Before WIENER, GRAVES, and HO, Circuit Judges. WIENER, Circuit Judge: A jury found Defendant-Appellant Ademola Babatunde Okulaja guilty of using two counterfeit passports to open bank accounts, in violation of 18 U.S.C. § 1543. Okulaja appeals his conviction on both counts, as well as his sentence. We affirm Okulaja’s conviction, but we vacate his sentence and remand for resentencing. Background In May 2016, Okulaja applied for a nonimmigrant visa to enter the United States. The next month, he opened a bank account at the Alief branch of the International Bank of Commerce (“IBC”) in Houston, Texas, in the name of Michael C. Millet. In doing so, he used a counterfeit passport in the name of Michael Charles Millet (the “Millet Passport”), purportedly issued by the United Kingdom of Great Britain and Northern Ireland, number 020338506. The photograph on the Millet Passport was the same as the one used on Okulaja’s visa application. The email address that Okulaja provided in opening this bank account (the “Millet IBC account”) was the same as the one he used for his visa application and his Texas driver’s license application. The IBC representative took a photo of Okulaja for the bank’s records when he opened the Millet IBC account, as is IBC’s practice if the equipment required to do so is functioning at the time. In November 2016, Okulaja opened a bank account at the Galleria IBC branch in Houston, Texas, in the name of David S. Allen. In doing so, he used a counterfeit passport in the name of David Samuel Allen (the “Allen Passport”), purportedly issued by the United Kingdom of Great Britain and Northern Ireland, number 794614923. The photograph on the Allen Passport was the same as that used on the Millet Passport and Okulaja’s visa application. The IBC representative did not take a picture of Okulaja when he opened this account (the “Allen IBC account”), likely because the required equipment was not working. The Texas address that Okulaja provided when opening the Allen IBC account was that of an unoccupied house on his own street. In March 2017, Okulaja opened a bank account at the Cesar Chavez IBC branch in Austin, Texas, in the name of Ronald D. Schnur, using a counterfeit passport in the name of Ronald Dean Schnur (the “Schnur Passport”), purportedly issued by the United Kingdom of Great Britain and Northern Ireland, number 020338506. The photograph on the Schnur Passport was the same as that used on the Millet Passport, the Allen Passport, and Okulaja’s visa application. The number on the Schnur Passport was the same as the number on the Millet Passport. The IBC representative took a photo of Okulaja when he opened this account (the “Schnur IBC account”). Charges In June 2018, Okulaja was charged in a two-count indictment. Count 1 alleged that, on or about June 6, 2016, Okulaja used the Millet Passport to open the Millet IBC account, in violation of 18 U.S.C. § 1543 (False Use of a Passport). Count 2 alleged that, on or about November 21, 2016, Okulaja used the Allen Passport to open the Allen IBC account in violation of 18 U.S.C. § 1543 (False Use of a Passport). The indictment did not charge Okulaja with bank fraud or any other crimes relating to various deposits he allegedly made as part of a larger scheme. Neither did it charge Okulaja with any crime relating to the Schnur IBC account. Trial At trial, the district court admitted two webcam photos over Okulaja’s objection that they were not sufficiently authenticated. The photos were offered to demonstrate the identity of the individual who opened the Millet IBC account and the Schnur IBC account. The government introduced the photos through the testimony of Shamsali Momin (“Ms. Momin”), an IBC officer with 14 years of experience at the bank, including as a branch manager. She testified that her experience included opening numerous customer accounts like the ones at issue here, as well as managing employees who open accounts. She had not been present when the accounts were opened, but she had reviewed IBC’s records related to those accounts, where she located the photos. Ms. Momin testified extensively regarding IBC’s normal business procedures for opening new accounts, explaining that a person must be physically present to open a new account. Ms. Momin stated that the IBC employee then takes that person’s photo with a webcam, scans his or her identification, and visually compares such person and the identification. Ms. Momin further testified that it is IBC’s general practice to take a photo of the person who opens a new account and explained that a photo is always taken if the equipment is working. She opined that, based on her knowledge of IBC’s practices, and because she located the relevant photos in the files for the Millet IBC and Schnur IBC accounts, the relevant photos were taken when the accounts were opened. For the photo taken when the Millet IBC account was opened, Ms. Momin identified the Alief IBC branch by the background in the photo since she is familiar with that branch. Also at trial, defense counsel attempted to admit Defendant’s Exhibit 2c, a photo of a fake driver’s license, found on Prince Ogunjimi’s phone, depicting a third party. The PSR explains that Ogunjimi was a friend of Okulaja and was involved in similar but unrelated incidents of fraud. The government objected to the relevance of Exhibit 2c because the fake ID depicted a man who was not Okulaja and featured a name unrelated to Okulaja or his offenses. Defense counsel suggested that the person in the photo in the fake ID looked “a lot” like Okulaja and “could pass for him” during a brief encounter. The district court sustained the government’s objection, apparently after determining that (1) defense counsel had not shown the relevance of Exhibit 2c and (2) the exhibit would significantly distract the jury. The jury convicted Okulaja on both counts. Sentencing The presentence report (PSR) listed as relevant conduct eleven other bank accounts, including the Schnur IBC account, that Okulaja opened with false passports. The indictment did not charge Okulaja with any crimes relating to these accounts. Also listed as relevant conduct was the intended loss attributed to Okulaja based on altered and counterfeit checks deposited into (1) those eleven bank accounts and (2) the two IBC accounts associated with the offenses of conviction. The intended loss described in the PSR totals $407,810.56. That number includes a $263,975 check deposited into the Schnur IBC account (the “Schnur Check”). The PSR determined Okulaja’s base offense level through a series of cross-references. It explains that: The Sentencing Guideline for false use of a passport—a violation of 18 U.S.C. § 1543—is U.S.S.G. § 2L2.2. Per § 2L2.2(c)(1)(A), U.S.S.G. § 2X1.1 applies because Okulaja used a passport in the commission of felony bank fraud and the resulting offense level was greater than it would be if it were determined under § 2L2.2.[1] Section 2B1.1 was applied per § 2X1.1 because it is the guideline for the underlying offense of fraud.[2] The base offense level for § 2B1.1 is six.[3] A 12-level increase was applied per § 2B1.1(b)(1)(G) because the intended loss attributed to Okulaja exceeded $250,000 but was not more than $550,000. Okulaja’s total offense level of 20 also included a two-level enhancement for obstruction of justice which was applied in a second addendum to the PSR. That offense level, combined with a criminal history category of I, produced an advisory guidelines range of 33 to 41 months of imprisonment. Among other things, Okulaja objected to the inclusion of the eleven uncharged bank accounts and the intended-loss amount associated with those accounts as relevant conduct. The government disagreed, as did the probation officer who maintained that the uncharged bank accounts were properly considered relevant conduct. Okulaja renewed this objection at sentencing. The district court determined that, although uncharged, Okulaja’s opening of the Schnur IBC account, and the subsequent deposit of funds into it, was relevant conduct, apparently because it was part of the same scheme as his opening of the Millet IBC and Allen IBC accounts. The district court found a total intended loss of $341,463, which supported the 12-level increase in offense level. That number included the Schnur Check. Absent the Schnur Check, the intended- loss amount would have supported only a six-level increase. The district court sustained counsel for Okulaja’s objection to the obstruction of justice enhancement and determined a total offense level of 18. That, combined with a criminal history category of I, resulted in a revised advisory guidelines range of 27 to 33 months imprisonment. The district court sentenced Okulaja within that guidelines range to 33 months imprisonment and three years supervised release on each count, to run concurrently. Okulaja timely appealed.[4] Standard of Review A “district court’s interpretation or application of the Sentencing Guidelines is reviewed de novo.”[5] A “‘district court’s determination of what constitutes relevant conduct for purposes of sentencing’ is a factual finding that ‘is reviewed for clear error.’”[6] “A factual finding is not clearly erroneous if it is plausible in light of the record as a whole.”[7] “The Court will find clear error ‘only if a review of all the evidence leaves us with the definite and firm conviction that a mistake has been committed.’”[8] A sentencing error may be reviewed for harmlessness. “To show a sentencing error is harmless, the government must ‘convincingly demonstrate both (1) that the district court would have imposed the same sentence had it not made the error, and (2) that it would have done so for the same reasons it gave at the prior sentencing.’”[9] “A district court’s decision to admit or exclude evidence is reviewed for abuse of discretion.”[10] “A trial court abuses its discretion when its ruling is based on an erroneous view of the law or a clearly erroneous assessment of the evidence.”[11] A ruling that photos were improperly admitted or excluded is not determinative by itself. “Any error in admitting the evidence is subject to harmless error review.”[12] “Unless there is a reasonable possibility that the improperly admitted evidence contributed to the conviction, reversal is not required.”[13] Analysis On appeal, Okulaja challenges his conviction on two separate grounds, and his sentence on a third. We address each challenge in turn. Conviction Admission of the Webcam Photos As noted, the district court admitted two webcam photos over Okulaja’s objection that they were not sufficiently authenticated. Okulaja renews this objection on appeal, contending that the photos should have been excluded because the government’s witness (1) was not present when the photos were taken, (2) did not recognize Okulaja, and (3) did not state explicitly that the photos “fairly and accurately” represented the customer who opened the relevant accounts. “To satisfy the requirement of authenticating or identifying an item of evidence, the proponent must produce evidence sufficient to support a finding that the item is what the proponent claims it is.”[14] “This is not a burdensome standard.”[15] Conclusive proof of authenticity is not required “before allowing the admission of disputed evidence.”[16] “Testimony by a witness with knowledge that the ‘matter is what it is claimed to be’ can be enough to prove the thing’s authenticity.”[17] Once this requirement has been met, “the trial court should admit the exhibit . . . in spite of any issues the opponent has raised about flaws in the authentication” because “[s]uch flaws go to the weight of the evidence instead of its admissibility.”[18] “The ultimate responsibility for determining whether evidence is what its proponent says it is rests with the jury.”[19] We have explained that “[a] witness qualifying a photograph need not be the photographer or see the picture taken; it is sufficient if he recognizes and identifies the object depicted and testifies that the photograph fairly and correctly represents it.”[20] The district court did not err, let alone abuse its discretion, in ruling that the photos were properly authenticated. Ms. Momin testified clearly that each picture “is what it is claimed to be,” a photo of the person who opened the relevant account.[21] That Ms. Momin did not identify Okulaja as the person in the photos is irrelevant because the photos were not offered to identify Okulaja; they were offered to identify the person who opened the relevant accounts. Jurors are capable of looking at a photo and determining whether it looks like the person in question. Like the properly admitted testimony in United States v. Rahim regarding the audio recordings kept as business records, Ms. Momin’s testimony described the process by which the photos were captured, stored, and produced.[22] In essence, Okulaja seeks to transform United States v. Clayton into a rigid rule that limits the ways in which a photograph may be authenticated by someone other than its photographer. Yet he offers no support for that rule. Furthermore, imposing such a requirement would be at odds with United States v. Barlow (which postdates Clayton by more than twenty-five years) and the nature of the Rule 901 authentication inquiry. Even if we were to treat Clayton as establishing such a rule (and even if a failure to recognize such a rule were an abuse of discretion), it would not undermine Okulaja’s conviction. Ms. Momin did “recognize and identif[y]” the Alief IBC branch by the background in the photo, noting that she was familiar with that branch. The second photo depicts the person who opened the Schnur IBA account. As discussed below, Okulaja was never charged with any crime relating to that account, so there is no reasonable possibility that the photo’s exclusion would have yielded a different result. Exclusion of the Photo of the Fake Driver’s License The district court sustained the government’s objection to Defense Exhibit 2c, a photo depicting a fake driver’s license found on Ogunjimi’s phone, issued to a third party. Okulaja claims that the district court abused its discretion when it failed to admit the exhibit. He contends that the main issue at trial was the identity of the person who used the counterfeit passports to open the Millet IBC and Allen IBC accounts. Okulaja insists that Exhibit 2c was relevant and not a distraction because it made it less probable that he was the one using counterfeit passports to open the bank accounts. Okulaja contends that the man pictured on the fake driver’s license featured in Exhibit 2c “ could pass” for him and could be the person who opened the Millet IBC and Allen IBC accounts. He acknowledges that the same fake ID depicted in Exhibit 2c is partially visible in Exhibits 2a and 2b, which were admitted into evidence, but he insists that admission of Exhibit 2c was necessary because the picture is larger and clearer. He adds that the exclusion of Exhibit 2c was harmful because it undermined his defense as to the issue of identity. As noted, the district court’s evidentiary rulings are, on appeal, “subject to harmless error review.”[23] “Any error made in excluding evidence . . . does not necessitate reversal unless it affected the defendant’s substantial rights.”[24] “In assessing any error, we must consider the other evidence in the case and determine whether the improperly excluded evidence, if admitted, would have had a substantial impact on the jury’s verdict.”[25] A trial court “may exclude relevant evidence if its probative value is substantially outweighed by a danger of one or more of the following: unfair prejudice, confusing the issues, misleading the jury, undue delay, wasting time, or needlessly presenting cumulative evidence.”[26] “A ‘trial court is afforded wide discretion in assessing the relevance and prejudicial effect of evidence.’”[27] Okulaja’s insistence that the district court abused its discretion in excluding the photo is unconvincing. Defense counsel wanted to offer the photo to support an alternative theory of the crimes—presumably that Ogunjimi and the man depicted in the fake ID photo both opened the accounts. But Ogunjimi was never linked to Okulaja’s crimes. There is no indication that he was present for or involved in the opening of any of the accounts. The only evidence that a second person was present for the opening of any account is a note in the bank records that “Mr. Allen” and “one of his good friends” were present at least once when Okulaja visited the branch. Neither was the relevant fake ID used in the opening of any of the accounts. There is nothing linking the fake ID to the opening of the accounts. That makes it difficult, if not impossible, for Okulaja to demonstrate that the district court’s decision was erroneous, let alone an abuse of discretion. Affirmance is warranted on this ground. But even if the decision to exclude that exhibit had been erroneous and an abuse of demonstration, there is no reasonable possibility that its admission would have had any impact on the jury, let alone a substantial one. There is ample evidence to support Okulaja’s conviction. The counterfeit Millet, Allen, and Schnur Passports all feature the photograph of Okulaja that was included in his visa application. Okulaja’s personal email address, used in his visa and Texas driver’s license applications, was also used to open the Millet IBC account. The Texas address that was provided in opening the Allen IBC account was that of an unoccupied house on the street where Okulaja lived. A photo was taken of the customer who opened the Millet IBC account. A photo was not taken when the Allen IBC account was opened, but IBC staff are trained to look at the person opening a new bank account and compare her or him to the identification that is provided. The jurors were able to compare these photos with Okulaja when they were asked at trial to take “a good strong look” at him. Given all this evidence, the jury rejected Okulaja’s contention that Ogunjimi or another person opened these bank accounts. Even if the exclusion of Exhibit 2c had been an abuse of discretion, it would have been a harmless one. That, too, merits affirmance. Guidelines Calculation We conclude, as noted above, that the district court clearly erred in calculating Okulaja’s guideline range. It did so when it used § 1B1.3(a)(2)’s broad definition of relevant conduct, even though that provision does not apply to Okulaja’s offenses of conviction. Applying that broad definition, the loss attributable to the charged offenses and relevant conduct was $341,463. Again, that sum includes the $263,975 Schnur Check. The district court applied a 12-level enhancement to the calculation of Okulaja’s guidelines range under § 2B1.1(b)(1), resulting in an advisory sentencing range of 27-33 months. If the district court had applied the correct definition of relevant conduct, the Schnur Check would not have qualified, and the loss would have yielded only a 6-level enhancement and an advisory sentencing range of 10-16 months. It is undisputed that, absent the Schnur Check, the appropriate intended loss would have been $75,438.50, the value of the checks deposited to the accounts that were opened as a result of the charged offenses. The government has not argued on appeal that the error was harmless and has therefore forfeited any such argument. Similarly, the government does not appear to seriously insist that the deposit of the Schnur Check could be seen as relevant conduct under § 1B1.3(a)(1), which does apply to Okulaja’s offenses of conviction. That is clear: The deposit did not “occur[] during the commission of the offense[s] of conviction, in preparation for th[ose] offense[s], or in the course of attempting to avoid detection or responsibility for” those offenses.[28] The sentence in this case thus turns on the language of § 1B1.3(a)(2). Before the district court applies any cross-references in the sentencing guidelines, it must identify the appropriate offense guideline provision and determine the scope of the relevant conduct. The Guidelines’ standard relevant conduct provision has two prongs, § 1B1.3(a)(1)(A) and § 1B1.3(a)(1)(B). Only the first is relevant to this appeal: Chapters Two (Offense Conduct) and Three (Adjust- ments). Unless otherwise specified, (i) the base offense level where the guideline specifies more than one base offense level, (ii) specific offense characteristics and (iii) cross references in Chapter Two, and (iv) adjustments in Chapter Three, shall be determined on the basis of the following: (1)(A) all acts and omissions committed, aided, abetted, coun- seled, commanded, induced, procured, or willfully caused by the defendant; and . . . that occurred during the commission of the offense of convic- tion, in preparation for that offense, or in the course of attempt- ing to avoid detection or responsibility for that offense[.][29] We recently explained that “the trailing (or hanging) clause of subsection (a)(1) . . . makes clear that either of the[] two categories of conduct must have ‘occurred during the commission of the offense of conviction, in preparation for that offense, or in the course of attempting to avoid detection or responsibility for that offense.’”[30] Importantly, however, we also clarified that “[b]ecause this clause trails behind the text of subsections (a)(1)(A) and (a)(1)(B), we treat it as located within subsection (a)(1)—and not within either subsections (a)(1)(A) or (a)(1)(B).”[31] That last point is important to the disposition of this case because of § 1B1.3(a)(2), the Guidelines’ other relevant conduct definition. That provision adds to the definition of relevant conduct: solely with respect to offenses of a character for which § 3D1.2(d) would require grouping of multiple counts, all acts and omissions described in subdivisions (1)(A) and (1)(B) above that were part of the same course of conduct or common scheme or plan as the offense of conviction.[32] As we explained in United States v. Deckert, the fact that the trailing language (which limits the scope of relevant conduct significantly) is included in (a)(1) but is not included in (1)(A) and (1)(B) means that it does not apply to (a)(2). The result is that, “when a defendant is convicted of a groupable offense, courts should look to all acts and omissions that were part of a similar course of conduct or common scheme as the offense of conviction.”[33] But, if the offense of conviction is not one for which § 3D1.2(d) would require grouping, acts and omissions are only relevant conduct if they “occurred during the commission of the offense of conviction, in preparation for that offense, or in the course of attempting to avoid detection or responsibility for that offense.”[34] The district court treated the Schnur Check as relevant conduct after determining that it was part of the “same scheme” as the offenses of conviction, and without indicating that the deposit of that check “occurred during the commission of the offense of conviction, in preparation for that offense, or in the course of attempting to avoid detection or responsibility for that offense.” That would likely be enough if (a)(2) applied to this case. But (a)(2) only applies when “a defendant is convicted of a groupable offense.”[35] And, as the government concedes, false use of a passport in violation of 18 U.S.C. § 1543, is not “of a character for which § 3D1.2(d) would require grouping of multiple counts.”[36] In fact, § 3D1.2(d)”[s]pecifically exclude[s]” the relevant guideline, § 2L2.2, from grouping.[37] Okulaja’s convictions therefore do not trigger § 1B1.3(a)(2), so the district court erred in using that definition of relevant conduct to justify considering the Schnur Check. The government acknowledges this problem but contends that it is not dispositive. The government claims that § 1B1.3(a)(2) applies to Okulaja’s convictions because the district court applied cross-references from the original, non-groupable guideline—§ 2L2.2—to eventually reach § 2B1.1, the fraud guideline, which is on the grouping list. If we were to take that approach we would proceed as though Okulaja had been convicted of bank fraud. The government notes that we have previously declined to conclude that such an approach is plain error. We explained in United States v. Jackson that “[i]f this [approach] was error, it was not plain error,” because the “government . . . provided a plausible argument.”[38] And, at that time, the argument was plausible. But, it has been undermined significantly by our more recent opinion in Deckert. As explained above, we made clear in Deckert that § 1B1.3(a)(2) applies only when the charged offense is groupable.[39] The government also claims that its approach is dictated by § 1B1.5(a) and (c), which state that, “[i]f the offense level is determined by a reference to another guideline . . ., the adjustments in Chapter Three (Adjustments) also are determined in respect to the referenced offense guideline, except as otherwise expressly provided.”[40] That suggests, the government contends, that § 3D1.2(d) should be “determined in respect to” § 2B1.1, the cross- referenced guideline. At first glance, that language seems compelling. But, although we are examining § 3D1.2(d), we are not applying it to determine the offense level. Rather, we are referencing it to apply Chapter One’s relevant conduct provisions. This distinction is underscored by the Sentencing Guidelines’ sequence of proceedings, found in § 1B1.1(a). A court’s first step when calculating a guideline range is to determine, pursuant to § 1B1.2, the offense guideline section “applicable to the offense of conviction.”[41] Section 1B1.2 dictates that, to accomplish this initial step, a court should first “[d]etermine the offense guideline section . . . applicable to the offense of conviction (i.e., the offense conduct charged in the count of the indictment . . . of which the defendant was convicted).”[42] Next, “[a]fter determining the appropriate offense guideline section pursuant to subsection (a) of this section, [the court should] determine the applicable guideline range in accordance with § 1B1.3 (Relevant Conduct).”[43] Only then, after the relevant conduct has been identified, is it time to “[d]etermine the base offense level and apply any appropriate . . . cross-references.”[44] Under the government’s approach, a district court would first apply cross-references and then identify relevant conduct. That is both backwards and at odds with the Sentencing Guidelines. To recap: The district court wrongly applied § 1B1.3(a)(2) by using a cross-referenced guideline instead of the actual offense of conviction. Doing so meant that it could treat Okulaja’s deposit of the Schnur Check as relevant conduct because it was “part of the same course of conduct or common scheme or plan as the offense of conviction,” even though it did not occur (1) during the commission of the offense of conviction, (2) in preparation for that offense, or (3) in the course of attempting to avoid detection or responsibility for that offense. Had the district court applied the correct definition—the one found in § 1B1.3(a)(1)—it could not have treated the Schnur Check as relevant conduct. Application of the correct definition would have left Okulaja with a guidelines range of 10-16 months for each count, rather than the 27-33 months for each count that the district court calculated. Conclusion Neither of Okulaja’s challenges to his conviction are meritorious, so we AFFIRM his conviction. But, because the district court clearly erred in calculating Okulaja’s guideline range, we VACATE the judgment below and REMAND for resentencing.
