In the
United States Court of Appeals For the Seventh Circuit
____________________
No. 18-3157 UNITED STATES OF AMERICA, Plaintiff-Appellee,
v.
ANTHONY HOWELL, Defendant-Appellant.
____________________
Appeal from the United States District Court for the Northern District of Illinois, Eastern Division. No. 1:13-cr-250 — Gary Feinerman, Judge.
____________________
ARGUED FEBRUARY 27, 2020 — DECIDED MAY 4, 2020
____________________
Before BRENNAN, SCUDDER, and ST. EVE, Circuit Judges. SCUDDER, Circuit Judge. On an afternoon in December 2012, the Chicago Police Department received an anonymous 911 call reporting a Hispanic man in a black sweater and black hat, carrying a bag, and climbing under a warehouse fence. Officers arrived and found someone who matched the description , but after stopping and frisking him, determined he was not engaged in any crime. The initial suspect then pointed the officers to someone else nearby who was crossing 2 No. 18-3157
the street and walking toward the police. This man, Anthony Howell, was white and wearing a black jacket and dark hat. When an officer approached to ask what was going on, Howell did not answer, looked panicked, and put his hands in his pockets. The officer reacted by patting down Howell and found a gun in his jacket. A federal gun charge followed, and Howell moved to suppress the gun as the fruit of an unconstitutional stop-and-frisk. The district court denied the motion , Howell proceeded to trial, and a jury found him guilty.
Howell now appeals from the denial of the suppression motion. In evaluating his position, we also confront a question about the proper scope of the record on review. The question is whether we limit our review to the pretrial record or expand our look to consider the arresting officer’s trial testimony as well. The answer matters because the facts in the pretrial record differed in a material way from those that emerged at trial, where the arresting officer testified that he decided to proceed with the pat down only after Howell ignored a directive to remove his hands from his pockets. In the end, we limit ourselves to the pretrial record, for that is the only source of facts the district court considered in denying Howell’s motion. Viewing that record as a whole, we conclude that police lacked reasonable suspicion to frisk Howell. We therefore reverse the denial of his suppression motion and vacate his conviction for possessing that gun.
Our reversal is only partial, however, because Howell was also convicted on a second gun charge. Three months after the December 2012 stop-and-frisk, police executed a warrant to search Howell’s apartment, where they found more guns and ammunition. There was ample evidence for the
No. 18-3157 3
jury to find that Howell possessed the guns in his apartment, so we affirm his conviction for this separate offense.
I
A
Around noon on December 4, 2012, an anonymous 911 caller reported that a Hispanic man wearing a black sweater, black hat, and black bag was climbing under a fence at a warehouse on South Artesian Avenue in Chicago’s Brighton Park neighborhood. Officers Sean Kelly and Christopher Miller arrived about five minutes later and saw a man, Eric Escobar, who matched the caller’s description and was walking on the sidewalk outside the warehouse. The officers stopped Escobar and immediately patted him down but found nothing suspicious . Escobar explained that he worked at the warehouse and had stepped outside to buy a drink and a snack for his manager. The manager emerged from the building and confirmed that account while also verifying Escobar’s identity.
While talking to police, Escobar noticed another person nearby—a white man wearing a black jacket and a dark hat who was walking toward the officers. He was later identified as Anthony Howell, who lived across the street. Upon first noticing Howell, Escobar remarked that he seemed to match the police’s account of the 911 caller’s description. Officer Kelly reacted to Escobar’s comment by calling out to Howell from across the street and asking, “What’s going on?” According to Kelly, Howell refused to answer and instead did a “quick double take,” had “a look of panic on his face,” and placed his hands in his pockets. Finding this reaction suspicious , Kelly approached Howell and immediately frisked him for weapons.
4 No. 18-3157
As soon as Officer Kelly began the frisk, he felt a hard object in Howell’s jacket pocket. When asked what it was, Howell replied, “protection.” When Kelly tried to retrieve the gun, Howell pulled away, started to run, but quickly slipped on gravel and fell. At some point in the ensuing scuffle, a .38 caliber Smith & Wesson revolver fell out of Howell’s pocket, and the police secured it and placed Howell under arrest.
B
A federal grand jury later charged Howell with unlawfully possessing a gun as a prior convicted felon, a violation of 18 U.S.C. § 922(g)(1). Before trial Howell moved to suppress the gun, arguing that the police violated his Fourth Amendment rights by stopping and frisking him without reasonably suspecting him of being engaged in criminal activity. Howell also sought an evidentiary hearing on the motion.
The district court denied both requests in an oral ruling. It first denied Howell’s request for a hearing on the ground that he had not shown a material factual dispute. While Howell’s brief in support of his motion contested the officers’ version of events—for example, he denied refusing to answer Officer Kelly’s question—the court emphasized that Howell had stopped short of submitting an affidavit swearing under oath to the same representations. Without such an affidavit, the district court reasoned, Howell failed to create a genuine factual dispute that warranted a hearing.
From there the district court relied on police paperwork and FBI reports of interviews with Officer Kelly and the other officers involved in the stop-and-frisk to rule on the merits of Howell’s Fourth Amendment challenge. Applying the familiar reasonable suspicion standard from Terry v. Ohio, 392 U.S.
No. 18-3157 5
1 (1968), the district court considered the totality of the circumstances and began by observing that Howell came close enough to matching the 911 caller’s description to authorize the stop—while not Hispanic, he was wearing a black jacket and a dark hat. The district court also emphasized that Howell reacted to Officer Kelly’s question about what was going on by refusing to answer, doing a double take, looking panicked , and putting his hands in his pockets. The combination of these reactions and circumstances, the court concluded, not only supplied the reasonable suspicion necessary to support Officer Kelly’s stop of Howell, but also suggested that he may have been concealing something—thereby authorizing the pat down.
Howell proceeded to a jury trial, where he renewed his motion to suppress at the close of evidence. The district court made quick work of the renewed motion, observing that the matter had been fully resolved pretrial. The court therefore denied the motion “for the reasons that [were] already given in addressing the defendant’s prior motions on the same topic.” Neither party said a word about any aspect of the trial evidence affecting or informing the court’s prior ruling.
The jury returned a guilty verdict. Howell then moved for a new trial or a judgment of acquittal, again arguing that the district court should have granted his prior motion to suppress . The district court construed Howell’s argument as renewing the motion to suppress for a third time. And the court reacted by referring to its pretrial ruling—reiterating that it had “already ruled on [the] Fourth Amendment issue”—and “incorporate[d] by reference everything that [it had] already said.” Throughout this colloquy nobody referred to the trial evidence.
6 No. 18-3157
Howell now appeals.
II
A
Before reaching the merits, we confront an interesting and challenging question about how to define the scope of the record on appeal when reviewing a motion to suppress. While preparing for oral argument, we noticed a material difference between the facts relevant to the suppression motion in the pretrial record and those elicited at trial. Specifically, at trial, Officer Kelly testified that he had “asked [Howell] to take his hands out of his pocket, and [Howell] didn’t respond to that, either.” But the pretrial record on which the district court based its denial of Howell’s motion made no mention of Officer Kelly giving such a direction or Howell then ignoring it. More to it, the police paperwork and FBI reports (prepared in anticipation of federal charges being brought against Howell) only referenced Howell putting his hands in his pockets.
The difference may be significant. If true, the additional fact that Howell disobeyed a police order to remove his hands from his pockets would have bolstered suspicions that he was armed and dangerous.
It is less clear, however, that we should consider trial testimony . The district court did not. It denied Howell’s motion to suppress (and post-trial renewal of the motion) entirely—100 percent—on the pretrial record (the police and FBI paperwork ), giving no effect whatsoever to any aspect of Officer Kelly’s trial testimony.
It warrants emphasis that the district court acted well within its discretion in handling Howell’s motions in this way. When a defendant chooses to renew a suppression
No. 18-3157 7
motion at or after trial, a district court is free to incorporate its past reasoning, as it did here, or alternatively to consider evidence introduced at trial. Unless the parties bring new evidence gleaned from trial to the court’s attention, the law does not compel either approach, and we leave the choice to the district court. Relatedly, the path the district court chooses may inform our own discretion as to which facts warrant consideration on appeal. See United States v. Hicks, 978 F.2d 722, 724–25 (D.C. Cir. 1992) (employing similar reasoning).
Recognizing the importance of these questions and considerations to our analysis, we asked the parties for supplemental briefs addressing the proper scope of the record.
B
Ordinarily we define the record on appeal by limiting the facts we review to those considered by the district court. Defining the appellate record in this way ensures that we perform our role as a court of review while also respecting the district court’s role as factfinder. See Midwest Fence Corp. v. U.S. Dep’t of Transp., 840 F.3d 932, 946 (7th Cir. 2016) (“As a general rule, we will not consider evidence on appeal that was not before the district court when it rendered its decision.”); see also WRIGHT & MILLER, FEDERAL PRACTICE & PROCEDURE § 3956.1 (5th ed.) (noting that “[t]he court of appeals’ concern with the scope of the record stems in part from the notion of the respective roles of the appellate and trial courts,” especially “the district court’s role as the first-instance finder of fact”).
But here the government presses a different approach, inviting us to consider both pretrial and trial evidence. No other appeal has required us to confront this question in much 8 No. 18-3157
depth, and what case law does exist sends differing signals. The government’s position finds support in cases like United States v. Parra, where we affirmed the denial of a defendant’s pretrial motion to suppress cocaine found in a search incident to arrest. 402 F.3d 752, 767 (7th Cir. 2005). In identifying the standard of review, we observed—without saying more— that “[i]n reviewing a denial of a suppression motion, we may consider evidence introduced both at the suppression hearing and at trial.” Id. at 764; see also United States v. Duguay, 93 F.3d 346, 350 (7th Cir. 1996) (stating without further elaboration that we may consider evidence presented at trial in reviewing a district court’s pretrial ruling denying a motion to suppress ).
On the other side of the ledger come cases like United States v. Smith, 80 F.3d 215 (7th Cir. 1996). The district court there denied the defendant’s pretrial motion to suppress marijuana found in his car during a traffic stop. See id. at 218–19. On appeal the defendant pointed to evidence elicited at trial to argue that no reasonable officer would have made the stop. See id. at 220. But we declined to consider the trial evidence, instead basing our review “solely on what the district court knew at the time of the ruling.” Id.; see also United States v. Fryer, 974 F.2d 813, 819 (7th Cir. 1992) (limiting our review of a suppression ruling to the pretrial record in the same way).
To be sure, even decisions like Parra and Smith, while limiting review to the pretrial record, say very little (if anything) about why that approach is proper. Perhaps our most fulsome treatment of the issue came in United States v. Longmire, 761 F.2d 411 (7th Cir. 1985). There we took guidance from the Supreme Court’s 1925 decision in Carroll v. United States, where the Court relied in part on trial evidence to affirm the denial
No. 18-3157 9
of a suppression motion. See 267 U.S. 132, 162 (1925). Following suit in Longmire, we observed that we had the discretion to consider trial evidence in reviewing a pretrial suppression ruling. See 761 F.2d at 418 (“[E]vidence adduced only at trial may be used to sustain the denial of a motion to suppress.”).
We took care, however, to sound caution in the exercise of that discretion, recognizing that consideration of the trial testimony presents both benefits and risks. See id. On the one hand, using that testimony “avoids a windfall reversal of the defendant’s conviction” where the trial record reveals that a police action was constitutional. Id. But just as importantly, “several problems may be presented by such use of trial testimony ,” including “[p]rejudice to the accused.” Id. A balance must be struck between these two interests—avoiding a windfall reversal of a conviction while also steering clear of unfair prejudice to the defendant.
Our decision in Longmire offered guidance for achieving that balance. In the ordinary course, we may consider trial testimony in reviewing a pretrial suppression ruling. See id. We opted to follow that baseline rule on the facts of Longmire because Darlene Longmire did not contest the trial testimony, let alone attempt to show that it prejudiced her. See id. at 420– 21 (“Longmire apparently believed that [the trial] testimony did not alter the correctness of the pretrial suppression ruling for she failed to ask the trial court to reconsider that ruling.”). Still, we recognized that consideration of the trial testimony would be inappropriate where a defendant shows that doing so would result in prejudice—for example, where “the credibility and veracity of a relevant government witness have been put into question by defense counsel” and that witness introduces new facts at trial. Id. at 418.
10 No. 18-3157
Considered collectively our prior cases show that the question presented—when we may consider trial evidence in reviewing a pretrial motion to suppress—does not lend itself to bright-line answers. Rather, we approach the inquiry on a case-by-case basis, taking account of all available information regarding the proceedings below. We read our case law (and the principles underpinning it) to at least establish that we retain the discretion to consider trial evidence bearing on a district court’s ruling on a motion to suppress where that evidence came into play in the district court’s consideration of the motion—where the defendant renewed the motion and thereby invited the district court to reevaluate its prior ruling in light of trial evidence or where the district court undertook such a reevaluation of its own accord. See Smith, 80 F.3d at 220 (limiting review to the pretrial record in part because the defendants did not renew their suppression motions at trial); see also Hicks, 978 F.2d at 725 (observing that when trial evidence casts doubt on a pretrial suppression ruling, the parties should “bring alleged errors to the trial court’s attention by making a proper objection or filing a motion”). The prior cases likewise counsel that one factor properly informing our exercise of discretion is whether considering the trial evidence would cause unfair prejudice to the defendant.
On balance we conclude that these principles tilt against consideration of the trial record here. Foremost, the district court itself never considered Officer Kelly’s trial testimony— neither in denying Howell’s pretrial motion, nor in denying the later renewals of the same motion. To the contrary, the district court made plain that it was denying the renewed motion for the same reasons given in the pretrial ruling. The district court never hinted that the trial evidence was even
No. 18-3157 11
relevant, much less that it in any way affected any dimension of the court’s prior reasoning.
So, too, is it clear that it would prejudice Howell to consider Officer Kelly’s trial testimony, as it contained a new, material representation—that Howell disregarded a clear direction from Officer Kelly to remove his hands from his pockets. At trial Howell had no reason to believe the district court would consider that testimony as part of revisiting its pretrial ruling. Even more, Howell may have had sound strategic reasons , when cross-examining and attempting to impeach Officer Kelly, not to draw attention to the new fact offered for the first time at trial. See Longmire, 761 F.2d at 418 (emphasizing this precise caution). Howell’s focus at trial was not on somehow seeking to relitigate the motion to suppress before the jury. That Howell renewed the motion during and after trial almost certainly reflected nothing more than an effort to ensure preservation of the issue for appellate review.
The upshot is that Howell had little incentive at trial to focus on factual details pertinent to a pretrial motion that the district court resolved before trial even began. Put another way, Howell had every reason to believe the trial would be all and only about whether the government presented evidence to prove beyond a reasonable doubt that he possessed a gun following a prior felony conviction.
We find it equally noteworthy that neither party’s briefing on appeal even identified the difference between the pretrial and trial records. Both parties addressed the district court’s denial of Howell’s motion to suppress by focusing strictly on the pretrial record. Only after we raised the question at oral argument and requested supplemental briefing did the 12 No. 18-3157
government seek to defend the district court’s ruling by relying on Officer Kelly’s trial testimony.
The law does not compel us to consider trial evidence in reviewing a suppression ruling; it merely affords us the discretion to do so. See Parra, 402 F.3d at 764; Longmire, 761 F.2d at 418. Under these circumstances, we decline to consider Officer Kelly’s trial testimony. Doing so would prejudice Howell in a material and unfair way. We therefore look only to the pretrial record in evaluating Howell’s motion to suppress.
III
A
A seizure occurs within the meaning of the Fourth Amendment if, in the totality of the circumstances, a reasonable person would not feel free to disregard the police and move along. See Florida v. Bostick, 501 U.S. 429, 434 (1991). Here the parties agree that at some point during Officer Kelly’s approach and questioning, Howell was seized. But they dispute whether the seizure was constitutional.
Under the Fourth Amendment, police may stop a person only if they have reasonable suspicion that he is engaged in criminal activity. See Terry, 392 U.S. at 21–22. Our focus on reasonableness “balanc[es] the need to search (or seize) against the invasion which the search (or seizure) entails.” Id. at 21. The inquiry is fact-intensive: we look to the totality of the circumstances to see whether police “ha[d] a particularized and objective basis for suspecting the particular person stopped of criminal activity.” United States v. Cortez, 449 U.S. 411, 417–18 (1981); see also United States v. Street, 917 F.3d 586, 593 (7th Cir. 2019) (explaining that reasonable suspicion must
No. 18-3157 13
be based on specific, articulable facts that would justify an intrusion on the suspect’s liberty and dignity).
The district court determined that Officer Kelly’s decision to stop Howell respected the Fourth Amendment. We agree. The police were responding to a 911 call reporting suspicious activity across the street from where Howell was walking. He roughly matched the caller’s description. Howell then reacted to that approach in a way Officer Kelly reasonably could have found suspicious—by doing a double take, taking on a panicked look, and refusing to respond. These facts and circumstances combined to give Officer Kelly sufficient reason to approach Howell, and in posing a question to him, to conduct an investigatory stop. We move, then, to the frisk.
B
A frisk—a limited pat down of the suspect’s outer clothing to search for weapons—is permissible under the Fourth Amendment only if a police officer can “point to specific and articulable facts” indicating “that criminal activity may be afoot and that the persons with whom he is dealing may be armed and presently dangerous.” Terry, 392 U.S. at 21, 24–25, 30. It is precisely because a frisk is more intrusive than a stop that the Fourth Amendment compels this additional armedand -dangerous inquiry. See id. at 27.
We begin where the police did—with the information from the 911 call. In assessing a stop-and-frisk based on a tip, we must assess the reliability of the information conveyed by the caller. See United States v. Lopez, 907 F.3d 472, 479 (7th Cir. 2018). Sometimes callers identify themselves and their doing so lends meaningful credibility to the information they provide . Other times callers remain anonymous and reliability 14 No. 18-3157
comes from independent sources corroborating the tipster’s account. See id. at 480. Along these lines, we have identified “a spectrum of knowledge and reliability that affects the reasonableness of police action taken pursuant to the tip”—at one end, “a tip from a known, trusted, and reliable source,” and at the other, “an anonymous tip without signs of reliability .” Id. at 479–80. “Tips that come from more trustworthy sources will require less independent corroboration than those obtained from more questionable sources.” Id. at 480.
The call here was anonymous. The Supreme Court has long recognized that “[a]n anonymous tip alone seldom demonstrates the informant’s basis of knowledge or veracity” because that basis is “by hypothesis largely unknown, and unknowable.” Alabama v. White, 496 U.S. 325, 329 (1990) (internal quotations omitted). Additionally, the tip itself contained no further indicia of the informant’s reliability. It also offered nothing but a barebones description of the suspect: the caller identified the suspect’s race (Hispanic), sex (male), and dress (black sweater, hat, and bag). What limited details the caller did supply fell short of describing Howell with significant accuracy. Above all else, the call mentioned a bag, which Howell did not have, and described Howell, who is white, as appearing Hispanic.
To justify a frisk of Howell, then, police needed some additional indicia of reliability or another source of corroboration beyond the limited information provided by an anonymous caller. But there was none. There were no additional calls to police, for example, or bystanders at the scene who said they had witnessed anything troubling. Remember that the police, upon arriving at the warehouse, spoke to both Eric Escobar and his manager—both of whom worked there and
No. 18-3157 15
neither of whom expressed any concern about an attempted burglary or other crime. Nor did police arrive to find any evidence of illegal activity themselves.
Consider, too, the nature of the reported offense. A call to police is less likely to support reasonable suspicion in the Terry analysis when it does not describe an ongoing crime or emergency. That precept follows directly from Terry itself. See 392 U.S. at 30 (holding that a frisk is justified where there is reasonable suspicion that both the suspect “may be armed and presently dangerous” and “criminal activity may be afoot”); see also Lopez, 907 F.3d at 485 (“[I]nvestigative stops related to completed crimes must be distinguished from investigative stops related to ongoing or imminent crimes.”). We recently emphasized much the same point in United States v. Watson, 900 F.3d 892 (7th Cir. 2018). There police received an anonymous call reporting that “boys” were “playing with guns” by a “gray and greenish Charger” in a parking lot. Id. at 893. Police went to the parking lot, saw a car matching the description, searched it, and found a gun. Id. at 894. We held that the police lacked reasonable suspicion to justify that search, however, because the caller was anonymous and “did not describe a likely emergency or crime.” Id. at 893.
In the same vein, the anonymous tip here merely reported someone climbing a warehouse fence. Nothing about it suggested that an emergency was underway or that anybody was in imminent danger. Not a word was said about weapons, an injured victim, or anyone being threatened. The alleged offense took place around noon—in broad daylight—and the record is devoid of any evidence that it took place in a high- crime area. By any reasonable measure, the 911 caller described a low-end, nonviolent offense—something that surely 16 No. 18-3157
warranted a police response, but by no means could be considered an emergency. See United States v. Goodwin, 449 F.3d 766, 769 (7th Cir. 2006) (interpreting the Supreme Court’s Fourth Amendment jurisprudence to apply a sliding scale approach where “the amount of permissible intrusion is a function . . . of the gravity of the crime being investigated”); see also United States v. Williams, 731 F.3d 678, 686 (7th Cir. 2013) (explaining that a frisk is a “more burdensome intrusion” on the suspect’s liberty and dignity than a stop alone).
The government sees things differently. It contends that although the district court characterized the offense as a mere “trespass,” a reasonable police officer could have interpreted the caller to be reporting a felony burglary. And if the reported offense was a felony, the government continues, it was serious enough to justify a frisk, even though the crime was no longer underway by the time police arrived. See United States v. Hensley, 469 U.S. 221, 233–34 (1985) (holding that police had reasonable suspicion to stop a suspect in a felony armed robbery, even though the crime was already complete and therefore no longer in progress).
The government did not raise this argument in the district court, and we are reluctant to entertain a position that Howell had no opportunity to contest. See In re Veluchamy, 879 F.3d 808, 821 (7th Cir. 2018). Regardless, we need not decide whether the conduct described by the 911 caller would constitute a misdemeanor or a felony to assess its seriousness. Nor do we expect police to categorize the tips they receive into misdemeanors or felonies before assessing the appropriate response to each situation. Sometimes the answer would be easy (shots fired and victim screaming); other times it may be next to impossible (suspicious person wearing stocking
No. 18-3157 17
cap, yelling, and parked in front of neighbor’s house). Requiring the police to place the calls they receive into a felony, misdemeanor , or some other bucket strikes us as unworkable. See United States v. Jones, 953 F.3d 433, 436–37 (6th Cir. 2020) (explaining that the federal courts have avoided a brightline rule for when police may investigate a completed nonfelony, in part because of “the elusive and evolving nature of the felonymisdemeanor distinction”). The better approach is to consider the totality of the information supplied by the 911 caller, including the nature of the reported crime, the reasonable inferences following from the caller’s information, and how the police responded.
Here what matters perhaps most is that the 911 call in no way suggested that the suspect was armed or dangerous. The caller did not so much as hint at violence, injuries, or weapons . Nor did such a threat arise after police responded to the call. The officers saw no crime in progress and encountered no victim or witnesses. To the contrary, they arrived to find a rather innocuous scene: Eric Escobar walking on the sidewalk outside a warehouse in broad daylight, on his way to grab snacks for his manager. Indeed, it was Escobar—not any officer —who saw Howell and suggested that perhaps he was the one the police were looking for. Then and only then did Officer Kelly turn his attention to Howell—who presented only as white, not Hispanic as the 911 caller described, and who was not carrying a bag of any kind. And even then, all that Officer Kelly reported seeing from across the street was Howell appear nervous and panicked, fall silent, and put his hands in his pockets. There were no suggestions that Howell was armed—nobody claims he had a bulge in his pocket or made any move to hide anything.
18 No. 18-3157
Finally, the government urges us to place substantial weight on Howell’s panicked look upon seeing police. We agree that this nervousness is relevant, but it must be considered against the full context of the circumstances facing the police. Nervousness alone, at least not as a categorical matter, does not create reasonable suspicion that a suspect is armed and dangerous. See Williams, 731 F.3d at 687 (recognizing that “[m]ost people, when confronted by a police officer, are likely to act nervous, avoid eye contact, and even potentially shift their bodies as if to move away from the area”). Nor does a suspect’s mere refusal to answer an officer’s questions, without more, create reasonable suspicion. See Illinois v. Wardlow, 528 U.S. 119, 125 (2000) (“[A]ny refusal to cooperate, without more, does not furnish the minimal level of objective justification needed for a detention or seizure.”).
Nervousness is more salient to the reasonable determination calculus when it accompanies other suspicious behavior or circumstances suggesting a risk to officer safety. In Wardlow , for example, the Supreme Court held that police had reasonable suspicion to stop and frisk a person who, while carrying a bag and upon making eye contact with the officers, suddenly started running away. See id. at 122. The suspect’s “[h]eadlong flight,” the Court concluded, was “the consummate act of evasion,” made all the more suspicious by the fact that his spontaneous bolt occurred in a high-crime area. Id. at 124.
Our case law follows a similar pattern. In United States v.
Brown, we concluded that police had reasonable suspicion to frisk a suspect who was stopped for speeding and then asked to exit the vehicle for a pat down. 188 F.3d 860, 865 (7th Cir. 1999). Not only did the suspect show “excessive nervousness”
No. 18-3157 19
when police stopped him, but several other facts contributed to reasonable suspicion: his car was under FBI surveillance for possible involvement in a large-scale drug operation, it reeked of marijuana, and he was stopped “in a high crime area where there had been drug activity, shootings, and gang violence.” Id.
More recently, in United States v. Adair, we held that police had reasonable suspicion to stop and frisk a suspect who tried to evade an officer by weaving through a crowd away from him. 925 F.3d 931, 933 (7th Cir. 2019). Critical to our conclusion , however, were the circumstances in which that happened : a woman had reported to police that a group of people she did not recognize were standing outside her apartment smoking, drinking, and engaged in “very suspicious activity .” Id. She described the suspect in specific terms and stated that he had a black gun in his front pocket. See id. When police arrived, an officer found a matching suspect with a bulge in his pocket who was trying to evade detection. Id. at 933, 937. This all transpired “late at night in a high-crime area.” Id. at 936. Considering these facts in combination, we held that police reasonably suspected that the suspect was armed and dangerous. See id. at 936–37.
Howell’s case presents far different facts and circumstances . His panicked look and silence in response to Officer Kelly’s question were not accompanied by any attempt to flee or any furtive movement. He responded to seeing Officer Kelly as many might, by appearing to want to move along and avoid a discussion with the police. What most concerns us is how Officer Kelly reacted. He did not respond by continuing to approach Howell or allowing more time for further questions —“Where do you live?”, “Do you know anything about 20 No. 18-3157
a burglary here?”, “Were you trying to climb under this fence?” and the like—but instead by immediately commencing a pat down. Put another way, Officer Kelly reacted to seeing Howell much like he did to observing Eric Escobar upon arriving at the scene—by seeing that he matched aspects of the caller’s description then instantly patting him down. But Terry teaches that frisks need to account for the totality of circumstances —they cannot be rote or reflexive—and here the circumstances required more before Officer Kelly’s encounter with Howell would permit a frisk.
The caution the Supreme Court sounded in Florida v. J.L., 529 U.S. 266 (2000), warrants underscoring. There an anonymous 911 caller reported that a young, black male wearing a plaid shirt at a bus stop was carrying a concealed gun. See id. at 271. Upon responding to the bus stop, the police saw someone who perfectly matched that description, patted him down, and discovered a gun. See id. The Court held the frisk unconstitutional, because the tip was anonymous and described only “a subject’s readily observable location and appearance .” Id. at 272. The Court explained that such a description is reliable only in that it helps police to identify the accused . See id. Reasonable suspicion required more: the tip needed to be “reliable in its assertion of illegality, not just in its tendency to identify a determinate person.” Id.
These same concerns weigh on us here. The Court found that the tip in J.L. was too barebones to support a frisk because it identified only “readily observable” traits such as race, sex, clothing, location, and age. Id. Here the caller provided even less information, all of which was readily observable: race, sex, clothing, and location. And unlike in J.L., the match was not exact. If the tip in J.L. did not create reasonable suspicion
No. 18-3157 21
to support a frisk, we find it hard to reach another conclusion here, especially considering the nature of the reported crime—someone climbing under a warehouse fence in broad daylight—and what the police encountered upon arriving, which was nothing suggesting any sort of burglary or trespass , much less a violent crime or anyone threatened or injured .
We are mindful that police, in carrying out their duties, often must react to potential threats quickly and under diffi- cult and uncertain circumstances. But having considered the entirety of the facts and circumstances here, we conclude that police did not have reasonable suspicion to frisk Howell. We therefore reverse the district court’s denial of his motion to suppress.
IV
A
A final issue remains for resolution. Following Howell’s initial encounter with police in December 2012, law enforcement obtained a warrant and arrested him in his home a few months later. In the course of the arrest, law enforcement conducted a protective sweep of Howell’s bedroom and found .32 caliber ammunition hidden inside a sock. The police then obtained a warrant to search the entire apartment.
The search revealed a few notable items. First, the police found a North American Arms .22 caliber revolver at the bottom of the apartment building’s internal air shaft—an open space just outside of Howell’s window. Police later traced that firearm to Howell’s father, Thomas Howell, who bought it in 1989 and passed away in 2010. Second, the police found a Frontier Derringer .22 caliber revolver, with Thomas Howell’s 22 No. 18-3157
name engraved on the handle, in a small safe in the living room. The safe also contained .22 and .32 caliber ammunition and a holster bearing a North American Arms logo.
The federal charges brought against Howell came in two counts (both alleging violations of 18 U.S.C. § 922(g)(1)) and covered a total of three guns. Count one addressed the gun the police found in his pocket in December 2012, and count two covered the two additional guns they recovered from his apartment in March 2013. Howell asks us to vacate his convictions on both counts. In doing so, he does not argue that the items recovered during the March 2013 search of his apartment were the fruit of the unconstitutional pat down that took place in December 2012. He contends in a much less direct way that the jury’s learning about the gun the police found in his pocket during that pat down impermissibly tainted its consideration of the evidence presented on count two.
In support of this argument, Howell points to a few events that unfolded at trial. He testified in his own defense on count two, stating that he did not know about the guns found in his apartment in March 2013. He emphasized that both guns were stored out of view and were traceable to his deceased father. On cross-examination, however, the government also asked Howell about the evidence against him on count one— the gun found in his pocket in December 2012. Howell refused to answer. He now insists that this refusal damaged his credibility in front of the jury—an outcome he would have avoided if the firearm on count one was never admitted.
Howell also points to a second event from his trial. Before returning a verdict, the jury posed two questions to the court: first, whether Howell was arrested in the room overlooking the air shaft where the police found the North American Arms
No. 18-3157 23
revolver, and, second, where other personal effects of Thomas Howell (a photograph and a memorial CD recovered by police ) were stored inside the apartment. The district court declined to answer, informing the jury that the evidence was closed and thus that they needed to decide the case on the record as it stood. Howell now argues that the jury’s questions revealed concern about whether he constructively possessed the guns in his apartment, given that they were out of view and traceable to his father.
B
A constitutional error requires reversal unless it was harmless beyond a reasonable doubt—“that is, [unless] no reasonable doubt exists that the error affected the jury’s verdict .” United States v. McKinney, 954 F.2d 471, 475 (7th Cir. 1992) (citing Chapman v. California, 386 U.S. 18, 24 (1967)). Where evidence was erroneously admitted at trial, we consider several factors in determining whether the evidence was harmless: the prejudicial effect of the evidence, how the government used the evidence at trial, the strength of the government ’s case outside that evidence, and “if there are any such indications from the verdict, how the jury likely received and considered the impermissible evidence.” United States v. Miller , 673 F.3d 688, 701 (7th Cir. 2012).
No doubt the wrongfully admitted gun was harmless as to count two. The government had robust evidence supporting that count. The jury heard about two revolvers tucked away in different corners of Howell’s apartment. It also heard that police found .22 caliber ammunition inside the living room safe—ammunition that matched the .22 caliber North American Arms pistol found in the air shaft just outside his window.
24 No. 18-3157
The district court also instructed the jury to consider each count separately. Even if the jury’s questions evinced some degree of reservation as to whether Howell constructively possessed the guns found in his apartment, we have no reason to think that the jury returned a guilty verdict on count two because of the evidence it heard on count one. The government presented ample independent evidence to support the jury’s verdict on count two. Right to it, the admission of the gun that served as the basis for count one, however wrongful, was harmless beyond a reasonable doubt as to count two.
* * *
The .38 Smith & Wesson revolver recovered from Howell’s pocket in December 2012 should have been suppressed as the fruit of an unconstitutional frisk. Any error in admitting that gun, however, was harmless as to Howell’s conviction on count two. We therefore REVERSE the denial of Howell’s motion to suppress the gun recovered in December 2012, VACATE his conviction on count one, and AFFIRM his conviction on count two.