United States v. Binyamin Stimler

PRECEDENTIAL
UNITED STATES COURT OF APPEALS
FOR THE THIRD CIRCUIT

________________

No. 15-4053
________________

UNITED STATES OF AMERICA

v.

BINYAMIN STIMLER,
Appellant

________________

No. 15-4094
________________

UNITED STATES OF AMERICA

v.

JAY GOLDSTEIN a/k/a Yaakov

JAY GOLDSTEIN,
Appellant
 ________________

No. 15-4095
________________

UNITED STATES OF AMERICA

v.

MENDEL EPSTEIN,
Appellant
________________

On Appeal from the United States District Court
for the District of New Jersey
(D.N.J. Nos. 3-14-cr-00287-005, 3-14-cr-00287-003 and
3-14-cr-00287-001)
District Judge: Honorable Freda L. Wolfson
________________

Argued on January 25, 2017

Before: CHAGARES, RESTREPO and ROTH, Circuit
Judges

(Opinion filed: July 7, 2017)

Nathan Lewin [Argued]
Lewin & Lewin
888 17th Street, N.W.
4th Floor
Washington, DC 20006

2
Gedalia M. Stern
541 Passaic Avenue
Clifton, NJ 07014
Counsel for Appellant Stimler

Aidan P. O’Connor [Argued]
Pashman Stein Walder Hayden
21 Main Street
Court Plaza South, Suite 200
Hackensack, NJ 07601
Counsel for Appellant Goldstein

Laura K. Gasiorowski
Robert G. Stahl
Stahl Farella
220 St. Paul Street
Westfield, NJ 07090

Peter Goldberger [Argued]
50 Rittenhouse Place
Ardmore, PA 19003
Counsel for Appellant Epstein

Mark E. Coyne
Office of United States Attorney
970 Broad Street
Room 700
Newark, NJ 07102

3
Norman Gross [Argued]
Glenn J. Moramarco [Argued]
Office of United States Attorney
Camden Federal Building & Courthouse
401 Market Street
Camden, NJ 08101
Counsel for Appellee

________________

OPINION
________________

ROTH, Circuit Judge

Binyamin Stimler, Jay Goldstein, and Mendel Epstein
are Orthodox Jewish rabbis who were charged with various
kidnapping-related offenses, stemming from their
involvement in a scheme through which they, along with
others,1 sought to assist Orthodox Jewish women to obtain
divorces from recalcitrant husbands. After a jury trial, all
three defendants were convicted of conspiracy to commit
kidnapping. The defendants now appeal various rulings made
by the District Court before, during, and after trial. Because
we find no merit in any of the defendants’ arguments, we will
affirm all three convictions.

1
The other rabbis associated with the ring were variously 1)
charged but pled guilty, 2) went to trial and were acquitted, or
3) had their charges dismissed by the government. Stimler,
Goldstein, and Epstein were the only three to be convicted
after trial.

4
 I2

In the Orthodox Jewish tradition, a married woman
cannot obtain a religious divorce until her husband provides
her with a contract called a “get” (pluralized as “gittin”),
which must, in turn, be signed by an “eid,” or witness. A
woman who attempts to leave her husband without obtaining
a get becomes an “agunah” (pluralized as “agunot”), which
subjects her to severe social ostracism within the Orthodox
Jewish community. Agunot may seek relief in a “beth din,” a
rabbinical court presided over by a panel of three rabbis. The
beth din may then issue “psak kefiah,” or contempt orders
authorizing sanctions, which include, but are not limited to,
the use of force against a husband to secure a get. To assist
an agunah to obtain a get is a “mitzvah,” or religious
commandment of the Orthodox Jewish faith. Starting in at
least 2009, Stimler, Epstein, and Goldstein participated in the
beth din process to help agunot obtain gittin. They worked
with “tough guys” or “muscle men” in exchange for money to
kidnap and torture husbands in order to coerce them to sign
the gittin.

In 2013, the FBI learned of the kidnapping ring and
began investigating the rabbis. As part of this investigation,
an FBI agent posed as an agunah and approached Epstein.
The agent met with Epstein at his home in New Jersey.
Epstein suggested that kidnapping would be appropriate in
the agent’s “situation,” promising that “what we’re doing is

2
As all three defendants were convicted, the facts and
evidence are taken in the light most favorable to the
government. United States v. Ozcelik, 527 F.3d 88, 93 (3d
Cir. 2008).

5
basically gonna be kidnapping a guy for a couple of hours and
beatin’ him up and torturing him . . ..”3 One month after this
meeting, Epstein and Goldstein found a potential location for
the kidnapping. Epstein then convened a beth din at which
he, Goldstein, and a third rabbi presided. Together, the rabbis
issued a psak kefiah authorizing the use of force against the
agent’s “husband.” Epstein and the agent subsequently
planned the details, including the date, location, and manner
of the kidnapping of the “husband.” On the day of the
kidnapping, a team of rabbis and “tough guys” assembled at
the agreed-upon location. Goldstein and Stimler arrived in
disguise and Stimler conducted counter-surveillance of the
area.

Once the kidnapping team had assembled, the FBI
arrested them. Epstein and Stimler were each charged with
one substantive kidnapping count, one count of attempted
kidnapping, and one count of conspiracy to commit
kidnapping. Goldstein was charged with two substantive
kidnapping counts, one count of attempted kidnapping, and
one count of conspiracy.

During its preparation for trial, the government applied
for a court order, pursuant to Section 2703(d) of the Stored
Communications Act (SCA), compelling AT&T to turn over
historic cell site location information (CSLI) generated by
Goldstein’s phone. CSLI is generated every time a cell phone
user sends or receives a call or text message; when the call or
message is routed through the nearest cell tower, the user’s
service provider generates and retains a record identifying the
particular tower through which the communication was

3
JA 4654.

6
routed.4 In more densely populated areas, cell towers are able
to triangulate an individual’s approximate location based on
the individual’s distance from the three nearest towers. Thus,
while less precise than traditional GPS systems, historic CSLI
records can nonetheless generate a rough profile of an
individual’s approximate movements based on the phone calls
that individual makes over a period of time. The order for
such records, covering a total of 57 days of Goldstein’s
location history, was issued by a magistrate judge on October
30, 2014.

II

The defendants filed numerous pretrial motions before
the District Court; we consider only the three which are
relevant to this appeal. First, Goldstein moved to suppress
the CSLI obtained pursuant to the SCA, arguing that cell
phone users have a reasonable expectation of privacy in such
metadata, implicating the Fourth Amendment’s warrant
requirement. The District Court denied Goldstein’s motion,
reasoning that collection of CSLI “does not involve physical
intrusion upon [Goldstein’s] property or any real time
tracking information” and did not “concern the search or
seizure of a cell phone, or the content of any
communication.”5

Second, all three defendants sought dismissal of the
indictment pursuant to the Religious Freedom Restoration Act
(RFRA), arguing that the government’s decision to prosecute

4
United States v. Epstein (Epstein II), No. 14 CR 287, 2015
WL 1646838, at *1 (D.N.J. Apr. 14, 2015).
5
Epstein II, 2015 WL 1646838, at *3.

7
them substantially burdened their sincerely held religious
beliefs and was not the least restrictive means of furthering
any compelling government interest. Stimler raised the
additional argument that RFRA required him to be severed
from the trial of Epstein and Goldstein. The District Court
rejected these arguments, finding that the government’s
decision to prosecute the defendants did not substantially
burden their religious exercise.6 In the alternative, the
District Court found that the government had a compelling
interest in the uniform prosecution of kidnapping laws and
that the prosecution of the defendants was the least restrictive
way of achieving that interest.7 The District Court summarily
rejected Stimler’s request for severance, reasoning that the
joint prosecution was not a substantial burden and that
“[t]here is nothing in [RFRA] which suggests that it can be
used to argue for severance.”8

Third, the defendants sought to introduce evidence of
their religious beliefs and, more broadly, of Orthodox Jewish
law in order to negate the motive element of the kidnapping
statute, or, in the alternative, to demonstrate consent on the
part of the husbands. The District Court refused to admit
such evidence, finding that it was irrelevant both to motive
and to the affirmative defense of consent. In the alternative,
the District Court held that the evidence would be unduly
prejudicial under Rule 403 of the Federal Rules of Evidence,

6
United States v. Epstein (Epstein I), 91 F. Supp. 3d 573,
582-83 (D.N.J. 2015).
7
Id. at 584-85.
8
Id. at 588.

8
as it “would carry a significant potential for jury
nullification.”9

At trial, the government introduced a variety of
evidence against the defendants. As relevant here, the
government introduced testimony from FBI agents placing
Stimler and Goldstein at the site of the proposed kidnapping
in a disguise. The agents stated that Stimler performed
counter-surveillance at the site. The government also
introduced the statements made about Goldstein by another
rabbi at a beth din convened to determine the validity of a get
obtained from one of the prior kidnappings. After both the
government and the defense rested, the judge instructed the
jury as to the elements of each charged offense and
sequestered the jury for deliberations. On the fourth day of
deliberations, the jury sent a question to the judge, inquiring
whether failure to intervene could make an individual liable
for kidnapping. The judge wrote back that, having interpreted
the question to refer only to the substantive kidnapping
counts, the answer was no. The defendants objected to this
response, arguing that it implied that failure to intervene
could support a conviction for the attempt and conspiracy
charges.

All three defendants were convicted. Stimler was
sentenced to 39 months incarceration, Goldstein to 96
months, and Epstein to 120 months. They appealed. We
consolidated the appeals for disposition.

9
Id. at 597.

9
 III10

This appeal presents eight issues, not all of which
apply to every defendant. As such, we treat each issue in
turn, and note which defendants have raised which claims.

A.

We first address whether the District Court erred in
denying Goldstein’s11 motion to suppress the CSLI evidence.
Because the parties agree as to all the relevant facts and
dispute only the legal implications thereof, our review is
plenary.12 Section 2703 of the SCA authorizes the
government to “require a provider of electronic
communication service . . . to disclose a record or other

10
The District Court had jurisdiction over the criminal
prosecutions of the three defendants pursuant to 18 U.S.C. §
3231. We have jurisdiction over the defendants’ appeal
pursuant to 28 U.S.C. § 1291.
11
To the extent Epstein also seeks to challenge the District
Court’s denial of the motion to suppress, he lacks standing to
do so because it does not appear that the government obtained
any CSLI about his whereabouts. See United States v. Nagle,
803 F.3d 167, 178 (3d Cir. 2015) (“Fourth Amendment rights
are personal rights which, like some other constitutional
rights, may not be vicariously asserted.”) (quoting Rakas v.
Illinois, 439 U.S. 128, 133-34 (1978)).
12
United States v. Lafferty, 503 F.3d 293, 298 (3d Cir. 2007)
(“We . . . exercise plenary review as to [a suppression
motion’s] legality in light of the [district] court’s properly
found facts.” (internal quotation marks and citation omitted)).

10
information pertaining to a subscriber to or customer of such
service . . . when the governmental entity . . . obtains a court
order for such disclosure . . ..”13 Such a court order “shall
issue only if the governmental entity offers specific and
articulable facts showing that there are reasonable grounds to
believe” that the records “are relevant and material to an
ongoing criminal investigation.”14 Notably, this “reasonable
grounds” requirement is a lesser burden than the “probable
cause” requirement of the Fourth Amendment.15

Recognizing that the SCA permits precisely what the
government here did, Goldstein argues that the SCA violates
the Fourth Amendment insofar as it authorizes the
government to require disclosure of historic CSLI without
obtaining a warrant. To the extent that historic CSLI records
allow the government to aggregate an individual’s movement
history over an indefinite period of time, Goldstein argues
that the Supreme Court has suggested that individuals have a
reasonable expectation of privacy in such information. In the
alternative, Goldstein asserts that the government failed to
meet even the relaxed “reasonable grounds” requirement of
the SCA. In response to Goldstein’s constitutional argument,

13
18 U.S.C. § 2703(c)(B). The parties do not dispute that cell
site location information is a record “pertaining to” the user
of a cell phone.
14
18 U.S.C. § 2703(d).
15
In re Application of the United States for an Order
Directing a Provider of Electronic Communication Service to
Disclose Records to the Government (In re Application), 620
F.3d 304, 315 (3d Cir. 2010) (“[T]he standard [of reasonable
grounds] is an intermediate one that is less stringent than
probable cause.”).

11
the government rests primarily on the third-party doctrine,16
arguing that cell phone users voluntarily turn over CSLI to
their service providers. With respect to Goldstein’s statutory
arguments, the government maintains that the detailed
descriptions of prior kidnappings allegedly committed and the
identification of specific periods of interest provided
“reasonable grounds” for disclosure.

We do not decide these questions on a blank slate.
Both parties agree that our decision in In re Application of the
United States for an Order Directing a Provider of Electronic
Communication Service to Disclose Records to the
Government (In re Application) addresses the same issues
presented in this appeal. In In re Application, we rejected the
applicability of the third-party doctrine to CSLI, holding that
the transmission of CSLI was not truly voluntary.17 We went
on to conclude, however, that the SCA’s disclosure regime
did not violate the Fourth Amendment because individuals
lack a reasonable expectation of privacy in CSLI.18
Pursuant to our Internal Operating Procedures, “the
holding of a panel in a precedential opinion is binding on

16
The third-party doctrine precludes defendants from
asserting any privacy interests in information which they
voluntarily disclose to third parties. United States v. Christie,
624 F.3d 558, 573-74 (3d Cir. 2010).
17
In re Application, 620 F.3d at 317.
18
Id. at 312-13 (explaining that CSLI does not implicate an
individual’s privacy interests because such interests “are
confined to the interior of the home” and holding that “CSLI
from cell phone calls is obtainable under a § 2703(d) order
and that such an order does not require the traditional
probable cause determination”).

12
subsequent panels.”19 Thus, as a general matter, we remain
obliged to follow our precedent absent en banc
reconsideration. A narrow exception to this rule exists,
however, where intervening legal developments have
undercut the decisional rationale of our precedent.20 We have
described this as an exacting standard; we generally will not
decline to follow our precedent unless it “no longer has any
vitality”21 or is “patently inconsistent”22 with subsequent legal
developments. Because In re Application has not been
overturned by this Court sitting en banc, we will continue to
follow it in its entirety unless the government demonstrates
that intervening legal developments have undermined In re
Application’s rejection of the third-party doctrine or
Goldstein can demonstrate that intervening changes in the law
have created a reasonable expectation of privacy in CSLI.
We conclude that neither the government nor Goldstein have
met their respective burdens.

1.

We begin with the government’s contention that
individuals voluntarily convey CSLI to their cell service

19
3d Cir. I.O.P. 9.1.
20
See, e.g., In re Krebs, 527 F.3d 82, 84 (3d Cir. 2008) (“A
panel of this Court may reevaluate the holding of a prior
panel which conflicts with intervening Supreme Court
precedent.” (citations omitted)).
21
West v. Keve, 721 F.2d 91, 93 (3d Cir. 1983) (emphasis
added).
22
United States v. Tann, 577 F.3d 533, 541 (3d Cir. 2009)
(quoting Cox v. Dravo Corp., 517 F.2d 620, 627 (3d Cir.
1975)).

13
providers. The government attempts to sidestep In re
Application by characterizing its rejection of the third-party
doctrine as dictum. However, the government placed the
issue of the third-party doctrine squarely before us in In re
Application by arguing that the doctrine prevented CSLI from
ever implicating Fourth Amendment concerns. We explicitly
considered and rejected this argument, reasoning that “[a] cell
phone customer has not ‘voluntarily’ shared his location
information with a cellular provider in any meaningful
way.”23 We noted, however, that § 2703(c) offers the judge
the option of requiring probable cause before CSLI is
released.24 By holding that magistrate judges presented with
requests for § 2703(d) orders retain discretion under §
2703(c) to “mak[e] a judgment about the possibility that such
disclosure would implicate the Fourth Amendment,”25 we
necessarily rejected the third-party doctrine. Had the
government been correct that CSLI records were records
voluntarily disclosed to third parties, CSLI could never
implicate an individual’s reasonable expectation of privacy
and thus would never be protected by the Fourth
Amendment’s warrant requirement. Thus, the rejection of the
third-party doctrine was necessary to the holding of In re
Application.

The government suggests, nevertheless, that this Court
“should join all of its sister circuits” in applying the third-
party doctrine to CSLI,26 apparently arguing that the

23
In re Application, 620 F.3d at 317.
24
Id. at 319.
25
Id. at 317.
26
See, e.g., United States v. Graham, 824 F.3d 421, 425 (4th
Cir. 2016) (en banc).

14
subsequent decisions of other circuits may constitute
intervening legal authority allowing departure from our
precedent. We have never so held, and we decline to do so
now. To the contrary, in declining to follow our decisions,
we have recognized intervening authority only from the
Supreme Court of the United States, Congress, or
administrative agencies.27 Accordingly, we continue to
adhere to our view, espoused in In re Application, that the
third-party doctrine does not apply because cell phone users
do not voluntarily disclose CSLI to their service providers
simply by signing a service contract.28

2.

We next address whether intervening changes in law
undermine In re Application’s holding that CSLI does not
implicate an individual’s reasonable expectation of privacy.
Goldstein argues that the Supreme Court’s subsequent
decisions in Riley v. California29 and United States v. Jones30

27
See, e.g., Baptiste v. Attorney Gen., 841 F.3d 601, 609 n.8
(3d Cir. 2016) (noting that “an intervening Supreme Court
decision . . . is also a ‘sufficient basis’ for us to reevaluate our
precedent” even without en banc review); United States v.
Adams, 252 F.3d 276, 286 (3d Cir. 2001) (noting that “a panel
may reevaluate a precedent in light of intervening authority
and amendments to statutes and regulations”).
28
For these reasons, to the extent that sections of the District
Court opinion may be read to suggest that the third-party
doctrine applies here, those sections are reversed. See, e.g.,
Epstein II, 2015 WL 1646838, at *3.
29
134 S. Ct. 2473 (2014).
30
565 U.S. 400 (2012).

15
render the underpinnings of In re Application untenable, and
warrant a departure from our precedent. In Goldstein’s view,
Riley and Jones, taken together, strongly imply that an
individual has a reasonable expectation of privacy in his or
her aggregated movements over a period of time, particularly
where cell phones are involved.

We are not persuaded by Goldstein’s readings of Riley
and Jones. In Riley, the Supreme Court held that officers’
warrantless search of data stored on an individual’s cell phone
ran afoul of the Fourth Amendment, noting that the diversity
and quantity of data stored on mobile phones today created a
reasonable expectation of privacy therein.31 However, Riley
focused primarily on protecting the contents of cell phones,
not metadata generated from cell phone usage.32 This
distinction is far from trivial; Fourth Amendment
jurisprudence has consistently protected only the contents of
an individual’s communications.33 We recently emphasized
this point in United States v. Stanley,34 rejecting the argument

31
Riley, 134 S. Ct. at 2494-95.
32
Id. at 2489 (discussing the amount and variety of data
stored on cell phones).
33
See, e.g., Katz v. United States, 389 U.S. 347, 353 (1967)
(“The Government’s activities in electronically listening to
and recording the petitioner’s words . . . constituted a ‘search
and seizure’ within the meaning of the Fourth Amendment.”)
(emphasis added); Smith v. Maryland, 442 U.S. 735, 741
(1979) (finding no Fourth Amendment violation in
warrantless application of a pen register because “pen
registers do not acquire the contents of communications”)
(emphasis in original).
34
753 F.3d 114 (3d Cir. 2014).

16
that an individual had a reasonable expectation of privacy in
his or her IP address routed through a third party’s wireless
router. Even though we acknowledged that obtaining an
individual’s IP address could roughly track his or her
location, we reasoned that such records “revealed only the
path of the signal establishing this connection [and] revealed
nothing about the content of the data carried by that signal.”35
Riley’s holding is thus an application of the Fourth
Amendment’s protection of content. Goldstein does not
argue that the CSLI at issue here is content, nor would we
find any such argument persuasive.36 Accordingly, Riley
provides little support for extending Fourth Amendment
protections to historic CSLI.

Goldstein’s argument finds better support in the
statements of the concurring opinions of Jones, in which the
Supreme Court held that warrantless placement of a GPS
tracker on an individual’s car for 28 days, and the resulting
aggregated movement history, violated the Fourth
Amendment.37 Although the majority rested on a trespass
theory to reach this conclusion, five justices—in two separate
concurrences—suggested that location tracking also

35
Id. at 122.
36
In the wake of Riley, we have adopted a flexible test for
determining whether data is content or ancillary by analyzing
whether the data is “part of the substantive information
conveyed to the recipient,” noting that “location identifiers
have classically been associated with non-content means of
establishing communication.” In re Google Inc. Cookie
Placement Consumer Privacy Litig., 806 F.3d 125, 136 (3d
Cir. 2015) (internal quotation marks and citation omitted).
37
Jones, 565 U.S. at 412-13.

17
implicated an individual’s reasonable expectation of privacy.
Justice Sotomayor, writing separately, expressed her view
that the Fourth Amendment inquiry turns on “whether people
reasonably expect that their movements will be recorded and
aggregated in a manner that enables the Government to
ascertain, more or less at will, their political and religious
beliefs, sexual habits, and so on.”38 Justice Alito’s
concurrence, which was joined by three other justices, echoed
this idea, focusing on “whether respondent’s reasonable
expectations of privacy were violated by the long-term
monitoring of the movements of the vehicle he drove.”39

We do not believe that either concurrence in Jones has
undercut In re Application in any meaningful way because of
the different technologies at issue. Jones dealt with GPS
tracking, not historic CSLI. In re Application expressly
considered the differential accuracy of CSLI and GPS,
holding that CSLI is less intrusive on individuals’ privacy
rights than GPS tracking. Jones made no suggestion that this
holding was erroneous.

Goldstein admits the inexact nature of CSLI. He
concedes that the tower which transmits the signal is
generally, but not always, the tower closest to the cell phone.
He further concedes that a phone may change from one tower
to another nearby tower during a call without the phone
having moved. Because of the less precise nature of CSLI

38
Id. at 416 (Sotomayor, J., concurring).
39
Id. at 419 (Alito, J., concurring in the judgment).

18
data, we are not persuaded that CSLI is sufficiently similar to
GPS to warrant departure from In re Application.40

Goldstein’s reading of Jones suffers from another
flaw; four of the five justices, who engaged in an analysis of
whether individuals have a reasonable expectation of privacy
in their movements, expressly limited their consideration to
areas in which Congress has not provided statutory
protection. Justice Alito’s concurrence noted that “[i]n
circumstances involving dramatic technological change, the
best solution to privacy concerns may be legislative.”41 This,
he explained, is because “[a] legislative body is well situated
to gauge changing public attitudes, to draw detailed lines, and
to balance privacy and public safety in a comprehensive

40
We do note some aspects of the testimony adduced at trial
that suggest that the line between GPS tracking and CSLI
records is blurring. It appears that the government used CSLI
to track Goldstein’s movement down various interstate
highways between New York and New Jersey. R. 3497a-
3502a. The government’s expert explained that CSLI is no
longer only generated at the beginning and end of each call,
but at every point at which an individual moves closer to a
different cell tower. R. 3497a. The government expert
further stated that, given the density of cell towers in New
York and New Jersey, CSLI generated there is relatively
precise. R. 3492a-94a. Finally, the expert noted that CSLI
records are generated far more frequently than they used to
be, including when an individual sends text messages or uses
certain applications. R. 3531a-32a.
41
Jones, 565 U.S. at 429 (Alito, J., concurring in the
judgment).

19
way.”42 Citing to the wiretapping statute as an example,
Justice Alito stated that, where Congress strikes a particular
balance between digital age privacy rights and government
investigative interests, “regulation . . . has been governed
primarily by statute and not by case law.”43 Justice Alito
expressly warned against judicial creation of new privacy
interests, cautioning that “judges are apt to confuse their own
expectations of privacy with those of the hypothetical
reasonable person . . ..”44

Here, Congress has expressly weighed the privacy
rights in digital information against government interest in
passing the SCA.45 Justice Alito’s concurrence suggests that
we should be wary of revisiting this balance that Congress
has struck.46

Accordingly, Goldstein’s readings of Jones and Riley
do not persuade us to reconsider our own precedent, nor do

42
Id.
43
Id. at 428.
44
Id. at 427.
45
In re Application, 620 F.3d at 313-15 (discussing the
legislative history of the SCA). Other circuits agree with this
determination as well. See, e.g., United States v. Carpenter,
819 F.3d 880, 889 (6th Cir. 2016) (“Congress has specifically
legislated on the question . . . and in doing so has struck the
balance reflected in the [SCA].”).
46
Moreover, in view of the balance reflected by the statutory
provisions for obtaining a court order under § 2703(d), we are
particularly loathe to disregard the holding in In re
Application based not on a direct holding of the Supreme
Court, but on two cobbled together concurrences in Jones.

20
we see any independent reason to do so. While the rapidly
evolving nature of CSLI may one day give us a reason to
reconsider the distinction between GPS and CSLI, we decline
to do so today. We continue to adhere to our view of In re
Application: the Fourth Amendment is not violated when the
government has shown “reasonable grounds to believe that
the . . . records . . . are relevant and material to an ongoing
criminal investigation.”47

3.

Goldstein argues, in the alternative, that the
government failed to meet the “reasonable grounds” standard
of the SCA. As noted above, the “reasonable grounds”
standard is a lesser burden than that of probable cause, and
“in essence is a reasonable suspicion standard.”48 We thus
look to the totality of the circumstances to determine whether
the government had a “particularized and objective basis” for
believing that the CSLI would assist its investigation, mindful
of the fact that agents are entitled “to draw on their own
experience and specialized training to make inferences from
and deductions about the cumulative information available to
them . . ..”49

47
18 U.S.C. § 2703(d).
48
United States v. Daniels, 803 F.3d 335, 351 (7th Cir. 2015);
accord In re Application of the United States of America for
an Order Pursuant to 18 U.S.C. Section 2703(d), 707 F.3d
283, 287 (4th Cir. 2013).
49
United States v. Arvizu, 534 U.S. 266, 273 (2002) (citations
omitted).

21
 In light of this threshold, we find that the
government’s application was more than satisfactory. The
government presented information about the kidnapping ring,
the charged kidnappings, and the alleged involvement of each
defendant. In addition,, the government stated that another
coconspirator had implicated the defendants in his statements
to agents.50 The government then explained that its request
was limited to CSLI records “during the time periods when
the alleged kidnappings and attempted kidnappings occurred”
in order to “identify the location of the alleged participants . .
..”51 Collectively, this information provided the government
with reasonable grounds to believe that the records would be
relevant to their investigation.52

Accordingly, we will affirm the District Court’s order
denying suppression of the CSLI records obtained pursuant to
the SCA.

B.

We next turn to the arguments raised by all three
defendants that various aspects of the prosecution violated
RFRA. Because the motion implicates the proper scope of

50
SA 9.
51
Id.
52
The government urges us to adopt a strict rule that
suppression of evidence is not among the remedies available
under the SCA. Two of our sister circuits have so held. See
Carpenter, 819 F.3d at 890; United States v. Guerrero, 768
F.3d 351, 358 (5th Cir. 2014). In light, however, of our
holding that the government’s application satisfied the SCA,
we need not and do not reach this question.

22
RFRA, we exercise plenary review.53 RFRA proscribes
government conduct which “substantially burden[s] a
person’s exercise of religion” unless the government can
demonstrate, inter alia, that the burden is the “least restrictive
means of furthering [a] compelling government interest.”54
This proscription extends to the government’s criminal
prosecutions under laws of general applicability; a defendant
“may raise RFRA as a shield in the hopes of beating back the
government’s charge.”55 The party invoking RFRA bears the
initial burden of making out a prima facie case by showing
that (1) it possesses a sincerely held religious belief, and (2)
the government’s conduct substantially burdened that belief.56
The burden then switches to the government to demonstrate
that its conduct is the least restrictive means of furthering a
compelling interest.57
1.

We agree with the District Court’s holding that the
defendants failed to satisfy their burden of establishing that
the government substantially burdened their religious beliefs
by prosecuting them for kidnapping. While the government’s
decision to prosecute the defendants undoubtedly constituted
a burden on their sincerely held religious beliefs, the District

53
See, e.g., Gonzales v. O Centro Espirita Beneficente Uniao
do Vegetal, 546 U.S. 418, 428 (2006).
54
42 U.S.C. § 2000bb-1.
55
United States v. Christie, 825 F.3d 1048, 1055 (9th Cir.
2016).
56
Tagore v. United States, 735 F.3d 324, 330 (5th Cir. 2013);
see also Adams v. Comm’r of Internal Revenue, 170 F.3d 173,
176 (3d Cir. 1999).
57
Tagore, 735 F.3d at 330; Adams, 170 F.3d at 176.

23
Court properly analyzed whether the burden was “substantial”
by looking to acceptable alternative means of religious
practice that remained available to the defendants.58 Here,
none of the defendants argue that they are unable to
participate in the mitzvah of liberating agunot without
engaging in kidnapping; as the District Court noted, “it is
unclear whether all non-violent methods were exhausted
before the alleged kidnappings took place here.”59 The
defendants do not challenge this determination on appeal. As
the defendants have failed to demonstrate that the prosecution
was a “substantial” burden, we will affirm the District Court’s
holding that prosecution under the federal kidnapping statute
did not violate RFRA.

Moreover, even if the defendants had demonstrated
that the government’s actions constituted a substantial burden
on their religious exercise, we would nonetheless affirm the
District Court’s determination that the government has a
compelling interest in uniform application of laws about
violent crimes and that no other effective means of such
uniformity existed. The Supreme Court has advised that the
government’s interest in preventing serious crimes “is both
legitimate and compelling.”60 The defendants fail to cite, nor
can we identify, any cases in which any court has allowed
RFRA to shield individuals in the commission of violent
crimes.

58
We have previously examined the adequacy of alternative
means of practice in determining whether a religious burden
is “substantial.” See, e.g., Washington v. Klem, 497 F.3d 272,
282-83 (3d Cir. 2007).
59
Epstein I, 91 F. Supp. 3d at 582.
60
United States v. Salerno, 481 U.S. 739, 749 (1987).

24
 Accordingly, we will affirm the District Court’s denial
of the motion to dismiss the indictment.

2.

We turn next to the argument raised only by Stimler
that his joinder with Epstein and Goldstein constituted an
independent RFRA violation. No court appears to have
answered the question of whether RFRA imposes further
limits on the government’s ability to structure a prosecution.
The Supreme Court has noted that RFRA requires only that
an individual face “serious disciplinary action”61 for acting on
their religious beliefs. This phrase encompasses sanctions
short of prosecution. However, Stimler’s briefing fails to
suggest that the joinder itself was any kind of sanction, nor
does it suggest that the joinder caused any burden on his
religious exercise; instead, it focuses entirely on whether the
prosecution itself worked a unique burden on Stimler’s
religious practice. Thus, we need not determine the exact
boundaries of RFRA here, as Stimler has not adequately
alleged that joinder violated his rights under RFRA.

C.

All three defendants challenge the District Court’s
decision under Rules 402 and 403 of the Federal Rules of

61
Holt v. Hobbs, 135 S. Ct. 853, 862 (2015) (second
alteration in original). Although Holt dealt with a claim
under the Religious Land Use and Institutionalized Persons
Act (RLUIPA), the Supreme Court has stated that RLUIPA
“imposes the same general test as RFRA . . ..” Burwell v.
Hobby Lobby Stores, Inc., 134 S. Ct. 2751, 2761 (2014).

25
Evidence to bar admission of evidence about Orthodox
Jewish marital law and the religious motivations for the
kidnappings, arguing that such evidence was relevant to
negating the specific intent required for conviction.
Alternatively, the defendants argue that evidence of Orthodox
Jewish laws about marital duties would be relevant to
showing that the husbands consented to the kidnappings. We
review a district court’s exclusion of evidence for an abuse of
discretion, but review de novo its interpretation of the Federal
Rules of Evidence.62

In relevant part, the federal kidnapping statute requires
that the kidnapping be committed “for ransom or reward or
otherwise . . ..”63 The District Court properly read this
language as encompassing a broad set of potential motives,
including the religious benefit of performing a mitzvah, while
reasoning, however, “that religious motivation simply cannot
negate the intent to commit a crime.”64 We agree that a
religious benefit can constitute a “benefit” under the statute.

We will also affirm the District Court’s determination
that the evidence of religious practices was not relevant to the
affirmative defense of consent. The defendants argue that, by
practicing Orthodox Judaism and signing a marriage contract,
the husbands consented to any use of force authorized by any
beth din. The District Court properly rejected this argument,
reasoning that “[w]hile consent can be a defense to
kidnapping, it has to be specific and cannot be prospective in

62
Hirst v. Inverness Hotel Corp., 544 F.3d 221, 225 (3d Cir.
2008).
63
18 U.S.C. § 1201(a).
64
Epstein I, 91 F. Supp. 3d at 595.

26
nature.”65 The defendants do not argue that the religious
evidence would demonstrate that the husbands gave specific
consent to their particular kidnappings; accordingly, we agree
that the religious evidence was not relevant to the affirmative
defense.

We further agree with the District Court that any
marginal relevance that the religious evidence may have had
was substantially outweighed by the prejudicial impact it
would have had on the trial. Suggesting that the defendants
acted for a religious purpose might have given rise to the
potential for jury nullification, which we have held is
substantially prejudicial.66

Accordingly, we find no abuse of discretion in the
District Court’s decision under Rules 402 and 403 to exclude
evidence about Orthodox Jewish marital law.

D.

We next turn to the argument, raised by Epstein and
Goldstein, that the District Court erred in three respects in
charging the jury. First, they argue that the District Court
failed to include the jurisdictional element of the kidnapping
offense in the conspiracy instruction. Second, they contend
that the District Court erred in refusing to charge the jury that
kidnapping requires the victim to be held for an appreciable

65
Id.
66
See United States v. DeMuro, 677 F.3d 550, 566 (3d Cir.
2012) (finding that it was not abuse of discretion to exclude
evidence that “had the potential for confusion and opened the
door to jury nullification.”).

27
period of time. Finally, the defendants claim that the District
Court’s instruction as to motive constituted a constructive
amendment of the indictment inasmuch as it did not include
the specific religious motives charged in the indictment.
Because the defendants did not object in the District Court,
our review is for plain error.

A district judge’s failure to instruct the jury as to a
necessary element of the offense “ordinarily constitutes plain
error”67 unless the instructions as a whole make clear to the
jury all necessary elements of the offense.68 In determining
whether there has been a plain error in jury instructions, we
“consider the totality of the instructions . . ., not focusing on a
particular paragraph in isolation.”69 Finally, even if the
instructions omitted a necessary element in a way that would
confuse the jury, we may nonetheless affirm if “no reasonable
jury could find that the element was not present.”70

We see no merit in the claim that the District Court
failed to make clear the interstate commerce element in
instructing the jury on the conspiracy charge. While the
conspiracy charge included no explicit jurisdictional

67
United States v. Haywood, 363 F.3d 200, 207 (3d Cir.
2004) (quoting United States v. Xavier, 2 F.3d 1281, 1287 (3d
Cir. 1993)) (emphasis in original).
68
See, e.g., United States v. De Lazo, 497 F.2d 1168, 1171
n.6 (3d Cir. 1974).
69
United States v. Kukafka, 478 F.3d 531, 539 (3d Cir. 2007)
(citing United States v. Coyle, 63 F.3d 1239, 1245 (3d Cir.
1995)).
70
United States v. Andrews, 681 F.3d 509, 527 (3d Cir. 2012)
(citation omitted).

28
requirement, it did state that the jury would have to find that
the defendants conspired to engage in “kidnapping, as alleged
in the indictment.”71 The parties agree that the District Court
included this element in the instructions as to the substantive
kidnapping counts and in the indictment. We therefore
conclude that the instructions as a whole made clear that the
conspiracy charge required the jury to find a conspiracy to
commit a kidnapping that would cross state lines.

We similarly decline to find error in the District
Court’s decision to not include a temporal element in the
kidnapping instruction. Seizing on just one line in Chatwin v.
United States, the defendants argue that federal kidnapping
requires holding the victim “for an appreciable period.”72 We
do not believe that the Supreme Court intended to create a
new limit on kidnapping liability—one not found anywhere in
the statutory text—in one line of dictum. Further, as the
government properly notes, we have upheld jury instructions
that do not refer to any temporal limit on kidnapping liability
even after Chatwin.73 Indeed, we have upheld convictions
when an individual was held for mere minutes.74 Moreover,
even if we were to see merit in the defendants’ assertions on
this point, we would nonetheless affirm. No reasonable juror

71
R. 4123a.
72
326 U.S. 455, 460 (1946).
73
See, e.g., United States v. Sriyuth, 98 F.3d 739, 750 (3d Cir.
1996).
74
Cf. Government of Virgin Islands v. Ventura, 775 F.2d 92,
96 (3d Cir. 1985). In fact, Ventura dealt with the Virgin
Islands aggravated kidnapping statute, under which we have
expressly held that we must consider “the duration of the
detention.” Id. at 95 (citation omitted).

29
would have failed to find that the seizure, blindfolding and
coercion by the defendants did not involve holding for “an
appreciable period.”

Finally, the District Court did not constructively
amend the indictment because, in instructing the jury, it
actually did include the specific motive charged in the
indictment. The defendants assert that the indictment alleged
that the defendants’ motive was “to threaten and coerce
Jewish husbands to give gets to their wives.”75 The District
Court expressly stated in its instructions that “[t]he indictment
alleges the defendants had a purpose of holding the individual
victims to coerce them into giving a get to the victim’s
wife.”76 In light of this instruction, and the entirety of the
evidence produced at trial, we cannot say that there is any
“substantial likelihood that the jury may have convicted the
defendant for an offense differing from the offense the
indictment . . . actually charged.” 77

Accordingly, we reject the defendants’ challenges to
the District Court’s jury instructions.

E.

Stimler and Goldstein next argue that the District
Court improperly responded to a question from the jury on
the third day of deliberations. The jury asked the District
Court whether “element #1 of kidnapping”—the “seizes,

75
R. 174a.
76
R. 4128a.
77
United States v. Daraio, 445 F.3d 253, 260 (3d Cir. 2006)
(emphasis added).

30
confines, inveigles, decoys, kidnaps, abducts, or carries away
and holds” requirement—could be satisfied “[i]f you know
that someone is being confined against their will and . . . do
not intervene . . ..”78 The District Court responded that it
“interpreted [the] question as referring to the kidnapping
counts, counts 2 and 3. If that is accurate, then the answer to
[the] question is no.”79 The court further told the jury that if
they were “inquiring about any other count, please so
indicate, so that I may more fully consider [the] question and
answer appropriately.”80 Stimler objected to this response,
asserting that it suggested that the conspiracy count could be
satisfied by failure to intervene. Because Stimler properly
objected to the response, we conduct plenary review.81 If we
determine that the response was improper, we nonetheless
may affirm if the error was harmless.82

We do not agree that the District Court’s response
improperly suggested that liability could be found upon a
failure to intervene. The District Court expressly stated that it
was speaking only to the substantive kidnapping counts, and
suggested nothing about the other counts, noting that it would
need to “more fully consider [the] question” to “answer
appropriately” if the jurors had questions about the other
counts. Accordingly, we find no error in the District Court’s
response to the jury’s question.

78
R. 4853a.
79
R. 4854a.
80
Id.
81
United States v. Waller, 654 F.3d 430, 434 (3d Cir. 2011)
(citation omitted).
82
Id.

31
 F.

Goldstein and Epstein next challenge the admission of
certain statements made by alleged co-conspirators, arguing
that the statements were inadmissible hearsay under the
Federal Rules of Evidence and violated the Confrontation
Clause. While we “review a nonconstitutional challenge to
the admission of hearsay for abuse of discretion,” we
“exercise plenary review over Confrontation Clause
challenges.”83 In distinguishing when to review admission of
evidence under the Confrontation Clause from when to
review under the Federal Rules of Evidence, the touchstone
of our inquiry is “whether the contested statement by an out-
of-court declarant qualifies as testimonial . . ..”84 “[W]here
nontestimonial hearsay is concerned, the Confrontation
Clause has no role to play in determining the admissibility of
a declarant’s statement.”85 A statement is testimonial only if
it meets two requirements: (1) it is a “solemn declaration or
affirmation made for the purpose of establishing or proving
some fact;”86 and (2) it was made primarily for the purpose of
“prov[ing] past events potentially relevant to later criminal
prosecution.”87 If the contested statement is testimonial, we

83
United States v. Berrios, 676 F.3d 118, 125 (3d Cir. 2012)
(internal quotation marks and citations omitted).
84
Id. at 127.
85
Id. at 126.
86
Melendez-Diaz v. Massachusetts, 557 U.S. 305, 310 (2009)
(quoting Crawford v. Washington, 541 U.S. 36, 51 (2004)).
87
Michigan v. Bryant, 562 U.S. 344, 361 (2011) (internal
quotation marks and citation omitted); see also Ohio v. Clark,
135 S. Ct. 2173, 2184 (2015) (Scalia, J., concurring in the
judgment) (formalizing two-part inquiry).

32
next must determine “if the defendant had a prior opportunity
to cross-examine [the declarant].”88

The statements at issue were testified to by Aryeh
Ralbag, another rabbi involved in the kidnapping ring. At
trial, Ralbag described the statements made before a beth din
which was convened when the alleged victim of one of the
charged kidnappings had challenged the validity of the get he
signed. Ralbag and two other rabbis presided at the beth din,
and four witnesses—including Goldstein and Stimler—
testified that Goldstein and Stimler had served as eid in
procuring the contested get.89

We have no trouble concluding that these statements
were not testimonial within the meaning of the Sixth
Amendment. It is clear that none of the individuals at the
beth din—all of whom were charged as part of the
conspiracy—would have reasonably believed that they were
making statements for the purpose of assisting a criminal
prosecution. Accordingly, the Confrontation Clause is
inapplicable to defendants’ challenge, and we only analyze
whether the statements were inadmissible hearsay under the
Federal Rules of Evidence.90

88
Giles v. California, 554 U.S. 353, 358 (2008).
89
The other two witnesses, and Ralbag himself, were arrested
and charged as part of the kidnapping ring. Ralbag was
granted immunity in exchange for his testimony, and the
other two witnesses pled guilty to lesser offenses.
90
We therefore decline to consider whether the defendants
had an adequate opportunity to cross-examine the two other
witnesses at the beth din, an issue which neither the
defendants nor the government has briefed.

33
 Rule 801 of the Federal Rules of Evidence explains
that a statement is not hearsay if “the statement is offered
against an opposing party and . . . was made by the party’s
coconspirator during and in furtherance of the conspiracy.” 91
The Rule thus imposes two predicate inquiries before a
statement will be admitted: (1) the statement must be made
by a coconspirator,92 and (2) the statement must be made
during the course of and in furtherance of the conspiracy.
Both requirements must be satisfied by a preponderance of
the evidence.93 Goldstein and Epstein assert that neither
requirement was met here. We disagree. The fact that the
two other individuals were present at the warehouse as part of
the kidnapping team, coupled with their knowledge of the
other kidnappings, was sufficient to demonstrate that they
were indeed coconspirators.94 Similarly, because the purpose
of the conspiracy was broadly to secure valid gittin from
husbands, statements by coconspirators to prove the validity
of the gittin were clearly made during the course and in
furtherance of the conspiracy, regardless of when they were

91
Fed. R. Evid. 801(d)(2)(E).
92
A declarant will be considered a “coconspirator” whenever
a conspiracy existed between the declarant and the party
against whom the statement is offered. See United States v.
Gambino, 926 F.2d 1355, 1360 (3d Cir. 1991).
93
Bourjaily v. United States, 483 U.S. 171, 175-76 (1987).
94
The mere fact that the two other individuals were allowed
to plead to violations of other statutes does not persuade us
otherwise; prosecutors have broad discretion in deciding what
to charge and what pleas to accept. See Bordenkircher v.
Hayes, 434 U.S. 357, 368 n.2 (1978) (Blackman, J.,
dissenting).

34
occurred.95 We thus find no error, constitutional or
otherwise, in the District Court’s decision to admit the
statements made at the beth din.

Even if we were to assume that the statements were
improperly admitted, however, we would nonetheless affirm
because any error was harmless. 96 As noted, Ralbag testified
as to statements made by four witnesses, including Goldstein
and Stimler. The defendants do not challenge the admission
of the statements they themselves made, which largely
paralleled those of the other two witnesses; the statements of
the other two witnesses, therefore, were largely duplicative,
and any error in their admission was harmless.97

Accordingly, we conclude that the District Court
properly admitted the evidence.

95
See, e.g., United States v. Weaver, 507 F.3d 178, 185-87
(3d Cir. 2007) (holding that statements made after the core
acts of the conspiracy were committed were in furtherance of
the conspiracy insofar as they were made to conceal the
unlawful acts).
96
See United States v. Moreno, 809 F.3d 766, 774, 776 (3d
Cir. 2016) (citations omitted) (noting that harmless error
review applies to both admission of hearsay evidence and
violations of Confrontation Clause, although the harmless
error inquiry under the Federal Rules of Evidence is a
“slightly less onerous standard”).
97
Cf. DeMuro, 677 F.3d at 564-65 (noting that exclusion of
duplicative evidence is well within the discretion of a trial
judge).

35
 G.

Stimler next challenges the sufficiency of the evidence
against him. We employ a “particularly deferential standard
of review” to appeals challenging the sufficiency of evidence
presented to the jury.98 In examining the sufficiency of the
evidence, we should not “weigh the evidence or . . . determine
the credibility of witnesses.”99 Rather, the defendant bears
the “very heavy burden” of showing that no rational trier of
fact could have found the essential elements of the crime
beyond a reasonable doubt.100

Stimler has failed to meet this burden. The
government presented uncontested evidence that Stimler was
present at the site of the proposed kidnapping, and wore a
disguise. The government next introduced evidence that
Stimler performed countersurveillance by walking around the
warehouse with a flashlight. On appeal, Stimler presents an
alternative explanation of these facts. Simply disagreeing
with the jury’s interpretation of the facts, however, is
insufficient. We believe that the jury made a reasonable
inference in finding that Stimler knew of the conspiracy and
took affirmative steps to help carry it out. We therefore
affirm his conviction.

98
United States v. Soto, 539 F.3d 191, 194 (3d Cir. 2008)
(quoting United States v. Dent, 149 F.3d 180, 187 (3d Cir.
1998)).
99
Id. (internal quotation marks and citation omitted).
100
Id. (citation omitted).

36
 H.

Finally, all three defendants challenge the FBI’s sting
operation as conduct so outrageous that it violated due
process. This claim is procedurally barred, as the defendants
failed to make the argument in the District Court, despite full
knowledge of the scope of the government’s investigation.
Nor do the defendants identify any new information that
supports their claim of outrageous government conduct. We
have made clear that such failure waives challenges to
allegedly outrageous government conduct.101

Moreover, even if this argument had been preserved,
we would see no merit to it. In reviewing claims of
outrageous government conduct, we “repeatedly have noted
that we are ‘extremely hesitant to find law enforcement
conduct so offensive that it violates the Due Process
Clause.’”102 We have rejected the argument that the
government’s invitation to engage in criminal activity rises to
this level where, as here, the defendants used their own
knowledge and connections to set up and carry out the
unlawful conduct.103 We have suggested that “the supply of
ingredients” to commit a crime would be insufficient to meet

101
United States v. Salahuddin, 765 F.3d 329, 350 (3d Cir.
2014).
102
United States v. Hoffecker, 530 F.3d 137, 154 (3d Cir.
2008) (quoting United States v. Voigt, 89 F.3d 1050, 1065 (3d
Cir. 1996)).
103
United States v. Lakhani, 480 F.3d 171, 182 (3d Cir.
2007).

37
this standard.104 Given that Epstein first suggested the use of
violence, and that the defendants assembled the kidnapping
team, chose a location, and acquired their own tools, we see
no due process violation here.

IV.

In our legal system, “liberty and social stability
demand a religious tolerance that respects the religious views
of all citizens . . ..”105 Respect for religious beliefs cannot,
however, trump all other legitimate, and sometimes
competing, government objectives. This appeal asks us to
clarify the balance between religious freedom and public
safety. The balance here clearly lies on the side of public
safety. For the reasons stated above, we will affirm the
District Court’s convictions of Mendel Epstein, Jay
Goldstein, and Binyamin Stimler.

104
See United States v. Twigg, 588 F.2d 373, 378 (3d Cir.
1978).
105
McCreary County, Kentucky v. American Civil Liberties
Union of Kentucky, 545 U.S. 844, 860 (2005) (internal
quotation marks and citation omitted).

38
RESTREPO, Circuit Judge, concurring in part and concurring
in the judgment.

I join parts I, II, III(A)(1), and III(B)-(H) of the
Opinion of the Court, which address the parties’ arguments
concerning the application of the third party doctrine to
historical cell site location information (“CSLI”); applications
of the federal Religious Freedom Restoration Act, 42 U.S.C.
§ 2000bb et seq., to discretionary trial procedure decisions of
the District Court; introduction of evidence about Orthodox
Jewish law; propriety of jury instructions and the District
Court’s response to jury questions; admission of co-
conspirator statements; sufficiency of evidence; and
outrageous government conduct. However, I concur only in
the judgment with respect to parts III(A)(2)-(3), because I
believe that the Government obtaining 57 days of aggregated
CSLI with only a § 2703(d) order supported by reasonable
suspicion is, in this case, a warrantless search that violates the
Fourth Amendment. I depart from the Majority because of
two Supreme Court opinions that have issued since our own
Court last considered this issue.

I

“[T]he holding of a panel in a precedential opinion”—
such as that in our Court’s most recent opinion on law
enforcement requests for CSLI, In the Matter of the
Application of the United States for an Order Directing a
Provider of Electronic Communication Service to Disclose
Records to the Government, 620 F.3d 304 (3d Cir. 2010) (“In
re Application”)—“is binding on subsequent panels.” Third
Circuit I.O.P. 9.1. This rule exists for good reason: it
maintains uniformity of law within the Circuit, and promotes

1
predictability for litigants. However, if Supreme Court
authority abrogates or calls existing Circuit precedent into
question, our Court has recognized that subsequent panels
may decline to follow the prior holding without reconsidering
the issue en banc. George Harms Constr. Co., Inc. v. Chao,
371 F.3d 156, 161 (3d Cir. 2004); see also Reich v. D.M.
Sabia Co., 90 F.3d 854, 858 (3d Cir. 1996).

This exception to our Internal Operating Procedures is
narrow. When our Court has declined to follow past
precedent on the basis of intervening Supreme Court
authority, we typically have declined to follow only the
specific portions of the prior precedent that the intervening
authority has called into question or abrogated. United States
v. Johnson, 587 F.3d 203, 207 n.4 (3d Cir. 2009); see also
Animal Science Prods., Inc. v. China Minmetals Corp., 654
F.3d 462, 467-68 (3d Cir. 2011). I agree with the Majority
that the third party doctrine holding of In re Application has
not been called into question by subsequent authority.
However, I take a different view on the ongoing vitality of the
reasonable expectation of privacy analysis in In re
Application, in light of Supreme Court opinions in United
States v. Jones, 565 U.S. 400 (2012), and Riley v. California,
134 S.Ct. 2473 (2014).

Jones and Riley have “sufficiently undercut the
decisional basis” of In re Application’s holding that
magistrate judges can issue a § 2703(d) order for aggregated
location information, rather than requiring a warrant, in many
instances. West v. Keve, 721 F.2d 91, 93 (3d Cir. 1983). In
re Application held that magistrate judges could require a
warrant for CSLI upon considering the individual surveillance
target’s reasonable expectation of privacy under the Fourth

2
Amendment. With the more recent guidance of Jones and
Riley, I would conclude that in at least most factual
circumstances—including those before us—magistrate judges
must require a warrant for the aggregated collection of
historical CSLI to comply with the Fourth Amendment.

In Jones, law enforcement officers surreptitiously
placed a tracking device on the bumper of a surveillance
target’s car without a valid warrant, and collected 28 days’
worth of global positioning system (“GPS”) location data.
Jones, 565 U.S. at 403. The lead opinion in Jones held that
this conduct violated the Fourth Amendment because of the
physical trespass committed. Id. at 404-05. In what has
come to be regarded as the “shadow majority” of Jones,1
however, five Justices in two concurrences disagreed with the
reasoning of the lead opinion, which “disregards what is
really important (the use of a GPS for the purpose of long-
term tracking).” Jones, 565 U.S. at 424 (Alito, J., concurring)
(emphasis in original). Those Justices found a constitutional
privacy interest implicated by aggregated tracking of an
individual’s location over time. Id. at 430 (Alito, J.,

1
See, e.g., In the Matter of the Application of the
United States of America for an Order Authorizing
Disclosure of Historical Cell Site Information for Telephone
Number [Redacted], 40 F. Supp. 3d 89, 92 (D.D.C. 2014)
(rejecting a § 2703(d) application absent a showing of
probable cause or more evidence about CSLI, because of
“serious questions about whether the Fourth Amendment
requires a warrant to obtain CSLI” in light of Jones).

3
concurring); id. at 413-14 (Sotomayor, J., concurring).2 That
reasonable expectation of privacy reflects the intrusion that
occurs when the Government can aggregate enough location
data on individuals to draw inferences about their private
lives and constitutionally protected activities. Id. at 416
(Sotomayor, J., concurring). The concurrences place more
weight on protecting the privacy interest itself, and, in
particular, considering the aggregation of information
obtained by the Government. Id. at 426 (Alito, J.,
concurring); id. at 414 (Sotomayor, J., concurring).

Historically, this interest has been protected in part by
resource constraints facing law enforcement agencies—but
those resource constraints no longer present an obstacle to
this type of aggregation. In the past, obtaining aggregated
location information on any individual by tracking him or her
“for any extended period of time was difficult and costly and
therefore rarely undertaken.” Id. at 429 (Alito, J.,
concurring). Constant monitoring of an individual’s location
is possible now, however, because of new technology
“available at a relatively low cost.” Id. at 416 (Sotomayor, J.,
concurring); id. at 429 (Alito, J., concurring). Those former
resource constraints, however, have shaped what the Jones

2
Although the lead opinion in Jones resolved the case
on the basis of physical trespass, it addressed the
concurrences’ position that the trespass theory provided
insufficient protection of an individual’s reasonable
expectation of privacy by acknowledging that “[i]t may be
that achieving the same result through electronic means,
without an accompanying trespass, is an unconstitutional
invasion of privacy.” Jones, 565 U.S. at 412.

4
shadow majority recognized as a reasonable expectation of
privacy. Historically, “society’s expectation has been that
law enforcement agents and others would not—and indeed, in
the main, simply could not—secretly monitor and catalogue
every single movement of an individual’s car for a very long
period.” Id. at 430 (Alito, J., concurring).

Here, in its § 2703(d) application to the Magistrate
Judge, the Government requested location information for 57
total days. Such a quantity of location information prompts
exactly the question Justice Sotomayor posed in Jones:
“whether people reasonably expect that their movements will
be recorded and aggregated in a manner that enables the
Government to ascertain, more or less at will, their political
and religious beliefs, sexual habits, and so on.” Jones, 565
U.S. at 416 (Sotomayor, J., concurring); see also id. at 419
(Alito, J., concurring). Five Justices in Jones would answer
that question in the negative—at least as to aggregation
exceeding “the 4-week mark.” Id. at 430 (Alito, J.,
concurring); see also id at 415 (Sotomayor, J., concurring).

The Majority declines to read the Jones concurrences
as undercutting “In re Application in any meaningful way” in
part “because of the different technologies at issue.” Maj.
Op. 15. To the extent that tracking an individual’s cell phone
by CSLI and tracking an individual’s car by GPS differ, the
privacy interest that protects an individual from the
Government aggregating that location information (without a
warrant) remains the same. If anything, the reasonable
expectation of privacy in aggregated location derived from an
individual’s use of a cell phone is stronger than the
reasonable expectation of privacy in aggregated location
derived from that same individual’s use of a car. Aggregating

5
data points from cell phone location into a comprehensive
record offers the Government more opportunity to infer
things about an individual, because cell phones accompany
individuals many places that cars do not. “Historic location
information . . . can reconstruct someone's specific
movements down to the minute, not only around town but
also within a particular building.” Riley, 134 S.Ct. at 2490.
Moreover, cell phones accompany individuals who travel by
public transit or otherwise not by car, regular drivers who
temporarily rent a different car, and those who ride in the cars
of others. And regardless of how an individual moves
through the world, “nearly three-quarters of smartphone users
report being within five feet of their phones most of the time,
with 12% admitting that they even use their phones in the
shower.” Id.

The Majority also reads the Jones shadow majority as
not undercutting In re Application in part because of the
distinction in precision between GPS data—at issue in
Jones—and the CSLI at issue here. This distinction has
nearly disappeared since we decided In re Application. By
the time of the events at issue in this case, CSLI had grown
quite precise, particularly in more densely-populated
jurisdictions.3 In cities where wireless providers have more
towers to provide service for more people packed into a given
area, the identifiable radius in which a subscriber would

3
Because we consider the case on the facts before us,
we consider CSLI as precise as it is in urban areas. It may be
that the less precise CSLI in rural areas is so dissimilar from
GPS location data as to make Jones inapplicable, but we need
not consider that question.

6
connect to one tower rather than another is substantially
smaller. In explaining this concept to the jury, the
Government’s expert at trial noted that the layout of “tightly
compacted towers in Brooklyn” “will reduce the coverage
area of” any one tower. App. 3494a. The number of towers
and antennas in Brooklyn, for instance, allowed the expert to
note proximity to an antenna on “the side of a building near
the intersection of Webster Avenue and . . . Coney Island
Avenue.” Id. The expert also used CSLI to describe an
individual’s “southbound movement on I-278.” App. 3499a-
500a.

By contrast, when the In re Application court
considered this issue, CSLI did “not provide information
about the location of the caller closer than several hundred
feet.” In re Application, 620 F.3d at 311. Since then,
wireless network improvements have included the
distribution of “hundreds of thousands of ‘microcells,’
‘picocells,’ and ‘femtocells,’” which function similarly to
hotspots and create CSLI that “can in some cases be more
accurate than GPS.” Stephanie K. Pell & Christopher
Soghoian, Can You See Me Now?: Toward Reasonable
Standards For Law Enforcement Access To Location Data
That Congress Could Enact, 27 Berkeley Tech. L.J. 117, 132
(2012). Even the proliferation of traditional cell towers has
resulted in smaller coverage areas and CSLI that is “far more
accurate—in some cases as good as GPS.” Id. at 133.

The reasonable expectation of privacy of an individual
in an urban area in the aggregated location information of his
or her CSLI is functionally indistinguishable from the
reasonable expectation of privacy of that same individual in
the aggregated location information of his or her GPS data.

7
Distinguishing Jones on the basis of the greater precision of
GPS ignores the current capabilities of CSLI, and indeed, the
use the Government made of it in this case.

Although the Majority distinguishes Riley from the
facts here by separating contents and metadata, Riley should
inform our analysis of the reasonable expectation of privacy
in CSLI, as well. The animating principle behind Riley is the
same as the principle behind Jones: the Government may
violate an individual’s reasonable expectation of privacy
when it obtains too much aggregated information without a
warrant. In Riley, decided two years after Jones, the
aggregation at issue merely took a different form. There, the
Supreme Court recognized that the aggregation of data
allowed by the increased capacity of digital storage helps law
enforcement agents make inferences that intrude on an
individual’s reasonable expectation of privacy. The Court
held unconstitutional a warrantless search of a cell phone, in
part because the types of information stored on the cell phone
in question “reveal[ed] much more in combination than any
isolated record.” Riley, 134 S.Ct. at 2489.

The Riley Court rejected applications of doctrine
created for older technologies that allowed for less
aggregation of historically protected information. The Court
distinguished call logs on modern cell phones from pen
registers in part on the basis that “call logs typically contain
more than just phone numbers; they include any identifying
information that an individual might add.” Id. at 2493.
While cataloguing the different types of data stored on cell
phones that had not historically been stored on landline
telephones, the Court explained that doctrines governing
“qualitatively different” pre-digital counterparts do not

8
compare well to modern technology in considering questions
of criminal procedure. Id. at 2490, 2493. Allowing
warrantless collection by analogy to older technologies would
instead cause “a significant diminution of privacy.” Id. at
2493.

Here, technological changes since In re Application in
the provision of wireless service mean that CSLI—like the
phone itself, in Riley—conveys a greater quantity of
information for the Government to aggregate than it did
previously. The Government’s application for CSLI
encompassed more data points than merely the location at the
time of incoming or outgoing telephone calls. In requesting
“[a]ll data about which ‘cell towers’ and ‘sectors’ received a
radio signal from each cellular telephone or device assigned
to the Account, including, but not limited to, per call
management data or return Time from Tower data,” App. 459
(emphasis added), the Government sought information that
would allow for essentially continuous location tracking,
rather than rare location snapshots. AT&T, from which the
government sought and obtained the information, collects
CSLI data upon call “hand-offs,” which occur when a person
moves while on a call, and the call switches to routing
through the next tower (or a different face of the same tower)
as the individual gets closer to it. App. 3497a. At trial, the
Government’s expert was able to use hand-offs during a
defendant’s 52-second call to describe the CSLI as
“consistent with southbound movement on I-278” from New
York to northern New Jersey. App. 3499a-3500a. Tracking
an individual through space during the course of a call
represents more data to aggregate—and a correspondingly
greater privacy intrusion—than simply collecting his or her
location only at the origination and termination of a call.

9
 The application also reflects the Government’s
capability to obtain data from use of a cell phone that it could
not typically have obtained from an individual’s use of a
telephone, which the Riley Court regarded as a reason to
require warrants for cell phone searches. Riley, 134 S.Ct. at
2489. Here, the Government requested location information
for text messages, as well. Indeed, the Government’s request
to AT&T may stretch even more broadly than calls and
texts—asking for data about each time a tower “received a
radio signal,” App. 459, from a phone could conceivably
encompass any time any application on a phone, even one
running passively in the background, connects to the network.
In re Application for Telephone Information Needed for a
Criminal Investigation, 119 F. Supp. 3d 1011, 1024 (N.D.
Cal. 2015) (affirming a magistrate judge’s denial of an
application for CSLI under § 2703(d)). CSLI may be
generated by an action as innocuous as a user’s email
application passively checking for mail in the background
without an active request that it do so by the user. Id.
Collecting data at every radio signal—whether the origin or
termination of a call, a call hand-off, a text message, or a data
connection by an application—threatens an individual’s
reasonable expectation of privacy more than collecting data at
the origination and termination of calls only.

For all of the foregoing reasons, I believe that the
Government obtaining the quantity of historical CSLI it did in
this case amounts to a search that, without a warrant,
infringes on an individual’s reasonable expectation of privacy
and violates the Fourth Amendment.

10
 II

All of this said, I would not suppress the CSLI
evidence in this case because of the good faith exception to
the warrant requirement. “Searches conducted in objectively
reasonable reliance on binding appellate precedent are not
subject to the exclusionary rule.” United States v. Katzin, 769
F.3d 163, 172-73 (3d Cir. 2014) (en banc) (quoting Davis v.
United States, 564 U.S. 229, 231 (2011)).4 In re Application
amounted to binding Circuit precedent that “specifically
authorize[d the] particular police practice” at issue here.
Davis, 564 U.S. at 241. As such, the CSLI in this case need
not have been excluded, and I concur with the Opinion of the
Court as to the judgment on this issue.

4
The Opinion of the Court and the dissents in Katzin
disagree as to how directly a prior case must authorize
particular law enforcement conduct to amount to “binding
appellate precedent” on which law enforcement officers could
rely. Compare Katzin, 769 F.3d at 174, with id. at 192-93
(Greenaway, Jr., J., dissenting). This case presents no such
question of directness. The Katzin dissents also pointed out
that limiting the application of the exclusionary rule might
lead to more occasions of law enforcement officers
conducting searches not sanctioned by judges. See id. at 189-
90 (Greenaway, Jr., J., dissenting). Here, agents still sought
out the imprimatur of a neutral magistrate judge for the search
(albeit under a lesser standard than probable cause).

11
 III

Despite applying the good faith exception to the
warrant requirement in this instance, I believe that obtaining
historical CSLI that approaches GPS-level precision,
aggregated over at least the four week period the Jones
shadow majority rejected—as in this case—should require a
warrant supported by probable cause rather than a § 2703(d)
order supported by reasonable suspicion. “Our cases have
historically recognized that the warrant requirement is an
important working part of our machinery of government, not
merely an inconvenience to be somehow weighed against the
claims of police efficiency.” Riley, 134 S.Ct. at 2493
(internal quotation marks omitted). Especially where
“[r]ecent technological advances . . . have . . . made the
process of obtaining a warrant itself more efficient,” we need
not sanction the rapid pace and expansive scope of
technological change eroding important constitutional
protections that we have enjoyed for centuries. Id. “The
[Fourth] Amendment and the common law from which it was
constructed leave ample room for law enforcement to do its
job. A warrant will always do.” U.S. v. Carloss, 818 F.3d
988, 1015 (10th Cir. 2016) (Gorsuch, J., dissenting).

For all of the foregoing reasons, I concur only as to the
judgment in parts III(A)(2)-(3) of the Opinion of the Court.

12