Nijhawan v. Atty Gen USA

Opinions of the United
2008 Decisions States Court of Appeals
for the Third Circuit

5-2-2008

Nijhawan v. Atty Gen USA
Precedential or Non-Precedential: Precedential

Docket No. 06-3948

Follow this and additional works at: http://digitalcommons.law.villanova.edu/thirdcircuit_2008

Recommended Citation
"Nijhawan v. Atty Gen USA" (2008). 2008 Decisions. Paper 1096.
http://digitalcommons.law.villanova.edu/thirdcircuit_2008/1096

This decision is brought to you for free and open access by the Opinions of the United States Court of Appeals for the Third Circuit at Villanova
University School of Law Digital Repository. It has been accepted for inclusion in 2008 Decisions by an authorized administrator of Villanova
University School of Law Digital Repository. For more information, please contact Benjamin.Carlson@law.villanova.edu.
 PRECEDENTIAL

UNITED STATES COURT OF APPEALS
FOR THE THIRD CIRCUIT
__________

No. 06-3948
__________

MANOJ NIJHAWAN,
Petitioner

v.

ATTORNEY GENERAL
OF THE UNITED STATES,
Respondent
__________

Petition for Review of an Order of the
United States Department of Justice
Board of Immigration Appeals
(BIA No. A39 075 734)
Immigration Judge: Walter A. Durling
__________

Argued December 11, 2007
 Before: RENDELL and STAPLETON, Circuit Judges,
and IRENAS*, District Judge.

(Filed May 2, 2008)

Thomas E. Moseley [ARGUED]
Suite 2600
One Gateway Center
Newark, NJ 07102
Counsel for Petitioner

Michelle G. Latour
Lyle D. Jentzer
Jamie M. Dowd [ARGUED]
U.S. Department of Justice
Office of Immigration Litigation
P. O. Box 878
Ben Franklin Station
Washington, DC 20044
Counsel for Respondent
__________

OPINION OF THE COURT
__________

__________________

* Honorable Joseph E. Irenas, Senior Judge of the United
States District Court for the District of New Jersey, sitting by
designation.

2
RENDELL, Circuit Judge.

Manoj Nijhawan appeals from the determination of the
Board of Immigration Appeals (“BIA”) that he had committed
an aggravated felony and was thus removable under 8 U.S.C. §
1101(a)(43)(M)(i) because his conspiracy conviction constituted
an offense involving fraud or deceit in which the loss to the
victims exceeded $10,000. Nijhawan challenges both aspects of
this finding, the “involving fraud” prong as well as the “loss”
aspect. As to the latter, he contends that, in order to satisfy the
qualifying language, the loss amount had to have been
adjudicated as part of his conviction, and was not. We reject
both challenges and will proceed to address each in turn.

The indictment involved a scheme by individuals who, it
was alleged, set out to deprive their victims, major banks, of
“hundreds of millions of dollars.” A.R. 229. Through a series
of misrepresentations, the banks were induced to make a number
of loans to the defendants’ companies, among them Allied
Deals, Inc. Nijhawan, who was the Deputy General Manager of
Allied Deals, Inc., was listed in Count 1, the overall conspiracy
count that contained the general loss allegation as to the entire
fraud scheme and involved conspiracy to commit bank fraud,
mail fraud, and wire fraud in violation of 18 U.S.C. § 371, and
in Count 30, which alleged conspiracy to commit money
laundering in violation of 18 U.S.C. § 1956(h). The remaining
counts were fraud counts implicating one or more of the other
defendants in specific fraudulent loans ranging from $163,441
to $2,568,526. The case was tried before a jury, which
convicted Nijhawan of all of the counts against him in the
indictment. The jury was not asked to, nor did it, determine the

3
amount of the loss attributable to any defendant.

Nijhawan entered into a stipulation for sentencing
purposes in which he agreed that, “because the loss from the
offense exceeds $100 million, the offense level is increased 26
levels.” A.R. 264. In entering the judgment of conviction, the
trial judge filled in the space for “loss” with the amount
“$683,632,800.23.” A.R. 281. The form footnoted the fact that
“findings for the total amount of losses are required under
Chapters 109A, 110, 110A, and 113A of Title 18.” A.R. 281.
Nijhawan was sentenced to 41 months of imprisonment and
ordered to pay restitution in the amount of $683,632,800.23. No
appeal was taken.

While Nijhawan was serving his sentence, he was
charged with removability under 8 U.S.C. § 1101(a)(43)(D) for
conviction of a money laundering offense under 18 U.S.C. §
1956 for which the amount laundered exceeded $10,000 and
under 8 U.S.C. § 1101(a)(43)(M)(i) for conviction of a crime
involving fraud or deceit in which loss to the victims exceeded
$10,000. The IJ sustained both charges, relying primarily on the
§ 1101(a)(43)(D) charge, and entered an order of removal on
February 22, 2006.

On appeal, the BIA rested its decision solely on the 8
U.S.C. § 1101(a)(43)(M)(i) charge. A.R. 2 (“We will affirm the
decision of the Immigration Judge insofar as he found the
respondent removable as an alien convicted of an aggravated
felony as defined in sections 101(a)(43)(M)(i) and (U) of the
Immigration and Nationality Act”). The BIA rejected
Nijhawan’s argument that fraud in the Immigration and

4
Nationality Act (“INA”) should be congruent with the common
law meaning of the term. As to the loss determination, the BIA
agreed that loss was not a necessary element of the offense for
which he was convicted, noting that the loss requirement “was
used as a qualifier, in a way similar to length of sentence
provisions in other aggravated felony subsections.” A.R. 4
(citing Singh v. Ashcroft, 383 F.3d 144, 161 (3d Cir. 2004)). It
reasoned that, although the jury had not found a specific dollar
amount in rendering its guilty verdict, the IJ could properly find
loss based on the stipulation of facts for sentencing and the
judgment of conviction stating that the loss involved is
$683,632,800.23, jointly and severally. A.R. 4-5. The BIA held
that the stipulation, judgment of conviction, and restitution order
were “sufficient to establish that the respondent’s conviction
renders him removable.” A.R. 5.

Nijhawan timely filed a petition for review, appealing the
BIA’s decision.1 On appeal, Nijhawan argues (1) that his
offense of conviction does not involve fraud or deceit as those
terms are used in the INA; and (2) that his conviction did not
establish that loss to his victims exceeded $10,000.

1. Did the offense “involve fraud”

1
After serving his sentence, Nijhawan risked being
immediately removed from the United States by United States
Immigration and Customs Enforcement. He, therefore, filed a
motion for a stay of removal, which we granted pending the
resolution of the present appeal.

5
 Nijhawan was convicted of conspiracy to commit fraud
in violation of 18 U.S.C. § 371. The INA provision under which
Nijhawan was charged with removability provides:

(43) The term “aggravated felony” means--

...

(M) an offense that--

(i) involves fraud or deceit in which the
loss to the victim or victims exceeds
$10,000.

8 U.S.C. § 1101(a)(43)(M)(i). Nijhawan contends that the
“fraud” and “deceit” in this provision should be given their
common law meaning, which requires actual reliance upon
allegedly fraudulent statements made and harm from that
reliance. Because actual reliance and harm from reliance are
not necessary legal elements of the federal fraud statutes under
which he was convicted, Neder v. United States, 527 U.S. 1
(1999), his conviction, Nijhawan urges, was not an aggravated
felony. We can easily dispense with this argument.

In Valansi v. Ashcroft, we examined the very section of
the INA at issue here and interpreted the language broadly. 278
F.3d 203 (3d Cir. 2002). We said:

we determine whether the phrase “offense
that-involves fraud or deceit” has a plain
meaning. The word “involves” means “to

6
 have within or as part of itself” or “to
require as a necessary accompaniment.”
Webster’s Third New International
Dictionary at 1191. Thus, an offense that
“involves fraud or deceit” is most naturally
interpreted as an offense that includes
fraud or deceit as a necessary component
or element. It does not require, however,
that the elements of the offense be
coextensive with the crime of fraud.

Id. at 209-10 (emphasis added); see also Bobb v. Att’y Gen., 458
F.3d 213, 218 (3d Cir. 2006) (“[W]e have held that subsection
(M)(i) covers all offenses that have as an essential element an
intent to defraud or deceive.”); Ki Se Lee v. Ashcroft, 368 F.3d
218, 222 (3d Cir. 2004) (“Subsection (M)(i) has a general
application–the gamut of state and federal crimes involving
fraud and deceit causing losses over $10,000.”).

Other circuits have followed our lead. See Conteh v.
Gonzales, 461 F.3d 45, 59 (1st Cir. 2006) (“We agree with the
Third Circuit. . . . An offense with a scienter element of either
intent to defraud or intent to deceive categorically qualifies as an
offense involving fraud or deceit.”); James v. Gonzales, 464
F.3d 505, 508 (5th Cir. 2006) (noting that “[w]e recognize that
‘[w]hether an offense “involves” fraud is a broader question
than whether it constitutes fraud’” and concluding that “[t]he
plain language of § 1344 . . . provides that a violation of either
subsection necessarily entails fraud or deceit”).

Here, the criminal statutes under which Nijhawan was

7
convicted require that fraud or false or fraudulent pretenses be
employed (mail fraud, wire fraud, and bank fraud). They
therefore “involve” fraud or deceit for the purposes of the INA.
Clearly, Nijhawan’s arguments to the contrary are foreclosed by
our precedent.

2. Was Nijhawan convicted of a fraud “in which
the loss to the victims exceeded $10,000”?

Nijhawan was convicted of conspiracy to commit fraud
and therefore is subject to removal under 8 U.S.C. §
1101(a)(43)(U), which provides that “an attempt or conspiracy
to commit” another aggravated felony constitutes an aggravated
felony. The precise aggravated felony provision at issue here
defines an aggravated felony as an offense that “involves fraud
or deceit in which the loss to the victim or victims exceeds
$10,000.” 8 U.S.C. § 1101(a)(43)(M)(i).

As we have noted above, the “involves fraud” language
of this provision permits the range of actual offenses to be
broader than common law fraud. The issue remains, however,
whether the language “in which the loss to the victim or victims
exceeds $10,000” requires that a jury have actually convicted
defendant of a loss in excess of $10,000, as Nijhawan contends,
or permits resort to the prior criminal record in order to
determine what loss was in fact occasioned by or attributable to
the offense of conviction.

We conclude that the language of § 101(a)(43)(M)(i)
does not require a jury to have determined that there was a loss
in excess of $10,000. To read the “in which” language as

8
requiring that what follows must have been proven as an
element of the crime would bring about an absurd result.
Clearly, the phrase is, as the BIA found, qualifying and does not
constitute a provable element. For example, what if the
language were “in which the victims were elderly” or “in which
three or more banks suffered losses”? Would the facts of these
qualifying phrases have to have been proven as part of the
offense? We suggest not.

To hold to the contrary would essentially gut every
deportability standard containing the “in which” or other
analogous qualifying language,2 for we cannot imagine previous

2
8 U.S.C. § 1101(a)(43)(M)(ii) (conviction for an offense
“that is described in section 7201 of Title 26 (relating to tax
evasion) in which the revenue loss to the Government exceeds
$10,000”); see also id. § 1101(a)(43)(D) (“an offense described
in section 1956 of Title 18 (relating to laundering of monetary
instruments) or section 1957 of that title (relating to engaging in
monetary transactions in property derived from specific
unlawful activity) if the amount of the funds exceeded
$10,000”).
As we noted in Singh, analogous provisions include all
subsections that limit convictions to those “for which the term
of imprisonment is at least one year.” Id. §§ 1101(a)(43)(F),
(G), (J), (P), (R), & (S). Also relevant are subsections that
exempt from the definition of aggravated felony “the case of a
first offense for which the alien has affirmatively shown that the
alien committed the offense for the purpose of assisting,
abetting, or aiding only the alien’s spouse, child, or parent (and

9
convictions in which an aspect of the crime that is not an
element has been proven by the jury. To hold to the contrary
would impose a totally impractical standard.

Notwithstanding our belief that reasonable minds could
not differ on this issue, we acknowledge that other courts of
appeals, and, indeed, Judge Stapleton, have reached a contrary
conclusion. They have done so based upon the very argument
that Nijhawan makes here, namely that the Taylor v. United
States, 495 U.S. 575 (1990), and Shepard v. United States, 544
U.S. 13 (2005), line of cases require it. We suggest that this
reasoning is flawed. The “in which” qualifying language
renders the analysis under § 101(a)(43)(M)(i) different from the
approach in Taylor and Shepard. In fact, we have already so
stated.

In Singh v. Ashcroft, Judge Becker explored the contours
of the applicability of the Taylor-Shepard approach to the
concept of “aggravated felony” in the INA. 383 F.3d 144 (3d
Cir. 2004). Both Taylor and Shepard involved the question of
which court documents or records can be consulted to determine
whether a prior conviction qualifies for a sentencing
enhancement in a subsequent criminal proceeding.3 These cases

no other individual) to violate a provision of this chapter.” Id.
§§ 1101(a)(43)(N) & (P).
3
In Taylor v. United States, the Supreme Court held that
an enhancement for a prior conviction for “burglary” under
§ 924(e) required that either the statutory definition substantially

10
set forth what have become known as the “categorical” and
“modified categorical” approaches to determining the crime of
which the defendant was previously convicted. The categorical
approach looks at the statute of conviction, comparing elements
of the offense to the requirements of the enhancing provision.
When the formal categorical approach of Taylor does not yield
an answer, two different types of inquiry may be called into
play. Judge Becker reviewed our jurisprudence and reasoned as
follows regarding the precise issue before us:

Our jurisprudence in the aggravated
felony area – twelve cases in all – is
not a seamless web. In order to

correspond to “generic” burglary or the record demonstrate that
the jury necessarily found all of the elements of generic burglary
in order to convict the defendant. 495 U.S. 575 (1990).
In Shepard v. United States, the issue was whether the
defendant’s prior plea of guilty to burglary, under a statute that
included generic burglary as well as nongeneric burglary such
as burglary of a boat or motor vehicle, was a conviction for the
violent felony of generic burglary under the Armed Career
Criminal Act. 544 U.S. 13 (2005). The Supreme Court
rejected the notion that police reports or complaint applications
could be used to show that the defendant had necessarily pled to
the qualifying type of burglary, ruling instead that the sentencing
court must look only at the “statutory definition, charging
document, written plea agreement, transcript of plea colloquy,
and any explicit factual finding by the trial judge to which the
defendant assented.” Id. at 16.

11
 resolve the appeal we have found it
necessary to analyze and synthesize
this body of case law, and we do so
at length. . . . As will appear, a
pattern emerges, causing us to
conclude that, while the formal
categorical approach of Taylor
presumptively applies in assessing
whether an alien has been
convicted of an aggravated felony,
in some cases the language of the
particular subsection of 8 U.S.C.
§ 1101(a)(43) at issue will invite
inquiry into the underlying facts of
the case, and in some cases the
disjunctive phrasing of the statute
of conviction will similarly invite
inquiry into the specifics of the
conviction.

Singh, 383 F.3d at 148 (emphasis added). Judge Becker thus
correctly drew the crucial distinction between deportability
language that, on the one hand, calls Taylor and Shepard into
play, inviting inquiry into the specifics of the conviction, and, on
the other, is essentially qualifying language not demanding a
categorical analysis, but requiring, instead, inquiry into the
underlying facts. Cases in which a court has recourse to the
modified categorical approach generally involve “divisible”
statutes, where the prior criminal offense, by statute, includes a
wide range of activity but the requisite enhancing provision –
such as violent felony or aggravated felony – requires one or

12
more particular elements that may or may not have been found
as part of the conviction. The modified categorical approach
entails scrutiny of the nature of the conviction itself and those
elements that the jury necessarily found through an examination
of judicial record evidence. If the jury did not necessarily find
that element, the “conviction” will not fit within the enhanced
category. Taylor-Shepard is thus implicated.

On the other hand, the instant enhancing provision is
different. The language does not state “convicted of a $10,000
fraud.” Rather, it reads, “involves fraud or deceit in which the
loss to the victim or victims exceeds $10,000.” 8 U.S.C. §
1101(a)(43)(M)(i). As Judge Becker specifically stated, the
provision before us here invites inquiry into “the underlying
facts of the case.” There is no issue here regarding which crime
was committed by the petitioner under a divisible statute, in
which event we would be limited to an examination of the
“specifics of the conviction” and would employ the modified
categorical approach of Taylor and Shepard.

Addressing the analysis required under the very provision
at issue here, Judge Becker made clear that:

[A] departure from the formal
categorical a pproach seem s
warranted when the terms of the
statute invite inquiry into the facts
underlying the conviction at issue.
The qualifier “in which the loss to
the victim or victims exceeds
$10,000” in 8 U.S.C.

13
§ 1101(a)(43)(M )(i) is the
prototypical example – it expresses
such a specificity of fact that it
almost begs an adjudicator to
examine the facts at issue. This
principle explains our holdings in
Nugent and Munroe. Another
example would be an enumerating
statute specifying crim es
“committed within the last two
years.” Such a statute could not be
read to cover only crimes which
have “within the last two years” as
an element; instead a court would
read “within the last two years” as
a limiting provision on crimes that
would otherwise qualify.

In contrast, cases interpreting
relatively unitary categorical
concepts – like “forgery” (Drakes),
“burglary” (Taylor itself) or “crime
of violence” (Francis and Bovkun)
– do not look to underlying facts
because the enumerating statute
does not invite any such inquiry.
Likewise, the hypothetical federal
felony trilogy (Steele, Gerber, and
Wilson) asks only whether the
elements of a federal criminal
statute can be satisfied by reference

14
 to the actual statute of conviction;
this presents no invitation to depart
from Taylor’s formal categorical
approach an d ex am ine the
underlying facts.

383 F.3d at 161 (emphasis added). Accordingly, our Court’s
precedent directs us to “examine the facts at issue,” because the
amount of loss is a “qualifier,” not an element.4

Our case law has consistently treated the amount of loss
as a qualifier rather than an element of the crime. In 2003 in
Munroe v. Ashcroft, we did not require that the defendant have
specifically pled guilty to a loss amount. 353 F.3d 225 (3d Cir.
2003). To the contrary, we stated that “the indictment alleged
that the loss exceeded this amount, and Munroe does not claim
that when he pled guilty, he admitted to only a lesser loss. Nor
is there any suggestion that the Superior Court ever found that
the amount of the loss was less than $10,000.” Id. at 227. For
the purposes of § 101(a)(43)(M)(i), we looked to the indictment,
which contained an averment as to loss in excess of $10,000,

4
Our Court’s view regarding the meaning of, and inquiry
permitted by, 8 U.S.C. § 1101(a)(43) has been referenced
approvingly by other courts. See, e.g., James v. Gonzales, 464
F.3d 505, 510 n. 26 (5th Cir. 2006); Conteh v. Gonzales, 461
F.3d 45, 55 (1st Cir. 2006). We have also applied its rationale
in interpreting other provisions of the INA. See Joseph v. Att’y
Gen., 465 F.3d 123, 127 (3d Cir. 2006); Knapik v. Ashcroft, 384
F.3d 84, 92 n. 8 (3d Cir. 2004).

15
rather than an amended restitution order, which reduced
defendant’s restitution to $9,999.5 However, we decided that
the amount of loss in a restitution order, which by its nature is
neither found by a jury nor specifically pled to by a defendant,
could be considered. Id. at 227 (“[T]he amount of restitution
ordered as a result of a conviction may be helpful to a court’s
inquiry into the amount of loss to the victim if the plea
agreement or indictment is unclear as to the loss suffered.”).

Nijhawan contends that more recent authority, namely,
our opinion in Alaka v. Attorney General, 456 F.3d 88 (3d Cir.
2006), contradicts Singh and Munroe and requires conviction of
the requisite amount of loss. In Alaka, the total loss averred in
the indictment as to the overall scheme exceeded $10,000.
However, Alaka pled guilty only to a single count in a plea
agreement that referenced a loss to the victim of $4,716.68. Id.
at 92. The other counts against Alaka were dismissed. We
concluded that Alaka’s offense did not qualify for treatment as

5
The dissent states that the holding in Munroe was based on
a loss amount “admitted in the plea agreement.” This is
incorrect as the district court’s opinion in that case makes clear.
Munroe v. Ashcroft, No. Civ. A. 02-2256, 2003 WL 21048961
(E.D. Pa. Jan. 16, 2003) (“In this case, the indictment stated that
the fraud involved caused a loss to the victim in excess of
$10,000.00. There is no evidence that the defendant pled guilty
to any facts other than as alleged in the indictment.”). The
holdings of both our court and the district court relied on the
amount alleged in the indictment and found by the sentencing
court, not an amount in the plea agreement.

16
an aggravated felony. Id. at 108.

Nijhawan urges that Alaka stands for the proposition that
the loss amount is an element to which the defendant must
plead, or of which the defendant must be convicted. We
disagree. To the contrary, Alaka stands for the unremarkable
proposition that one who has admitted to a loss of less than
$10,000 as part of a guilty plea cannot later be said to have been
convicted of an offense involving fraud in which the loss to the
victim exceeds $10,000. Where there is a plea agreement that
sets forth the loss it is to that agreement we must look to
determine the loss. Alaka does not require that the defendant
plead to a specific loss amount; it requires only that, if he has,
that amount is controlling. Alaka does not limit the inquiry if
no loss is stated in a plea agreement or submitted to a jury. In
fact, in Alaka we concluded that “the IJ properly considered the
factual finding in the sentencing report.” Id. at 105, 106. Alaka
requires only that we “focus narrowly on the loss amounts that
are particularly tethered to the convicted counts.” Id. at 107
(quoting Knutsen v. Gonzales, 429 F.3d 733, 739-40 (7th Cir.
2005)).

The only real issue in the case before us is whether the
“tether” of a loss in excess of $10,000 to Count 1, the count of
conviction, is sufficiently strong. We have not previously
opined as to the nature of the nexus required, or the breadth of
the inquiry into the facts as authorized by Singh, and, here, we
need only determine whether the record is sufficiently clear that
the loss resulting from the convicted conduct exceeds $10,000.

Here, Count 1 of the indictment charged a conspiracy,

17
alleging that defendants “engaged in a fraudulent scheme to
obtain millions of dollars in loans” from the victim banks and
setting forth the scheme and roles of the co-conspirators.
Nijhawan was convicted of Count 1, but the jury did not, and
was not asked to, determine the amount of loss to the victims.
However, in a stipulation for the purposes of sentencing on
Count 1, Nijhawan agreed that the loss exceeded $100 million.
And, in entering the judgment of conviction, the District Court
made a finding of “Total Loss” in the amount of
$683,632,800.23. As in Munroe, here we have no argument, let
alone anything in the record, that Nijhawan was convicted of an
offense involving less than $10,000. This is not a case where
the jury’s findings contradict the restitution order or loss was
calculated on the basis of uncharged or unconvicted conduct.
All the documents and admissions support a finding that the loss
amounted to hundreds of millions of dollars.

We need not decide whether any of the “facts” here,
standing alone, would suffice as a “tether,” as we conclude that,
taken together, the indictment, judgment of conviction, and
stipulation provide clear and convincing evidence that the
requisite loss was tied to Nijhawan’s offense of conviction.

We note that we are not the only court of appeals to have
viewed the inquiry into the record of conviction to permit
examination of loss not specifically admitted in the plea
colloquy or agreement or found by a jury as part of the

18
conviction.6 The Court of Appeals for the First Circuit also has
indicated that a court should look to loss occasioned by the
conviction, rather than loss as an element found by the jury or
explicitly incorporated in the plea agreement. Conteh v.
Gonzales, 461 F.3d 45 (1st Cir. 2006). Exercising care in
interpreting the “loss exceeds” language in § 1101(a)(43)(M)(i),
the court recognized that “the distinction between conviction for
and commission of an aggravated felony is an important one;
because the BIA may not adjudicate guilt or mete out criminal
punishment, it must base removal orders on convictions, not on
conduct alone.” Id. at 56. Thus, the court found it improper for
the BIA to rely on a narrative statement in the PSI report, but did
approve the BIA’s reliance on the indictment, which alleged
specific losses exceeding $10,000, and the final judgment,
which included a finding of loss and restitution order. Id. at 59
(quoting Shepard, 544 U.S. at 21). As here, an indictment and
judgment, indicating loss and restitution, were available and
were a sufficiently reliable indication of the loss of which the
petitioner had been convicted.

6
Other courts permit a broader inquiry and have allowed
loss amount to be established by reference to conduct that
formed part of the same conspiracy as the convicted conduct, a
broader inquiry than that we have here. See Khalayleh v. INS,
287 F.3d 978 (10th Cir. 2002) (where alien pleaded guilty to one
count of the indictment which listed a check in the amount of
$9,308 but agreed to pay restitution as determined by the
sentencing court, the loss from the total scheme to defraud
involving other checks could be counted); see also James, 464
F.3d at 511-12 (following Khalayleh).

19
 In Knutsen v. Gonzales, a case upon which we relied in
Alaka, the Court of Appeals for the Seventh Circuit similarly
reasoned that “consistent with the statute . . . the court should
focus narrowly on the loss amounts that are particularly
tethered to convicted counts alone.” 429 F.3d 733, 739-40 (7th
Cir. 2005) (emphasis added). In that case, the petitioner had
pled guilty to one count of a multi-count indictment, which
listed a loss amount less than $10,000; for the purposes of the
Sentencing Guidelines, however, he entered into a stipulation
with the government in which he acknowledged that “the total
loss from the offense of conviction and relevant conduct
exceeded $20,000.” Id. at 736. Because the stipulation included
relevant conduct and was not limited to the loss connected to or
caused by the offense of conviction, the court found that the IJ
erred by relying on it, but did not require the plea colloquy to
have included the specific loss. Id. at 739. The loss was not
sufficiently “tethered” to the offense of conviction so as to
constitute clear and convincing evidence that the petitioner had
been convicted of an aggravated felony under §
1101(a)(43)(M)(i).

The decision of the Court of Appeals for the Eleventh
Circuit in Obasohan v. Attorney General further substantiates
our interpretation of this provision. 479 F.3d 785 (11th Cir.
2007). In that case, the petitioner had been ordered to pay
restitution, due to fraudulent charges on other credit cards that
were not the subject of the indictment or the plea agreement. Id.
at 789-90. The court found it particularly significant that the
petitioner objected to the PSI’s assertion of loss due to
additional conduct and “therefore did not admit, adopt, or assent
to the factual findings that formed the basis of the restitution

20
order.” Id. at 790. Given that the restitution order was based
entirely on other unconvicted conduct, was not admitted by the
petitioner, and was the only evidence that such loss had
occurred, the IJ could not find loss by clear and convincing
evidence. Id. at 790 (gathering cases and citing Knutsen,
Munroe, and Conteh with approval). A restitution order could
be evidence of the loss amount, but only if it was “based on the
conspiracy charge to which Obasohan pled guilty, []or on the
overt acts to which Obasohan admitted by pleading guilty,” not
“on additional conduct that was alleged only in the PSI.” Id. at
789-90.

We should note that neither we nor these other courts
have abandoned the Taylor-Shepard approach. Indeed, we still
resort to it at the initial phase of our analysis because §
101(a)(43)(M)(i) instructs us to decide whether the alien has
been convicted of a crime involving fraud or deceit. Employing
the formal categorical approach and looking to the statute of
conviction, we determined that Nijhawan’s conviction involved
fraud or deceit and thus was a proper predicate offense within
the “aggravated felony” definition. Once this conclusion is
reached, our case law then requires an “inquiry into the
underlying facts of the case” to ascertain whether the “in which”
qualifying loss provision is satisfied.

Nijhawan urges that we should depart from our case law
and follow those courts of appeals that have interpreted the loss
requirement in INA § 101(a)(43)(M)(i) in a more restrictive
way. In particular, he urges that we should adopt the reasoning
of the Court of Appeals for the Second Circuit, which has set
forth a rule that the loss requirement must be established by

21
“facts actually and necessarily found beyond a reasonable doubt
by a jury or judge in order to establish the elements of the
offense, as indicated by a charging document or jury
instructions.” Dulal-Whiteway v. U.S. Dep’t of Homeland Sec.,
501 F.3d 116, 131 (2d Cir. 2007).7 There, the court determined
that, because restitution was not necessarily found by the jury,
a restitution order was insufficient to establish that the fraud
conviction was one “in which the loss to the victims exceeds
$10,000.” Id. at 130. The Court of Appeals for the Ninth
Circuit also applied the “modified categorical” approach to the
loss requirement in Li v. Ashcroft and found that it was improper
to rely on the charging document, which described specific loss
amounts, and the judgment of conviction for those counts,
because it had “in the record no jury instructions, verdict form,
or other comparable document suggesting that the jury actually
was called on to decide, for example, that Petitioner’s false
claims were for a particular amount.” 389 F.3d 892, 898 (9th
Cir. 2004) (expressing no opinion however “as to whether a
defendant’s admission of a specific sentencing fact would
suffice”). As we noted above, we conclude that this treatment
of the qualifying language as setting forth an element of the
offense is uncalled for and makes little sense. While our
dissenting colleague urges that express conviction of the loss
amount will lend certainty and ease to the analysis, we do not

7
In the case of pleas of guilty, the dissent’s rule restricts
inquiry to “facts to which a defendant actually and necessarily
pleaded in order to establish the elements of the offense, as
indicated by a charging document, written plea agreement, or
plea colloquy transcript.” 501 F.3d at 131.

22
think this justifies our embracing an interpretation of the
language that will render the provision toothless.

Moreover, our case law clearly rejects the restrictive
interpretation of INA § 101(a)(43)(M)(i)’s loss requirement
adopted by the Second Circuit in Dulal-Whiteway and the Ninth
Circuit in Li. Munroe, Singh, and Alaka make clear that the loss
amount need not be found specifically by the jury or set forth in
the plea agreement or colloquy.8 Rather, as we have said, the
loss requirement invites further inquiry into the facts underlying
the conviction, and that inquiry is satisfied if the amount of loss
is sufficiently tethered to the fraud conviction.

Had our prior precedent not compelled our conclusion,
we still would firmly disagree with the restrictive interpretation.
For, our decision actually fosters the principles the Second
Circuit identified in Dulal-Whiteway and best comports with the
text and purpose of the INA’s aggravated felony provision. In
Dulal-Whiteway, the Second Circuit noted that the words of the
INA provision render deportable one who has been convicted of
an aggravated felony, not one who has committed an aggravated
felony. Id. at 132. We do not disagree with this and, much like
the Court of Appeals for the First Circuit in Conteh, we endorse
careful consideration of the record to determine whether it is
sufficiently clear that the loss connected to the crime of
conviction exceeded $10,000. As Judge Becker noted in Singh,

8
In order to reach a contrary result, the dissent labels salient
portions of our prior precedent “dicta.” See dissenting op., n. 9
& 11.

23
the specific words “in which the loss to the victims exceeds
$10,000” suggest just such an inquiry into the facts underlying
the conviction. The requirement that we set forth today that the
loss amount be sufficiently tied or tethered to the offense of
conviction both responds to the Second Circuit’s concern that a
restitution order based upon conduct of which the defendant was
not convicted should not be relied on, and does not arbitrarily
cabin the inquiry.9

The difficulty in saying that the court will limit inquiry to
the precise “record of conviction” used in the Armed Career
Criminal context for purposes of determining loss under §
101(a)(43)(M)(i) is made manifest in the decisions of the Court
of Appeals for the Ninth Circuit. That court appears to adopt

9
The dissent posits that our opinion permits consideration
of loss caused by “relevant conduct” rather than the conduct of
conviction. This is not correct. By requiring that loss be
tethered to the convicted conduct, we are excluding
consideration of relevant conduct, as did the Court of Appeals
for the Seventh Circuit in Knutsen and the Court of Appeals for
the Eleventh Circuit in Obasohan. In fact, we use the word
“relevant” only in discussing these courts’ opinions.
Furthermore, there is no conduct in this case other than
that underlying the conviction. The dissent incorrectly states
that the conduct in Nijhawan’s sentencing stipulation pertinent
to the Guidelines enhancement and the restitution order includes
relevant, as well as convicted conduct, as in Obasohan. It does
not. In fact, this very clearly distinguishes Obasahan and
Knutsen, cases with which we agree.

24
the requirement that the petitioner had to have been convicted of
the loss, but then looks beyond what the jury found in order to
determine loss amount. For example, in Ferreira v. Ashcroft,
the court cited our decision in Munroe with approval and
reasoned that there was no rule prohibiting immigration judges
from looking to a restitution order to determine loss amount.
390 F.3d 1091 (9th Cir. 2004) (relying on Munroe and Chang v.
INS, 307 F.3d 1185 (9th Cir. 2002)). Although the court has
insisted that it is using the modified categorical approach, it has
actually engaged in a broader inquiry.

Our holding today is consistent with the different
evidentiary standards used in criminal, sentencing, and
immigration proceedings, respectively. In Dulal, the Court of
Appeals for the Second Circuit criticized the approach we
endorse because, it believed, it “would permit the government
to order an alien removed in the absence of the clear,
unequivocal and convincing evidence required by [immigration]
law.” 501 F.3d at 132. However, its holding raises the standard
of proof to beyond a reasonable doubt while our holding actually
adheres to the “clear and convincing” standard. Accord Conteh,
461 F.3d at 56 (rejecting “the implicit proposition that the INA’s
use of the word “convicted” in 8 U.S.C. § 1227(a)(2)(A)(iii)
elevates the government’s burden in aggravated felony cases
from clear and convincing evidence to proof beyond a
reasonable doubt (that is, proof that facts were necessarily found
by a criminal jury or admitted by the alien qua criminal
defendant)”).

Most fraud statutes, including the federal statutes at issue
here, do not contain loss as an element or require that a jury find

25
loss or a defendant plea to a specific loss amount. As we noted
above, insistence on loss as part of the conduct would render §
1101(a)(43)(M)(i) largely inoperative, for rarely will a defendant
be convicted of a fraud offense with loss as an element found by
the jury or explicitly admitted to in a guilty plea. Under the rule
adopted in Dulal-Whiteway which the dissent embraces, a
finding beyond a reasonable doubt would be required, not
merely the allegation of a specific loss amount in a criminal
indictment.10 A jury would have to be charged as to loss amount
and make a specific and additional finding.11

Unlike the Second Circuit, we find no “‘daunting’
practical difficulties” associated with looking to a wider array of
records that possess a high indicia of reliability. It is well within
the competence of a court to examine the record for clear and
convincing evidence of loss caused by the conduct of

10
In fact, Li, upon which the dissent relies, did not consider
the charging document which listed specific loss amounts and
the judgment of conviction on those counts to be sufficient to
prove the loss amount precisely because the jury was not
required to find a loss amount to a guilty verdict. 389 F.3d at
898. Here the prosecutor did in fact include the loss amount in
the criminal indictment.
11
It would necessarily be the prosecutor who would request
this charge, for, if the rule espoused in Dulal-Whiteway applies,
defense counsel would be content not to have the loss found by
the jury. We must wonder why the prosecutor would ever ask
the jury to find a fact not relevant to the conviction.

26
conviction. Indeed, we believe there are far greater practical
difficulties inherent in attempting to bend the “modified
categorical approach” of Taylor and Shepard to apply to a
finding of the requisite minimum loss caused by fraud or deceit,
which is rarely found by a jury or explicitly included in the plea
agreement, because it is a qualifier, not an element of the
offense. Moreover, we should not raise an aspect of an
immigration statute to the level of an element of a criminal
offense, as the dissent urges, merely because requiring that it be
a part of the conviction eases a court’s decision-making process.

Accordingly, because the petitioner was previously
convicted of conspiracy to commit “an offense that involves
fraud or deceit in which the loss to the victim or victims exceeds
$10,000,” he committed an aggravated felony, and we will deny
his petition for review.

27
Nijhawan v. Attorney General – No. 06-3948

STAPLETON, J., dissenting:

I agree with the Court that Nijhawan’s conviction for
conspiracy to commit bank fraud, mail fraud and wire fraud
constituted a conviction for conspiracy to commit an offense
“that involves fraud or deceit” as defined by the INA. I
therefore join Section 1 of the Court’s opinion. I disagree,
however, with the Court’s conclusion that prior decisions of this
Court compel the approach to the § 1101(a)(43)(M)(i) loss
element that the Court adopts, and I believe that our Court
should retain the INA’s conviction requirement for that element.
I would therefore grant the petition for review.

Under the Immigration and Naturalization Act (“INA”),
“[a]ny alien who is convicted of an aggravated felony at any
time after admission is deportable.” 8 U.S.C. §
1227(a)(2)(A)(iii) (emphasis added). The term “aggravated
felony” is defined in 8 U.S.C. § 1101(a)(43) to include, inter
alia, “an attempt or conspiracy to commit” “an offense that -- (i)
involves fraud or deceit in which the loss to the victim or
victims exceeds $10,000.” Id. at §§ 1101(a)(43)(M)(i),
1101(a)(43)(U). Therefore, under the plain language of the
INA, petitioner is removable only if he was “convicted” of a
conspiracy to commit “an offense that . . . involves fraud or
deceit in which the loss to the victim or victims exceeds

28
$10,000.” Id.

Several Courts of Appeals, including ours, presumptively
apply some variant of the “categorical approach” first articulated
by the Supreme Court in Taylor v. United States, 495 U.S. 575
(1990), and further explained in Shepard v. United States, 544
U.S. 13 (2005), to determine whether an alien’s prior conviction
qualifies as an “aggravated felony.” Courts of Appeals have
diverged, however, regarding how a reviewing court should
determine whether an alien’s prior conviction satisfies the
$10,000 loss requirement of § 1101(a)(43)(M)(i). Although all
Courts of Appeals permit the reviewing court to look beyond
Taylor’s “formal” version of the categorical approach – a simple
comparison of the elements of the prior statute of conviction to
the INA definition – and allow recourse to the “record of
conviction” to some degree, courts disagree regarding the
precise nature of that further inquiry. The Courts of Appeals for
the Second and Ninth Circuits, and, as I read its precedent, the
Eleventh Circuit, have adopted a “modified categorical
approach” in which the reviewing court looks to the record of
conviction in order to determine the facts upon which the
petitioner’s prior conviction actually and necessarily rested.12 In
contrast, the Court of Appeals for the First Circuit allows a

12
Dulal-Whiteway v. U.S. Dep’t of Homeland Security, 501
F.3d 116, 128 (2nd Cir. 2007); Li v. Ashcroft, 389 F.3d 892,
895-98 (9th Cir. 2004); Obasohan v. Attorney General, 479
F.3d 785, 788-89 (11th Cir. 2007).

29
broader inquiry under which immigration courts may scrutinize
other facts, gleaned from the alien’s record of conviction, to
independently determine, by clear and convincing evidence,
whether the crime resulted in a loss greater than $10,000.13 I
find the approach of the Courts of Appeals for the Second and
Ninth Circuits to be the better reasoned approach.

The Supreme Court articulated the Taylor-Shepard
categorical approach when reviewing 18 U.S.C. § 924(e), which
provides for a sentencing enhancement if a defendant has been
convicted of certain enumerated prior offenses. The Courts of
Appeals have transplanted that categorical approach into the
INA because of obvious similarities between the two inquiries.
The plain language of the INA, like § 924(e), mandates that the
alien was “convicted” of the prior offense designated in the INA
as an “aggravated felony.” It is not sufficient for the BIA to
independently conclude that the alien “has committed” that prior
offense. Therefore, the INA, like § 924(e), requires a
comparison of the prior conviction to the generic definition of
the pertinent aggravated felony – in this case, §§
1101(a)(43)(M)(i) and (U).

The rationale is not just a textual one, however. Courts
have adopted categorical approaches for the INA also because

13
See Conteh v. Gonzales, 461 F.3d 45 (1st Cir. 2006).

30
the INA inquiry involves the same sorts of practical difficulties
and fairness concerns underlying the Supreme Court’s decisions
in Taylor and Shepard. As the Second Circuit explained, “the
BIA and reviewing courts are ill-suited to readjudicate the basis
of prior criminal convictions.” Dulal-Whiteway, 501 F.3d at
132. See also id. (“we decline the invitation to piece together an
underlying attempt conviction by weighing evidence and
drawing conclusions in a manner appropriate only for a criminal
jury”) (quoting Sui v. I.N.S., 250 F.3d 105, 119 (2nd Cir. 2001));
Shepard, 544 U.S. at 23 (a purpose of the categorical approach
is the “avoidance of collateral trials”). As the Second Circuit
also recognized, the categorical approach promotes basic
precepts of fairness. Id. at 133 (“‘[I]f the guilty plea to a lesser,
[non-removable] offense was the result of a plea bargain, it
would seem unfair to [order removal] as if the defendant had
pleaded guilty to [a removable offense].’ [Taylor, 495 U.S.] at
601-02. By permitting the BIA to remove only those aliens who
have actually or necessarily pleaded to the elements of a
removable offense, our holding promotes the fair exercise of the
removal power”).14 In sum, I agree with the Court of Appeals

14
The Court of Appeals for First Circuit found such fairness
concerns less than compelling because Shepard had emphasized
that, in the context of sentencing enhancements under § 924(e),
those concerns also raise Sixth Amendment problems, and such
constitutional concerns are inapplicable in civil removal
proceedings. Conteh, 461 F.3d at 55. However, Taylor and
Shepard were rooted in basic notions of fairness that extend
beyond the protections of the Sixth Amendment, and we, like
the Second Circuit, began to adopt categorical approaches for

31
for the Second Circuit that the same practical and fairness
difficulties identified by Taylor and Shepard would attend an
interpretation of the INA that allowed immigration courts to
reopen the factual record of prior criminal convictions and
undertake new factual findings, utilizing a different standard of
proof, to determine whether a required element (a $10,000 loss)
was met. Indeed, if the loss requirement is not subject to the
conviction requirement, why limit the evidentiary net to the
prior record of conviction at all? Absent the conviction
requirement, the standards become arbitrary.15

the INA before Shepard articulated its Sixth Amendment
rationale. Dulal-Whiteway, 501 F.3d at 132-33. See Shepard,
544 U.S. at 20 (“certainly, ‘the practical difficulties and
potential unfairness of a factual approach are daunting,’ no less
in pleaded than in litigated cases”) (internal citation omitted).
15
The Court concludes that the loss must merely be found by
the Immigration Judge and BIA under their “clear and
convincing evidence” standard and be “tethered” to the
conviction. The Court does not define the “tethered” test further
but merely holds that it is satisfied by the facts of this case. The
holding provides no guidance to the Immigration Judges who
will apply Sections 1227(a)(2)(A)(iii) and 1101(a)(43)(M)(i).
Under the standard the Court adopts, for example, would a
future IJ be permitted to conclude (under its clear and
convincing evidence standard) that the $10,000 loss is
established, and is “tethered” to the alien’s conviction, by
looking to facts in a pre-sentence investigation report (“PSI”),
or to facts in a police report, or to select evidence presented in

32
the criminal trial, or to new testimony or documents introduced
at the removal hearing? The task of defining the “tethered”
inquiry will fall to future panels of this Court, and with the loss
element divorced from the conviction requirement, the task will
not be an easy one.
The First Circuit, the only other court to have deviated
from the modified categorical approach, sought to provide
answers to these questions in Conteh, but that opinion
demonstrates the analytical difficulty of defining the loss inquiry
once it is divorced from the conviction requirement. Conteh
made two fundamental rulings regarding the loss inquiry.
Conteh first ruled, as does the Court today, that the INA does
not require a convicted loss but rather merely a determination by
the IJ, under its ordinary clear and convincing evidence
standard, that the loss requirement is satisfied. Conteh, 461 F.3d
at 55-56. This ruling allowed it to conclude that the IJ did not
err by relying on a restitution order, which could have included
“relevant” but un-convicted conduct and facts found by a mere
preponderance of the evidence. Id. at 59. Conteh next,
however, joined every Court of Appeals to have addressed this
issue by ruling that the inquiry is limited to the “record of
conviction.” Id. at 57. In reaching this latter ruling the Court
“emphasize[d] that the difference between [its] approach and
that of the Ninth Circuit [which the Second Circuit subsequently
joined] is only a matter of degree,” id. at 56, and it agreed that
“because the BIA may not adjudicate guilt or mete out criminal
punishment, it must base removal orders on convictions, not on
conduct alone.” Id. Based on this second ruling, the Court
concluded that the IJ did err by looking to a PSI and to

33
 Because of the plain language of the INA, as well as the
practical and fairness concerns that I have discussed, I am wary
of permitting immigration courts to undertake de novo factual
inquiries, under the “clear and convincing evidence” standard,
into facts merely “relevant to,” or “tethered to,” an alien’s prior
conviction. I would permit immigration courts to look to the
record of conviction, but only to establish “that a prior
conviction ‘necessarily’ involved ([or] a prior plea necessarily

testimony presented in the removal hearing: the Court reasoned
that restitution orders (memorialized in the final judgment) were
part of the “record of conviction,” but that the other two types
of evidence were not. Id. at 57-59. The Court allowed recourse
to restitution orders by ruling, as does the Court today, that the
alien need not have been actually convicted of a loss; however,
the Court rejected the IJ’s other two sources of evidence because
they fell outside of the “record of conviction” as that Court
defined it, a limit which must derive from the conviction
requirement. In other words, the Court found that the INA’s
conviction requirement applies to the loss inquiry in some
respects but does not apply to it in other respects. Certainly no
such line appears in § 1101(a)(43)(M)(i). I also note that
allowing unqualified reliance upon restitution orders would
allow future IJs to look to facts a prior sentencing court may
have found by a mere preponderance of the evidence and to
elevate those facts to the higher “clear and convincing evidence”
standard, without the benefit of having the underlying evidence
before it.

34
admitted) facts equating to [the generic offense in the INA
statute].” Shepard, 544 U.S. at 24. See also Dulal-Whiteway,
501 F.3d at 128 (“while the issue of statute divisibility and
reliance upon the record of conviction are theoretically
separable, in practice they demand a single inquiry: has an alien
been actually and necessarily convicted of a removable
offense?”); Li, 389 F.3d at 895-98. The “necessarily” pleaded
or convicted requirement explains and defines the “record of
conviction” inquiry: once the court determines that the statute
of conviction proscribes both conduct that would constitute an
“aggravated felony” and conduct that would not, the court
consults the record of conviction to determine the type of
conduct the conviction necessarily includes. Dulal-Whiteway,
501 F.3d at 131; Li, 389 F.3d at 895-96.

In this case, loss was not an element of the crime of
conviction. The conspiracy count of the indictment did assert a
fraudulent scheme to obtain “hundreds of millions of dollars” in
loans from major banks, but the Court in petitioner’s criminal
trial instructed the jury that it need not find any loss in order to
convict. A.R. at 150, 156, 158. We thus know that despite the
averment of the indictment, the jury’s verdict does not establish
that petitioner was convicted by it of conspiracy to commit fraud
occasioning any particular amount of loss. The BIA and our
Court acknowledge as much. As a result, they point not to the
indictment and verdict to support their conclusion, but rather the
record of the subsequent sentencing proceedings. Specifically,
they focus attention on (1) the sentencing judge’s order that all
defendants be jointly and severally liable for restitution in

35
excess of $10,000; and (2) the petitioner’s stipulation with the
government that a correct application of the U.S. Sentencing
Guideline to petitioner’s convictions on Counts 1 (conspiracy to
commit fraud) and 30 (conspiracy to commit money laundering)
produced a base offense level of 38, an offense level including
an enhancement “[b]ecause the loss from the offense exceeds
$100,000,000.” A.R. at 264. Neither portion of the sentencing
record, however, establishes that petitioner has been “convicted”
of causing a $10,000 loss.

With respect to the sentencing judge’s restitution order,
I agree with the Second and Eleventh Circuits that it does not
support a conclusion of removability. As the Dulal-Whiteway
Court put it in the context of a guilty plea case:

The restitution set by a judge is based on a
loss amount established by a preponderance of the
evidence and need not be tied to the facts
admitted by defendant’s plea. . . . In other words,
the amount of the restitution is not constrained by
facts upon which the plea “necessarily” rested.

Dulal-Whiteway, 501 F.3d at 130. See also Obasohan v.
Attorney General, 479 F.3d 785 (11th Cir 2007) (“[W]hile a
sentencing court in the criminal context may order restitution

36
not only for convicted conduct but also for a broad range of
relevant conduct, the plain language of the INA requires that an
alien have been convicted of an aggravated felony to be
removable.”). I also agree with those courts that a contrary
conclusion would put one facing removal and lifetime exclusion
in a difficult and unfair position.

We note that if the immigration court were
authorized to base a finding of an aggravated
felony on conduct and victim losses that were not
charged, proven or admitted, it would be
impossible for a criminal defendant to evaluate
the immigration consequences of a guilty plea at
the time of entering that plea, because those
consequences would be known only at the time of
sentencing. Where loss amounts are charged and
proven or admitted, however . . . no such concern
arises.

Obasohan, 479 at 791, n.12.

For much the same reasons, I would reach the same
conclusion with respect to the propriety of the BIA consulting
the sentencing stipulation of the parties in this case. The
stipulation with respect to the application of the Sentencing

37
Guidelines in this case is not the equivalent of a plea or plea
agreement admitting to an element of the offense of conviction.
This stipulation came both after petitioner’s conviction and in
the context of a sentencing regime that requires consideration of
losses from relevant as well as convicted conduct.16

It is true, as the Court stresses, that retention of the
convicted conduct requirement will result in the BIA being able
to remove fewer aliens on the ground that they have been
convicted of an aggravated felony. I do not find that
problematic because that appears consistent with the
Congressional intent reflected in 8 U.S.C. § 1227(a)(2)(A)(iii).
If there is a problem, however, I would reserve it for legislative
correction. Furthermore, the modified categorical approach
does not, as the Court suggests, elevate the government’s burden

16
The Court suggests that neither petitioner’s sentencing
stipulation nor the sentencing court’s restitution order involved
consideration of relevant conduct. It fails to explain, however,
how it knows this to be true. The stipulation was solely for the
purpose of a guideline regime that requires consideration of
losses from relevant as well as convicted conduct and, there
being no limitation to the later, the stipulation clearly applied to
both. See U.S.S.G. § 1B1.3, Application Notes 1-2. The
restitution regime, like the Guidelines, also allows the Court to
consider losses from relevant conduct, and nothing I have found
in the record suggests that petitioner’s sentencing court focused
on the distinction.

38
of proof in immigration cases from “clear and convincing
evidence” to “beyond a reasonable doubt.” It merely requires
the government to prove, by clear and convincing evidence, that
the alien was actually “convicted” of the asserted “aggravated
felony.” See Obasohan v. Attorney General, 479 F.3d 785, 790
(11th Cir. 2007) (“There was no basis in this record from which
the IJ could have found by ‘clear, unequivocal and convincing’
evidence that the restitution order was based on convicted or
admitted conduct.”).

This Court has never before found an alien deportable for
conduct the alien was neither convicted of nor pled guilty to; the
Court’s approach, therefore, will significantly expand the reach
of the INA’s “aggravated felony” provisions in this Circuit. As
the Court emphasizes, in Singh v. Ashcroft, 383 F.3d 144 (3d
Cir. 2004), we reviewed our “aggravated felony” jurisprudence
and concluded that we had failed to follow the “formal”
categorical approach in three cases, all of which applied §
1101(a)(43)(M)(i).17 That provision, the Court stated, “begs an
adjudicator to examine the facts at issue.” Id. at 161. Singh did
not explain precisely which facts were “at issue.” However, it
suggested a “further inquiry” much like the one I would adopt.
Singh was decided prior to the Supreme Court’s opinion in

17
Singh itself merely held that, when applying a different
“aggravated felony” definition, “sexual abuse of a minor,” 8
U.S.C. § 1101(a)(43)(A), this Court should follow the strict
categorical approach. Singh, 383 F.3d at 163-64.

39
Shepard, and the Court reviewed our prior case law only to
determine when we had applied the “formal” version of the
categorical approach described in Taylor. As the Singh Court
explained,

“[u]nder that approach, an adjudicator ‘must look
only to the statutory definitions of the prior
offenses,’ and may not ‘consider other evidence
concerning the defendant’s prior crimes,’
including, ‘the particular facts underlying [a]
conviction.”

Singh, 383 F.3d at 148 (quoting Taylor, 495 U.S. at 600). That
“formal” approach is essentially the first step of the two-step
inquiry of the Courts of Appeals for the Second and Ninth
Circuits. The Singh Court concluded that “a departure from the
formal categorical approach seems warranted when the terms of
the [INA’s definition of an “aggravated felony”] invite inquiry
into the facts underlying the conviction,” Singh, 383 F.3d at 148
(emphasis added), and that § 1101(a)(43)(M)(i) is such a statute.
Singh did not, however, suggest divorcing the §
1101(a)(43)(M)(i) “qualifier” from the INA’s conviction
requirement entirely.18 The Supreme Court offered further

18
Singh recognized that either (1) a statute of conviction
containing a disjunctive element under which one part of the

40
guidance on the categorical approach in Shepard, less than a
year after we decided Singh. Shepard reemphasized that the
inquiry is not limited to a formal comparison of statutory
elements but rather should focus on identifying the facts upon
which the prior conviction “necessarily” rested.19 Singh’s

disjunctive would render the alien removable and one would
not, a statute it termed “divisible,” or (2) an element of the
“generic” definition of the prior offense designated by the INA
as an “aggravated felony,” might force an IJ to look beyond the
“formal” categorical approach. However, I do not read Singh to
say that the former situation invokes Taylor and Shepard, while
the latter authorizes the IJ to undertake a broad factual inquiry.
Singh simply recognized that both are instances where the
statute of conviction sweeps more broadly than the INA’s
definition. A statute of conviction containing a disjunctive
element under which one part of the disjunctive would render
the alien removable and one would not is “divisible,” and
similarly a statute of conviction containing no loss element is
“divisible” under § 1101(a)(43)(M)(i) into (1) convictions for
aggravated felonies where the loss is more than $10,000 and (2)
other convictions where it is less than $10,000. In either
instance, the nature of the inquiry does not change. The Second
Circuit properly interpreted Singh in this manner. Dulal-
Whiteway, 501 F.3d at 127-28.
19
Shepard held that a guilty plea constitutes a “conviction,”
and that a reviewing court may look to a “transcript of plea
colloquy or [the] written plea agreement presented to the court,
or by a record of comparable findings of fact adopted by the

41
conclusion that § 1101(a)(43)(M)(i) invites further inquiry
beyond the formal approach in order to determine “the facts
underlying the conviction” is entirely consistent with Shepard’s
admonition to focus on the facts “a prior conviction
‘necessarily’ involved.” Shepard, 544 U.S. at 24. And, those
inquiries are essentially the “modified” or second step of the
categorical approach of the Courts of Appeals for the Second
and Ninth Circuits.

Our opinion in Munroe v. Ashcroft, 353 F.3d 225, 227
(3d Cir. 2003), also did not abandon the INA’s conviction
requirement for the § 1101(a)(43)(M)(i) loss element. In
Munroe, we merely held that an immigration court should not
rely on the restitution order to establish the loss when the

defendant upon entering the plea,” to determine precisely what
conduct the defendant pled guilty to. Shepard, 544 U.S. at 20.
In so doing the Court reemphasized that, when the conviction
resulted from a jury verdict, the Court is not limited to a
comparison of the statutory elements – the “formal” version of
the categorical approach upon which Taylor had largely focused
– but also may undertake an analogous inquiry, looking to
“charging documents[] and jury instructions to determine
whether an earlier conviction after trial was for [the generic
enumerated offense].” Shepard , 544 U.S. at 16. In either
instance, the inquiry is to determine whether the conviction
“had ‘necessarily’ rested on the fact identifying the [prior crime]
as [the enumerated offense].” Id. at 21.

42
convicting court’s original restitution order had been based on
the convicted loss, but the court subsequently reduced the
restitution from just above, to just below, $10,000 only to affect
subsequent deportation proceedings. Munroe, 353 F.3d at 227.
We emphasized that the alien had pled guilty to two counts in
the indictment, each of which specified a precise loss amount,
and we concluded:

“We agree . . . that the amount of loss involved in
that conviction was greater than $10,000. The
indictment alleged that the loss exceeded this
amount, and Munroe does not claim that, when he
pled guilty, he admitted to a lesser loss.”

Id. This holding is based on a convicted loss amount (admitted
in the plea agreement) and is therefore entirely consistent with
cases such as Shepard and Dulal-Whiteway.20

20
Although the Munroe Court opined that, in different
circumstances, the amount of restitution ordered “may be
helpful” to determine the loss amount, id., I do not find that
dicta controlling in this case. The Court’s holding was that the
restitution order should not have been relied upon in that case.
I interpret the Court’s statement as merely declining to adopt
any broad-based rule regarding restitution orders and instead
limiting the Court’s holding to the (somewhat unusual) facts of

43
 Our opinion in Alaka v. Attorney General, 456 F.3d 88
(3d Cir. 2006), is also consistent with this approach.21 Alaka
stated that “the formal categorical approach properly may be
abandoned . . . when the terms of the statute on which removal
is based invite inquiry into the facts of the underlying
conviction,” id., and that (M)(i) “invites further inquiry.” Id.
However, much like Singh, Alaka stated that the “further
inquiry” is to identify “the facts underlying the conviction,” id.,
and the Court further explained that “[a] focus on the conduct
that resulted in a conviction is thus our analytical starting point.”
Id. at 107. Indeed, Alaka expressly rejects reliance upon
“relevant” but unconvicted losses calculated for sentencing
purposes; to do so, the Court explained, “would divorce the
$10,000 loss requirement from the conviction requirement . . .
because relevant conduct for sentencing purposes need not be
admitted, charged in the indictment or proven to a jury.” Id. at
108 (internal quotation marks and citations omitted). That is
precisely what the Court’s approach does: the Court finds that
the § 1227 conviction requirement applies to the “fraud or
deceit” component of § 1101(a)(43)(M)(i), but that the loss

that case.
21
As the Court emphasizes, Alaka simply held that, if an alien
pleads guilty to one count in an indictment, he or she cannot be
deported for conduct alleged in a different, unpled and
unconvicted count of the indictment. Id. at 106. However,
Alaka’s reasoning supports the approach I would adopt.

44
element is merely a “qualifier” not subject to that conviction
requirement, thus divorcing the two.22

Because I would join those Courts of Appeals which
require that removability under § 1227 and § 1101(a)(43)(M)(i)
be predicated on convicted conduct, and because the record does
not demonstrate that petitioner was actually and necessarily

22
Although Alaka did state that the IJ could consider factual
findings in the sentencing report, id. at 105, I would not rely on
that dicta because to do so here would be contrary to Alaka’s
clear rationale. Alaka does not explain precisely when a court
may look to facts found in a sentencing report, but the Court’s
holding did not rely on any such facts: the Court emphasized
that, “as was the case with Knutsen and Chang, Alaka
unmistakably pled guilty to one count, and the plea agreement
plainly documented that loss at less than $10,000.” Alaka, 456
F.3d at 108.
Alaka’s reference to the sentence may have been a
recognition that, for “aggravated felonies” other than the one at
issue in this case, the INA expressly directs courts to look to the
sentence, and therefore a per se rule that courts can never look
to facts found in a sentencing report is certainly not appropriate.
See Singh, 383 F.3d at 162 (8 U.S.C. § 1101(a)(43)(G) directs
courts to look to the sentence actually imposed because that
definition states “ a theft offense . . . for which the term of
imprisonment [imposed is] at least one year,” whereas other §
1101(a)(43) definitions include the qualifier “for which a
sentence of one year imprisonment or more may be imposed”)
(bracketed text in original; emphasis added).

45
convicted of any particular loss, I would grant the petition for
review.

46