United States v. Jonathan Kirschner

PRECEDENTIAL UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT __________ No. 20-1304 __________ UNITED STATES OF AMERICA v. JONATHAN KIRSCHNER, a/k/a Jonathan Kratcher, Appellant __________ On Appeal from the United States District Court for the District of New Jersey (District Court No. 1:18-cr-00360-001) District Judge: Honorable Robert B. Kugler __________ Argued January 26, 2021 Before: RESTREPO, BIBAS, and PORTER, Circuit Judges (Filed: April 22, 2021) __________ Mark E. Coyne Office of United States Attorney 970 Broad Street, Room 700 Newark, NJ 07102 Molly S. Lorber [ARGUED] Office of United States Attorney Camden Federal Building & Courthouse 401 Market Street Camden, NJ 08101 Counsel for Appellee Justin T. Loughry [ARGUED] Loughry & Lindsay 330 Market Street Camden, NJ 08102 Counsel for Appellant __________ OPINION OF THE COURT __________ RESTREPO, Circuit Judge. In 2018, Jonathan Kirschner pleaded guilty to one count of impersonating an officer acting under the authority of the United States and one count of importing counterfeit coins and bars with intent to defraud. At sentencing, the District Court applied to Kirschner’s sentence three enhancements pursuant to the U.S. Sentencing Guidelines—a 2-level enhancement be- cause Kirschner’s fraud used sophisticated means; another 2- level enhancement because Kirschner abused a position of public trust to facilitate his crimes; and a 22-level enhancement because the “loss” attributable to his scheme was greater than $25 million but less than $65 million, even though it grossed only about one one-thousandth of that. Kirschner appeals the District Court’s judgment of sentence and challenges the three enhancements it applied. For the rea- sons that follow, we will vacate Kirschner’s sentence and re- mand for resentencing. While the District Court was well within its discretion to apply the abuse-of-trust and use-of-so- phisticated-means enhancements, we conclude it clearly erred in applying the 22-level enhancement for loss, and we cannot say that the error was harmless. I. BACKGROUND In 2017, Jonathan Kirschner earned $30,105 by importing counterfeit coins and bullion and then, posing as a federal law 2 enforcement agent, selling them as genuine articles to unsus- pecting customers. Eventually, several customers discovered the coins or bullion they bought were fakes and alerted federal authorities—the real federal authorities. Pretending to be pro- spective buyers, federal agents set up a sting and snared Kirschner. Searching his home and interdicting packages des- tined for him, agents discovered and seized thousands of coun- terfeit coins and bullion that, according to the government’s expert, would have been worth approximately $46.5 million if genuine. Agents also seized a plethora of numismatic and bul- lion paraphernalia meant to mimic the idiosyncratic packaging materials used by the authentication agencies. Kirschner subsequently pleaded guilty to one count of im- personating a federal agent, in violation of 18 U.S.C. § 912, and one count of importing counterfeit coins and bars with in- tent to defraud, in violation of 18 U.S.C. § 485. In the plea agreement, the parties agreed that the sentencing judge may impose any “reasonable sentence” up to the statutory maxi- mums—three years for the § 912 count, and fifteen years for the § 485 count. The parties also agreed on the framework for calculating an appropriate sentence under the Guidelines. While acknowledging that the District Court retained “sole dis- cretion” to impose Kirschner’s sentence, App. 55, the parties agreed that, under the Guidelines, Kirschner’s base offense level would be 6; the base offense level would be enhanced for “intended losses” by either 14 (according to Kirschner) or 22 (according to the government); and the resulting offense level would be reduced by 3 for Kirschner’s acceptance of responsi- bility. The parties thus agreed that the total offense level ought to be 17 (if Kirschner’s proposed enhancement for loss was adopted) or 25 (if the government’s was). The parties did not contemplate any further sentencing enhancements, reductions, or deviations. Prior to sentencing, a Presentence Investigation Report (“PSR”) was prepared, employing the Guidelines. The PSR calculated Kirschner’s base offense level to be 6, and included a 22-level enhancement for loss and two 2-level enhancements for use of sophisticated means and abuse of trust. It also 3 included a separate 2-level enhancement because Kirschner’s crime involved ten or more victims. According to the PSR, then, Kirschner’s total offense level was 34. Because of prior offenses, Kirschner had a criminal history category of III. Thus, the PSR calculated a Guidelines range of 188 to 235 months. In their sentencing memoranda, the parties tracked the con- tours of their plea agreement. Both agreed that Kirschner’s base offense level was 6 and that he should receive a 3-level reduction for acceptance of responsibility. They further agreed that Kirschner was due no further enhancement for the number of victims, the use of sophisticated means, or the abuse of a position of trust. And they continued to disagree over the amount of loss at- tributable to Kirschner’s scheme. Kirschner argued that he in- tended to cause losses of only about $1.2 million, warranting a 14-level enhancement and a total offense level of 17 (base of- fense level of 6 plus 14 for loss less 3 for acceptance of respon- sibility). The government, by contrast, argued that he intended to cause losses of approximately $36 million, equaling a 22- level enhancement and a total offense level of 25 (base offense level of 6 plus 22 for loss less 3 for acceptance of responsibil- ity). At sentencing, the District Court adopted the PSR’s recom- mendations in part. It adopted the base level offense, and the enhancements for loss, use of sophisticated means, and abuse of trust. The District Court also granted Kirschner a 3-level decrease for acceptance of responsibility, resulting in a total offense level of 29. With his criminal history category of III, Kirschner’s Guidelines range was 108 to 135 months. The Dis- trict Court sentenced Kirschner to 126 months’ imprisonment, and this appeal followed. II. DISCUSSION The District Court had jurisdiction under 18 U.S.C. § 3231, and we have jurisdiction pursuant to 28 U.S.C. § 1291 and 18 4 U.S.C. § 3742(a). This Court reviews factual findings relevant to the Guidelines for clear error and exercises plenary review over a district court’s interpretation of the Guidelines. United States v. Grier, 475 F.3d 556, 570 (3d Cir. 2007) (en banc). A. The District Court’s increase of Kirschner’s offense level for intended loss was clear error Kirschner first argues that the District Court clearly erred in applying a 22-level enhancement for intended loss. We agree. The applicable advisory Guidelines provision is U.S.S.G. § 2B1.1(b)(1), which enhances fraud sentences based on “loss.” Under a Guidelines comment, a court must compare the pecuniary harm the defendant “actual[ly]” inflicted on his victims with the losses he “intended” to result, with “intended loss” defined as “the pecuniary harm that the defendant pur- posely sought to inflict[.]” See U.S.S.G. § 2B1.1 cmt. n.3(A)(i)-(ii). A court must then identify the greater figure, the actual or intended loss, and enhance the defendant’s offense level accordingly. Only this comment, not the Guidelines’ text, says that defendants can be sentenced based on the losses they intended. By interpreting “loss” to mean intended loss, it is possible that the commentary “sweeps more broadly than the plain text of the Guideline.” United States v. Nasir, 982 F.3d 144, 177 (3d Cir. 2020) (en banc) (Bibas, J., concurring). But Kirschner assumes the comment is correct, and so we will too. Relevant for our purposes, § 2B1.1 says that if the greater of the actual or intended loss falls between $550,000 and $1.5 million, the defendant’s offense level should increase by 14 levels. But if the greater loss figure is more than $25 million but less than $65 million, the defendant’s offense level should increase by 22 levels. There is no question that the actual losses Kirschner in- flicted on his victims were far less than the losses he intended to cause. Compare App. 259-62 (invoices revealing actual losses of approximately $30,000), and Dist. Ct. Dkt. 72 (dis- trict court ordering $14,600 in restitution to identified victims), with App. 236-37 (government estimating intended losses of approximately $36 million), and App. 190, 271-72 (Kirschner 5 estimating intended losses of $1.2 million). But the parties vo- ciferously dispute what unrealized losses he intended. Before the District Court, the government argued that Kirschner intended to cause losses of approximately $36 mil- lion and so merited a 22-level offense level increase. The gov- ernment’s $36 million calculation involved these steps. To begin, relying on a numismatic expert, the government esti- mated and summed the fair market value of the genuine version of each counterfeit seized from Kirschner, and concluded that the total fair market value of all seized counterfeits was ap- proximately $46.5 million. Next, the government analyzed in- voices documenting Kirschner’s successful sales and deter- mined that Kirschner historically had been able to sell his counterfeits at about 77 percent of the estimated fair market value of the counterfeits’ genuine counterparts. Finally, the government multiplied the total market value of the genuine versions of wares seized from Kirschner ($46.5 million) by Kirschner’s historic yield (77 percent) to arrive at an intended loss of approximately $36 million. Id. Roughly algebraically, the government’s intended-loss methodology was as follows: 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 𝐹𝐹𝐹𝐹𝐹𝐹 𝑜𝑜𝑜𝑜 𝑡𝑡ℎ𝑒𝑒 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 = � (0.77) ∗ � 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑜𝑜𝑜𝑜 � 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 𝑒𝑒𝑒𝑒𝑒𝑒ℎ 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 Kirschner objected to the government’s loss figure. He ar- gued that the government’s estimation of total intended loss did not address what losses he “intended” to cause, that is, did not address what losses he purposely sought to inflict on future vic- tims. To understand the force of Kirschner’s objection, we must unpack the contents of the government’s methodology. Over the course of the government’s investigation, agents seized from Kirschner well over one-thousand counterfeit coins and bars. The majority (something like 80 to 90 percent) were fake bullion or counterfeits of common collector coins. Estimating the fair market value of the genuine versions of these counterfeits appears to have been straightforward. The market for common collector coins is robust, so to estimate the price their genuine counterparts could fetch on an open market, 6 the government’s expert reviewed auction sales and sales re- ported in a popular trade publication. And valuing the bullion was simpler still. Like any fungible commodity, the govern- ment’s expert approximated the counterfeit bullion’s grade, measured its weight, and valued it accordingly. The expert tes- tified that valuing wares in this way was commonly accepted in the numismatic community. App. 110. After discounting the total estimated fair market value by Kirschner’s historic yield, the government estimated the intended losses attributa- ble to these counterfeits to be approximately $1.2 million ($1.5 million genuine fair market value discounted at 77 percent). The remaining minority of wares seized from Kirschner, however, were counterfeits of six exceptionally rare coins. Agents seized three counterfeits of the 1933 Saint-Gaudens Double Eagle; nine counterfeits of the 1804 Bowed Liberty; four counterfeits of the 1794 Flowing Hair; six counterfeits of the 1866-S coin, fifty-three Morgan Dollars (without the motto, making them rarer); and one-hundred-twenty-seven Morgan Dollars (with the motto, and thus more common). To illustrate the exceptional rarity of these coins (counterfeits though they were), the government’s expert identified the 1933 Saint-Gaudens Double Eagle as “one of the most valuable coins in United States history.” App. 128. Indeed, only one 1933 Saint-Gaudens Double Eagle has been sold publicly, and that one was sold at auction in 2002 for $8 million. See Langbord v. U.S. Dep’t of Treasury, 832 F.3d 170, 177 (3d Cir. 2016) (en banc). And it had been owned by King Farouk of Egypt. Despite their extreme rarity, the government attempted to estimate the fair market value of these coins just as it estimated the fair market value of the others. It purportedly referenced auction records and recent trades or sales among dealers and collectors. But the uniqueness of the Saint-Gaudens Double Eagle suggests that the market for these wares was signifi- cantly less robust and more specialized than the market for common collector coins. Nevertheless, employing the same methodology, the government estimated that their intended 7 losses accounted for approximately $34.5 million, or about 97 percent of the total estimated intended loss, as follows: Num. FMV of Total FMV Discounted Coin Seized Each Coin of All Coins at 77 Percent 1933 Saint- Gaudens 3 $8,000,000 $24,000,000 $18,480,000 1804 Bowed Liberty 9 $2,000,000 $18,000,000 $13,860,000 1794 Flowing Hair 4 $95,000 $380,000 $292,600 1866-S Twenty Dollar 6 $82,000 $492,000 $378,840 Morgan Dol- lar (no motto) 53 $37,587 $1,992,111 $1,533,925 Morgan Dol- lar (motto) 127 $993 $126,111 $97,105 TOTAL $44,990,222 $34,642,471 The District Court adopted the government’s intended loss methodology and resulting loss figure. On appeal, Kirschner challenges the inclusion of the in- tended losses associated with the six high-value counterfeits. 1 Kirschner contends that the District Court never found by a preponderance of the evidence that he “purposely sought to in- flict” the losses the government claims he intended to inflict. He says that he never had access to the markets presupposed by the government’s “fair market value” methodology, nor did he attempt to access such markets. He further says that he had no knowledge of the prices for which genuine versions of his counterfeits had sold. Finally, he says that even if he knew the prices the genuine versions had sold for, the government 1 Before the District Court and on appeal, Kirschner challenges only the intended losses associated with the six sets of high- value counterfeits. He does so, he says, because he does not believe challenging the intended losses associated with the common collector coins would bring his total intended loss be- low the next relevant numeric threshold, which is $550,000. See U.S.S.G. § 2B1.1(b)(H). 8 introduced no evidence that he would have attempted to sell them at those prices discounted at 77 percent. We agree. While the District Court “focus[ed] [] on what [Kirschner] intended to do” with the high-value counterfeits, App. 156, it never found that Kirschner intended to sell the coins as counterfeits for the prices the government claimed. At the sentencing hearing, the District Court’s findings on in- tended loss were dedicated to rebuffing Kirschner’s ill-fated argument that he intended to sell the high-value counterfeits as replicas, not as the real thing. But the first-order inquiry is not whether Kirschner intended to sell the high-value counterfeits as replicas or the real thing. The first-order inquiry is whether the government proved, by a preponderance of the evidence, that Kirschner intended to sell these coins at the prices the gov- ernment claims. The District Court reasonably rejected Kirschner’s sell- them-as-replicas argument. As the Court found, “[h]is entire behavior was that he was importing these things in order to sell them and get over on people and defraud people,” and none of the counterfeit coins seized “were marked in any way to indi- cate that they were replicas.” App. 156-57. But the Court made no other finding supporting the intended losses associ- ated with the high-value coins. Indeed, the District Court ad- mitted that it did not know “[w]hat [Kirschner] intended to do specifically, logistically, with these valuable coins,” only that, “by a preponderance of the evidence, at some point he was go- ing to try to sell those coins as legitimate coins.” App. 157. On this record, we cannot say that the government carried its burden to show that Kirschner had the requisite purpose to sell the high-value counterfeits for the prices it suggested. The past coin sales detailed in the government’s complaint were of a particular type—advertised on the website Facebook; con- summated in grocery-store parking lots and local bank lobbies; and priced on the order of $100 per coin. It is not clear whether Kirschner intended to evolve his operation to attempt the type of rarefied sales contemplated by the government’s loss fig- ures. Nor can we say the District Court’s error adopting the 9 government’s methodology and resulting loss figure was harm- less. The possibility exists that the District Court, on remand, may find that the government fails to prove by preponderance of the evidence that Kirschner intended to sell the high-value counterfeits at those prices. And again, the principal question is not whether Kirschner could have sold the high-value coun- terfeits at the prices claimed by the government. The question is whether he intended to. While we have not considered how to estimate intended loss in this context, we have addressed how to do so in similar circumstances. In United States v. Diallo, the defendant trav- eled throughout the mid-Atlantic using stolen credit cards. 710 F.3d 147, 148 (3d Cir. 2013). Through the scheme, the defend- ant caused actual losses of $160,000. Id. at 149. But a Secret Service case agent testified that the total intended loss was ap- proximately $1.6 million. Id. To arrive at that figure, the agent had contacted the financial institutions associated with each stolen credit card to request the card’s credit limit, and was told the total credit limit for the 327 stolen credit cards was $1.6 million. Id. Over the defendant’s objections, the government argued that this figure should be the “intended loss” for the purposes of sentencing. Id. The district court accepted the government’s arguments, and increased the defendant’s sen- tence accordingly. On appeal, we considered “how sentencing courts should calculate what pecuniary harm was intended to result from credit card fraud when the fraud’s perpetrator did not know the credit limit, which is the potential loss amount from the stolen credit card.” Id. at 150 (internal quotation marks omitted). Looking to United States v. Geevers, 226 F.3d 186 (3d Cir. 2000), we held that “‘[w]hile intended loss may not be auto- matically determined based on what the potential loss is, in- tended loss may still equal potential loss.’” Diallo, 710 F.3d at 151 (quoting Geevers, 226 F.3d at 192). We noted, however, that “‘[i]t is clear that a district court errs when it simply equates potential loss with intended loss without deeper analy- sis.’” Id. (quoting Geevers, 226 F.3d at 192). Instead, a bur- den-shifting framework is appropriate in such cases. 10 “[T]hough the government bears the burden of proof in guide- lines cases, the burden of production may shift to the defendant once the government presents prima facie evidence of a given loss figure.” Id. (quoting Geevers, 226 F.3d at 188). But we reiterated that “the government always bears the burden of proving by a preponderance of the evidence that the facts sup- port a sentencing enhancement, and the defendant does not have to prove the negative to avoid the enhanced sentence.” Id. (emphasis added) (internal quotation marks omitted). With respect to the defendant’s sentencing in Diallo, we found that the district court had not engaged in the requisite “deeper analysis.” We noted that: It is possible that the District Court relied on the Secret Service agent’s testimony that the search of his car un- covered a skimming device; the evidence that Diallo has traveled from Virginia to New York in order to use the fraudulent credit cards; that Diallo had already spent $160,000 and was continuing to make additional pur- chases; or that at the time of his arrest, Diallo had re- turned to a store where he had made $2,600 in purchases just one day prior. It is also conceivable that the District Court agreed with the Government’s argument that Di- allo intended to charge up to the credit limit on every credit card number found in his possession. On the other hand, the District Court might simply have incor- rectly presumed that the aggregate credit limit alone can make out a prima facie case for intended loss amount in credit card fraud. . . . [W]e would be speculating as to what evidence or argument was the basis for the District Court’s finding that $1.6 million was Diallo’s intended loss amount. Id. at 153. Accordingly, we vacated the sentence and re- manded for reconsideration of intended loss and whether a de- parture was warranted based on the intended loss amount over- stating or understating the seriousness of the offense. Id. at 154. It is true that our decision to vacate and remand in Diallo primarily was caused by the district court’s failure to justify its 11 conclusion that the defendant intended to commit the maxi- mum potential loss. Here, the District Court did not commit that error. Following the government, the Court discounted the estimated fair market value of the counterfeits’ genuine coun- terparts by Kirschner’s historic yield of 77 percent. But our admonition in Diallo that district courts must con- duct a “deeper analysis” before inferring that a defendant in- tended to cause a particular loss applies to any loss methodol- ogy the government elects to adopt. So, in Diallo, the district court’s decision to equate intended loss with maximum poten- tial loss without justification was an example of a sentencing court failing to conduct the requisitely deep analysis. A district court similarly runs afoul of Diallo if, as here, it adopts an in- tended-loss methodology without demonstrating that the de- fendant’s “purpose” was to inflict the losses the government claims he intended to inflict. See U.S.S.G. § 2B1.1 cmt. n.3(A); see also United States v. Manatau, 647 F.3d 1048, 1050-51 (10th Cir. 2011) (Gorsuch, J.) (detailing the mental state requirement of the intended loss analysis). In sum, to estimate the losses a defendant intended to cause his victims under U.S.S.G. § 2B1.1 cmt. n.3(C), a district court must conduct a “deeper analysis” to make sure the defendant purposely sought to inflict each component of the losses the government claims he intended to inflict. In conducting that analysis, the court is free to make reasonable inferences about the defendant’s mental state from the available facts. The gov- ernment need only prove a defendant’s intent by a preponder- ance of the evidence, and the court need only make a “reason- able estimate” of the intended loss. See Diallo, 710 F.3d at 150; U.S.S.G. § 2B1.1(b)(1) cmt. n.3(C). If the losses associ- ated with the defendant’s past conduct are easily extrapolated to losses the defendant intended to cause in the future (as in stealing credit cards and charging the same amount on each be- fore discarding), the district court is free to estimate intended losses using one governing methodology (say, number of sto- len but unused cards times dollar amount charged on used and discarded cards). But if the losses associated with the defend- ant’s past conduct do not neatly map to intended future losses, 12 the district court must finely tune its methodology to ensure it does not underestimate or overestimate a defendant’s intended losses. Here, on remand, the District Court should examine whether Kirschner purposely sought to inflict the losses asso- ciated with the six high-value counterfeits. Among other in- quiries, the District Court may examine whether Kirschner had knowledge of the fair market values claimed by the govern- ment; whether Kirschner intended to sell the high-value coun- terfeits at the claimed market price with only a minor discount, or whether he took any steps to sell counterfeits of those types at all; whether Kirschner had sold any of the high-value coun- terfeits in the past and, if so, whether that sale price is not the most sensible intended loss for future sales of that coin, see Kirschner Br. at 27 (government claiming $37,000 per coin losses for coin previously sold for $20); and whether Kirschner was aware of various idiosyncrasies that would cause the value of a counterfeit to go up or down, as with the motto on the Morgan Dollars. The court may also consider whether the comment is right that intended loss matters, and whether Kirschner has preserved this issue. Cf. United States v. Nasir, 982 F.3d 144, 157-59 (3d Cir. 2020) (en banc). But if intended loss does matter and Kirschner has preserved the issue, the Dis- trict Court cannot simply estimate “intended loss” by accepting the government’s slightly discounted fair market values with- out any finding that the defendant intended to inflict losses rea- sonably approaching those figures. 2 2 Two additional points warrant discussion. First, the parties spend a good deal of their briefs discussing whether Kirschner’s selling the high-value counterfeits was even possi- ble, and whether and how to count losses that would have been impossible or unlikely to occur. Before the District Court, Kirschner had argued that selling the high-value counterfeits at genuine prices would have been impossible and, therefore, the losses associated with them should be excluded. The Court rejected that argument, noting that “if you look at Application Note 3, intended loss means that pecuniary harm that the de- fendant purposely sought to inflict and includes intended 13 pecuniary harm that would have been impossible or unlikely to occur.” App. 156. It “underst[oo]d the argument that it would have been almost impossible to sell these $8 million coins or the 1804 silver dollars,” but “[impossibility] doesn’t matter un- der the Sentencing Guidelines.” Id. It is true that under the Guidelines comment, intended loss “includes intended pecuni- ary harm that would have been impossible or unlikely to occur (e.g., as in a government sting operation, or an insurance fraud in which the claim exceeded the insured value).” U.S.S.G. § 2B1.1 cmt. n.3(A)(ii). But as the insurance fraud and sting examples illustrate, counting an “impossible or unlikely” loss still requires the defendant to have possessed the requisite men- tal state. Further, as we explained in Geevers, that inflicting certain losses would have been “impossible” is not a defense to a finding of purposeful intent. Instead, “impossibility” bears on whether a defendant had purposeful intent to begin with. See Geevers, 226 F.3d at 195 (“Impossibility bears on what is reasonable for the person to have intended to do, . . . adven- turesome or deluded criminals are [not] to have their actions interpreted differently than those of their more sober counter- parts.”). Here, on remand, if the District Court finds that Kirschner “purposely” tried and failed to sell any of the high- value counterfeits at the prices the government suggested, Kirschner cannot defend against a § 2B1.1 enhancement by claiming that such sales were impossible or unlikely to begin with. But Kirschner may argue that he did not possess the req- uisite mental state to sell the high-value counterfeits at the prices the government suggested in part because such sales would have been impossible or unlikely. Second, and related, Kirschner suggests that had the govern- ment’s sting operation involved the attempted sale of a Saint- Gaudens Double Eagle for $4 or $5 million or more, “if Kirschner has accepted the proposition and made arrangements to make the sale, of course the sentence must reflect the four or five or even eight million dollar loss[.]” Kirschner Reply Br. 7-8. But we have left this issue open. Anytime a government sting operation triggers not-yet-triggered mandatory mini- mums or Guidelines enhancements, courts may be concerned with so-called sentencing entrapment or sentencing factor ma- nipulation. See United States v. Sed, 601 F.3d 224, 229-31 (3d Cir. 2010). Sentencing entrapment focuses on the defendant’s 14 B. The District Court’s increase of Kirschner’s offense level for abuse of a position of trust was not clear error Kirschner next argues that the District Court erred in apply- ing the abuse-of-trust enhancement. We disagree. That en- hancement, set forth in Chapter 3 of the Sentencing Guidelines, tells a sentencing court that, “[i]f the defendant abused a posi- tion of public or private trust . . . in a manner that significantly facilitated the commission or concealment of the offense, in- crease [the offense level] by 2 levels.” U.S.S.G. § 3B1.3. In United States v. Douglas, 885 F.3d 124 (3d Cir. 2018) (en banc), we adopted a two-step inquiry to determine whether to apply the abuse-of-trust enhancement. Under Douglas, we ask, first, “whether the defendant actually occupied a position of public or private trust.” Id. at 130. To aid that inquiry, we ask whether the defendant’s position included “the power to make decisions substantially free from supervision based on (1) a fiduciary or fiduciary-like relationship, or (2) an authori- tative status that would lead his actions or judgment to be pre- sumptively accepted.” Id. at 133. If we conclude that the defendant did hold such a position, we reach the second step. There, the question is whether the defendant abused his position of trust “in a manner that signif- icantly facilitated his crime.” Id. at 130 (citation and internal predisposition—did law enforcement entrap him into commit- ting crimes beyond what he was predisposed to commit? Sen- tencing factor manipulation, by contrast, focuses on the gov- ernment’s conduct. According to “its broadest formulation,” it is a violation of the Due Process Clause, and occurs when the government engages in a longer-than-needed investigation or extraordinary sting operation to unfairly exaggerate the de- fendant’s sentencing range. See id. at 231. But we have “nei- ther adopted nor rejected the doctrines of sentencing entrap- ment and sentencing factor manipulation.” Id. at 229. And because the agent in this case sought to buy from Kirschner wares similar to ones Kirschner had sold previously and other- wise committed herself with professionalism, we need not con- sider that issue today. 15 quotation marks omitted). In answering that question, courts should consider, among other relevant and case-specific fac- tors, “whether the defendant’s position allowed him to commit a difficult-to-detect wrong”; whether the defendant’s position granted him “authority vis-à-vis the object of the wrongful act”; or “whether the victim relied on the defendant’s integrity, such that the victim became a more susceptible target for the defendant.” Id. at 134. Here, the District Court heard sufficient evidence to support the abuse-of-trust enhancement. Kirschner admitted at his plea hearing and does not contest on appeal that he occupied a po- sition of trust relative to his victims. 3 The record also 3 Though not at issue here, under Guidelines commentary, the abuse-of-trust enhancement “also applies in a case in which the defendant provides sufficient indicia to the victim that the de- fendant legitimately holds a position of private or public trust when, in fact, the defendant does not.” U.S.S.G. § 3B1.3 cmt. n.3. That appears to align with the enhancement’s text. Read- ing the enhancement naturally, to “abuse” a position of trust, a defendant can actually hold such a position and then leverage the formal powers granted to that position for nefarious ends or pretend to hold one in order to exploit the perception of power a victim bestows on that position. So the abuse-of-trust enhancement has applied to an investment broker employed by a real brokerage house who, using certain powers granted to him by the brokerage house and the governing law, embezzles an investor’s money. See, e.g., United States v. Rivernider, 828 F.3d 91, 114 (2d Cir. 2016). And the abuse-of-trust enhance- ment also has applied to an individual who leads investors to believe he is a legitimate investment broker employed by a real brokerage house in order to steal their money. See, e.g., United States v. Reyes-Rivera, 812 F.3d 79, 86-87 (1st Cir. 2016). If this reading holds, then at the first step of Douglas, where the question is whether the defendant “actually occupied” a posi- tion of public or private trust, 885 F.3d at 130, what we want the sentencing court to ask is, Did the defendant hold a position of public or private trust relative to the victim? This formula- tion clarifies that Douglas’s threshold requirement that a de- fendant “actually occupy” a position of trust encompasses not just defendants who hold positions of trust in fact but also 16 demonstrates that Kirschner’s impersonation of a law enforce- ment agent “significantly facilitated the commission or con- cealment of the offense.” See Douglas, 885 F.3d at 134. A buyer of Kirschner’s counterfeits testified that she thought Kirschner may have been perpetuating “one of those scams you hear about all the time,” but appeared to overcome that fear in part because Kirschner represented to her that he was a federal law enforcement agent. See App. 164. Several other victims similarly alleged that Kirschner’s apparent position as a federal law enforcement agent aided his fraud. One said he “person- ally felt more secure making the $11,000 purchase from Kirschner because Kirschner represented himself to be an ATF Special Agent.” PSR, at 13. Another explained how Kirschner used his position as a federal agent to explain how he came to possess the valuable coins. To that victim, Kirschner said that as a federal agent, he “ha[d] access to variety of decommis- sioned silver and gold coins.” PSR, at 19. And another said that when Kirschner flashed his badge and explained he was undercover, “I believed him,” and “[h]e seemed credible to me.” PSR, at 20. Kirschner contends he only impersonated a law enforce- ment agent so that his buyers would not try to rob him. Kirschner Br. 30; App. 174. But that contention is belied by the record. During his plea colloquy, Kirschner admitted that he falsely claimed to be a federal agent so that his victims would “feel more at ease.” App. 92. Kirschner’s fear-of-get- ting-robbed contention also is belied by logic. If impersonat- ing a federal law enforcement officer protected his counterfeits or the money earned from them from robbery, then by defini- tion the impersonation played a significant role in the execu- tion of his scheme. As this Court has held previously, we do not today hold that every fraud committed by law enforcement or those masquer- ading as them will justify an upward adjustment under § 3B1.3. defendants who sufficiently represented to their victims that they did. 17 Cf. United States v. Brann, 990 F.2d 98, 103 (3d Cir. 1993). Further, at the second step of Douglas, a defendant’s intention to assume and then abuse a position of trust to significantly facilitate a crime may not always warrant a Guidelines en- hancement. A defendant could pretend to be a police officer to facilitate some fraud, but the record could reveal that the de- fendant’s apparent position played no role in the fraud’s suc- cess or failure. Sentencing courts, acting pursuant to the Guidelines, must always determine whether the defendant’s apparent position significantly facilitated the commission of the crime or concealment of it. Under the facts of this case, the District Court reasonably found that Kirschner’s pretending to be a federal law enforcement agent contributed substantially to his ability to defraud his victims. Kirschner finally argues that the District Court’s abuse-of- trust enhancement double-counts his impersonation charge un- der § 912. But as the District Court found, there was no double counting because of the Guidelines’ grouping rules. App. 160; see also App. 61, ¶ 9 (agreeing in plea agreement to group of- fenses). Under those rules, Kirschner’s fraud and impersona- tion counts were grouped for sentencing purposes because U.S.S.G. § 2B1.1 governed sentencing for each 4 and because 4 Provision 2B1.1 expressly governed Kirschner’s fraud count. See § 2B1.1 (cross-referencing 18 U.S.C. § 912, the charged statute). Provision 2B1.1 also governed Kirschner’s imperson- ation count. While § 2J1.4 generally governs impersonation counts, § 2J1.4(c)(1) says that if the impersonation count “was to facilitate another offense,” the district court should apply the guideline for that other offense “if the resulting offense level” for the other offense “is greater than the offense level” for the impersonation offense. Here, the parties agreed that Kirschner’s impersonation of a federal agent “was to facilitate” his importation and sale of counterfeit goods. App. 59-60 (plea agreement). And “the resulting offense level” from the fraud count was “greater than the offense level” from the imperson- ation count. Kirschner’s impersonation of a federal agent car- ried an adjusted offense level of 8. App. 60, ¶ 5 (plea agree- ment); see also § 2B1.1(b)(1)(B). But the parties stipulated that the fraud count carried an adjusted offense level of at least 18 “the offense behavior was ongoing and continuous in nature.” App. 61, ¶ 9 (citing the grouping rules in § 3D1.2(d)). Kirschner’s sentence therefore did not account for his imper- sonation of a federal agent because sentencing for that crime was subsumed by sentencing for the counterfeiting crime. In any case, even if there was double counting, in this context double counting would have been fine. See United States v. Fisher, 502 F.3d 293, 309 (3d Cir. 2007). C. The District Court’s increase of Kirschner’s offense level for use of sophisticated means was not clear error Kirschner finally argues that the District Court erred in ap- plying the use-of-sophisticated-means enhancement. We dis- agree. That enhancement, set forth in Chapter 2 of the Sen- tencing Guidelines, tells a sentencing court that, if “the offense otherwise involved sophisticated means and the defendant in- tentionally engaged in or caused the conduct constituting so- phisticated means, increase [the offense level] by 2 levels.” U.S.S.G. § 2B1.1(b)(10)(C). The commentary to § 2B1.1 elaborates on what conduct might constitute “sophisticated means.” That note says, “‘sophisticated means’ means espe- cially complex or especially intricate offense conduct pertain- ing to the execution or concealment of an offense.” Id. at cmt. n.9(B). Like with the abuse-of-trust enhancement, the District Court heard sufficient evidence to find that Kirschner’s fraud employed sophisticated means. While Kirschner did not him- self create the counterfeits he sold or the intricate packaging in which he sold them, he deployed various other strategies to conceal his fraud. He used a pseudonym to conceal his iden- tity, and he created fake businesses, social media accounts, and sale invoices to give his scheme the veneer of legitimacy. App. 158-60. 20. App. 61, ¶ 8 (plea agreement). Section 2Bl.1 therefore governed both the impersonation and fraud counts. 19 Kirschner appears to oppose the enhancement on two grounds. Neither is decisive. First, Kirschner claims that his “criminal conduct was in fact highly unsophisticated” because it “involv[ed] selling out of the trunk of his car in a public park- ing lot, or in a public bank lobby.” Kirschner Br. 18-19. But that argument could only prove that Kirschner’s conduct could have been more sophisticated, which, without more, is not enough to upset the District Court’s finding. See, e.g., United States v. Lewis, 93 F.3d 1075, 1082 (2d Cir. 1996) (reversing finding that tax evasion scheme was not sophisticated, even though it did not employ shell corporations, defendant did not conceal his identity, and a “reasonably competent IRS auditor” would have uncovered it). Second, Kirschner observes that he did not create any of the packaging material used to sell his counterfeits himself; he just bought them on the internet. That argument again is inapt. For one, like with the locations he used for his sales, that Kirschner could have delivered his wares in more sophisticated packag- ing is not enough to upset the District Court’s finding. For an- other, that argument does not address the District Court’s find- ing that Kirschner’s fake business entities and the like played major roles in the commission and concealment of his crimes. III. CONCLUSION For the foregoing reasons, we will vacate the District Court’s judgment of sentence and remand for reconsideration of the loss amount. The District Court need not reconsider the two 2-level sentence increases for use of sophisticated means and abuse of trust. 20