United States v. Michael Lustig
830 F.3d 1075
9th Cir.
2016
Check Treatment
Docket

UNITED STATES of America, Plaintiff-Appellee, v. Michael LUSTIG, AKA George, Defendant-Appellant.

No. 14-50549

United States Court of Appeals, Ninth Circuit.

Filed July 29, 2016

830 F.3d 1075

Argued and Submitted January 8, 2016 Pasadena, California

ty to explore fully a factual issue at resentencing simply because it failed to do so during the first proceeding as a result of an erroneous legal ruling.“). The record on remand shall be limited accordingly.

III. CONCLUSION

We reverse the district court‘s denial of Thomsen‘s Motion For Judgment Of Acquittal as to Counts 33 and 34, and vacate those convictions. We reverse the order of restitution and sentence of incarceration, and remand for redetermination of both.

AFFIRMED IN PART, REVERSED IN PART, AND REMANDED.

Timothy A. Scott (argued) and Nicolas O. Jimenez, Coleman, Balogh & Scott LLP, San Diego, California, for Defendant-Appellant.

Helen H. Hong (argued), Assistant United States Attorney; Peter Ko, Chief, Appellate Section, Criminal Division; Laura E. Duffy, United States Attorney; United States Attorney‘s Office, San Diego, California; for Plaintiff-Appellee.

Before: PAUL J. WATFORD and MICHELLE T. FRIEDLAND, Circuit Judges and J. FREDERICK MOTZ,** Senior District Judge.

Concurrence by Judge WATFORD

OPINION

FRIEDLAND, Circuit Judge:

The United States Supreme Court held in , that the Fourth Amendment requires law enforcement officers to obtain a warrant before they may search an arrestee‘s cell phone. Approximately two years before that decision, an officer arresting Michael Lustig conducted warrantless searches, incident to the arrest, of cell phones found in Lustig‘s pockets. We must determine whether pre-Riley precedent provided a reasonable basis to believe such searches were constitutional. Because we hold that binding appellate precedent at the time of the searches did provide a reasonable basis to believe the searches were constitutional, the good-faith exception to the exclusionary rule applies to the evidence obtained from those searches. In addition, we must determine the effect of a concededly erroneous denial of a motion to suppress evidence obtained from separate searches of other cell phones found in Lustig‘s car. To do so, we first adopt our sister circuits’ test for evaluating harmlessness in the context of a conditional guilty plea. Because the Government has not met its burden of establishing harmlessness under that test, Lustig must be given an opportunity to vacate his guilty plea if he so wishes. We thus affirm in part, reverse in part, and remand.1

** The Honorable J. Frederick Motz, Senior District Judge for the U.S. District Court for the District of Maryland, sitting by designation.

I

In June 2012, a task force consisting of local and federal law enforcement agencies conducted a sting operation to obtain evidence of prostitution offenses. To effectuate the operation, an undercover officer posed as a prostitute and placed listings on a classified advertisements website. Defendant-Appellant Michael Lustig responded to the advertisements and agreed to meet the undercover officer at a hotel in Encinitas, California. Lustig was arrested at the hotel for soliciting prostitution in violation of California law. Upon the arrest, Deputy Sheriff Chase Chiappino seized and searched cell phones found on Lustig‘s person and in his car.

Two cell phones were seized from Lustig‘s pockets incident to his arrest (the “Pocket Phones“). One was an Apple iPhone, which Chiappino, upon its seizure, unlocked by swiping across the screen. Chiappino observed that the phone opened to the website where the fake advertisement was posted, and he located the phone‘s number on its settings page. The other Pocket Phone was a Kyocera flip phone. Chiappino searched the Kyocera phone by viewing its call history and text messages and identifying its phone number. The search revealed text messages suggesting further involvement with prostitution.

Officers seized additional cell phones from Lustig‘s car, which was in the parking lot of the hotel (the “Car Phones“). At the scene, Chiappino searched those phones and found additional text messages regarding prostitution.

Four days later, Chiappino returned to searching the phones. He downloaded content from the phones and searched the phones’ contacts in law enforcement databases. The parties dispute whether Chiappino searched the Car Phones or the Pocket Phones first, and whether evidence discovered in one set of phones motivated searches of the other.

In one of the Car Phones, Chiappino found text message exchanges suggesting prostitution activity with a contact named “Dominick.” He searched that contact‘s phone number in law enforcement databases but found no match. He also found a contact named “Dominick” in one of the Pocket Phones (the iPhone), searched that phone number, and discovered a match to a twelve-year-old minor female, whom the officers thereafter referred to as “MF1.”

In his investigation of the Kyocera Pocket Phone, Chiappino found a series of messages discussing libraries and bookstores with a contact named “Andrew.” He searched for that contact‘s phone number in law enforcement databases and matched it to a fourteen-year-old minor female, “MF2.”

Officers then located and interviewed MF1 and MF2 separately, and both confirmed that they had engaged in commercial sex activity with Lustig. According to a declaration filed by Chiappino but disputed by Lustig, MF2 also directed officers to a motel, where the officers eventually obtained video surveillance of Lustig entering and leaving a room with a female whom officers identified as MF1.

No warrants were obtained prior to any of these cell phone searches. Sixteen months later, however, the officers did obtain warrants to search two of the already searched Car Phones.

Lustig was indicted in the United States District Court for the Southern District of California on two counts of child sex trafficking in violation of 18 U.S.C. §§ 1591(a) and (b), based on his conduct with MF1 and MF2. During pretrial proceedings, Lustig moved to suppress the evidence found through the searches of the phones. He argued that the seizure of the Car Phones, and the searches of both the Car Phones and Pocket Phones, violated the Fourth Amendment.2 The district court, after declining to hold a hearing, denied the motion approximately three months before the Supreme Court issued its decision in .

Regarding the Pocket Phones, the district court held that the searches were unconstitutional. It reasoned that “searching an arrestee‘s phone [without a warrant], beyond what is in plain view, is an unreasonable search under the Fourth Amendment... where the crime charged is a misdemeanor,” as Lustig‘s charge was at the time of arrest.3 Nevertheless, the district court went on to conclude that the evidence found in the searches was admissible pursuant to the good-faith exception to the exclusionary rule. The court explained that at the time of the searches, the California Supreme Court in , had held that warrantless searches of cell phones seized from an arrestee‘s person incident to arrest did not violate the Fourth Amendment. The district court also noted that there were “no binding decisions to the contrary from the federal courts.”

As to the Car Phones, the district court held that they were constitutionally seized, but that the warrantless searches of the phones’ content were unconstitutional. The district court nevertheless declined to suppress evidence obtained from the Car Phones. Because the Government eventually attained—16 months later—a search warrant for the Car Phones, the district court reasoned that the evidence would inevitably have been discovered.

Lustig filed two motions to reconsider these suppression rulings, each of which the district court denied. Lustig subsequently entered a conditional guilty plea pursuant to Federal Rule of Criminal Procedure 11(a)(2). Under the plea agreement, Lustig pled guilty to three counts of violating 18 U.S.C. § 1952(a)(3) by using a cell phone to facilitate a prostitution offense under 18 U.S.C. § 1591, involving only MF2, rather than the original indictment‘s two counts for child sex trafficking involving both MF1 and MF2. The conditional guilty plea preserved Lustig‘s right to appeal the Fourth Amendment issues related to his motions to suppress.

After the plea was entered, the Government filed as part of its sentencing submissions the aforementioned declaration from Chiappino, which asserted that evidence concerning MF2 “was wholly untainted by” evidence from the Car Phones, and that officers “would have inevitably discovered” MF1 even if not for the Car Phone searches.

Lustig now appeals the denial of his suppression motions.

II

We review a district court‘s denial of a motion to suppress evidence de novo. . We review a district court‘s factual findings for clear error and its application of the good-faith exception de novo. .

III

Lustig advances two primary contentions on appeal. First, he argues that pre-Riley authority provided no reasonable basis for Chiappino to search without a warrant the contents of the Pocket Phones, and that the district court therefore erred in holding that the fruit of those searches was admissible under the good-faith exception to the exclusionary rule. Second, Lustig argues that the district court erred in declining to suppress the Car Phone evidence. On appeal, the Government concedes that the district court erred as to the Car Phone evidence, but argues that the error was harmless because it did not affect Lustig‘s counts of conviction. We address each issue in turn.

A

In , the Supreme Court unanimously held that warrantless searches of cell phones seized incident to arrest violate the Fourth Amendment. Id. at 2495. There is thus no question that the searches of Lustig‘s Pocket Phones were unconstitutional. The question on appeal is instead whether the good-faith exception to the exclusionary rule nevertheless makes admissible the evidence found in the Pocket Phone searches. We hold that it does.

The Fourth Amendment protects the “right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures.” U.S. Const. amend. IV. To deter Fourth Amendment violations, courts apply the exclusionary rule to suppress evidence that has been unconstitutionally obtained. . In circumstances in which “suppression fails to yield ‘appreciable deterrence,‘” however, the Supreme Court has held that “exclusion is ‘clearly... unwarranted.‘” , 131 S.Ct. 2419 (alteration in original) (quoting ). “[W]hen the police act with an objectively ‘reasonable good-faith belief that their conduct is lawful... the ‘deterrence rationale loses much of its force,‘” and therefore the exclusionary rule does not apply. , 131 S.Ct. 2419 (quoting ). In Davis, the Supreme Court held that such a “reasonable good-faith belief” exists when searches are conducted “in objectively reasonable reliance on binding appellate precedent.” , 131 S.Ct. 2419.

Davis involved a vehicle search during which the arrestee, Davis, was out of reaching distance of the car. Davis moved to suppress a revolver found inside the vehicle. , 131 S.Ct. 2419. The Eleventh Circuit had long approved of such searches, understanding the Supreme Court‘s decision in , “to establish a bright-line rule authorizing substantially contemporaneous” automobile searches incident to arrest. , 131 S.Ct. 2419 (citing ). The district court denied Davis‘s motion, consistent with the Eleventh Circuit‘s law at the time. While Davis‘s appeal was pending, however, the Supreme Court held in , that vehicle searches pursuant to arrest are generally forbidden when the arrestee is out of reaching distance of the vehicle. , 131 S.Ct. 2419. The Supreme Court in Davis held that, although Gant made the search of Davis‘s car unconstitutional, the good-faith exception applied because the search had been “in strict compliance with” “binding appellate precedent.” , 131 S.Ct. 2419.

Here, the Government argues that, like the officers in Davis, Chiappino reasonably relied on then-binding appellate precedent authorizing his search of Lustig‘s Pocket Phones. The Government specifically points to , in which the Supreme Court held, seemingly as a categorical matter, that “in the case of a lawful custodial arrest a full search of the person is not only an exception to the warrant requirement of the Fourth Amendment, but is also a ‘reasonable’ search under that Amendment.” , 94 S.Ct. 467. The Supreme Court emphasized that “[t]he authority to search the person incident to a lawful custodial arrest... does not depend on... the probability in a particular arrest situation that weapons or evidence would in fact be found upon the person of the suspect,” because once there is probable cause to arrest, “a search incident to the arrest requires no additional justification.” Applying this broad principle, the Supreme Court held that an officer had not violated the Fourth Amendment by searching a crumpled package of cigarettes in the arrestee‘s pocket without a warrant, or by seizing the heroin capsules hidden therein. , 94 S.Ct. 467.

We agree with the Government that, before Riley, it was objectively reasonable to have interpreted Robinson to announce a bright-line rule authorizing any search incident to arrest of any item found in an arrestee‘s pocket.

1

As a threshold matter, we recognize the obvious fact that Robinson did not involve searches of cell phones, and indeed could not have, given the state of technology at the time. Lustig argues that Robinson‘s lack of factual equivalence to his case is alone sufficient to preclude application of the good-faith exception under Davis. But, as the Third Circuit has accurately observed, “[n]o two cases will be factually identical.” , cert. denied, . The Third Circuit has explained that the Davis inquiry “is not answered simply by mechanically comparing the facts of cases and tallying their similarities and differences. Rather, [it] involves a holistic examination of whether a reasonable officer would believe in good faith that binding appellate precedent authorized certain conduct.” The relevant determination is thus whether “the rationale underpinning the [binding appellate precedent]... clearly authorized the [officers‘] conduct.” (emphasis added); see also (considering whether the purported binding precedent “provide[s] a rationale to justify [the officer‘s] search“); (“[I]t is the legal principle of [the precedent], rather than the precise factual circumstances, that matters.“), cert. denied, ; (rejecting the contention that binding appellate precedent must be “specific to the facts at hand“).

Our own case law is consistent with this approach to applying the good-faith exception. In , we held that the good-faith exception applied when officers relied on Supreme Court precedent4 that was silent on the key fact motivating the suppression motion. There, the defendant challenged as unconstitutional a drug-detection dog‘s touching of his vehicle during a dog-sniff inspection of the vehicle—an inspection that resulted in the discovery and seizure of marijuana. . The defendant relied on two Supreme Court cases decided after the seizure in question for the proposition that the dog‘s physically touching his vehicle was an unconstitutional trespass prohibited by the Fourth Amendment. . We held that, whether or not the dog‘s physical contact with the car violated the Fourth Amendment under these later cases, Supreme Court precedent at the time of the incident categorically authorizing dog-sniff inspections at vehicle stops made the evidence admissible under the good-faith exception to the exclusionary rule. (citing (“A dog sniff conducted during a... lawful traffic stop that reveals no information other than the location of a substance that no individual has any right to possess does not violate the Fourth Amendment“) (alteration in original), and (“[A]n exterior sniff of an automobile” is permissible because it “does not require entry into the car.“)). Although Caballes and Edmond did not address physical contact with a vehicle, we nevertheless held that the good-faith exception applied because neither case “so much as hinted that officers were to avoid contact” between the dog and the vehicle‘s exterior. . Because the binding case law at the time of the inspection “specifically authorize[d] a

particular police practice“—exterior dog-sniffs at vehicle stops—“the absence of a previously expressed limit” on the categorical rule, rather than a prior endorsement of a particular subset of factual circumstances, was dispositive of the good-faith analysis. Id. (quoting , 131 S.Ct. 2419).

Following this approach, we reject Lustig‘s contention that the good-faith exception cannot apply here because, at the time of his arrest, there had not been any decision by this Circuit or the Supreme Court directly authorizing warrantless cell phone searches incident to arrest. If precedent had to constitute a factual match with the circumstances of the search in question for the good-faith exception to apply, it would make the good-faith exception a nullity because the exception would only apply when the search was necessarily constitutional under existing precedent.

Considering, then, the legal principles established by Robinson and not merely its specific facts, we conclude that Robinson was binding appellate authority that made it reasonable to search Lustig‘s Pocket Phones. Even the Supreme Court in Riley, which “decline[d] to extend Robinson” from physical objects to cell phone data, acknowledged that Robinson had established a “categorical rule,” and that “a mechanical application of Robinson might well support” cell phone searches. .5

Lustig argues, however, that the law governing warrantless searches of cell phones was unsettled at the time of the search, thus precluding objectively reasonable reliance on Robinson. In support, Lustig cites a handful of federal district court decisions and an Ohio Supreme Court decision pre-dating the searches here, which had held that cell phone searches incident to arrest were unconstitutional. See, e.g., ; .

The Davis inquiry, however, is focused on binding appellate authority, which Lustig‘s cases are not. See (looking to Supreme Court and Ninth Circuit precedent in applying Davis); see also (noting that courts applying Davis look to “circuit-level binding appellate precedent,” but that “[c]ircuits without local precedent rel[y] on... Supreme Court” precedent); (for the Davis good-faith exception to apply, “officers performing a particular investigatory action... must strictly comply with binding appellate precedent governing the jurisdiction in which they are acting“); (“binding precedent” under Davis “refers to the precedent of this Circuit and the Supreme Court“). Even if state appellate court decisions could serve as “binding appellate precedent“—a question we do not decide, see infra n.10—an Ohio state court appellate decision had no “binding” effect on the officers’ searches of Lustig‘s phones in California. Lustig‘s contrary argument would suggest that police could not rely on Supreme Court precedent that seems to authorize the search in question if any district court or state court anywhere in the country disagreed about the breadth of that precedent. We decline to impose on law enforcement an obligation to constantly search for non-binding authority across all jurisdictions and to curtail their otherwise authorized activities as soon as any court casts existing precedent into doubt.6

Lustig contends that application of the good-faith exception here is precluded by our decision in , which he argues has already held that the good-faith exception does not apply to pre-Riley warrantless cell phone searches. Lustig misconstrues Camou, which dealt only with the timing of searches following an arrest. In Camou, United States Border Patrol agents had stopped the defendant‘s truck at an inspection checkpoint and discovered an undocumented immigrant hiding in the truck. . The defendant was placed under arrest and agents seized his truck as well as a cell phone found in the cab of the truck. Id. One hour and twenty minutes after the defendant‘s arrest, an agent searched the cell phone and found photographic images of child pornography. . The defendant was indicted on child pornography charges and moved to suppress the images found on his cell phone. Id. The district court denied the motion and we reversed. . We held, inter alia, that the search of the phone was not incident to arrest because it was conducted at a time too remote from the arrest, and that the good-faith exception did not apply because the “governing law at the time of the search made clear that a search incident to arrest had to be contemporaneous with the arrest.” (citing ).

Although Camou, in its broadest outlines, is a post-Riley case holding that the good-faith exception did not apply to a pre-Riley cell phone search, it did not address the central issue here—whether, when a cell phone is found during an otherwise unquestionably valid search incident to arrest, it may be searched during the arrest without a warrant.7 Because Camou said nothing about the question we face here—and indeed never mentioned Robinson at all, let alone its relationship to Riley—it does not foreclose application of the good-faith exception to the searches of Lustig‘s Pocket Phones.

Lustig‘s reliance on our recent decision in —another post-Riley case that declined to apply the good-faith exception to a pre-Riley search of a cell phone—is similarly unavailing. Lara concerned a warrantless search of a probationer‘s cell phone pursuant to a probation agreement that included a “Fourth Amendment waiver.” . The Fourth Amendment waiver provided that the probationer would submit his “person and property, including any residence, premises, container or vehicle... to search and seizure at any time” by any officer, “with or without a warrant, probable cause, or reasonable suspicion.” Id. The probationer was ultimately charged with being a felon in possession of a firearm and ammunition based on evidence discovered from a search of his cell phone. . The probationer moved to suppress that evidence. Id. On appeal, we held that the waiver language did not clearly authorize the search in that case, and that the search was not otherwise a reasonable probation search. .

In addition, although we rejected the government‘s reliance on the Davis good-faith exception,8 we specifically distinguished cases involving searches of cell phones incident to arrest: “It hardly needs saying that a search incident to arrest is not the same thing as a warrantless, suspicionless, probation search. Nor is a case dealing with an incidental search on all fours with a probation search.” . Given this language in Lara making clear that the questions it addressed were distinct from the questions posed by searches incident to arrest, Lara does not help Lustig resist application of the good-faith exception here.

Finally, Lustig suggests that Riley tacitly rejected applying the good-faith exception to cell phone searches. He points to the fact that the Supreme Court in Riley unanimously rejected the argument that Robinson extended to cell phone searches as evidence that it was never reasonable to think that Robinson authorized such searches. But the Supreme Court suggested exactly the opposite when it observed, as noted above, that “mechanical application of Robinson might well support the warrantless searches at issue here.” .9

Because Robinson, by its terms, “specifically authorize[d]” the search incident to arrest of an object found on the arrestee‘s person, the good-faith exception makes admissible the evidence obtained during the searches of the Pocket Phones incident to Lustig‘s arrest.10 ,

131 S.Ct. 2419 (emphasis omitted). As the First Circuit observed in discussing another line of precedent, even though this bright-line rule “turned out not to be as categorical as [it] seemed... that is not a reason to penalize the police for applying [it] faithfully before [that] clarification[] occurred.” .

2

Lustig contends that even if the good-faith exception saves the searches of the Pocket Phones conducted at the hotel, the delay between those initial searches and the more comprehensive stationhouse searches undertaken four days later rendered the stationhouse searches unconstitutional. We disagree.

In , we held that once an item “has been lawfully seized and searched, subsequent searches of that item, so long as it remains in the legitimate uninterrupted possession of the police, may be conducted without a warrant.” . In specifically holding that a brief search of a purse incident to arrest and a more detailed warrantless search of the same purse later at the stationhouse were both constitutional, we emphasized the “necessarily reduced expectation of privacy one holds in his person after being placed under arrest” and the “necessarily reduced” expectation of privacy in an item already validly searched incident to arrest. “Requiring police to procure a warrant for subsequent searches of an item already lawfully searched would in no way provide additional protection for an individual‘s legitimate privacy interests.” This reasoning applies to the searches here, whether delayed by four hours or four days. Because the Pocket Phones were lawfully seized from Lustig‘s person and immediately searched incident to arrest, Burnette fully authorizes the later searches. At the very least, it was reasonable for Chiappino to believe that four days was a permissible delay.

Lustig argues to the contrary, contending that the four-day delay is “far more egregious” than the one hour and twenty minute delay at issue in Camou. See . Camou, however, did not consider how a preliminary search at the time of arrest might affect a later search of the same item. In Camou, there was no search of the cell phone incident to arrest, so the delayed warrantless search was the initial search. Camou thus has no bearing here.

In sum, Robinson made it objectively reasonable to believe that the searches of the Pocket Phones were constitutional. We further conclude that Burnette authorized the subsequent stationhouse searches of the Pocket Phones, or at least provided a basis for a good-faith belief that those searches were lawful. We therefore affirm the denial of Lustig‘s suppression motion as to the Pocket Phones.

B

Lustig also challenges the denial of the motion to suppress evidence obtained through the Car Phone searches. In its Answering Brief, the Government concedes, citing ,11 that it did not present sufficient evidence to show

because we may affirm on any ground supported by the record, , we need not decide whether state court decisions such as Diaz have any relevance to the good-faith analysis here.

that the 16-month delay between the seizure of the Car Phones and the officers’ obtaining a warrant to search them was reasonable under the Fourth Amendment, and that the district court therefore erred in denying Lustig‘s motion. The Government argues, however, that the error was harmless. It contends that because Lustig pled guilty only to charges involving MF2, whom the Government asserts was identified exclusively through information obtained from the Pocket Phones, any evidence derived from the Car Phones was “immaterial to Lustig‘s conviction.” We are not persuaded that this is the relevant harmlessness inquiry. Rather, as our sister circuits have held, the relevant question in the conditional plea context is whether the erroneous suppression ruling could have affected Lustig‘s decision to plead guilty. Because it could have, we reverse the suppression ruling on the Car Phones.

1

As an initial matter, we agree with the Government‘s contention that harmless error review applies here. The Federal Rules of Criminal Procedure specifically provide, under the heading “[h]armless [e]rror,” that “[a]ny error, defect, irregularity, or variance that does not affect substantial rights must be disregarded.” Fed. R. Crim. P. 52(a). And the Supreme Court has held that, generally, constitutional errors in criminal proceedings must be disregarded if the government can prove that they are harmless “beyond a reasonable doubt.” (quoting ).

Consistent with these general principles, our prior decisions have applied harmless error review in the Rule 11(a)(2) conditional plea context.12 In , for instance, we held that a frisk of the defendant violated his Fourth Amendment rights and that the district court erred by failing to suppress the hashish oil discovered as a result of that frisk. Nevertheless, we held that the error did not mandate reversal because, even without the hashish oil, the officers had “sufficient [evidence] to establish probable cause to search [the defendant‘s] truck“—a search that ultimately led to discovery of the marijuana plants that formed the basis of the defendant‘s conviction. . Therefore, “[a]ny error by the district court in failing to suppress the hashish oil was harmless.” ; see also, e.g., , amended, (reviewing district court errors for harmless error on appeal from a conditional plea).

Lustig‘s contention that harmless error review does not apply in the Rule 11(a)(2) context, and that any error, however slight or tangential, requires reversal with the opportunity to withdraw the plea, is incorrect in light of this precedent. Lustig rests his argument entirely on a statement in a footnote in our decision in , that “[i]f any ruling that forms a basis for the conditional plea is found to be erroneous, we are required to permit the defendant to withdraw his plea.” . In context, it is clear that this sentence was not stating a general proposition

but responded instead to the particular facts of that case.

Mejia concerned two motions to suppress, relating to a confession and consent to search a home, respectively. Both the confession and the consent to search arose out of an allegedly unconstitutional interrogation. The error we held the district court to have made related to a continuance denial that prevented the defendant from presenting testimony needed to resolve material fact disputes about the interrogation. We explained in the same footnote that:

given the fact that the [two] motions [to suppress] were heard together, that they related to the same interrogation and involved overlapping issues, that the failure to give a Miranda warning can be a consideration when determining questions of consent, and that the court‘s error as to both motions was identical, we would conclude that, under all the circumstances, a showing of prejudice as to either would be sufficient to require a finding of error and a new hearing as to both.

This factual context shows that the statement Lustig relies upon cannot be interpreted to broadly foreclose harmless error review in all instances. Instead, it refers to the interrelated nature of the two motions and the conditional plea at issue in that case. Indeed, the need to show that an error was prejudicial in order for that error to trigger the right to vacate a plea was clarified in the same footnote by the phrase “a showing of prejudice as to either [motion] would be sufficient to require a finding of error and a new hearing as to both.” (emphasis added).

This understanding of Rule 11(a)(2) is consistent with the approaches of other circuits, which likewise have applied harmless error type principles in the conditional plea context. See, e.g., ; see also ; .13 We thus agree with the Government that harmless error review applies to the district court‘s failure to suppress the fruit of the Car Phone searches.

2

Having established that harmless error review applies in Rule 11(a)(2) appeals, we must now determine the standards that govern that review. The Government urges us to adopt a standard that defines an error as harmless when we can conclude, beyond a reasonable doubt, that the evidence erroneously admitted was “immaterial to [the defendant‘s] conviction.” Our cases have not directly addressed this issue, but Rule 11(a)(2) itself and authority from our sister circuits cause us to believe that the correct standard is instead whether the government has proved beyond a reasonable doubt that the erroneously denied suppression motion did not contribute to the defendant‘s decision to plead guilty.

The critical event for a defendant in a conditional plea context is the decision to plead guilty after considering what a trial would entail in light of the failed pretrial motions. Rule 11(a)(2) allows a defendant, having lost certain pretrial motions, to plead guilty while reserving the

right to appeal those pretrial rulings. See Fed. R. Crim. P. 11, advisory committee‘s note to 1983 amendment (stating that the purpose of subsection (a)(2) is to avoid forcing “a defendant who has lost one or more pretrial motions” to “go through an entire trial simply to preserve the pretrial issues for later appellate review“). As the Tenth Circuit has held, unlike in cases decided by a jury, in which constitutional error will be harmless if the court concludes “beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained,” (quoting , 87 S.Ct. 824), for convictions based on conditional guilty pleas, the test must be “reformulated to determine whether there is a reasonable possibility that the error contributed to the plea,” (emphasis added) (quoting ); see also (considering whether suppression of the contested evidence “would have affected [the defendant‘s] decision to plead guilty“). It is thus whether the evidence at issue in an erroneously denied suppression motion could have affected the defendant‘s decision to plead guilty, not whether the evidence was material to a charge to which the defendant pled, that determines whether the suppression error was harmless in a conditional plea context.

The relevant inquiry in this case is thus whether there is a “reasonable possibility”14 that the erroneously admitted Car Phone evidence contributed to Lustig‘s decision to plead guilty. . This “reasonable possibility” standard is necessarily hard for the government to meet. This is because, as the Tenth Circuit has explained, “in the context of a plea, the record will be unlikely to contain enough information for an appellate court” to conclude beyond a reasonable doubt that the evidence did or did not contribute to the defendant‘s plea decision. A “defendant‘s decision to plead guilty may be based on any factor inside or outside the record,” (quoting ), and “only the defendant is in a position to evaluate the impact of a particular erroneous refusal to suppress evidence,” (citation omitted) (quoting ). “Accordingly, ‘an appellate court will rarely, if ever, be able to determine whether an erroneous denial of a motion to suppress contributed to the defendant‘s decision [to plead guilty], unless at the time of the plea he states or reveals his reason for pleading guilty.‘” (alteration in original) (quoting ).

Applying these principles, the Tenth Circuit in Benard rejected the government‘s argument that the suppression error there was harmless because the key firearm evidence supporting the firearm conviction that determined the defendant‘s ultimate sentence was not affected by the error. Instead, the Tenth Circuit held that it was unable to “conclude beyond a reasonable doubt that the district court‘s error did not contribute to [the defen-

dant‘s] decision to plead guilty. The record does not indicate why [the defendant] decided to plead guilty, what other defenses or evidence he might have produced on his behalf, or how the altered bargaining positions of the parties might have affected his decision if his post-arrest statements had been properly suppressed.” .

Further, the Tenth Circuit rejected the government‘s contention in Benard that, on remand, the case should be limited to the defendant‘s firearm conviction because the suppression error implicated only the defendant‘s drug conviction. The Tenth Circuit explained that a reviewing court should consider the error‘s effect on the “bargaining positions of the parties” in light of “the aggregate strength of all the incriminating evidence accumulated by the government,” including evidence on other counts. (alteration omitted) (quoting ). “[F]inding the suppression error to affect only some counts of a multi-count indictment would interfere with the defendant‘s ‘prerogative to personally decide whether to stand trial or to waive his rights by pleading guilty’ to the various counts of the indictment.” (quoting , overruled on other grounds by ). Because the Tenth Circuit could not conclude beyond a reasonable doubt that the defendant “would still have agreed to waive his right to a jury trial as to either or both of the counts of conviction absent the district court‘s error,” it remanded “both counts of conviction under Rule 11(a)(2).” .

Other circuits are in accord with these principles. The Sixth Circuit in , for example, articulated a standard substantially similar to the Tenth Circuit‘s for determining when a defendant would be entitled to withdraw his plea, requiring consideration of “the probability that the excluded evidence would have had a material effect on the defendant‘s decision to plead guilty.” (emphasis added). The D.C. Circuit has adopted a similar test. See (quoting Leake to conclude that the defendant was entitled to withdraw his plea); see also (addressing conditional pleas prior to Rule 11(a)(2) and holding that failure to suppress evidence was harmless error because suppression “would not have altered appellant‘s decision to plead guilty“).

Recently, the First Circuit arguably applied a harmlessness standard even harder (or impossible) for the government to satisfy when it remanded a case to allow the defendant to withdraw his guilty plea despite noting that “it is highly unlikely that the suppression of [the statements in question] regarding drug trafficking activity... would have affected [the defendant‘s] decision to plead guilty.” . The First Circuit explained that determining whether the defendant would have pled guilty absent the error was “not our decision to make.... [A] court has no right to decide for a defendant that his decision [to plead guilty] would have been the same had the evidence the court considers harmless not been present.” Id. (second alteration in original) (quoting ). The defendant “is entitled to determine for himself whether he still wishes to plead guilty given the suppression of the drug-trafficking-related statements.” Id.

Insofar as Molina-Gomez may be read to mandate remand on any error without considering harmlessness,15 our precedents applying harmless error review, described above, foreclose adopting such a blanket rule. See, e.g., ; see also (holding that a three-judge panel is bound by prior circuit precedent unless “clearly irreconcilable... intervening higher authority” “effectively overrule[s]” the precedent). We instead adopt the rule articulated by a plurality of the circuit courts, under which we must consider whether an erroneous denial of a motion to suppress contributed to the defendant‘s decision to plead guilty, and under which it is only the “rare” case in which we may definitively make the harmlessness determination necessary to preclude remand. ; see also (recognizing Benard‘s standard for allowing the defendant to vacate the plea but concluding beyond a reasonable doubt that any error did not contribute to the defendant‘s decision to plead guilty because “[t]he government unequivocally represented to [the defendant] and the court that it would not seek to admit [the defendant‘s] statements at trial” and thereby took the contested statements “off the table“).

Contrary to the Government‘s arguments, our precedent is not inconsistent with a Rule 11(a)(2) inquiry that looks to

the decision to plead guilty rather than the relationship of the wrongfully admitted evidence to the conviction. Although we noted in , that the evidence wrongfully admitted was “immaterial to [the defendant‘s] conviction,” that case did not mention, much less consider, the essential distinction between evidence of underlying guilt and evidence that could contribute to a plea decision in the Rule 11(a)(2) context. Moreover, even if it had, it is unlikely that that distinction would have made a difference to the outcome of that particular case. Sines was an example of the “rare” case in which it was clear that the wrongfully admitted evidence made no difference either to the decision to plead guilty or to the conviction. The evidence at issue in Sines was the defendant‘s passport, which the prosecution could have used to corroborate a witness‘s testimony that the defendant was in Thailand at the relevant time. Id. We determined, however, that the passport was entirely unnecessary for that purpose because other ample and admissible evidence served the same function. Id. Furthermore, the prosecution did not even mention the passport as part of the evidence against the defendant during the defendant‘s nolo contendere plea colloquy, despite mentioning all of the other evidence that proved his presence in Thailand. Id. The passport was thus unambiguously not a factor in the case.16

In sum, in light of the purpose and effect of Rule 11(a)(2) and our existing case law, we agree with the approach taken by at least the plurality of our sister circuits for analyzing whether an error is harmless under Rule 11(a)(2). If it is beyond a reasonable doubt that the error did not contribute to the decision to plead guilty, it will be considered harmless. Otherwise, the error will require a remand to provide an opportunity for the defendant to vacate the guilty plea.

3

Applying this framework to the present case, we conclude that the Government has not met its burden of establishing harmless error. See (“The burden of proving a constitutional error harmless beyond a reasonable doubt rests upon the government.“). The Government asks and answers the wrong question when it argues that admission of the Car Phone evidence was harmless because it was “immaterial to Lustig‘s conviction.” The relevant inquiry is whether the erroneous admission of the Car Phone evidence was immaterial to Lustig‘s decision to enter a guilty plea. Given the dearth of factual clarity in the record as to Lustig‘s plea considerations, and indeed as to what evidence, exactly, was derived from the Car Phones, we cannot conclude beyond a reasonable doubt that the Car Phone evidence did not contribute to Lustig‘s decision to plead guilty.

The Government centers its argument on its assertion that all of the evidence pertaining to MF2—which formed the basis for the only charges to which Lustig ultimately pled guilty—was obtained solely from the Kyocera pocket phone rather than from the Car Phones. But this argument “ignores the fact that the guilty plea was entered as part of an agreement involving all of the counts of the [indictment],” (alteration in original) (quoting )—an indictment that initially included not just counts related to MF2, but also counts related to MF1, about whom some evidence was found in the Car Phones. The express terms of Lustig‘s conditional plea make explicit that “[i]n exchange for Defendant‘s guilty plea... the United States agrees to dismiss the Indictment without prejudice at the time of sentencing.” Thus, Lustig‘s “decision to plead guilty to [the MF2-related counts in the superseding information] was not made in a vacuum independent of the evidence on [the MF1-related counts].” . Considering the “bargaining positions of the parties” in light of the “aggregate strength of all the incriminating evidence accumulated by the [government],” (alteration in original) (quoting ), the Car Phone evidence could have had some effect on Lustig‘s decision to plead guilty even if that evidence may not have supported the MF2-related counts of conviction.

An additional and independent reason to reject the Government‘s harmlessness argument is that it is unclear what evidence may have constituted the fruit of the Car Phone searches. The Government relies entirely on a single declaration by Deputy Chiappino for its assertion that none of the Car Phone evidence was used to locate the evidence needed to support the MF2-related charges. But this declaration was submitted at sentencing, long after the suppression motions were litigated, and Lustig never had an opportunity to challenge Chiappino‘s statements through cross-examination. Indeed, at oral argument before this court, the Government conceded that Chiappino‘s statements were not “tested below.” Lustig, for his part, raises factual questions as to the order of the searches of the various phones and had asked the district court to hold a hearing to identify the fruit of the searches. Because the district court denied Lustig‘s motion to suppress the Car Phones, it never had occasion to hold such a hearing or to make a determination as to the exact fruit of the searches.

For these reasons, we simply cannot know “how the altered bargaining positions of the parties might have affected [Lustig‘s] decision [to plead guilty] if [the Car Phone evidence and any fruit thereof] had been properly suppressed.” . We certainly cannot conclude, as the Government urges, that the Car Phone evidence was analogous to the redundant, essentially useless passport that the prosecution disclaimed as evidence in Sines. See .17 In light of the Government‘s failure to satisfy its burden of proving beyond a reasonable doubt that the district court‘s suppression error was harmless, we remand to allow Lustig the opportunity to withdraw his guilty plea.18

IV

The district court‘s denial of the motion to suppress evidence from the Pocket Phones is AFFIRMED. We REVERSE the district court‘s denial of the motion to suppress evidence from the Car Phones and REMAND for further proceedings consistent with this opinion.

WATFORD, Circuit Judge, concurring:

I join the court‘s opinion but write separately to highlight one aspect of the court‘s reasoning that I cannot fully embrace. The court holds that, even though we are reversing in part the district court‘s denial of Lustig‘s motion to suppress, he‘s not automatically entitled to withdraw his guilty plea. Instead, the court concludes that we must engage in “harmless error review” to determine whether the district court‘s partially erroneous denial of the motion “contributed to [Lustig‘s] decision to plead guilty.” Maj. op. at 1087, 1090.

I do not think Federal Rule of Criminal Procedure 11(a)(2)—or basic principles of contract law, which govern plea agreements—permit any such inquiry. In my view, if our court does anything other than affirm in full the district court‘s denial of Lustig‘s suppression motion, he is entitled to withdraw his guilty plea without more. The harmless error analysis the court engages in has no place in this context.

That conclusion is dictated by the plain language of Rule 11(a)(2), a short, two-sentence provision added in 1983. The first sentence authorizes a new type of guilty plea—the “conditional” plea—that had not previously been sanctioned by rule or statute: “With the consent of the court and the government, a defendant may enter a conditional plea of guilty or nolo contendere, reserving in writing the right to have an appellate court review an adverse de-termination of a specified pretrial motion.” Adding this provision was necessary because a number of courts had held, prior to the Rule‘s amendment in 1983, that conditional pleas were not permissible. A defendant either had to plead guilty unconditionally and waive appellate review of adverse pre-trial rulings, or, if the defendant sought to preserve such review, he had to plead not guilty and proceed through trial, an often wasteful exercise with a foregone conclusion. See Fed. R. Crim. P. 11 advisory committee‘s note to 1983 amendments. Rule 11(a)(2) avoids that undesirable state of affairs by allowing a defendant, with the government‘s and the court‘s consent, to plead guilty on the condition that a specified adverse ruling is ultimately affirmed on appeal. , overruled on other grounds by . The second sentence of the Rule states what happens if the condition does not hold: “A defendant who prevails on appeal may then withdraw the plea.”

In a case in which the defendant reserves the right to challenge a single adverse ruling and that ruling ultimately gets reversed in full on appeal, the application of Rule 11(a)(2)‘s second sentence is simple. The defendant has obviously “prevail[ed] on appeal” and as a result must be afforded an opportunity to withdraw his plea. . I think the application of the Rule is just as simple when, as in this case, the defendant prevails in part on appeal. As we said in Carrasco, a Rule 11(a)(2) plea is “conditioned on the appellate court‘s affirmance of the adverse pretrial ruling.” . If the appellate court does anything other than affirm the specified ruling (or rulings) in full, then the condition is not satisfied. That means, under basic contract law principles, that the defendant is entitled to withdraw from his end of the bargain. See . There is no place for an appellate court to decide that the partial victory the defendant won on appeal is too insignificant to warrant the defendant‘s backing out of the deal. See .

Here, Lustig agreed to plead guilty on the condition that the ruling denying his motion to suppress would be affirmed on appeal. It didn‘t get affirmed; it got reversed in part with respect to the car phones. Thus, the one condition Lustig placed on his agreement to plead guilty wasn‘t satisfied, and only he gets to decide whether the partial victory he won on appeal is too inconsequential to justify backing out of the deal.

The parties, of course, could have struck a different deal. Nothing in the language of Rule 11(a)(2) precludes a defendant and the government from agreeing that the defendant‘s guilty plea will stand unless he wins reversal in full of a particular adverse ruling. Or, in cases in which the defendant challenges several distinct adverse rulings, that his guilty plea will stand unless he wins reversal of all of them. That‘s why the drafters of Rule 11(a)(2) inserted the requirement that a conditional plea may be entered only with the government‘s consent—to ensure that the government could not insist upon reserving the right to appeal some inconsequential pre-trial ruling, the reversal of which would not have any appreciable impact on the outcome of the case. See Fed. R. Crim. P. 11 advisory committee‘s note to 1983 amendments (“As for consent by the government, it will ensure that conditional pleas will be allowed only when the decision of the court of appeals will dispose of the case either by allowing the plea to stand or by such action as compelling dismissal of the indict-ment or suppressing essential evidence.“). But the key point here is that if the parties do not qualify the condition that a particular adverse ruling must be affirmed on appeal in order for the plea to stand, then a defendant who wins even partial reversal will be entitled under the terms of the agreement to withdraw his plea. See .

There is a place for harmless error review in the context of conditional pleas, but it differs from the kind of harmless error review the court engages in here. Appellate courts always have the authority to determine that, even though the district court‘s reasoning was flawed in some respect, the district court‘s bottom-line ruling is nonetheless correct and should be affirmed. Or, in like fashion, that the district court‘s ruling on a subsidiary issue was erroneous, but that the court‘s bottom-line decision to deny a suppression motion is still correct, albeit for reasons that differ from those given by the district court. See, e.g., . In those circumstances we say the district court‘s errors are “harmless” in the sense that they do not affect the ultimate disposition of the appeal—the district court‘s bottom-line ruling still gets affirmed.

That kind of harmless error review is perfectly proper in the context of Rule 11(a)(2) pleas. See . It allows the court to determine, as the court did in Davis, that the defendant ultimately won no victory on appeal—not even a partial one—and thus that he cannot be said, in the language of Rule 11(a)(2), to have “prevail[ed] on appeal.” In such cases, the court uses harmless error review to affirm in full the ruling that the defendant reserved the right to challenge on appeal. See, e.g., . The defendants in cases like Davis are not entitled to withdraw their conditional guilty pleas because the condition attached to their pleas is satisfied. This case has to come out differently because the condition attached to Lustig‘s plea was not satisfied. We did not affirm in full the district court‘s ruling on Lustig‘s motion to suppress.

In short, I agree with the court that Lustig‘s convictions must be vacated, and on remand he must be afforded an opportunity to withdraw his guilty plea. In my view, though, that result follows from an application of the plain language of Rule 11(a)(2) and basic contract law principles, not from an application of harmless error review.

Notes

1
In a concurrently filed memorandum disposition, we address and reject several secondary arguments Lustig raises in his briefing.
2
Lustig conceded that the Pocket Phones were properly seized incident to arrest.
3
Lustig was initially arrested for soliciting prostitution in violation of California Penal Code § 647(b). The state charge against Lustig was eventually dismissed.
4
Specifically, the defendant relied on (holding that attaching a GPS device to a car constituted a Fourth Amendment search), and (holding that a dog-sniff conducted in the curtilage of the defendant‘s home was a Fourth Amendment search).
5
The Fifth Circuit—before Riley and before the Pocket Phone searches at issue here—similarly understood Robinson to authorize searches of cell phones incident to arrest. See (holding that, under Robinson, a valid custodial arrest permits a warrantless search of an individual‘s cell phone, including its call records and text messages). In , also decided before the Pocket Phone searches here, the Seventh Circuit likewise held that looking in a cell phone for the cell phone‘s number did not exceed what Robinson allows. Lustig is correct that the Seventh Circuit went on to discuss the unique features of cell phones, but it explicitly left “for another day” the constitutionality of a “more extensive search of a cell phone without a warrant.” Id. The First Circuit eventually held that a search incident to arrest does not authorize the warrantless search of data on a cell phone seized from an arrestee‘s person, but it did so after the searches at issue here. See , aff‘d sub nom. .
6
A sufficient body of district court or state appellate court decisions could perhaps create enough uncertainty about the scope of prior appellate precedent to make it unreasonable to rely on that precedent. See , 131 S.Ct. 2419 (Sotomayor, J., concurring in the judgment) (arguing that when the “law in the area” is “unsettled,” law enforcement officials should “err on the side of constitutional behavior“) (quoting ). We need not determine here whether that is so or precisely what would be required to create enough uncertainty because it is clear that, in light of Robinson‘s seemingly broad and categorical holding, the handful of decisions that Lustig cites were not enough to make reliance on Robinson unreasonable.
7
Lustig also challenges the resumption of that initial search four days later, which we address below.
8
Because in Lara the government had not sought application of the good-faith exception in the district court, we held that the argument had not been preserved on appeal. . We nevertheless proceeded to explain that we would have rejected the argument on the merits even if not waived. Id.
9
Lustig also argues that because Riley affirmed the First Circuit‘s decision in Wurie, which rejected the government‘s good-faith exception arguments, Riley must have done so as well. But Wurie concluded that the government had waived the good-faith exception, not that the exception was inapplicable on the merits. See (holding that because the government “did not invoke the exception before the district court,” it “entirely failed to carry [its] burden“).
10
The Government argues that the California Supreme Court‘s decision in , supports the conclusion that Chiappino could reasonably believe that Robinson authorized the Pocket Phone searches. Diaz held that, under Robinson, searches of cell phones discovered directly from an arrestee‘s person comported with the Fourth Amendment. . Lustig responds that Diaz is irrelevant because it is not binding federal appellate authority, and the searches of his phones were conducted by officers cross-designated as federal agents. Because we hold that Robinson provides the applicable binding appellate authority creating a reasonable basis for the Pocket Phone searches here, and
11
This version of the Sullivan opinion cited by the Government was subsequently withdrawn and superseded by a revised opinion. See . The relevant portion remained substantively unchanged.
12
Federal Rule of Criminal Procedure 11(a)(2) provides: “With the consent of the court and the government, a defendant may enter a conditional plea of guilty or nolo contendere, reserving in writing the right to have an appellate court review an adverse determination of a specified pretrial motion. A defendant who prevails on appeal may then withdraw the plea.”
13
Peyton and Leake framed the issue as whether the defendant had “prevail[ed] on appeal” for purposes of Rule 11(a)(2), rather than whether the district court error was “harmless.” See ; . However framed, the ultimate question is the same: when is a defendant entitled to withdraw his plea due to the district court‘s error? If an error is deemed harmless, then the defendant will not have “prevail[ed] on appeal,” and vice versa.
14
In the Tenth Circuit‘s formulation, which we adopt here, concluding that there is a “reasonable possibility” that the error contributed to the plea decision is the opposite of concluding “beyond a reasonable doubt that the error did not contribute” to the plea decision. (emphasis added). In other words, an error will be harmless for Rule 11(a)(2) purposes if an appellate court can conclude beyond a reasonable doubt that the error did not contribute to the defendant‘s decision to plead guilty, but will not be harmless if there is a reasonable possibility that the error did contribute to the decision to plead guilty.
15
It is unclear to what extent, if any, the First Circuit intended to adopt a different standard than that articulated in Benard, given that it relied in Molina-Gomez on the same authority as Benard to establish an appellate court‘s limited role in determining harmless error under Rule 11(a)(2). See (quoting , and noting that Weber “adopt[ed] the rationale of the Seventh Circuit and numerous state courts,” namely , , and , all of which Benard also relied upon).
16
The Government also relies on , to argue for a harmlessness standard that looks solely at the relationship between the evidence in question and the charges of conviction. But nowhere in Richard Davis did we discuss the import of the suppression error on either the defendant‘s ultimate conviction or his decision to plead guilty.
17
Although the district court indicated that Lustig‘s motion to reconsider the Car Phone suppression ruling was “moot” due to the Government‘s self-suppression of the Car Phone evidence, the Government never actually stated that it would refrain from using the Car Phone evidence to prosecute its case. Instead, it stated that “to some extent we don‘t intend to use the evidence from the cell phones seized in the car.” This is a far cry from disavowing the Car Phone evidence altogether.
18
On remand, before Lustig is required to make a decision on whether to vacate his plea, Lustig should be given an opportunity to renew his motion to exclude any fruit of the Car Phone searches. See (“Because the question of taint was not fully explored below, we must remand for resolution of the remaining factual questions.“).
Read the detailed case summary