Reply Brief — Clemente Avelino Pereida, Petitioner v. Robert M. Wilkinson, Acting Attorney General

No. 19-438

IN THE

Supreme Court of the United States

_________________________________

CLEMENTE AVELINO PEREIDA,

v.

Petitioner,

WILLIAM P. BARR, ATTORNEY GENERAL,

Respondent.

_________________________________

ON WRIT OF CERTIORARI TO THE UNITED STATES

COURT OF APPEALS FOR THE EIGHTH CIRCUIT

_________________________________

REPLY BRIEF FOR PETITIONER

_________________________________

David V. Chipman

Raul F. Guerra

MONZÓN, GUERRA &

ASSOCIATES

1133 H Street

Lincoln, NE 68508

Thomas M. Bondy

Benjamin P. Chagnon

Monica Haymond

ORRICK, HERRINGTON &

SUTCLIFFE LLP

1152 15th Street NW

Washington, DC 20005

Brian P. Goldman

Counsel of Record

Thomas King-Sun Fu

Kory DeClark

ORRICK, HERRINGTON &

SUTCLIFFE LLP

405 Howard Street

San Francisco, CA 94105

(415) 773-5700

brian.goldman@orrick.com

E. Joshua Rosenkranz

ORRICK, HERRINGTON &

SUTCLIFFE LLP

51 West 52nd Street

New York, NY 10019

Counsel for Petitioner

i

TABLE OF CONTENTS

Page

TABLE OF AUTHORITIES ..................................... ii

INTRODUCTION ..................................................... 1

ARGUMENT ............................................................. 3

I.

The Government’s View Is Inconsistent

With This Court’s Cases And The First

Principles Of The Categorical Approach. .......... 3

A. The least-acts-criminalized

presumption governs all aspects of the

categorical approach and leaves no

work for a burden of proof to do. .................. 3

B. The government’s argument flouts the

first principles of the categorical

approach. ....................................................... 9

II. The Government’s Arguments Regarding

Practical Difficulties Lack Merit. .................... 16

CONCLUSION ........................................................ 24

ii

TABLE OF AUTHORITIES

Page(s)

Cases

Matter of A-G-G-,

25 I. & N. Dec. 486 (BIA 2011) ............................22

Apprendi v. New Jersey,

530 U.S. 466 (2000) ..............................................15

Carachuri-Rosendo v. Holder,

560 U.S. 563 (2010) ..............................................17

Cooper v. Harris,

137 S. Ct. 1455 (2017)..........................................21

Descamps v. United States,

570 U.S. 254 (2013) .................................... 5, 12, 16

Dir., Office of Workers’ Compensation

Programs v. Greenwich Collieries,

512 U.S. 267 (1994) ..............................................20

Esquivel-Quintana v. Sessions,

137 S. Ct. 1562 (2017)............................................7

Guerrero-Lasprilla v. Barr,

No. 18-776 (Mar. 23, 2020) ....................................9

Johnson v. United States,

559 U.S. 133 (2010) .............................. 2, 5, 6, 8, 19

Johnson v. United States,

135 S. Ct. 2551 (2015)..........................................12

iii

Marinelarena v. Barr,

930 F.3d 1039 (9th Cir. 2019).................. 17, 19, 21

Mathis v. United States,

136 S. Ct. 2243 (2016).................... 7, 10, 11, 15, 16

Mellouli v. Lynch,

135 S. Ct. 1980 (2015)......................................7, 14

Moncrieffe v. Holder,

569 U.S. 184 (2013) ...................................... passim

In re S-K-,

23 I. & N. Dec. 936 (BIA 2006) ............................22

Shepard v. United States,

544 U.S. 13 (2005) .............................. 10, 12, 15, 18

Taylor v. United States,

495 U.S. 575 (1990) ..............................................15

United States ex rel. Guarino v. Uhl,

107 F.2d 399 (2d Cir. 1939) ...................................9

United States ex rel. Mylius v. Uhl,

210 F. 860 (2d Cir. 1914) .................................4, 12

Matter of Vivas,

16 I. & N. Dec. 68 (BIA 1977) ..............................21

Young v. Holder,

697 F.3d 976 (9th Cir. 2012)................................17

iv

Statutes

Immigration and Nationality Act,

8 U.S.C. § 1101 et seq.

8 U.S.C. § 1229a(c)(3)(B) .....................................20

8 U.S.C. § 1229a(c)(3)(C) ...............................20, 21

8 U.S.C. § 1229a(c)(4) ............................................4

8 U.S.C. § 1229a(c)(4)(A) .................................1, 20

8 U.S.C. § 1229a(c)(4)(B) .....................................22

8 U.S.C. § 1229b(b)(1)(C) .................................1, 23

8 U.S.C. § 1326(b) ................................................14

Neb. Rev. Stat. § 28-608(1)(c) (2008) ........................12

Rules and Regulations

8 C.F.R. § 1003.47(b)(5) ............................................21

8 C.F.R. § 1003.47(e) .................................................21

8 C.F.R. § 1240.8(d) ................................... 4, 11, 21, 22

70 Fed. Reg. 4743 (Jan. 31, 2005).............................21

Fed. R. Evid. 301 .........................................................8

v

Other Authorities

Executive Office for Immigration Review,

U.S. Dep’t of Justice, Form EOIR-42B,

Application for Cancellation of Removal

and Adjustment of Status for Certain

Nonpermanent Residents (Rev. Oct. 2008),

http://cdn.ca9.uscourts.gov/datastore/libra

ry/2013/02/26/Young_eoir42b.pdf ........................22

2 McCormick on Evidence § 339 (8th ed.

2020) .....................................................................13

2 McCormick on Evidence § 344 (8th ed.

2020) .......................................................................8

U.S. Br., Johnson v. United States,

No. 08-6925 (U.S. Aug. 7, 2009) ......................8, 19

U.S. Br., Voisine v. United States,

No. 14-10154 (U.S. Jan. 19, 2016) .......................19

1

INTRODUCTION

The government’s brief makes clear that the parties’ dispute is not about the meaning of the immigration laws’ burden of proof provisions, but instead

about the operation of the modified categorical approach. There is no disagreement that a noncitizen

applying for relief from removal “has the burden of

proof to establish” that he “satisfies the applicable eligibility requirements,” 8 U.S.C. § 1229a(c)(4)(A), including that he “has not been convicted of an offense”

listed in the INA, § 1229b(b)(1)(C). The question is

just what it means to have been “convicted of an offense” for federal purposes, and thus how a noncitizen

may show that a past state conviction does not rise to

that level.

The categorical approach and its modified variant

answer that question. They provide that a state conviction is not a “conviction” for immigration purposes

unless “the record of conviction of the predicate offense necessarily establishes” the elements of the federally defined offense. Moncrieffe v. Holder, 569 U.S.

184, 197-98 (2013) (emphasis added). And they make

clear that a conviction necessarily establishes only the

“minimum conduct criminalized by the state statute,”

unless the record of conviction reveals something

more with certainty. Id. at 191; Pet. Br. 15-19. So a

noncitizen demonstrates that he “has not been convicted of” an enumerated offense, § 1229b(b)(1)(C)

(emphasis added), when the record shows only that he

was convicted under an overbroad statute and does

not specify which prong formed the basis for his conviction; under the least-acts-criminalized presumption, such a conviction does not correspond to a federal

2

predicate offense. He need not go the additional step

of “prov[ing] that his conviction was for a version of

the offense that is not disqualifying,” Gov’t Br. 14 (emphasis added).

The government’s contrary view turns on its assertion that the least-acts-criminalized presumption

does not apply to the modified portion of the categorical approach. It says that determining which prong of

a divisible statute a noncitizen was convicted of is instead a factual question with no presumptive answer,

so uncertainty on that score must be resolved by background burdens of proof. But the Court’s cases say

precisely the opposite, and the government outright

ignores the case that most clearly proves our point,

Johnson v. United States, 559 U.S. 133 (2010). Infra

§ I.A. The government’s vision of the modified categorical approach also departs sharply from the decades-old first principles of the categorical approach.

Infra § I.B. And the government’s defense of its rule

on practical grounds does not withstand scrutiny.

Where relief is ultimately in the government’s discretion anyway, there is no reason to think that Congress

intended an approach that would make eligibility for

humanitarian relief turn on the happenstance of state

and local recordkeeping practices. Infra § II.

3

ARGUMENT

I.

The Government’s View Is Inconsistent With

This Court’s Cases And The First Principles

Of The Categorical Approach.

A. The least-acts-criminalized presumption

governs all aspects of the categorical

approach a nd lea ves no work for a

burden of proof to do.

1. Much is common ground here. The government

does not dispute three key premises of our argument:

First, courts analyzing a potential predicate offense under the categorical approach must presume

that the “conviction ‘rested upon nothing more than

the least of the acts’ criminalized,” because that is all

the “conviction necessarily involved.” Moncrieffe v.

Holder, 569 U.S. 184, 190-91 (2013). See Pet. Br. 1617; Gov’t Br. 32-34, 36.

Second, a presumption that provides a default answer to a question “supplants a generally applicable

burden of proof” by leaving no ambiguity for a burden

of proof to resolve. Pet. Br. 30-32.

Third—putting those propositions together—

wherever the least-acts-criminalized presumption applies, it operates the same regardless of who bears the

burden of proof. Gov’t Br. 35-38; see Pet. Br. 23-26.

Accordingly, because the government agrees that

the least-acts-criminalized presumption applies to indivisible statutes, it concedes that “an overbroad indivisible crime is not disqualifying” when a noncitizen

4

applies for relief from removal, even though he bears

the burden of proof. Gov’t Br. 36. The government

does not argue, for example, that the presumption

and the burden of proof are inversely related, such

that a conviction would be presumed to rest on the

greatest acts criminalized any time the noncitizen

bears the burden of proof. See Gov’t Br. 35-36; see also

Pet. Br. 25 (citing Moncrieffe, 569 U.S. at 194-95,

204); Pet. Br. 41 (citing United States ex rel. Mylius v.

Uhl, 210 F. 860, 862 (2d Cir. 1914) (exclusion case)).

2. The government parts ways with us on only a

narrower, threshold question: whether the least-actscriminalized presumption applies to the modified portion of the categorical approach—i.e., the step of assessing which alternative prong of a divisible statute

gave rise to a conviction. If (as we say) the presumption applies to the whole categorical analysis, then

background burdens of proof have no role to play with

the modified portion, just as everyone agrees they do

not in the pure categorical approach. If (as the government says) “the least-acts presumption is irrelevant” to the “antecedent inquiry of which crime under

a divisible statute the defendant was convicted of

committing,” then the presumption does not resolve

any ambiguity on that question, but “the burden of

proof can and does.” Gov’t Br. 27, 34-35, 38. Virtually

everything the government says hinges on this one

contention.

So the dispute is not whether 8 U.S.C.

§ 1229a(c)(4) or 8 C.F.R. § 1240.8(d) contain any “unstated exception” for criminal convictions, nor

whether we are seeking to “shift the burden of

proof … from the alien to the government.” Gov’t Br.

5

17, 41. Rather, it is just whether the modified categorical approach is a “meaningfully different” analysis

from the ordinary categorical approach, such that

burdens of proof affect the former even though they

undisputedly do not affect the latter. Gov’t Br. 35.

The government is wrong to divorce the categorical and modified categorical approaches and treat

them as fundamentally different inquiries. As we

demonstrated (Br. 27-29), the least-acts-criminalized

presumption governs every aspect of the analysis, including the “which prong” question under the modified categorical approach. That is just how this Court

applied the presumption in Johnson v. United States,

559 U.S. 133 (2010)—the case that defined the presumption in the first place. Johnson required the

Court to consider whether a state battery conviction

corresponded to a “violent felony” under the Armed

Career Criminal Act. Id. at 137. The state statute was

divisible into three different battery crimes. Id. at

136-37; see Descamps v. United States, 570 U.S. 254,

264 n.2 (2013) (recognizing that Johnson “rested on

the explicit premise that the [state statute] ‘contain[ed] statutory phrases that cover several different

... crimes’”). But “nothing in the record of Johnson’s …

battery conviction” clarified which statutory prong he

was convicted of. Johnson, 559 U.S. at 136-37. So the

Court presumed that his conviction “rested upon” only

“the least of these acts”—i.e., the lowest-level alternative crime under the statute. Id. at 137. And it then

went on to see whether even that crime satisfied the

definition of a federal “violent felony.” Id. at 137-42.

The government hazards no response at all to

Johnson. But Johnson defeats the government’s

6

argument. The government says, for example, that

the “least-acts presumption applies only … to classify

an already-identified crime under the categorical approach.” Gov’t Br. 33 (emphasis added). It posits that

when a record of conviction does not identify a specific

statutory prong, an “adjudicator is no more able to

discern” the meaning of the conviction “than if the

judgment had specified no particular statute of conviction at all.” Gov’t Br. 27; see also id. at 24. It therefore insists, over and over, that the comparison of

elements under the categorical approach simply “cannot” proceed “without first identifying which crime”

under the statute an individual “was convicted of committing.” Gov’t Br. 36; see also id. at 24, 34. Johnson

proves otherwise. It applied the least-acts-criminalized presumption, so that the categorical analysis

could proceed, precisely when the “absence of records”

specifying a particular statutory prong meant that

the specific crime could not be identified under “the

modified categorical approach.” 559 U.S. at 145.

Moncrieffe likewise refutes the government’s position. It explained that the presumption holds even

for a conviction under a “state statute[] that contain[s] several different crimes”—i.e., in precisely the

circumstance when the government says it cannot—

unless the record of conviction makes clear “which

particular offense the noncitizen was convicted of.”

569 U.S. at 191. That is, a clear answer to the “which

prong” inquiry may overcome the baseline presumption that a conviction establishes only the “minimum

conduct criminalized by the state statute.” Id. But

identifying one specific statutory prong is not a necessary “antecedent” step, as the government says (Br.

34-35).

7

As a result, where a conviction record leaves ambiguous “what crime, with what elements” gave rise

to a particular conviction under a statute, Mathis v.

United States, 136 S. Ct. 2243, 2249 (2016), the categorical inquiry does not grind to a halt. Instead, Johnson and Moncrieffe prescribe a default rule: Presume

the least, because the most minor prong of the statute

is all that that conviction “necessarily” establishes.

Moncrieffe, 569 U.S. at 190-91. That allows the adjudicator to move ahead with comparing the elements

of that crime with those of the federal offense. Id. And

because the least-acts-criminalized presumption has

already resolved the “[a]mbiguity,” no burden of proof

ever comes into play—which is why no such burden

features in any of this Court’s categorical or modified

categorical approach cases. Id. at 194-95; see also Pet.

Br. 23-24, 26.

3. The government suggests that Moncrieffe,

along with Mellouli v. Lynch, 135 S. Ct. 1980 (2015),

and Esquivel-Quintana v. Sessions, 137 S. Ct. 1562

(2017), instead support its view. It notes that those

decisions did not rely on the least-acts-criminalized

presumption to resolve uncertainty as to “what crime”

was at issue; instead they “invoked that presumption

to determine the legal scope and significance … of

past convictions for crimes that had already been

identified.” Gov’t Br. 35. But the fact that the presumption applies in that instance, too, does not mean

that it applies only there. As the government recognizes, “[i]n each” of those cases, “no serious question

existed” as to “what crime” was at issue, so the Court

had no occasion to apply the presumption to that

question. Id. But Moncrieffe, at least, described how

the presumption would be applied in that

8

circumstance—just as it was in Johnson, where the

“which prong” question was addressed. See

Moncrieffe, 569 U.S. at 190-91.

Similarly beside the point is the government’s observation that a noncitizen convicted under a divisible

statute “must have been convicted of one of [the] distinct crimes” contained within, for he “cannot be convicted of a divisible statute in its entirety” just as he

“could not be convicted of violating Title 18 of the

United States Code.” Gov’t Br. 23, 36. That is true

enough. But no one is arguing otherwise. The question is just how a court treats a conviction when the

record does not specify which particular crime a

noncitizen was convicted of. And while a conviction

record reciting only “Title 18” is fanciful, records identifying only a particular criminal code section—but no

specific prong within that statute—are “common[],”

Johnson, 559 U.S. at 145. As the government told the

Court in Johnson, in many states “charging documents generally track the language of the statute”

without specifying particular statutory prongs. U.S.

Br. at 42, Johnson v. United States, No. 08-6925 (U.S.

Aug. 7, 2009). That is the case here. See Pet. Br. App.

7a-8a (C.A.R. 165).

4. Our argument that the least-acts-criminalized

presumption governs the entire analysis holds regardless of whether the “which prong” question is

viewed as a question of law or a question of fact (see

Pet. Br. 26-27); presumptions operate on questions of

fact too, and they can be used to satisfy generally applicable burdens of proof. See, e.g., 2 McCormick on

Evidence § 344 (8th ed. 2020); Fed. R. Evid. 301. We

note, however, that if that question is purely legal,

9

then that is further reason why it cannot be subject to

a burden of proof. See Pet. Br. 20-22. Indeed, the government does not dispute that “a ‘purely legal inquiry’” is not affected by a burden of proof, Gov’t Br.

37-38; it just says that the “which prong” question is

factual. But the only reason it offers for why that is so

is that the inquiry is “record-specific” and “relies on

documentary evidence.” Gov’t Br. 38-39. That fails to

answer our explanation that an inquiry is not “factual” just because it involves consulting records. Pet.

Br. 32-35. Nor does the government dispute that the

BIA’s application of the modified categorical approach

is judicially reviewable—and indeed, reviewed de

novo—just like any “question of law.” Pet. Br. 36; cf.

Guerrero-Lasprilla v. Barr, No. 18-776 (Mar. 23,

2020), slip op. 4-5.

B. The government’s a rgument flouts the

first principles of the ca tegorica l

approach.

The government’s position is also incompatible

with the categorical approach’s first principles.

1. Courts “must presume that [a] conviction

‘rested upon nothing more than the least of the acts’”

that could support it “[b]ecause” the categorical and

modified categorical approaches focus on “what [a]

state conviction necessarily involved.” Moncrieffe, 569

U.S. at 190-91 (emphasis added); United States ex rel.

Guarino v. Uhl, 107 F.2d 399, 400 (2d Cir. 1939) (L.

Hand, J.) (examining what a conviction “‘necessarily’”

established by looking to the least criminal conduct

punished by the statute). That focus reflects a “demand for certainty” about the meaning of past

10

convictions before they will be allowed to restrict individuals’ liberty for a second time, in sentencing or

immigration proceedings years or decades later.

Mathis, 136 S. Ct. at 2257. Insistence “on a legal ‘certainty’” about prior convictions has animated the categorical approach “from the Court’s earliest

decisions.” Id. at 2255 n.6.

The government does not dispute that the demand for certainty always requires presuming the

least acts that could satisfy a given statutory element

when it is identified. See Gov’t Br. 32-35; Pet. Br. 1617. But the government argues that no such certainty

is needed when the relevant statutory alternative is

not identified. There is no basis for driving a wedge

between the two, though. The only thing that is “certain[]” about what a past conviction stands for—all

that it “‘necessarily’ involved”—is the least of the acts

that could have resulted in it. Shepard v. United

States, 544 U.S. 13, 24 (2005) (plurality opinion);

Moncrieffe, 569 U.S. at 190-91. That is true whether

a court is faced with an indivisible statute or a divisible one; the “minimum conduct criminalized by the

state statute” is the same whether it has just one set

of elements or encompasses multiple alternatives.

Moncrieffe, 569 U.S. at 191.

This Court’s cases confirm that the demand for

certainty is not cabined to cases where the specific

statutory alternative is already identified. In Shepard, for instance, the Court invoked the “demand for

certainty” in prescribing which record materials may

be consulted to identify the statutory prong under the

modified categorical approach (jury instructions and

plea agreements, yes; police reports, no). 544 U.S. at

11

21-22. In Mathis, the Court explained that the “demand for certainty when determining whether a[n individual] was convicted of” a certain offense meant

that statutory alternatives could be treated as elements, rather than means, only if the record materials “plainly” indicate the former. 136 S. Ct. at 2257

(internal quotation marks omitted). Those cases make

clear that the need for “certainty” broadly anchors the

entire analysis of prior convictions, not just the ultimate comparison of elements. And because a past conviction is not a “conviction” under the INA unless it

establishes the elements of the federal offense with

certainty, a background burden of proof does not bear

on the analysis. See Pet. Br. 24.

2. The categorical approach’s demand for certainty also highlights another oddity with the government’s position. The government says that 8 C.F.R.

§ 1240.8(d)’s “preponderance of the evidence” standard governs the “which prong” inquiry under the modified categorical approach. Gov’t Br. 20, 41-42. So, in

the government’s view, “[t]he alien … must show it is

‘more likely than not’ that he was convicted of a nondisqualifying version of the offense.” Gov’t Br. 26 (internal citation omitted). But the government never

explains how that standard is supposed to operate in

this context, where “certainty” is the watchword. And

this Court has already rejected every conceivable approach that would make the inquiry a probabilistic

one.

An immigration judge may not, for example, ask

whether it is more likely than not that the conduct

alleged in a charging document (e.g., using a fake social security card to obtain employment) would have

12

been prosecuted under one of the non-disqualifying

prongs of the criminal statute (e.g., “[c]arr[ying] on

any … occupation without … authorization required

by law,” Neb. Rev. Stat. § 28-608(1)(c) (2008)); that

would run afoul of the prohibition on looking to the

alleged facts underlying the offense. See Descamps,

570 U.S. at 268, 270. Nor could an IJ take testimony

about which statutory prong an individual pleaded

guilty to years before; that would contravene the prohibition on unreliable “minitrials” about the prior offense “conducted long after the fact.” Moncrieffe, 569

U.S. at 200-01; Mylius, 210 F. at 863; Pet. Br. 23, 40.

An IJ similarly could not weigh clues in some conviction documents against others; the question of which

prong formed the basis of a past conviction must be

answered “free from any inconsistent, competing evidence.” Shepard, 544 U.S. at 21-23; id. at 25 (plurality

opinion). 1 Nor could an IJ try to gauge which statutory alternative is most often prosecuted in the jurisdiction, and treat that as the one that was more likely

than not at issue in the specific case; that would just

revive the “ordinary case” inquiry that the Court recently found unworkable. Johnson v. United States,

135 S. Ct. 2551, 2557 (2015).

1 The government suggests that we take issue with the

“‘strict limitations’ … on the types of documents that may be considered to identify the offense of conviction.” Gov’t Br. 45-46. Not

at all. Those well-established limitations make good sense, for

all the reasons we described (Pet. Br. 45) and the government

says. Our point is simply that those limitations confirm that the

modified categorical approach is a narrow inquiry into the least

that a conviction “necessarily” establishes, not an independent

factual inquiry that must resolve which statutory prong was involved in each case.

13

The “preponderance of the evidence” standard has

no work to do precisely because there is no place in

the categorical approach, including its modified portion, for a freewheeling exercise in “weigh[ing]” competing evidence to see if a party has met its burden of

proof. 2 McCormick on Evidence § 339 (8th ed. 2020)

(preponderance of the evidence); see Pet. Br. 29-30.

3. Eliminating the demand for certainty from one

stage of the categorical approach would also unduly

complicate the analysis. Our approach answers the

“which prong” question the same way all questions

under the categorical approach are answered: If “the

record of conviction of the predicate offense necessarily establishes” the elements of the federally defined offense, then the conviction counts; if not, then

not. Moncrieffe, 569 U.S. at 197-98 (emphasis added).

In contrast, the government would replace that

single question, rooted in a constant presumption,

with a three-part decision tree: If “(A) a conviction [is]

under an indivisible statute that is overbroad relative

to the relevant generic offense,” then apply the standard least-acts-criminalized presumption; if “(B) a conviction [is] under a divisible statute that covers

multiple crimes, some but not all of which are disqualifying, where the record does not indicate which crime

formed the basis of the conviction,” then apply the relevant burden of proof to resolve the “which prong”

question; and if (C) “the defendant was convicted under a divisible statute but the least acts criminalized

by the elements of every crime the statute covers

would (or would not) be disqualifying,” then do not

“conduct a threshold inquiry into which specific version of the offense the defendant was convicted of

14

committing,” but instead revert to applying the standard least-acts-criminalized presumption. Gov’t Br. 33

n.2, 35.

That cumbersome approach is unwarranted, and

nothing in this Court’s cases or the categorical approach’s underlying justifications support deviating

from the usual “demand for certainty” in this one, narrow context. The regime the government envisions

would also confuse the analysis of past convictions—

and eviscerate the categorical approach’s promise of

“predictability in the administration of immigration

law,” Mellouli, 135 S. Ct. at 1986-87—by having the

very same conviction simultaneously not count as a

CIMT at the removal phase (where the government

bears the burden of proof) and then count as one at

the relief phase (where the noncitizen does).

4. Last, the government’s treatment of the “which

prong” question as a factual question that sits outside

the categorical approach’s presumptions cannot be

squared with the approach’s separate Sixth Amendment underpinnings. 2 The government stresses that

the “which prong” inquiry must be a factual one because this Court has referred to the “fact that the defendant had been convicted” of an offense. Gov’t Br.

38 (emphasis added). But the Court’s cases make

clear that the moniker was not intended literally. A

2 Those concerns apply equally to the treatment of a past

“conviction” under the INA because the categorical approach also

governs criminal proceedings under the INA (see 8 U.S.C.

§ 1326(b)), and the government does not dispute that the categorical approach must operate the same in both its criminal and

noncriminal applications. See Pet. Br. 35 & n.8.

15

determination that an individual has a prior conviction for a certain offense is exempt from the Sixth

Amendment’s jury-trial guarantee, which applies to

all other facts that increase a mandatory sentence.

Apprend i v. New Jersey, 530 U.S. 466, 490 (2000).

That is because, unlike any other fact, the “fact” of a

prior conviction can always be known with “certainty,” id. at 488—i.e., “conclusive[ly]” from “a prior

judicial record,” Shepard, 544 U.S. at 24-25 (plurality

opinion). The Court has often signaled this differentiating feature of prior convictions by placing the “fact”

label in scare quotes, Apprendi, 530 U.S. at 488, and

attaching modifiers like “mere” or “simple” to the

“fact” of a prior conviction, Taylor v. United States,

495 U.S. 575, 602 (1990); Mathis, 136 S. Ct. at 2252.

The least-acts-criminalized presumption is the

only reason the modified categorical approach

squares with Apprendi at all: It guarantees that the

“which prong” question is determined with “certainty,” by requiring courts to treat convictions as

standing for no more than what they “‘necessarily’ involved”—i.e., the least of the acts established by the

record of conviction. Shepard, 544 U.S. at 24-25 (plurality opinion). Where that record shows only a statutory section but no subsection, then (as in Johnson)

the modified categorical approach says that the individual has only been “convicted” of the least criminal

prong.

5. Ultimately, the government’s insistence that

the modified categorical approach’s “which prong”

question is “meaningfully different” and governed by

“separate inquiries and … principles,” Gov’t Br. 35, is

just the latest effort to expand the modified

16

categorical approach into a much broader, distinct inquiry. This Court has twice rejected such arguments,

instead holding fast to the view that the modified categorical approach “acts not as an exception, but instead as a tool” that “merely helps implement the

categorical approach.” Descamps, 570 U.S. at 263.

And the analysis’s “central feature” and “basic

method” remain constant, whether a statute is divisible or not. Id.; Mathis, 136 S. Ct. at 2253-54 & n.4.

The Court should reject this reimagining of the modified categorical approach as well.

II. The Government’s Arguments Regarding

Practical Difficulties Lack Merit.

Our brief explained (Br. 43-50) how the government’s rule would yield grave practical difficulties for

both immigration courts and noncitizens—particularly individuals who are convicted of low-level misdemeanors like Mr. Pereida’s “attempted criminal

impersonation” offense. See also NACDL Br. 7-16; Br.

of Immigration Law Professors 21-23. The government does not dispute our illustration of how its rule

will often deprive noncitizens of the benefit of their

plea bargains (Pet. Br. 47-48) and deny humanitarian

relief from removal to individuals who have not committed a disqualifying offense, based on the happenstance of state and local recordkeeping practices (Pet.

Br. 43-46). Instead, it offers three defenses for those

results.

A. First, the government says that the modified

categorical approach presents both sides with the

challenge of having to prove the unprovable, because

when “documents that clarify the basis of a conviction

17

are lost or never generated,” it always means “one

party or the other … will not prevail.” Gov’t Br. 46.

So, the government asserts, our rule would just mean

that the government would suffer “those problems”

instead. Gov’t Br. 46.

But, as the Court has twice reminded the government, the consequences of “those problems” are not

remotely the same both ways. See Pet. Br. 49-50 (citing Moncrieffe, 569 U.S. at 204, and CarachuriRosendo v. Holder, 560 U.S. 563, 581 (2010)). The government does not lose when a conviction cannot be

treated as a mandatory bar to relief, contra Gov’t Br.

46, because the government ultimately makes the (judicially unreviewable) discretionary decision whether

to grant relief as a matter of executive grace.

In contrast, under the government’s rule, when a

noncitizen is convicted under a non-disqualifying

prong of a statute, he is nevertheless guaranteed to be

removed from the country if conviction records were

never created or have been lost to time. The Attorney

General would be barred from even considering

whether to grant him discretionary relief on the facts

of his case. See Pet. Br. 48-49. The government offers

no reason to think that “Congress’s considered judgment,” Gov’t Br. 46, included the “absurd” and “inherent[ly] unfair[]” result of denying humanitarian relief

simply because a noncitizen cannot prove a negative

using only documents that he neither creates nor

maintains. Young v. Hold er, 697 F.3d 976, 993 (9th

Cir. 2012) (en banc) (Ikuta, J., concurring in part and

dissenting in part), overruled by Marinelarena v.

Barr, 930 F.3d 1039, 1042 (9th Cir. 2019) (en banc).

18

B. Second, the government questions whether

ambiguity in conviction records is actually beyond the

noncitizen’s control. It suggests that “the alien … has

a strong incentive to ensure that relevant records

(such as a plea-colloquy transcript) are created and

retained in the first place.” Gov’t Br. 46-47; see also

id. at 31 (same).

The problem is not one of incentive, however, but

ability: State and local courts often do not record the

particulars of a conviction as a matter of course, especially in cases involving misdemeanors. See NACDL

Br. 7-13; Pet. Br. 44-46. And, contrary to the government’s suggestions (Br. 31, 46-47), noncitizens have

little ability to force them to do otherwise. NACDL Br.

7-13. Here, for example, the government notes that an

“online-docket printout … states that ‘[t]he parties

ha[d] agreed to a plea bargain,’” and it wonders why

“petitioner did not submit a plea agreement or colloquy in his removal proceedings before the IJ.” Gov’t

Br. 30-31 (quoting C.A.R. 210). But there was no written plea agreement—only an oral one, as is common

in Nebraska criminal practice. That is why the

“online-docket printout” reflects no filing of a plea

agreement. And, more generally, it is not uncommon

for misdemeanor pleas to be taken “off the record,” or

“without a colloquy” entirely. NACDL Br. 7-9.

Both the Court and the government have

acknowledged these realities. In Shepard itself, the

“record [wa]s silent on the generic element” because

there was “no plea agreement or recorded colloquy in

which Shepard admitted the generic fact.” 544 U.S. at

25. And in Johnson, the Court recognized that it is

“common[]” that “state and local records” will be

19

“incomplete,” such that the modified categorical approach comes up empty. 559 U.S. at 136-37, 145. The

government has cautioned the Court about this realworld dynamic as well: “[R]ecords from closed misdemeanor cases are often … incomplete.” U.S. Br. at 45,

Voisine v. United States, No. 14-10154 (U.S. Jan. 19,

2016). “And plea colloquies, which are not always

transcribed or otherwise available, may reflect ... only

that the defendant pleaded guilty to the offense as

charged in the information or indictment,” without

specifying a statutory prong. U.S. Br. at 43, Johnson

v. United States, No. 08-6925 (U.S. Aug. 7, 2009).

C. Third, the government says that a different

scenario is more concerning—one in which clarifying

records were created, and remain available, but reveal

that a conviction was under a disqualifying prong. In

that case, the government fears, our rule “w[ould] encourage aliens to withhold and conceal evidence.”

Gov’t Br. 47 (quoting Marinelarena, 930 F.3d at 1065

(Ikuta, J., dissenting)); see also Gov’t Br. 31. But any

whiff of “conceal[ing] evidence” could be grounds to

deny relief at the discretionary phase, when an immigration judge decides if an eligible noncitizen should

be granted relief. And there is a high likelihood that

such concealment would be detected, as the government must conduct its own criminal-records check anytime a noncitizen applies for relief from removal. See

infra at 20-22. Moreover, at the discretionary phase,

the IJ is not limited to the categorical approach and

may consider any other evidence of criminal conduct.

See Moncrieffe, 569 U.S. at 204; Pet. Br. 49. So concealing Shepard documents is unlikely to be a common practice, and certainly not a successful one.

20

The government’s concern is also misplaced for a

more fundamental reason: It incorrectly assumes that

noncitizens have an obligation to produce conviction

documents in the first place—i.e., that noncitizens

bear not only the burden of proof but also the burden

of production. Gov’t Br. 30-31.

The government conflates “two distinct concepts”—the burden of persuasion and the burden of

production—that do not automatically travel together. Dir., Office of Workers’ Compensation Programs v. Greenwich Collieries, 512 U.S. 267, 272, 27576 (1994). When Congress uses “the term ‘burden of

proof,’” as it did in § 1229a(c)(4)(A), it is understood

“to mean the burden of persuasion” only. Greenwich

Collieries, 512 U.S. at 272, 275-76. With respect to

“the burden of production or the burden of going forward with the evidence,” id. at 274, however, other

features of the statute make clear that Congress intended that burden to lie with the government.

Section 1229a(c) describes the “[b]urden on [the]

Service” (i.e., the government), as including the introduction of documents that “the Service” and “Service

official[s]” could receive “from a State or court” to use

as “proof of … criminal conviction[s].” 8 U.S.C.

§ 1229a(c)(3)(B)-(C). Importantly, that obligation applies “[i]n any proceeding” under the INA. Id. 3

3

We do not disagree with the government that

§ 1229a(c)(3)(B)’s list of permissible documents applies globally.

See Gov’t Br. 42-43. Our point is that Congress’s decision to provide guidance only to “Service official[s]” (and not noncitizens)

about how to submit certain conviction records suggests that

Congress expected the government to be the one producing

21

Immigration regulations likewise require the government “to initiate … law enforcement … investigations

or examinations concerning the alien or beneficiaries

promptly” following “any application for immigration

relief filed in the proceedings,” including cancellation

of removal. 8 C.F.R. § 1003.47(b)(5), (e); see Br. of Immigrant Defense Project 22-23, 25-26. It must therefore search criminal records to “determine whether an

alien in proceedings has been convicted of any disqualifying crime.” Background and Security Investigations in Proceedings Before Immigration Judges

and the Board of Immigration Appeals, 70 Fed. Reg.

4743, 4743 (Jan. 31, 2005). 4

That scheme is consistent with 8 C.F.R.

§ 1240.8(d), which, as the government acknowledges,

the statutory burden-of-proof provision meant to codify. Gov’t Br. 5-6. That regulation assumes that the

government will initially produce relevant documents

and then, only “[i]f the evidence” submitted by the

conviction records in “any proceeding” in immigration court.

§ 1229a(c)(3)(C).

4 That the government always seeks out such records from

state authorities is an additional reason the burden of production properly rests with the government. “The relative ease with

which the opposing parties can gather evidence is a familiar consideration in allocating the burden of production.” Cooper v. Harris, 137 S. Ct. 1455, 1491 (2017) (Alito, J., concurring in the

judgment in part and dissenting in part); see Matter of Vivas, 16

I. & N. Dec. 68, 70 (BIA 1977). As between noncitizens (who are

often detained and unrepresented), and the government (which

cooperates with state counterparts to obtain the records anyway), the government is better positioned to gather those records. See Moncrieffe, 569 U.S. at 201; Marinelarena, 930 F.3d at

1053 n.10; Br. of Immigrant Defense Project 11-18, 22-23, 25-26.

22

government “indicates that one or more” bars to relief

“may apply,” 8 C.F.R. § 1240.8(d), does “the burden of

proof … shift[] to the [noncitizen] to show by a preponderance of the evidence that the bar is inapplicable.” In re S-K-, 23 I. & N. Dec. 936, 939 (BIA 2006).

Said otherwise, although the noncitizen bears the

burden of persuasion, the government bears the “initial burden” to “secure and produce direct evidence”

that some ground of ineligibility applies. Matter of AG-G-, 25 I. & N. Dec. 486, 501 (BIA 2011). 5

The upshot is that if, after examining the applicant’s criminal background, the government believes

that he has been convicted of an offense that would

disqualify him from relief, it will raise the issue to the

IJ by moving to pretermit the application for relief.

And that is exactly what happened here. The government obtained and submitted the conviction records

to the immigration court. See Pet. Br. App. 2a (C.A.R.

162) (certificate of Immigration and Customs

5 The government suggests that § 1229a(c)(4)(B) places the

burden of producing conviction documents on noncitizens because it mandates that an applicant “comply with the applicable

requirements to submit information or documentation in support of the applicant’s application for relief.” Gov’t Br. 30. But

the application form that Mr. Pereida completed imposed no obligation to produce conviction documents. All it required was

that Mr. Pereida “‘give a brief description of each offense including the name and location of the offense, date of conviction, [and]

any penalty imposed.’” C.A.R. 128; see Executive Office for Immigration Review, U.S. Dep’t of Justice, Form EOIR-42B, Application for Cancellation of Removal and Adjustment of Status for

Certain

Nonpermanent

Resi

d ents

(Rev.

Oct.

2008),

http://cdn.ca9.uscourts.gov/datastore/library/2013/02/26/

Young_eoir42b.pdf. Mr. Pereida complied with that instruction.

See C.A.R. 104, 128, 204-18.

23

Enforcement officer who obtained the records); C.A.R.

161 (“DHS Filing” of “Conviction Records” in immigration court); Gov’t Br. 9. It then moved to pretermit

Mr. Pereida’s application on the ground that the offense constituted a disqualifying CIMT. See Pet. App.

21a; C.A.R. 150-59.

This makes perfect sense. To prove that he “has

not been convicted of an offense” listed in the INA, 8

U.S.C. § 1229b(b)(1)(C), a noncitizen need not show

the absence of conviction records in every jurisdiction

from Autauga County, Alabama, to Weston County,

Wyoming. Rather, the government puts the issue in

play by producing the records it believes establish a

disqualifying offense—and because the government

seeks out those records, there is nothing for the

noncitizen to “conceal.” The noncitizen then bears the

burden of showing that the conviction is not disqualifying—but, for all the reasons already explained, that

exercise usually involves making only a legal argument about the conviction, given the nature of the categorical inquiry and its governing presumption. And

if, at the end of that inquiry, “the record of conviction

of the predicate offense” has not “necessarily establishe[d]” a disqualifying offense, Moncrieffe, 569 U.S.

at 197-98 (emphasis added), then the noncitizen “has

not been convicted of” a disqualifying offense, as a

matter of law. § 1229b(b)(1)(C) (emphasis added). He

may then proceed with his application for relief,

which the Attorney General may grant or deny as he

sees fit.

24

CONCLUSION

The Court should reverse the judgment of the

Eighth Circuit.

Respectfully submitted,

David V. Chipman

Raul F. Guerra

MONZÓN, GUERRA &

ASSOCIATES

1133 H Street

Lincoln, NE 68508

Brian P. Goldman

Counsel of Record

Thomas King-Sun Fu

Kory DeClark

ORRICK, HERRINGTON &

SUTCLIFFE LLP

405 Howard Street

San Francisco, CA 94105

(415) 773-5700

brian.goldman@orrick.com

Thomas M. Bondy

Benjamin P. Chagnon

Monica Haymond

ORRICK, HERRINGTON &

SUTCLIFFE LLP

E. Joshua Rosenkranz

1152 15th Street NW

ORRICK, HERRINGTON &

SUTCLIFFE LLP

Washington, DC 20005

51 West 52nd Street

New York, NY 10019

March 30, 2020