prerequisites for a § 2241 habeas petition under the § 2255(e) savings clause, which is available if “the remedy by motion is inadequate or ineffective to test the legality of his detention.” Dhinsa challenges two convictions on counts of murdering a potential witness or informant in violation of

(a)(1)(C). He asserts his innocence under the Supreme Court’s intervening decision in Fowler v. United States,

(2011), which requires the government to show in a § 1512(a)(1)(C) prosecution that the murder victim was “reasonably likely” to have communicated with a federal official had he not been murdered. Upon review of the trial record, we conclude that Dhinsa has failed to demonstrate his innocence under Fowler. Dhinsa’s extensive racketeering enterprise represents a type of criminal activity that is commonly investigated and prosecuted by federal officials. A juror could reasonably find that each of Dhinsa’s victims was “reasonably likely” to have communicated with federal officials. The District Court was thus correct to conclude that Dhinsa failed to meet the requirements of the § 2255(e) savings clause. Because this test is jurisdictional, however, the District Court erred insofar as it entered a judgment denying the petition on the merits. We therefore vacate that portion of the judgment denying the petition and remand to the District Court with instructions to enter a judgment dismissing the petition for lack of jurisdiction.

VACATED AND REMANDED. ______________

MARC FERNICH (Giuliana Graham, on the brief), Law Office of Marc Fernich, New York, NY, for Petitioner‐Appellant.

2

NICHOLAS J. MOSCOW (James D. Gatta, Jo Ann M. Navickas, Assistant United States Attorneys, on the brief), for Richard P. Donoghue, United States Attorney for the Eastern District of New York, Brooklyn, NY, for Respondent‐Appellee. ______________

SUSAN L. CARNEY, Circuit Judge:

Petitioner‐appellant Gurmeet Singh Dhinsa appeals from a judgment entered in

the United States District Court for the Eastern District of New York (Edward R.

Korman, Judge) denying his petition for a writ of habeas corpus. Dhinsa claims that he is

entitled to bring a petition under

because the standard remedies under

are “inadequate or ineffective to test the legality of his detention,”

thereby entitling him to invoke the “savings clause” codified in

(e).

In 1999, a jury convicted Dhinsa on numerous counts arising out of his

leadership role in an extensive racketeering enterprise, including two counts of

murdering a potential witness or informant in violation of

(a)(1)(C)

(“informant murder”).1 He is now serving six concurrent life sentences, including a life

1 The District Court and the parties refer to this crime as “obstruction of justice murder.” We find this term inadequately precise. The “Obstruction of Justice” chapter of the United States Code, chapter 73 of title 18, contains multiple provisions that prohibit the killing, threatened killing, or attempted killing of particular types of victims for particular illicit purposes. See, e.g.,

(a) (prohibiting “threats or force” intended to influence, intimidate, or impede, or to retaliate against jurors or officers of the court); 1512(a) (prohibiting a killing or attempted killing intended to prevent any person from testifying, producing evidence, or communicating with law enforcement or judicial officers); 1513(a) (prohibiting a killing or attempted killing intended as retaliation against a witness or informant). For these reasons, we refer to the instant crime of conviction as “informant murder.”

3

term on each count of informant murder. In his habeas petition, he challenges only

those two convictions, and so the relief he seeks would not affect the length of his

incarceration. This circumstance raises the question whether he has standing to

maintain his petition. We conclude that he does, in light of the $100 special assessment

that attaches to each of the challenged convictions. If Dhinsa’s challenge is successful as

to either conviction, he may be entitled to vacatur of the corresponding sentence,

including the special assessment. Although the special assessments are slight in

amount, we think that each is a concrete, redressable injury sufficient to establish

Dhinsa’s Article III standing to challenge these two convictions.

Having confirmed Dhinsa’s constitutional standing to assert the petition, we next

consider whether Dhinsa is entitled to bring a § 2241 habeas petition under the § 2255(e)

savings clause. A petitioner who invokes the savings clause must demonstrate, on the

existing record, that he is innocent of the crime of conviction under a legal theory that

was not previously available to him. Here, Dhinsa asserts that he is innocent of both

counts of informant murder under the Supreme Court’s decision in Fowler v. United

States,

(2011), which requires the prosecution to show that the murder

victim was “reasonably likely” to have communicated with a federal official.

.

We conclude that, on the existing record, Dhinsa has failed to make the threshold

showing of innocence. He ordered the killing of the two victims after they began

investigating his multi‐year, multimillion‐dollar racketeering enterprise, through which

Dhinsa committed acts of violence, fraud, and bribery. Racketeering enterprises of this

nature are commonly investigated and prosecuted by federal officials. Thus, a juror

could reasonably find that each of Dhinsa’s victims was “reasonably likely” to have

communicated with federal officials had Dhinsa not ordered him killed. The District

4

Court was thus correct to conclude that Dhinsa has failed to make the required

threshold showing under § 2255(e). This threshold test is jurisdictional, however, and so

the District Court erred insofar as it entered a judgment denying the petition on the

merits. Accordingly, we VACATE that portion of the judgment and REMAND the case

with instructions for the District Court to enter a judgment dismissing the petition for

lack of jurisdiction.

BACKGROUND

I. Dhinsa’s criminal trial and direct appeal

In 1999, a jury convicted Dhinsa on twenty‐one counts arising out of Dhinsa’s

role in a racketeering enterprise. The evidence adduced during Dhinsa’s four‐month

trial before the District Court is described at length in this Court’s opinion ruling on his

multi‐faceted direct appeal. See United States v. Dhinsa (“Dhinsa I”),

(2d Cir.

2001). We set forth those facts here only as necessary to understand and resolve the

instant appeal.2

Dhinsa was indicted on twenty‐nine counts in connection with his role as “the

self‐professed leader of the ‘Singh Enterprise,’ a vast racketeering organization built

around a chain of fifty‐one gasoline stations that Dhinsa owned and operated

throughout the New York City metropolitan area under the name ‘Citygas.’”

.

2 Because Dhinsa’s direct appeal challenged a jury’s guilty verdicts, this Court considered the trial evidence “in the light most favorable to the government, crediting any inferences that the jury might have drawn in its favor.” Dhinsa I,

(internal quotation marks omitted). That approach is equally appropriate for our review of Dhinsa’s habeas claims in the instant appeal, as explained below in Discussion Section II.A.

5

Dhinsa installed “pump‐rigging” technology that overcharged consumers for gas,

generating tens of millions of dollars of excess revenues over a period of a decade.

He protected the enterprise by bribing public officials, purchasing weapons, and

threatening, kidnapping, or murdering individuals whom he viewed as threats to his

operations.

at 643–46.

The convictions at issue in this appeal arise out of the 1997 killings of Manmohan

Singh (“Manmohan”) and Satinderjit Singh (“Satinderjit”), each of whom was

“murdered on Dhinsa’s orders” after they began independently questioning the

activities of the Singh Enterprise.

. As we found when this case was last before

us:

Prior to his death, Manmohan made numerous inquiries, and confronted various members of the Singh Enterprise, about the July 1995 disappearance of his brother Kulwant, a Citygas employee. Dhinsa ordered Satinderjit murdered after learning that he was cooperating with police regarding, inter alia, Kulwant’s disappearance, Manmohan’s murder, and the enterprise’s pump‐rigging scheme.

The jury convicted Dhinsa on multiple counts in connection with these murders,

including two counts of informant murder (because of Manmohan and Satinderjit’s

“active or potential cooperation with the police”), along with two counts of murder in

aid of racketeering and two counts of conspiracy to murder in aid of racketeering, in

violation of

(a)(1), (5).

. Overall, the jury convicted Dhinsa on

twenty‐one of the twenty‐nine counts in the superseding indictment.

. On

appeal, we vacated three of Dhinsa’s convictions and affirmed the remaining eighteen,

6

including all convictions on the counts arising out of the murders of Manmohan and

Satinderjit.

at 677–78.

Dhinsa’s incarceratory sentence, as adjusted after the appeal, consists of six

concurrent terms of life imprisonment—four of which arise out of his convictions for

informant murder and murder in aid of racketeering in connection with the murders of

Manmohan and Satinderjit—plus several lesser concurrent and consecutive sentences of

varying lengths. The District Court also imposed financial penalties, including a

statutorily mandated special assessment of $50 to $100 for each count of conviction. See

(a) (specifying the amount of money courts “shall assess on any person

convicted of an offense against the United States”).

II. Collateral proceedings

In 2002, after Dhinsa’s convictions became final, he filed two habeas petitions

asserting claims of ineffective assistance of counsel: a petition under

in

the Central District of California, where he was then incarcerated, and a motion to

vacate his sentence under

in the District Court in the Eastern District of

New York. Both petitions were rejected on procedural grounds.

In May 2012, Dhinsa filed a new § 2241 habeas petition asserting that he is

innocent of both counts of informant murder under the standard articulated in the

Supreme Court’s 2011 decision in Fowler.3 The District Court transferred the § 2241

3 Dhinsa filed the 2012 § 2241 petition in the Middle District of Pennsylvania, his then‐current district of incarceration. Shortly thereafter, he asserted substantially identical arguments in a § 2255 motion filed in the District Court in Eastern District of New York. The District Court transferred that motion to this Court on May 31, 2012, for possible certification as a second or

7

petition to this Court for possible certification as a second or successive § 2255 motion

under

(h). In May 2013, a motions panel concluded that Dhinsa’s

motion did not qualify as a successive § 2255 motion. The Court remanded, however,

with instructions for the District Court to consider: (1) whether Dhinsa is entitled to file

a § 2241 petition via the savings clause in § 2255(e), and if so, (2) whether Dhinsa is

entitled to relief under § 2241.

On remand, after the parties filed briefs addressing those two questions, the

District Court requested additional briefing on the so‐called “concurrent sentence”

doctrine, under which “an appellate court affirming a conviction and sentence” may, in

its discretion, choose “not [to] hear a challenge to a conviction on another count if [that

second conviction] carries a sentence that is equal to or less than the affirmed

conviction.” Concurrent‐Sentence Doctrine, Black’s Law Dictionary (10th ed. 2014); see

generally United States v. Vargas,

, 956–60 (2d Cir. 1980). In response, Dhinsa

acknowledged that even if he receives all the relief sought in his habeas petition, he will

remain incarcerated pursuant to his four additional concurrent life sentences.

On March 2, 2017, the District Court issued an order denying Dhinsa’s petition.

Dhinsa v. Krueger (“Habeas Order”),

(E.D.N.Y. 2017). The District

Court concluded, first, that the petition should be dismissed under the concurrent

sentence doctrine because Dhinsa had not identified any meaningful collateral

consequences arising specifically out of the two challenged convictions.

at 427–33. In

successive § 2255 motion. While that matter was pending, the parties agreed to transfer Dhinsa’s § 2241 petition from the Middle District of Pennsylvania to the Eastern District of New York, at which time Dhinsa voluntarily dismissed the pending § 2255 motion.

8

conducting that analysis, the District Court cited extensively to cases discussing subject

matter jurisdiction and Article III standing, but did not expressly articulate any

jurisdictional conclusions regarding this case. Next, the District Court responded to the

questions posted in this Court’s mandate and concluded that Dhinsa is neither eligible

to file a § 2241 petition nor entitled to any habeas relief. Id. at 433–38. The court entered

a judgment denying the petition, but issued a certificate of appealability on all issues.

Dhinsa timely appealed.

DISCUSSION

We review de novo the District Court’s conclusions of law as to both subject

matter jurisdiction and the merits of Dhinsa’s § 2241 habeas petition. Adams v. United

States,

(2d Cir. 2004). The District Court’s analysis raises questions as

to Dhinsa’s Article III standing, and so we begin by assuring ourselves that Dhinsa has

standing to bring the petition. We next conclude, however, that Dhinsa is not entitled to

pursue a § 2241 habeas petition via the savings clause in § 2255(e), and that the District

Court therefore lacked jurisdiction over the petition. On the existing record, a juror

could reasonably conclude, as required under Fowler v. United States,

(2011), that each of Dhinsa’s victims was “reasonably likely” to communicate with a

federal official had he not been murdered. The District Court reached a similar

conclusion, but entered a judgment denying the petition on the merits. We therefore

vacate that portion of the judgment and remand for entry of an amended judgment

dismissing the petition for lack of jurisdiction. Because we resolve Dhinsa’s appeal on

jurisdictional grounds, we need not consider the government’s invitation to dispose of

9

the case under the discretionary concurrent sentence doctrine.4 Cf. Tavarez v. Larkin,

(2d Cir. 2016).

I. Article III standing

The District Court concluded that Dhinsa’s petition should be dismissed under

the concurrent sentence doctrine, but the court’s reasoning was consistent with a

conclusion that Dhinsa lacked Article III standing to assert the petition at all. We have

“a special obligation” to assure ourselves that the District Court properly exercised

jurisdiction over the petition, whether or not that issue has been raised by the parties on

appeal. Wright v. Musanti,

(2d Cir. 2018). If we determine that District

Court “was without jurisdiction,” our appellate jurisdiction is limited to the “purpose of

correcting the [District Court’s] error” in reaching the merits of the petition.

(internal

quotation marks omitted). We conclude that Dhinsa has Article III standing based on

the $100 special assessment that attaches to each of his challenged convictions, and

therefore that the District Court properly exercised jurisdiction over the petition.

4 In Ray v. United States,

(1987) (per curiam), the Supreme Court vacated an appellate court’s invocation of the concurrent sentence doctrine in a direct appeal. The concurrent sentence doctrine was unavailable, the Court explained, because “the District Court imposed a $50 assessment on each count” of conviction, and so the defendant’s “sentences are not concurrent” in all respects.

. Because all federal convictions carry a mandatory special assessment, see

(a), Ray is understood to have “abolished the [concurrent sentence] doctrine for direct review of federal convictions.” 7 Wayne R. LaFave et al., Crim. Proc. § 27.5(b) (4th ed.). We decline to decide whether Ray similarly precludes application of the concurrent sentence doctrine in the habeas context.

10

A. Legal standard: Habeas petitions and standing

Article III of the Constitution limits the federal courts’ power to the resolution of

“Cases” and “Controversies.” U.S. Const. art. III, § 2. A litigant who invokes federal

jurisdiction therefore must demonstrate “[s]tanding to sue,” a requirement “rooted in

the traditional understanding” of what it means to assert a “case or controversy” under

Article III. Spokeo, Inc. v. Robins,

(2016). The “‘irreducible

constitutional minimum’ of standing consists of three elements”: the individual

initiating the suit “must have (1) suffered an injury in fact, (2) that is fairly traceable to

the challenged conduct of the defendant, and (3) that is likely to be redressed by a

favorable judicial decision.”

(quoting Lujan v. Defs. of Wildlife,

, 560‐61

(1992)). “[S]tanding doctrine evaluates a litigant’s personal stake as of the outset of

litigation.” Klein, on behalf of Qlik Techs., Inc. v. Qlik Techs., Inc.,

(2d Cir.

2018) (quoting Altman v. Bedford Cent. Sch. Dist.,

, 70 (2d Cir. 2001)).

B. Analysis

The District Court concluded that Dhinsa’s petition should be dismissed under

the concurrent sentence doctrine based on the court’s determination that “the success of

[Dhinsa’s] current petition will not alter the length or nature of his custody,” and

further, that Dhinsa had failed to demonstrate any redressable collateral consequences

arising specifically from the two convictions challenged in his petition. Habeas Order,

11

238 F. Supp. 3d at 431–32. The District Court thus implicitly concluded that Dhinsa

lacked standing to bring his petition.5

We need not revisit the District Court’s analysis of collateral consequences

because we conclude that Dhinsa has established Article III standing based on a

redressable injury‐in‐fact: the $100 special assessment that attaches to each of his two

challenged convictions.6 We offer here a more extended discussion of special

assessments and Article III standing, and in particular, of how our standing analysis in

5 The District Court applied the legal framework developed in Nowakowski v. New York,

(2d Cir. 2016), and, in so doing, implicitly concluded that Dhinsa’s petition is not justiciable. In Nowakowski, we articulated a burden‐shifting framework for determining whether a habeas petition is rendered moot by a petitioner’s release from custody. We explained that the petitioner bears the burden, in the first instance, of “identify[ing] some continuing collateral consequences that may flow from his criminal conviction.”

. Once the petitioner has identified specific collateral consequences, the government may attempt to “prove . . . that there is ‘no possibility’ such consequences will attach to [the petitioner’s] conviction.”

(quoting Sibron v. New York,

(1968)). The District Court concluded here that the government “easily met” its burden of showing that Dhinsa’s informant murder convictions would not produce any future collateral consequence distinct from those arising from his other, unchallenged convictions. Habeas Order,

. We need not apply the Nowakowski framework in this case, however, because Dhinsa remains in custody and no “intervening circumstance” has deprived him “‘of a personal stake in the outcome of the lawsuit, at any point during litigation’ after its initiation” so as to render his petition moot. Klein, on behalf of Qlik Techs., Inc.,

(2d Cir. 2018) (quoting Genesis Healthcare Corp. v. Symczyk,

(2013)).

6 Nowakowski, by way of contrast, involved a petitioner who was unable to assert standing based on financial collateral consequences. Robert Nowakowski’s state conviction carried a $100 fine, but before Nowakowki filed his habeas petition under

, the state criminal court vacated the fine based on Nowakowski’s inability to pay. See Nowakowski,

.

12

the habeas context differs from the question whether a habeas petitioner is “in custody”

when he files his petition.

We note, as a preliminary matter, that “monetary loss” is a quintessential injury

in fact, and “even a small financial loss suffices” to establish standing. Carter v.

HealthPort Techs., LLC,

(2d Cir. 2016) (alteration and internal quotation

marks omitted). An “identifiable trifle is enough,” the Supreme Court has explained,

because standing doctrine’s purpose is “to distinguish a person with a direct stake in

the outcome of a litigation—even though small—from a person with a mere interest in

the problem.” United States v. Students Challenging Regulatory Agency Procedures

(SCRAP),

, 689 n.14 (1973) (internal quotations marks omitted). We thus

conclude that each of Dhinsa’s special assessments “constitutes a concrete injury” that

was “caused by the [challenged] conviction.” Spencer v. Kemna,

(1998). The

sole remaining question for purposes of standing, then, is whether a special assessment

is “redressable by invalidation of the [underlying] conviction.”

We conclude that it

is.

Dhinsa contends in his habeas petition that two of his convictions are unlawful.

If his petition is granted, he may be entitled to vacatur of the judgments of conviction,

the corresponding sentences, or both. See

(authorizing courts to

“dispose of” a § 2241 habeas petition “as law and justice require”);7 United States v.

7 This “broad” provision has been interpreted to authorize relief beyond immediate “discharge of the applicant from physical custody.” Carafas, 391 U.S. at 239. For additional guidance, we may consult the text of § 2255, which directs that a court, upon granting a habeas petition, “shall vacate and set the judgment aside,” and shall also “discharge the prisoner or resentence him . . . or correct the sentence as may appear appropriate.”

(b); see also Hill v. United

13

Triestman,

(2d Cir. 1999) (“Courts hearing § 2241 motions have

traditionally had the power to vacate or reduce . . . sentences” that were entered in

violation of “the Constitution or federal law.”); see also, e.g., Willette v. Fischer,

(2d Cir. 2007) (affirming, as habeas relief, vacatur of sentences entered pursuant

to convictions that violated the Double Jeopardy Clause). The special assessments are

part of Dhinsa’s sentence, as listed in the criminal judgment that was entered in the

District Court. Thus, if Dhinsa succeeds in securing habeas relief as to one or both

challenged convictions, he may be entitled to seek return of the related special

assessments. See Prost, 636 F.3d at 582 n.3.

This is not to say, however, that Dhinsa could file a habeas petition that directly

challenges the imposition of the special assessments. He could not, for example, seek

habeas review based on a claim that the District Court incorrectly calculated the amount

of a given special assessment. It is well‐established that courts may not entertain a

petition “addressed solely to noncustodial punishments.” Kaminski v. United States,

(2d Cir. 2003). Sections 2241, 2254, and 2255 authorize courts to adjudicate a

petition filed by an individual who is “in custody” and who challenges the basis for that

custody.8 We have characterized this statutory “custody” requirement as

“jurisdictional,” and therefore mandatory and non‐waivable. United States v. Rutigliano,

States,

(1962) (Section 2255 “was intended [] to provide in the sentencing court a remedy exactly commensurate with that which had previously been available by habeas corpus [under § 2241] in the court of the district where the prisoner was confined.” (emphasis added)).

8 See, e.g.,

(c)(3) (“The writ of habeas corpus shall not extend to a prisoner unless . . . [h]e is in custody in violation of the Constitution or laws or treaties of the United States.”).

14

(2d Cir. 2018) (internal quotation marks omitted). “[T]he monetary

components of criminal sentences, such as fines and restitution orders, generally do not

restrict liberty so severely as to satisfy the custody requirement,” and so generally do

not fall within the purview of a permissible collateral challenge.9

at 104‐05. (citing

Kaminski, 339 F.3d at 86–87). Nor may a party “avoid this conclusion by challenging

both the custodial and noncustodial parts of his sentence in the same [habeas petition].”

.

Here, however, Dhinsa does not challenge the special assessments themselves.

Rather, he challenges the two underlying convictions for informant murder, pursuant to

which the District Court imposed both special assessments and life terms of

incarceration. We have no trouble concluding that, when Dhinsa filed his § 2241

petition, he was “in custody in relation to the claims” asserted therein. Kaminski,

; see also Spencer,

(noting that the “custody” requirement must be

satisfied only at the time the petition is filed). That Dhinsa’s custody was independently

justified under concurrent, coterminous sentences does not preclude him from

satisfying the “custody” requirement. See Maleng v. Cook,

, 490–91 (1989)

(per curiam) (specifying that a “habeas petitioner [must] be ‘in custody’ under the

conviction or sentence under attack at the time his petition is filed”); see also, e.g.,

9 Although this Court has “left open the possibility of a [monetary penalty] imposing such a severe restraint on individual liberty as to equate to custody, [we have] also observed that such situations will likely be rare.” Rutigliano,

(internal quotation marks omitted). Here, Dhinsa expressly waived any argument “that the burden of a special assessment itself amounts to custody.” Appellant’s Br. 19 (internal quotation marks and alterations omitted) (emphasis in original).

15

Tavarez,

(reaching the merits of a habeas petition challenging only one of

two concurrent sentences, with no discussion of the jurisdictional “custody”

requirement); cf. Garlotte v. Fordice,

(1995) (explaining that a habeas

petitioner “remains ‘in custody’ under all of his sentences until all are served,” and so

permitting a § 2254 habeas petitioner to challenge an expired sentence because he

remained incarcerated under additional consecutive sentences).

Although Dhinsa may not assert a direct collateral attack on his special

assessments, the assessments are nonetheless redressable injuries for the purposes of

Article III standing. “Where a habeas challenge to incarceration results in the

overturning of a conviction or of a whole sentence so that the defendant must later be

resentenced in toto, the defendant on resentencing may . . . question the basis of any

[monetary penalty] that might then be assessed.” Kaminski,

n.3. Indeed,

this principle extends beyond monetary penalties; if Dhinsa successfully vacates one or

more of his convictions, he may be entitled to resentencing on his remaining convictions

if he can show that the District Court’s sentencing determinations depended in some

way on the now‐vacated conviction. See Triestman,

(holding that “the

district court had . . . the power to resentence [the § 2241 petitioner] to the overall term

that he would have received on his interdependent sentencing package absent his

unlawful § 924(c) conviction”).

For these reasons, we conclude that the special assessment attached to each

conviction constitutes a concrete, redressable injury sufficient to establish Article III

standing. See Prost, 636 F.3d at 582 n.3. We therefore conclude that Dhinsa has standing

to bring the petition notwithstanding his unchallenged concurrent sentences.

16

II. Jurisdiction under § 2241 and the § 2255(e) savings clause

Dhinsa seeks to collaterally attack his conviction on two counts of informant

murder. His prior motion under § 2255, the typical path for a collateral attack by a

federal prisoner, leaves him unable to assert these claims in a second such motion

absent our permission, which we earlier denied. Dhinsa argues, however, that his

claims fall within the “savings clause” in § 2255(e) and that he may therefore bring a

habeas petition under § 2241. A petitioner who seeks to establish habeas jurisdiction

under this savings clause must make a threshold showing, based on the existing record,

that he is innocent of his crime of conviction under a legal theory that was not

previously available to him. Here, Dhinsa argues that the conduct for which he was

convicted is insufficient to establish guilt under the intervening standard for informant

murder articulated in Fowler v. United States,

(2011).

We disagree. Rather, we find persuasive the District Court’s analysis concluding

that Dhinsa has failed to meet the requirements of the savings clause.10 Because the

District Court entered a judgment denying the petition on the merits, however, we

vacate that portion of the judgment and remand for entry of a judgment dismissing the

petition for lack of jurisdiction.

The District Court also suggested that Dhinsa’s unchallenged concurrent life sentences make 10

him per se ineligible to assert a § 2241 petition under the savings clause in § 2255(e). See Habeas Order,

(discussing Triestman v. United States,

(2d Cir. 1997)). We need not review the District Court’s reasoning on this point because we affirm the District Court’s conclusion regarding the savings clause on other grounds. See Lotes Co. v. Hon Hai Precision Indus. Co.,

(2d Cir. 2014) (“[T]his Court may affirm on any basis for which there is sufficient support in the record.” (internal quotation marks omitted)).

17

A. Legal standard

In general, federal prisoners who seek to collaterally attack the basis for

imposing a sentence—including by challenging the underlying conviction—must move

“to vacate, set aside or correct the sentence” under

(a). See

(e), 2244(a); Adams,

. To challenge “the execution of a sentence,”

including challenges to disciplinary actions, prison conditions, or parole decisions, by

contrast, an application under

provides the “proper means.” Adams,

. In his present petition, Dhinsa challenges the sufficiency of the evidence

as to two of his convictions. Section 2255 thus at first glance appears to be the

appropriate vehicle for Dhinsa’s claims. The remedies under § 2255 are unavailable to

him, however: he filed an initial § 2255 motion in 2002, and this Court has already

determined that the instant application does not satisfy the requirements of § 2255(h)

for allowing a “second or successive” motion. Faced with that obstacle, Dhinsa invokes

the savings clause in § 2255(e), which, in some circumstances, confers federal habeas

jurisdiction where it would not otherwise exist: a federal prisoner may pursue a petition

under § 2241 if it “appears that the remedy [under § 2255] is inadequate or ineffective to

test the legality of his detention.”

(e).

A petitioner cannot claim that § 2255 is “inadequate or ineffective” merely

because his prior § 2255 motion was denied, however. Id. We have interpreted the

savings clause of subsection (e) to authorize a § 2241 petition only when § 2255 is

unavailable and the petition is filed by an individual who (1) “can prove ‘actual

innocence on the existing record,’” and (2) “‘could not have effectively raised [his]

claim[] of innocence at an earlier time,’” perhaps due to an intervening change in the

18

governing interpretation of the statute of conviction. Cephas v. Nash,

(2d

Cir. 2003) (quoting Triestman v. United States,

(2d Cir. 1997)). To make a

threshold showing of “actual innocence” under the first prong, a petitioner “must

demonstrate that, in light of all the evidence, it is more likely than not that no

reasonable juror would have convicted him.” Bousley v. United States,

(1998) (internal quotation marks omitted). The petitioner must “state a colorable . . .

claim,” in other words, that the trial evidence was insufficient to support his conviction

on a correct understanding of the law. Cephas,

. The court must thus

determine whether, “after viewing the evidence in the light most favorable to the

prosecution, any rational trier of fact could have found the essential elements of the

crime beyond a reasonable doubt.” Coleman v. Johnson,

(2012) (per

curiam) (emphasis in original) (internal quotation marks omitted).

Because the savings clause articulates a jurisdictional requirement, a court

adjudicating a § 2241 petition must confirm that the savings clause can be applied at all

before proceeding with a full merits review of the petitioner’s claims. See

(e) (providing that a § 2241 petition seeking the type of relief available under

§ 2255 “shall not be entertained . . . unless [the savings clause applies]” (emphasis added));

Adams,

(“[I]f a prisoner moving under § 2241 cannot make the necessary

showing that a § 2255 motion is ‘inadequate or ineffective,’ the district court ‘must

either dismiss the habeas petition for lack of jurisdiction or recast it as a § 2255 motion.’”

(quoting Cephas,

)). In many cases—this one included—such analysis will

19

overlap with analysis on the merits.11 The threshold savings clause analysis does not

replace a full merits analysis, however; they are not co‐extensive. For example, the

district court may in the merits proceeding provide the government an opportunity to

“rebut” the factual basis for the petitioner’s claim by “present[ing] any admissible

evidence of petitioner’s guilt[,] even if that evidence was not presented” during the

underlying criminal proceedings.” Bousley, 523 U.S. at 623‐24 (1998) (cautioning that

“‘actual innocence’ means factual innocence, not mere legal insufficiency”). Thus, a

petitioner who makes the threshold showing of legal innocence on the existing record

will not necessarily be entitled to habeas relief. We therefore proceed here with our

scrutiny of Dhinsa’s “actual innocence” claims.

B. Analysis

Section 1512(a)(1)(C) criminalizes the act of killing a person with the intent to,

inter alia, prevent someone from “communicat[ing] . . . information relating to the

commission or possible commission of a Federal offense” to a federal “law enforcement

officer.”

(a)(1)(C). A jury convicted Dhinsa on two counts of informant

murder based on evidence that he ordered the murders of Manmohan Singh and

Satinderjit Singh. Shortly after the trial—but before judgment was entered—this Court

declared in United States v. Diaz,

(2d Cir. 1999), that § 1512(a)(1)(C) requires

the “government [to] prove that at least one of the law‐enforcement‐officer

communications which the defendant sought to prevent would have been with a

The District Court reached the same substantive conclusion that we do but denied the petition 11

on the merits. Because the jurisdictional inquiry must precede merits analysis, however, we find that dismissing the petition on jurisdictional grounds is the correct result.

20

federal officer.” Diaz,

(internal quotation marks omitted). In his direct

appeal, Dhinsa did not challenge the evidentiary basis for his § 1512(a)(1)(C)

convictions as insufficient under Diaz’s standard. In 2011, a decade after this Court

affirmed Dhinsa’s convictions, the Supreme Court specified in Fowler that a defendant is

guilty of informant murder only if the victim was “reasonably likely” to have

communicated with a federal law enforcement officer.

. We need not

decide whether Fowler materially altered this Circuit’s standard for informant murder,

as Dhinsa claims, because we conclude that Dhinsa has failed to satisfy the first prong

of the savings clause analysis: demonstrating his innocence on the existing record.

Triestman,

. In our view, a juror could reasonably conclude that both

Manmohan and Satinderjit were “reasonably likely” to have communicated with

federal officials, as Fowler requires.

The trial evidence showed that Manmohan and Satinderjit were both “murdered

on Dhinsa’s orders” after they started making inquiries into the Singh Enterprise’s

criminal activities. Dhinsa I,

. Manmohan was inquiring about the

disappearance of his brother Kulwant, who had worked at one of Dhinsa’s Citygas

stations.

Testimony showed that Dhinsa believed that Manmohan also saw Dhinsa

and an accomplice “shoot somebody,”

, making him a potential witness against

Dhinsa regarding a second crime. After Manmohan was killed in March 1997,

Satinderjit began cooperating with the New York State Police investigation of the Singh

Enterprise’s role in, inter alia, Kulwant’s disappearance, Manmohan’s murder, and

Citygas’s fraudulent pump‐rigging technology. See

. As part of that

cooperation, Satinderjit participated in a May 1997 raid of the Citygas offices in

21

Brooklyn during which state police seized firearms and arrested four individuals,

including Dhinsa himself.

The following month, in June 1997, Satinderjit was

murdered.

We are aware of no record evidence suggesting that either Manmohan or

Satinderjit communicated directly with any federal law enforcement officials before he

died. Dhinsa acknowledges, however, that federal authorities began “investigat[ing] the

Singh Enterprise in relation to the seized firearms” in July 1997, only one month after

Satinderjit’s death and four months after Manmohan’s. Appellant’s Br. 5. Shortly

thereafter, federal prosecutors indicted Dhinsa on twenty‐nine counts, several of which

related to the conduct that Manmohan and Satinderjit had been investigating before

they were killed: Kulwant’s kidnapping, Manmohan’s murder, and the pump‐rigging

technology.12

In Dhinsa’s view, this record fails to establish that either Manmohan or

Satinderjit was “reasonably likely” to communicate with a federal law enforcement

officer, notwithstanding any communications they made, or might have made, to state

police officers. We disagree.

We conclude, first, that Dhinsa’s crimes would likely have been investigated and

prosecuted by federal officials even if Dhinsa had not murdered Manmohan and

Satinderjit. We are conscious that our “federal nexus inquiry” under § 1512(a)(1)(C)

should not place “undue weight [on] the mere fact” that the defendant was, in fact,

The jury acquitted Dhinsa on the kidnapping charges related to Kulwant, but convicted him 12

on charges related to Manmohan’s murder and the pump‐rigging technology scheme.

22

“convicted of [] federal crime[s]” beyond the informant murders themselves. United

States v. Veliz,

(2d Cir. 2015). “[T]he frequent overlap between state and

federal crimes” means that many unremarkable criminal acts will constitute violations

of federal law, even though, “as a practical matter,” the “crimes, investigations, and

witness tampering . . . are purely state in nature.” Fowler,

. Dhinsa’s

crimes, though, were hardly unremarkable. As described above, Dhinsa headed a multi‐

year, multimillion‐dollar racketeering enterprise through which he defrauded

consumers, bribed public officials, and committed heinous acts of violence and murder.

Moreover, the government suspected Dhinsa of fraud and kidnapping that involved

activity outside New York State.13 See Dhinsa I, 243 F.3d at 645–46; App’x 96–97. A juror

could reasonably conclude that a racketeering enterprise of this interstate scope and on

such a large scale was likely to be investigated and prosecuted by federal officials. See,

e.g., Veliz,

(finding federal involvement reasonably likely where defendant

“committed multiple related crimes across multiple states, with multiple accomplices

that a jury found to constitute an association‐in‐fact RICO enterprise”).

13 The jury convicted Dhinsa on two counts related to interstate kidnapping, but this Court vacated those convictions on procedural grounds. Dhinsa I,

, 677–78. For our purposes in this appeal, it is immaterial whether Dhinsa was found guilty on the kidnapping charges. We are concerned with the likelihood that federal law enforcement officials would see fit to investigate Dhinsa’s activities or prosecute Dhinsa’s case in federal court. That likelihood increases when federal prosecutors believe they have sufficient evidence to charge a defendant with interstate criminal activity. See Veliz,

(noting the “interstate nature of the offenses” as a relevant factor when determining the likelihood of “federal involvement” in a criminal investigation).

23

This conclusion does not end our inquiry, however. As Dhinsa points out, “[i]t is

not enough that federal authorities simply become involved in an investigation.

Rather, . . . the government must prove it reasonably likely that a relevant

communication would have been made to a federal officer.” Appellant’s Reply Br. 5

(emphases in original); see

(a)(1)(C) (criminalizing a killing intended to

“prevent [a] communication . . . to a [federal] law enforcement officer or judge”

(emphasis added)). In arguing that the government failed to meet this standard, Dhinsa

emphasizes that state police took the lead in investigating the Singh Enterprise, while

federal officers—at least initially—played a more limited role by, for example,

providing technical assistance related to tracking firearms.

We are unpersuaded. Dhinsa fails to recognize that § 1512’s reference to “law

enforcement officer[s]” encompasses not only the officials who investigate the case, but

also those who prosecute it. See Fowler, 563 U.S. at 675–76 (Section 1512 “appl[ies] where

it is plain that federal officers would have been involved in investigating and prosecuting

the offense.” (emphasis added)); United States v. Snyder,

(7th Cir. 2017)

(“[I]f the underlying crime . . . would have been prosecuted in federal court, then it

is reasonably likely that the witness would have spoken with a federal officer during

the course of that prosecution.”). The government may satisfy its burden, then, by

showing that a victim was reasonably likely to have served as a witness in a federal

criminal trial or otherwise to have communicated with a federal officer as the

prosecution prepared its case.

A juror could reasonably find that the government met that burden here, and,

therefore, Dhinsa cannot demonstrate actual innocence on the existing record. As we

24

first observed some years ago, “[t]he [trial] record amply demonstrates that Dhinsa

murdered Manmohan and Satinderjit to deprive the government of potential

witnesses.” Dhinsa I,

(internal quotation marks and alterations omitted).

We have already concluded that a federal prosecution was reasonably likely, and, as

noted above, Dhinsa’s superseding indictment included several charges related to the

very criminal conduct that Manmohan and Satinderjit had been investigating. Both

victims had been looking into Kulwant’s disappearance. Dhinsa believed that

Manmohan had personally witnessed Dhinsa shoot someone. Satinderjit was

cooperating with a state investigation of the Singh Enterprise and had participated in a

raid that resulted in Dhinsa’s arrest. Given those facts, a juror could reasonably find that

Manmohan and Satinderjit’s likelihood of speaking with a federal official was far from

“remote, outlandish, or simply hypothetical.” Fowler,

.

Dhinsa has failed to state a colorable claim that “no reasonable juror would have

convicted him” on either count of informant murder. Bousley,

. Thus, the

District Court correctly concluded that Dhinsa has fallen short of satisfying the

threshold analysis required under the § 2255(e) savings clause and that he is therefore

not entitled to bring a § 2241 habeas petition. Although the savings clause imposes a

jurisdictional threshold, the District Court entered a judgment denying the petition on

the merits. Accordingly, we remand that portion of the judgment so that the District

Court may enter an amended judgment that dismisses the petition for lack of

jurisdiction.

25

CONCLUSION

Dhinsa’s § 2241 habeas petition challenges only two of six concurrent life

sentences. Thus, a successful petition would not affect the length of his confinement. He

has standing to bring his petition, however, because the special assessment that attaches

to each of the challenged convictions presents a concrete, redressable injury sufficient to

establish Article III standing. We are nonetheless precluded from reaching the merits of

Dhinsa’s habeas claims because he has failed to show that his petition satisfies the

jurisdictional requirements of the savings clause contained in § 2255(e). We conclude

that, on the existing record, a juror could reasonably find that Dhinsa is guilty of both

charged counts of informant murder pursuant to

(a)(1)(C). Dhinsa

therefore is not entitled to bring a habeas petition under § 2241.

Because the District Court entered a judgment denying the petition on the merits,

we VACATE that portion of the judgment that denied the petition and REMAND the

cause with instructions for the District Court to enter a judgment dismissing the petition

for lack of jurisdiction.

26