United States v. Darra Shephard

892 F.3d 666

• Court: Court of Appeals for the Fourth Circuit
• Decided: June 15, 2018
• Docket: 17-4148
• Status: Published

Opinion

DIAZ, Circuit Judge:

Darra Lee Shephard conspired to defraud U.S. residents through a telemarketing sweepstakes scheme. From a call center in Costa Rica, Shephard and others would tell their unwitting victims that they had won a sweepstakes prize. The catch? Winners could receive their prize money only after paying a "refundable insurance fee." Victims of the scheme wired thousands of dollars to the call center, but of course, no fees were refunded and no prize money was ever paid.

Shephard pleaded guilty to conspiracy to commit wire and mail fraud, conspiracy to commit money laundering, eight substantive counts of wire fraud and aiding and abetting, and four substantive counts of money laundering and aiding and abetting. She was sentenced to ninety-six months in prison and ordered to pay $7,215,695.20 in restitution. On appeal, she challenges (1) the district court's application of the vulnerable victim sentencing enhancement and (2) its calculation of the actual loss amount. We affirm.

I.

A.

From 2007 through February 2015, Shephard "participated in, and otherwise facilitated," a sweepstakes telemarketing call center in Costa Rica for the purpose of defrauding U.S. residents by convincing them to pay money in order to claim a fake sweepstakes prize. J.A. 158.

The call center required "openers" to call residents in the United States and inform them that they had won second prize in a sweepstakes in the amount of $350,000 to $450,000. The "openers"-often falsely representing themselves as U.S. government employees-would explain that in order to receive the prize money, the U.S. resident must pay a "refundable insurance fee" by wiring money via Western Union, MoneyGram or an international bank. Victims then wired money to a "runner," who would collect wire transfers at banks in Costa Rica, or a "bridge," who would receive the funds in the United States and forward them along to the call center in Costa Rica.

Once someone fell prey to the scam and made an initial payment, a "loader" would call again and convince the victim to send more money. The "loader" would tell the victim that a mistake had been made and he had actually won first prize, which was more than $3,000,000. Of course, the victim would first need to pay thousands of additional dollars in increased insurance fees to claim this larger prize. If the victim fell for the scheme again and made another payment, the "loader" would continuously raise the sweepstakes prize and ask for more money from the same victim. This process is known as "reloading." At various times throughout her tenure, Shephard worked in the call center as an "opener," a "runner," and a "loader."

B.

Shephard was arrested in Costa Rica in July 2015 and then extradited to the United States. She pleaded guilty to all fourteen counts of her indictment without the benefit of a plea agreement. Shephard also agreed to a written factual basis for her plea, which described the sweepstakes scheme and her involvement. Shephard reserved her right to object to two proposed findings at sentencing: (1) that the loss amount was over $3,500,000 and (2) that she knew that many of the victims were over the age of fifty-five and vulnerable.

Shephard's presentence report ("PSR") relied on the factual basis, as well as two written statements from Shephard, to calculate the appropriate Guidelines range. The PSR applied a base offense level of 7 for wire fraud, a number of enhancements, and a three-level downward adjustment for acceptance of responsibility. For our purposes, only two of the enhancements matter: First, the PSR applied a two-level enhancement because the offense involved vulnerable victims. U.S.S.G. § 3A1.1(b)(1). Second, it applied an eighteen-level increase for a loss amount of more than $3,500,000 but less than or equal to $9,500,000. U.S.S.G. § 2B1.1(b)(1)(J). The PSR relied on the financial analysis prepared by the government, which explained that during Shephard's involvement the loss amount collected from victims via Western Union was $5,876,814.27 and the loss amount collected via MoneyGram was $1,338,880.93, resulting in a total loss amount of $7,215,695.20.

Shephard objected to the vulnerable victim enhancement and the loss amount calculation. She argued there was no evidence that the scheme targeted unusually vulnerable victims or that Shephard knew it targeted such victims. As for the loss amount, she claimed it should be less than $3,500,000 because the victims' losses were not adequately documented and her involvement in the call center was sporadic. The court overruled the objection to the vulnerable victim enhancement "based on responses from the government, consistent with my other rulings in other [related] call center cases that ... vulnerable victim enhancements were appropriate in the nature of how this fraud activity worked." J.A. 71. And it affirmed the loss amount as a "conservative estimate" based on "solid methodology." J.A. 70. The court further explained that Shephard was responsible for the criminal activity of her coconspirators even after her personal participation ended because she never affirmatively withdrew from the conspiracy.

The district court adopted the PSR's Guidelines calculation, which applied a total offense level of 34 and a criminal history of I for a Guidelines range of 151 to 188 months' imprisonment. The court then departed and varied downward to arrive at a sentence of ninety-six months' imprisonment and $7,215,695.20 in restitution.

Shephard filed a timely appeal.

II.

We review criminal sentences for reasonableness using an abuse of discretion standard. Gall v. United States , 552 U.S. 38 , 51, 128 S.Ct. 586 , 169 L.Ed.2d 445 (2007). A sentence based on an improperly calculated Guidelines range is procedurally unreasonable. United States v. Davis , 679 F.3d 177 , 182 (4th Cir. 2012). In reviewing whether a sentencing court properly calculated the Guidelines range, we review the court's factual findings for clear error and its legal conclusions de novo. Id.

The Sentencing Guidelines allow for a two-level enhancement "[i]f the defendant knew or should have known that a victim of the offense was a vulnerable victim." U.S.S.G. § 3A1.1(b)(1). A vulnerable victim is a person "(A) who is a victim of the offense of conviction and any conduct for which the defendant is accountable ... and (B) who is unusually vulnerable due to age, physical or mental condition, or who is otherwise particularly susceptible to the criminal conduct." U.S.S.G. § 3A1.1 cmt. n.2.

For the vulnerable victim enhancement to apply, the record must show that (1) the victim of the offense was unusually vulnerable and (2) the defendant knew or should have known of the victim's unusual vulnerability. United States v. Llamas , 599 F.3d 381 , 388 (4th Cir. 2010). The enhancement thus "requires a fact-based explanation of why advanced age or some other characteristic made one or more victims unusually vulnerable to the offense conduct, and why the defendant knew or should have known of this unusual vulnerability." Id. (internal quotation marks omitted).

In this case, the factual basis and Shephard's written statements to the probation officer explain that the conspiracy involved "reloading" victims and that Shephard herself worked as a "loader." The practice of "reloading" involves targeting people who have already fallen victim to the scheme at least once, if not repeatedly. The question, then, is whether this practice of reloading, at least in the context of this particular telemarketing sweepstakes conspiracy, is sufficient to support the vulnerable victim enhancement.

We have not yet addressed this issue, 1 but many of our sister circuits have, concluding that evidence of reloading may support the enhancement. See United States v. Lloyd , 807 F.3d 1128 , 1172-73 (9th Cir. 2015) ; United States v. Hoffecker , 530 F.3d 137 , 201-02 (3d Cir. 2008) ; United States v. Day , 405 F.3d 1293 , 1295-96 (11th Cir. 2005) ; United States v. Coe , 220 F.3d 573 , 582 (7th Cir. 2000) ; United States v. Brawner , 173 F.3d 966 , 972-73 (6th Cir. 1999) ; United States v. O'Neil , 118 F.3d 65 , 75-76 (2d Cir. 1997). 2

As these courts have explained, a victim's decision to wire money in response to the initial fraudulent phone call is evidence that they are "particularly susceptible to the criminal conduct" and that is why fraudsters then repeatedly target that victim through the process of "reloading." For example, in United States v. Jackson , which addressed a similar fraudulent telemarketing scheme, the Seventh Circuit held the vulnerable victim enhancement applied because the defendants "reloaded" victims. 95 F.3d 500 , 502-03 (7th Cir. 1996). The court reasoned:

Whether these persons are described as gullible, overly trusting, or just naive (and whatever the cause of this condition), ... their readiness to fall for the telemarketing rip-off, not once but twice (and in [one] case, four times), demonstrated that their personalities made them vulnerable in a way and to a degree not typical of the general population.... The structuring of the reloading process, the designating of customers as "reloads," ... the charging of higher sums to reloads, and the systematic re-victimizing of people ... amply demonstrate that the defendants believed the reloaded persons to be especially vulnerable and that they specifically targeted them because of this vulnerability.

Id. at 508 .

In other words, the reloading process was sufficient to support a finding that the victims were unusually vulnerable and that the defendants targeted them because of their vulnerability. The dissent contends that the district court there made findings about one particular victim. But the court did not just find one or two victims to be unusually vulnerable. Instead, the court "focussed on the technique of reloading" and concluded "those who were successfully reloaded" were unusually vulnerable. Id. at 507-08 . Likewise, the Third and Eleventh Circuits affirmed applications of the vulnerable victim enhancement based on the simple fact that reloading-the repeated targeting of a victim-"constitutes evidence that the defendant knew the victim was particularly vulnerable to the fraud scheme." Day , 405 F.3d at 1296 ; Hoffecker , 530 F.3d at 201 (quoting Day ).

We believe the reasoning of our sister circuits is sound and applies here. The record shows that the more experienced telemarketers in this fraudulent sweepstakes scheme, including Shephard, were specifically designated as "loaders." The "loaders" called victims who had already fallen for the scheme and tried to get them to make additional payments. Shephard explained that she and her coconspirators "would reload the victim as long as the victim was willing and able to continue to transfer funds" and that "[m]any victims sent tens of thousands of dollars" in response to these calls. J.A. 442. Further, the individual victim impact statements attached to the presentence report demonstrate, as one would suspect, that the victims were in fact unusually vulnerable. Thus, we find no clear error in the district court's determination that some victims were particularly vulnerable and that Shephard was aware of this vulnerability. It is why she and her coconspirators purposefully targeted such victims again and again. 3

Shephard also argues that the district court did not sufficiently explain its decision to apply the vulnerable victim enhancement. A district court must provide a "sufficient explanation of its rationale" to enable review, United States v. Wilkinson , 590 F.3d 259 , 269 (4th Cir. 2010), and must "place on the record an individualized assessment based on the particular facts of the case before it," United States v. Carter , 564 F.3d 325 , 330 (4th Cir. 2009) (internal quotation marks omitted). The district court here overruled Shephard's objection to the vulnerable victim enhancement "based on responses from the government, consistent with my other rulings in other [related] call center cases that there were vulnerable victim enhancements were appropriate in the nature of how this fraud activity worked. [sic]" J.A. 71.

Although the district court's explanation was terse, we find it sufficient. First, the court based its decision to apply the enhancement on the government's response. The government's response to Shephard's objection was that "the practice of reloading victims evinces an offense reliant on unusually vulnerable victims." J.A. 380. The government fully explained the argument and cited a number of cases. Second, the court said the enhancement was appropriate given "the nature of how this fraud activity worked," which further confirms that the court found the practice of reloading sufficient to support the enhancement.

We therefore affirm the district court's decision to apply the vulnerable victim enhancement.

Shephard also objects to the loss amount used to justify an eighteen-level enhancement. Shephard argues that the district court clearly erred in finding a loss amount of $7,215,695.20. In her view, the real loss amount is under $3,500,000, which should have resulted in a sixteen rather than eighteen-level enhancement. See U.S.S.G. § 2B1.1(b)(1)(I)-(J).

We review a district court's calculation of loss amount for clear error. United States v. Jones , 716 F.3d 851 , 859-60 (4th Cir. 2013). "Loss" is "the greater of actual loss or intended loss." U.S.S.G. § 2B1.1 cmt. n.3(A). Here, the court relied on "actual loss," defined as "the reasonably foreseeable pecuniary harm that resulted from the offense." U.S.S.G. § 2B1.1 cmt. n.3(A)(i). The government must establish the amount of loss by a preponderance of evidence. United States v. Catone , 769 F.3d 866 , 876 (4th Cir. 2014). But the court need only make a "reasonable estimate" of the loss. U.S.S.G. § 2B1.1 cmt. n.3(C).

In the case of a conspiracy, loss is attributable to a defendant "if it results from the conduct of others so long as the conduct was in furtherance of, and reasonably foreseeable in connection with the criminal activity." United States v. Otuya , 720 F.3d 183 , 191 (4th Cir. 2013) (internal quotation marks omitted). "A defendant's membership in a conspiracy is presumed to continue until he withdraws from the conspiracy by affirmative action. Withdrawal must be shown by evidence that the defendant acted to defeat or disavow the purposes of the conspiracy." United States v. Allmendinger , 706 F.3d 330 , 341 (4th Cir. 2013) (internal quotation marks omitted). "[M]ere cessation of activity in furtherance of the conspiracy is insufficient." United States v. Walker , 796 F.2d 43 , 49 (4th Cir. 1986). The defendant must point to "[a]ffirmative acts inconsistent with the object of the conspiracy and communicated in a manner reasonably calculated to reach co-conspirators." United States v. U.S. Gypsum Co. , 438 U.S. 422 , 464-65, 98 S.Ct. 2864 , 57 L.Ed.2d 854 (1978). The burden of proving withdrawal rests with the defendant. Walker , 796 F.2d 43 , 49 (4th Cir. 1986).

Shephard first argues that the loss amount is wrong because it includes MoneyGram wire transfers even though the substantive wire fraud and money laundering counts in the indictment speak only of Western Union transfers. But this argument fails because the factual basis and Shephard's written statements to the probation officer clearly establish that MoneyGram transfers were part of the conspiracy described in counts 1 and 10 of the indictment.

Next, Shephard argues that the calculation wrongly includes losses accrued through February 2015 even though she stopped working at the call center in December 2012. Shephard, however, stipulated in the factual basis supporting the plea that she participated in, or otherwise facilitated, the conspiracy from 2007 through February 2015. And although Shephard reserved the right to object to the loss amount and whether she knowingly targeted vulnerable victims, she twice told the court that she agreed "with everything else" in the factual basis. J.A. 49-50. This express acknowledgment of her involvement in the conspiracy through February 2015 is a sufficient basis for including losses through February 2015 in the loss calculation. See United States v. Gilliam , 987 F.2d 1009 , 1013 (4th Cir. 1993) (finding government may use defendant's acknowledgements during the Rule 11 colloquy or sentencing proceedings to establish facts for sentencing purposes).

Further, even if Shephard stopped actively participating in the conspiracy in December 2012, there is no evidence that she acted to defeat or disavow the purposes of the conspiracy. Withdrawal requires more than the mere cessation of activity in furtherance of the conspiracy. Walker , 796 F.2d at 49 . The record indicates Shephard's involvement in the conspiracy was "on and off" and that, at times, she held legitimate jobs. Without evidence of an affirmative withdrawal, we don't know if Shephard was taking another temporary hiatus from the call center or if she had really abandoned the criminal enterprise.

Finally, Shephard claims the district court didn't make any particularized findings that the MoneyGram transfers or post-2012 Western Union transfers were within the scope of the criminal activity Shephard agreed to jointly undertake. No such findings are necessary, however, because it's clear on the face of the indictment, as well as the factual basis, that the conspiracy involved MoneyGram transfers. Further, the factual basis establishes Shephard's involvement in the conspiracy through February 2015 and, even if she did leave earlier, there's no allegation that the conspiracy changed in any unforeseeable way after December 2012. And as we've noted, absent proof of an affirmative withdrawal from the conspiracy, Shephard is liable for the losses her coconspirators continued to inflict. 4

III.

For the reasons given, we affirm the district court's judgment.

AFFIRMED

WYNN, Circuit Judge, dissenting in part:

I agree with the majority opinion's approach to calculating the loss amount in this case, and I join fully in that portion of the majority opinion. I write separately, however, because I disagree with the majority opinion's holding regarding the vulnerable victim enhancement, U.S.S.G. § 3A1.1(b)(1). I do not believe the district court sufficiently explained its rationale for imposing that enhancement, which would therefore lead me to vacate the Defendant's sentence and remand for resentencing.

The majority correctly cites the controlling test for whether the vulnerable victim enhancement applies, although the majority's paraphrasing of the law misses a key aspect of the inquiry. This Court has said that applying the vulnerable victim enhancement requires satisfying a two-part test. See United States v. Llamas , 599 F.3d 381 , 388 (4th Cir. 2010) [" Llamas I "]. "First, a sentencing court must determine that a victim was unusually vulnerable. Second, the court must then assess whether the defendant knew or should have known of such unusual vulnerability." Id. (internal citation omitted). This discussion makes clear that the district court must identify at least one specific victim who serves as a basis for applying the enhancement-not that the record must merely reflect such possible applicability-and then also find that the defendant knew or should have known of that victim's unusual vulnerability. The next sentence in Llamas I makes this requirement all the more clear: "In other words, applying the vulnerable victim adjustment 'requires a fact-based explanation of why advanced age or some other characteristic made one or more victims unusually vulnerable to the offense conduct, and why the defendant knew or should have known of this unusual vulnerability.' " Id . (quoting United States v. Vega-Iturrino , 565 F.3d 430 , 434 (8th Cir. 2009) (discussing importance of "evidentiary support" particularized to the specific victim or victims, such as a victim who "had received only an eighth grade education [and] had no experience or understanding of investments outside of connection with the defendant") ).

Additionally, as the majority opinion correctly recognizes, when a district court applies a sentencing enhancement-like the vulnerable victim enhancement-the "district court must provide a 'sufficient explanation of its rationale' to enable review." Ante at 672 (quoting United States v. Wilkinson , 590 F.3d 259 , 269 (4th Cir. 2010) ). And a district court "must 'place on the record an individualized assessment based on the particular facts of the case before it.' " Id. (quoting United States v. Carter , 564 F.3d 325 , 330 (4th Cir. 2009) (emphasis added) (internal quotation marks omitted) ). In light of this standard, the majority opinion acknowledges that the district court's explanation for applying the vulnerable victim enhancement in this case was "terse." Id. at 672. Nonetheless, the majority opinion finds the district court's rationale "sufficient" to pass muster. Id. I disagree.

To begin, I want to emphasize just how "terse" the district court's rationale for imposing the enhancement was. The district court's entire explanation for imposing the enhancement reads as follows:

I will overrule that objection based on responses from the government, consistent with my other rulings in other Bonner call center cases that there were vulnerable victim enhancements were [sic] appropriate in the nature of how this fraud activity worked

J.A. 71. Parsing out this rationale makes clear that it is not sufficiently "tailored to the particular case at hand ... to permit 'meaningful appellate review.' " Carter , 564 F.3d at 330 (quoting Gall v. United States , 552 U.S. 38 , 50, 128 S.Ct. 586 , 169 L.Ed.2d 445 (2007) ).

I start with the latter portion of the district court's rationale: that the decision to apply the enhancement was "consistent with my other rulings in other Bonner call center cases that ... vulnerable victim enhancements were appropriate in the nature of how this fraud activity worked." J.A. 71. Here, the district court alludes to Defendant's co-defendants, Jeffrey Bonner (who ran the call center) and another individual who worked there, as well as other defendants in related cases. But the government did not have transcripts prepared of the sentencings for Defendant's co-defendants, nor has the government provided us with transcripts from any of the related cases. The district court's explanations in those cases are thus not before this Court. Consequently, although the district court knew what reasoning was provided for sentencing Defendant's co-defendants-reasons that may have included significant factual findings-that information has not been conveyed to this Court in the record presented to us. Accordingly, the district court's reference to the other cases is a quintessential example of when a record does not allow for "meaningful appellate review." Carter , 564 F.3d at 330 . From the documents provided, this Court simply has no way of knowing what rationale the district court may have used in those other cases.

The closest the district court comes to explaining its rationale in the previous cases was to say that the enhancement was appropriate, given "the nature of how this fraud activity worked." J.A. 71. This statement, albeit vague, corresponds with the other basis listed by the district court for its holding-the "responses from the government." Id. The government laid out its position on the enhancement in two places: the "United States' Response to Defendant's Objections to the Draft Presentence Investigation Report" and the Final Presentence Report. J.A. 379-81, 453-454. * At the sentencing hearing, no party offered substantive argument on the issue, instead relying on their written submissions. J.A. 71. So, when the district court refers to the "responses from the government," it presumably is referring to these written filings, both of which, like the district court's comment about the "nature" of the fraud, relied on a generalized argument about the structure of the sweepstakes scheme. Id.

To give greater detail, in these filings, the government argued that Defendant "would seek out individuals gullible enough to believe her false promises and send money overseas, and then repeatedly victimize those same individuals until they had nothing left to send." J.A. 379-80. The government then argued that, "[i]t is well established that the practice of reloading victims evinces an offense reliant on unusually vulnerable victims." J.A. 380. To support that proposition, the government listed one case from the Fourth Circuit, two from the Ninth Circuit, and one case from the Seventh Circuit. See id. These cases do not bear the weight claimed, however.

The sole Fourth Circuit case cited by the government for that proposition deserves some introduction. The government cited United States v. Llamas , 473 Fed.Appx. 176 (4th Cir. 2012) [" Llamas II "], an unpublished opinion rendered by a panel of this Court after Llamas I vacated the defendant's sentence on grounds that the district court failed to adequately support the application of the vulnerable victim enhancement. The facts of the Llamas cases are strikingly similar to those here. The defendant in Llamas was part of a sweepstakes scheme based out of Costa Rica that operated in a manner almost indistinguishable from the call center in this case. Llamas I , 599 F.3d at 383-85 . At sentencing, the district court identified two specific victims that it viewed as unusually vulnerable: one had Alzheimer's disease; the other "had been expecting a call from a doctor regarding her daughter's health," which put her in a vulnerable, high-stress state of mind. Id. at 386 . Critically, however, the district court did not make any findings about how the defendant knew or should have known about those vulnerabilities. Id. at 389 . And for that reason, this Court found that the district court did not provide adequate reasons for applying the enhancement and vacated the defendant's sentence. Id.

Thus, the district court in Llamas I actually gave more reasons than the district court did in this case-and this Court nonetheless vacated the defendant's sentence. The district court in Llamas I identified specific victims and gave a detailed explanation as to why those victims were vulnerable. Those vulnerabilities were individualized to specific victims, not just blanket assertions premised on how the general structure of the fraudulent scheme targeted gullible people. In fact, had record evidence of a reloading scheme been sufficient, by itself, to apply the enhancement-as the majority opinion appears to conclude-the Llamas I Court would have upheld the application of the vulnerable victim enhancement, because the scheme in that case relied on a reloading system almost identical to the one here. See id. at 384 . Instead, this Court vacated the sentence because the district court's findings were insufficient. See id. at 389 .

Here, in contrast with Llamas I , the district court did not identify any specific victims as unusually vulnerable or make any findings as to why Defendant had knowledge of that vulnerability. The district court merely referenced the government's written submissions, which themselves relied on a vague theory that reloading schemes target gullible people. I do not discount the possibility that an individual could be so gullible to qualify as unusually vulnerable. But, at the very minimum, the Guidelines require the district court to identify the specific victims it found to be so vulnerable that they would serve as a basis for applying the enhancement. As the case stands, however, nothing in the record indicates which victims the government or the district court believed supported application of the enhancement. We simply know that some individuals (although who specifically, we do not know) gave money more than once to the scheme, which, according to the government, makes them unusually vulnerable (despite not knowing any more details about why these victims contributed to the scheme multiple times or any other information supporting this alleged vulnerability).

The approach to the enhancement I am describing also best accords with the enhancement's application note. That commentary states: "The adjustment would apply, for example, in a fraud case in which the defendant marketed an ineffective cancer cure or in a robbery in which the defendant selected a handicapped victim. But it would not apply in a case in which the defendant sold fraudulent securities by mail to the general public and one of the victims happened to be senile." U.S.S.G. § 3A1.1 cmt. n.2. The case at issue here is analogous to the situation described in the second part of the application note-at least as the case is currently described by the government and the district court. Defendant's co-conspirators called members of the general public and sought money. Presumably, many would decline, while some would provide the requested money. As the application note example contemplated, however, some of the people called would be gullible enough to fall for the scheme. But, at least upon initial phone contact, the coconspirators had no way of differentiating the gullible from the savvy. Once someone gave to the scheme, however, then the co-conspirators would come back to that person over and over again. Returning to previous victims potentially could cross the line to exploiting a vulnerable victim, but the key problem is that we do not know who those specific vulnerable victims were in this case. The record does not explain which victims were targeted multiple times, what actions the conspiracy took toward those specific individuals, and whether or how Defendant knew or should have known about those particular victims' vulnerabilities.

The majority opinion claims the victim impact statements demonstrate that there were unusually vulnerable victims, but the statements do not provide that level of detail. For example, the statements do not ask questions that would indicate who was reloaded, which, according to the government, is the key indicator of vulnerability. The other documents in the record are similarly vague on this point. For example, the government provided the district court with various charts detailing the total amounts of money taken from each victim, but these documents do not also reveal who was reloaded, because they do not show discrete transactions. An appellate court reviewing this record is left to guess who was reloaded based upon the total money an individual lost to the scheme. Furthermore, even if this Court could discern who specifically was reloaded from these documents, we have no additional details explaining what may have motivated an individual to contribute to the scheme more than once and how Defendant would have known of that person's alleged vulnerability. Without directing us toward specific information about particular victims, the district court's findings do not satisfy the standard established by this Court in Llamas I . See 599 F.3d at 387-89 .

That backdrop-the binding precedent of Llamas I and its impact on this case-sets the stage for the government's filings below, in which it cited the unpublished opinion in Llamas II to support the allegedly "well established" principle of applying the vulnerable victim enhancement to reloading schemes. J.A. 380. Llamas II was issued after the district court resentenced the defendant and again applied the vulnerable victim enhancement. 473 Fed.Appx. at 177-78 & n.1. Significantly, the defense did not challenge the application of the enhancement in the second appeal; it only challenged the district court's consideration of the sentencing factors found in 18 U.S.C. § 3553 (a) for procedural and substantive reasonableness. See id. at 179 . In fact, this Court specifically said, "[b]ecause Llamas does not challenge his Guidelines calculation, we do not need to address that stage of the resentencing." Id. at 178 n.2. Nonetheless, in its filings below, the government described this case as "affirming [a] sentence incorporating [the] vulnerable victim adjustment against a similarly situated Costa Rican telemarketing fraud defendant." J.A. 380. Although technically true, the government's characterization omits the crucial fact that the panel in Llamas II expressly declined to opine on the appropriateness of the enhancement's application, because that issue was not appealed. Thus, even if it were binding precedent, Llamas II does not carry the weight the government tries to place on it. The Fourth Circuit has not previously applied the vulnerable victim enhancement in the way the majority opinion appears to do here. To the contrary, the binding case on this issue- Llamas I -undercuts the government's argument by requiring detailed findings from the district court. A panel of this Court has since reiterated that holding. See United States v. Johnson , 480 Fed. App'x. 186, 189 (4th Cir. 2012) (quoting Llamas I for the proposition that "our precedent stresses the importance of an adequate explanation" for applying the vulnerable victim enhancement and then finding application of the enhancement procedurally unreasonable because the district court "failed to make explicit findings on the record as to the victims, their unusual vulnerability, and [the defendant's] knowledge of such vulnerability"). Yet the government appears to take great pains to avoid citing a published case that diminishes its argument, instead relying on a misleading characterization of a subsequent unpublished opinion to make its point.

The government's filings below also relied on three other cases: one from the Seventh Circuit and two from the Ninth Circuit. The Seventh Circuit case, like Llamas , does not help the government's argument. See United States v. Jackson , 95 F.3d 500 (7th Cir. 1996). The Seventh Circuit in Jackson upheld the application of the vulnerable victim enhancement in a case factually similar to the one here. See id. at 502-04, 506-08 . Unlike in this case, however, the district court in Jackson identified at least one particular victim as vulnerable, detailed the conspiracy's actions to take advantage of that person, and explained why the defendant should have known about that particular victim's vulnerability. See id. at 507-08 . Thus, the court in Jackson made the critical factual findings absent in this case.

Contrary to the majority opinion's characterization of Jackson , the vulnerability of specific victims played a key role in the outcome of that case. The Seventh Circuit articulated the enhancement's requirements as being that "the district court must find both 1) that a victim of the defendant's crime was unusually vulnerable in some way, and 2) that the defendant targeted that victim because of this vulnerability." 95 F.3d at 507 (emphases added). The court's subsequent explanation follows that standard. For example, throughout the opinion, the court discusses J.Z., one of the scheme's victims. See id. at 504, 507-08 . The opinion details how J.Z. was defrauded into contributing to the scheme four times and the amounts he contributed on each occasion. See id. at 504 . Thus, although the Seventh Circuit may have appeared to speak in general terms-which the majority is quick to emphasize-the court's analysis was firmly grounded in the specifics of a particular victim's case. See, e.g. , id. at 508 ("[T]he district court clearly determined that [the victims'] readiness to fall for the telemarketing rip-off, not once but twice (and in J.Z.'s case, four times), demonstrated that their personalities made them vulnerable in a way and to a degree not typical of the general population."); id. ("[T]he systematic re-victimizing of people like J.Z. amply demonstrate[s] that the defendants believed the reloaded persons to be especially vulnerable.").

Only the Ninth Circuit cases cited by the government may support its argument, because it is unclear from the appellate opinions whether the district court made findings about particular victims. See United States v. Durant , 89 Fed.Appx. 56 , 58 (9th Cir. 2004) ; United States v. Randall , 162 F.3d 557 , 560-61 (9th Cir. 1998). Yet the import of that precedent is unclear, because, for example, a more recent Ninth Circuit opinion goes into significant detail about the specific findings made about particular victims and why the defendant in a reloading scheme knew about the vulnerabilities of those victims. See United States v. Lloyd , 807 F.3d 1128 , 1172-73 (9th Cir. 2015).

To the extent that my position splits with that of the Ninth Circuit in Durant and Randall , I believe the Ninth Circuit's approach, as described in the cases cited by the government below, are at odds with this Circuit's binding precedent in Llamas I . The majority opinion also cites cases from other circuits to support the proposition that "evidence of reloading may support the enhancement." Ante at 671 (collecting cases). The cases cited do support the proposition that a reloading scheme may support applying the vulnerable victim enhancement-a conclusion with which I agree. My dispute with the majority opinion stems from whether a reloading scheme is sufficient to warrant applying the enhancement. The majority appears to argue that it is, thereby creating a broad rule that would apply the enhancement to any case where a victim falls prey to a similar scheme more than once-without any limiting principle in sight. In contrast, I argue that a district court must make adequate factual findings about at least one specific victim to justify applying the enhancement. The cases cited by the majority, like the Ninth Circuit opinions in Durant and Randall , leave unclear the extent to which the relevant district courts may have made specific findings about particular victims. See United States v. Hoffecker , 530 F.3d 137 , 201-02 (3d Cir. 2008) ; United States v. Day , 405 F.3d 1293 , 1295-96 (11th Cir. 2005) ; United States v. Coe , 220 F.3d 573 , 582 (7th Cir. 2000) ; United States v. Brawner , 173 F.3d 966 , 972-73 (6th Cir. 1999) ; United States v. O'Neil , 118 F.3d 65 , 75-76 (2d Cir. 1997). Thus, as the majority opinion notes, these cases stand merely for the proposition that a reloading scheme can support applying the enhancement; they do not address the key distinction between my view and that of the majority. That said, to the extent any of those cases indeed uphold application of the enhancement absent particularized findings about specific victims, I believe that approach conflicts with this Circuit's binding precedent, as articulated in Llamas I .

Furthermore, the approach taken by the Ninth Circuit in Durant and Randall diverges from that of several other circuits. The Second Circuit, for example, has spoken particularly well on the matter in United States v. Dupre , 462 F.3d 131 (2d Cir. 2006). In that case, co-conspirators targeted evangelical Christians by using Christian imagery and rhetoric to ask them to provide money again and again to a fraudulent investment scheme. See id. at 134-36, 144-45 . Although such "[s]chemes ... inevitably attract the gullible," the Second Circuit rejected applying the vulnerable victim enhancement in the case solely based on such a generalized argument. Id. at 145 . The court specifically said that "a criminal successfully asking a fraud victim for additional money does not itself demonstrate that the victim is especially vulnerable." Id. at 146 . The court further elaborated:

We have no reason to believe that evangelical Christians as a class are 'unusually susceptible' to fraud. The application notes reiterate that the vulnerable victims enhancement is improper except in cases where defendants 'should have known' of their victims' unusual vulnerability, such as 'in a fraud case in which the defendant marketed an ineffective cancer cure'.... Absent findings by the District Court that any specific victim was especially gullible because of his religion, we cannot conclude that evangelical Christian victims were susceptible to fraudulent schemes involving religious imagery in a manner analogous to how desperate cancer patients might be susceptible to con artists selling placebos.

Id. at 145 (emphasis added) (internal footnote omitted) (quoting U.S.S.G. § 3A1.1 cmt. 2). This case presents the same situation. "Absent findings by the District Court that any specific victim was especially gullible," we cannot meaningfully review this record and find that the enhancement was appropriately applied. Id. (emphasis added).

My good colleagues in the majority fail to grasp the basis of my reliance on Llamas I and Dupre. Indeed, I do not dispute that evidence of a reloading scheme can support application of the vulnerable victim enhancement, nor do I argue that Llamas I and Dupre bar reliance on such a scheme as a basis for applying the enhancement. Rather, I rely on Llamas I and Dupre for the proposition that a court may not apply the vulnerable victim enhancement unless the government introduces victim-specific , as opposed to generalized, evidence. Notably, like Llamas I and Dupre , several other circuit opinions have vacated sentences applying the vulnerable victim enhancement when a district court failed to make factual findings regarding specific victims. See, e.g. , United States v. Angeles-Mendoza , 407 F.3d 742 , 747-48 (5th Cir. 2005) (vacating application of the enhancement when the district court's "scant, generalized findings" mentioned only "general characteristics commonly held" by the victims and "inherent" in the class of victims targeted); United States v. Anderson , 349 F.3d 568 , 573 (8th Cir. 2003) (vacating application of the enhancement when the "general, not thorough" explanation provided below gave "no detailed analysis of why th[e] evidence established that [the defendant] knew or should have known that any victim was unusually vulnerable").

Accordingly, our binding precedent in Llamas I required the district court to identify at least one specific victim who served as the basis for applying the vulnerable victim enhancement and to explain why the defendant knew or should have known of that particular victim's unusual vulnerability. The district court did not do so in this case. Therefore, I would vacate the Defendant's sentence and remand her case for resentencing. Because the majority opinion reaches a different conclusion, I respectfully dissent.

1 The dissent claims that our decision in Llamas is controlling and that if record evidence of a reloading scheme were sufficient to uphold the application of the vulnerable victim enhancement, we would have done so there. In Llamas , however, we rejected the district court's application of the enhancement because, after discussing how one victim took medication for Alzheimer's disease and another fell for the scheme when she was expecting a call about her daughter's health, the court provided no explanation as to why the defendant knew or should have known about these two victims' unique vulnerabilities. Llamas I , 599 F.3d at 388-89 .

When the government on appeal argued the enhancement was appropriate because Llamas "reloaded" persons who had already fallen for the sweepstakes scheme, we did not reject the argument-we just noted that the district court never adopted it. We said the enhancement "cannot be justified simply because there might be some evidence in the record-not addressed by the sentencing court-supporting [it]." Id. at 389 . Here, by contrast, the sentencing court relied on the practice of reloading as evidence to support its application of the vulnerable victim enhancement. See infra at 672-73.

2 The dissent cites the Second Circuit's decision in United States v. Dupre , 462 F.3d 131 (2d Cir. 2006). There, the court did note that simply asking a fraud victim for more money does not warrant application of the vulnerable victim enhancement, but it expressly left open the possibility that "repeated targeting of certain victims" could justify the enhancement, as it did in O'Neil . Id. at 146 .

3 We need not (and do not) decide whether the mere practice of reloading (without more) will always justify applying the vulnerable victim enhancement. There may be unique factual scenarios that rebut the strong inference that defendants who reload victims target them because they are unusually vulnerable. But here, the district court's determination to apply the enhancement was entirely appropriate.

4 The parties dispute whether the spreadsheet of wire transfers that the government entered as an exhibit at sentencing was complete. We are satisfied that the exhibit as presently constituted, in conjunction with the rest of the record, establishes a total loss of $7,215,695.20 based on sound methodology. Accordingly, we deny the government's motion to supplement the Joint Appendix and also deny as moot Shephard's motion to strike the same.

The latter version filed with the Final Presentence Report quoted almost entirely from the earlier response and is not different from the prior version in any material way. Thus, for the sake of consistency, this opinion will quote from the first version, since that is the version quoted by the majority opinion.