APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF PUERTO RICO

[Hon. John A. Woodcock, U.S. District Judge] [Hon. Gustavo A. Gelpí, U.S. District Judge]

Before

Rikelman, Lynch, and Howard, Circuit Judges.

Alejandra Bird-López, Assistant Federal Public Defender, with whom Héctor L. Ramos-Vega, Interim Federal Public Defender, District of Puerto Rico, and Franco L. Pérez-Redondo, Assistant Federal Public Defender, Supervisor, Appeals Section, were on brief, for appellant.

Katherine Twomey Allen, with whom W. Stephen Muldrow, United States Attorney, Mariana E. Bauzá-Almonte, Assistant United States Attorney, Chief, Appellate Division, Gregory B. Conner, Assistant United States Attorney, and United States Attorney's Office were on brief, for appellee.

November 14, 2025

Page 2RIKELMAN , Circuit Judge . Carlos Maldonado appeals his securities fraud conviction and sentence. He contends that the district court improperly admitted at trial "summaries" of his bank records under Federal Rule of Evidence 1006, even though they were rife with hearsay, and that insufficient evidence supported the jury's verdict. He also argues that the court miscalculated his sentence because it incorrectly determined the number of victims and their total losses. Finally, he claims that the court was wrong to order him to pay restitution based on harm to individuals not identified in the indictment. Our review reveals no basis for disturbing the jury's verdict, the sentence imposed, or the restitution order. As a result, we affirm.

In 2005, Maldonado formed a company called Business Planning Resource International Corporation (BPRIC). BPRIC's purported mission was to raise money to develop various businesses and to help individuals "facilitate the resources [they] need to maximize [their] productive development." Promoters for BPRIC, some of whom were insurance agents, persuaded clients 1 to enter Page 3into financial agreements with the company. These agreements were generally called "Productive Development Contracts." Each client made an initial monetary contribution and received the promise of a fixed rate of return. But the size of the initial contributions, the amount of earnings guaranteed, and the duration of the agreements varied. For example, one client agreed to give BPRIC $15,000 in exchange for 17% annual earnings for five years. Another client agreed to pay BPRIC $50,000 in exchange for 50% annual earnings for one year.

Ultimately, BPRIC failed to fulfill its contractual obligations to its clients. In 2016, a grand jury indicted Maldonado on one count of securities fraud and fifteen counts of bank fraud. See 15 U.S.C. §§ 78j(b), 78ff(a); 18 U.S.C. §§ 2, 1344. The government alleged in the indictment that Maldonado ran a Ponzi scheme, meaning that he recycled money from later investors to pay earlier investors, even though BPRIC never turned a profit. The indictment listed twelve transactions involving eight alleged clients (some of which were households). Each of these individuals (or a representative from each household) subsequently testified at trial. The amounts they invested ranged from $10,000 to $164,000. The indictment also alleged that, "by depositing the investors' checks in the bank account for BPRIC," Maldonado committed bank fraud against the individuals' financial institutions.

Page 4Maldonado's case proceeded to trial, which lasted nine days and included twenty-three witnesses for the government. One of the government's first witnesses was an attorney who testified generally about Ponzi schemes. He explained that a Ponzi scheme involves shuffling money from one investor to reimburse another, instead of conducting profit-generating activity.

The government presented fifteen witnesses who had provided money to Maldonado.² It introduced each witness's Productive Development Contract, checks they issued to and received from BPRIC, BPRIC promotional materials they obtained, and their correspondence with BPRIC and Maldonado. Each witness explained that they believed they were "investing" their money in BPRIC. Specifically, many witnesses shared an understanding that BPRIC would use their money to develop other companies. Most witnesses testified that they received information about specific companies that BPRIC would invest in or that they could invest in directly. Those companies included, among others, Datavos, Pet Card Systems, and Glorimar Fashions and Tailoring LLC. At various points, Maldonado also touted to prospective and current clients the potential profits that could be earned through ventures such as investments in "Chinese bonds" and mining operations in Brazil.

Page 5Some witnesses testified that Maldonado told them BPRIC would grow their money through real property or stock market investments.

Alwin Díaz, one of Maldonado's promoters, also testified for the government. Díaz said that Maldonado conducted a training for promoters and claimed to operate "five or six other companies" "under" BPRIC, including Pet Card Systems and Datavos. According to Díaz, Maldonado showed the promoters "how each one of those companies was supposed to operate" so that they could share this information with potential clients. Specifically, Maldonado taught promoters to tell prospective clients that BPRIC would "distribute [their] money among each one of the companies that were under" BPRIC in order to generate earnings. Díaz also testified that Maldonado instructed promoters not to use the word "investment," because "this word would entail other things," and to use the phrase "capital accumulation" instead.

To demonstrate how Maldonado used the clients' money, the government introduced the testimony of Mirelis Domínguez, a forensic accountant for the Federal Bureau of Investigation. Domínguez analyzed approximately 10,000 transactions in Maldonado's bank accounts. She used this information to create a spreadsheet, Exhibit 327, which the district court admitted into evidence. The spreadsheet included a "general category" column, in which Domínguez assigned a label to each transaction -- for example, "Investor," "Promoter," or "Investment." When the Page 6government asked Domínguez how she categorized transactions, she explained that she consulted "supporting documentation," including "interviews" -- apparently referring to government interviews of potential victims, which she testified that she "reviewed" and in some cases "participated in."

Domínguez then used this spreadsheet to generate several additional tables and graphs, including pie charts and flow charts. For example, one table, Exhibit 306, listed each individual whom Domínguez had categorized as an investor (not just those who had testified at trial), their cash inflows and outflows, and their net gains or losses. It also included "comments," such as "[i]nterviewed by AUSA." A separate pie chart exhibit showed that 93% of the money entering Maldonado's accounts came from clients. Another pie chart exhibit showed that over half of the money leaving Maldonado's accounts went to clients and promoters. Domínguez also examined bank accounts with "just a few transactions," including accounts for Datavos, Glorimar Fashions, and Pet Card Systems. For these accounts, Domínguez created smaller tables summarizing their activity.

The government introduced Domínguez's exhibits under Rule 1006 of the Federal Rules of Evidence, which permits "summaries" of voluminous evidence. See Fed. R. Evid. 1006. Maldonado objected to several of the exhibits, including Exhibits 306 and 327. He argued that, by relying on interviews with alleged Page 7victims, Domínguez impermissibly incorporated hearsay into the exhibits. Maldonado also contended that several exhibits included Domínguez's assumptions and embellishments and therefore those exhibits were not neutral summaries as required under the evidence rules. The district court overruled the objections.

The jury convicted Maldonado on all counts, and the parties then prepared for sentencing. Ultimately, the U.S. Probation Office drafted a Presentence Investigation Report (PSR), and Maldonado objected to several of its findings. Specifically, he objected to the probation officer's application of an 18-level enhancement for "loss exceed[ing] $3,500,000" and to the application of a six-level enhancement "[b]ecause the offense resulted in substantial financial hardship to twenty-five or more victims." According to Maldonado, in calculating the total loss, the district court could consider only the financial harm to "the alleged victims that testified at trial." The court rejected this argument, labeling it a "blanket" objection, and explained that it determined a "loss amount based on both victims who testified during trial as well as those others mentioned in the PSR." In the end, the court sentenced Maldonado to 135 months' imprisonment.

The district court also ordered Maldonado to pay restitution in the amount provided for by the PSR, which listed forty-four victims with a collective loss of just over $2.1 million. The court explained that it would require restitution Page 8under the Mandatory Victim Restitution Act of 1996 (MVRA), 18 U.S.C. § 3663A, which applies to defendants convicted of certain crimes, including bank fraud. See id. § 3663A(a)(1). It noted, however, that the MVRA did not apply to Maldonado's securities fraud conviction. Accordingly, the court ordered restitution for the securities fraud conviction as a condition of Maldonado's supervised release under 18 U.S.C. § 3583. See id. §§ 3583(d), 3563(b)(2).

Maldonado timely appealed. In March 2025, less than two weeks before we heard oral argument in this case, the government filed a Rule 28(j) letter indicating that it was no longer pursuing Maldonado's bank fraud convictions and agreed that we should vacate those convictions.³

We now analyze each of Maldonado's legal claims, moving chronologically through the stages of his prosecution. First, Page 9Maldonado argues that the district court improperly admitted evidence summarizing his bank account activity under Federal Rule of Evidence 1006. We conclude that any error in admitting this evidence was harmless. Second, Maldonado challenges the jury's finding that the Productive Development Contracts were securities, contending that insufficient evidence supported this conclusion. We disagree, both because Maldonado misreads our precedent defining "investment contracts" and because sufficient evidence supported the jury's verdict. Finally, Maldonado challenges the district court's loss calculation and its restitution order. We conclude that the court's loss calculation was fully supported by reliable evidence, and that Maldonado has failed to show that the restitution award was legally invalid.

A. Evidence Admitted Under Rule 1006

We review evidentiary rulings for abuse of discretion. See United States v. Burgos-Montes, 786 F.3d 92, 114 (1st Cir. 2015). Under this umbrella standard, we evaluate questions of law de novo. See id. And if we hold that an abuse of discretion occurred, we ask whether the court's error was nonetheless harmless. See United States v. Abdelaziz, 68 F.4th 1, 70 (1st Cir. 2023).

Maldonado argues that the district court abused its discretion by improperly admitting certain exhibits and accompanying testimony into evidence under Federal Rule of Page 10Evidence 1006.⁴ Rule 1006 permits the introduction of "summary" materials. Generally, "[a]n original writing, recording, or photograph is required in order to prove its content." Fed. R. Evid. 1002. But Rule 1006 provides an exception, allowing a party to "use a summary, chart, or calculation to prove the content of voluminous writings . . . that cannot be conveniently examined in court." Fed. R. Evid. 1006.⁵ For a summary to be admissible under Rule 1006, it must meet several criteria. For one, the underlying source material must be admissible. See United States v. Kumar, 112 F.4th 30, 38 (1st Cir. 2024); United States v. Milkiewicz, 470 F.3d 390, 396 (1st Cir. 2006). And, "[t]he proponent must show that the voluminous source materials are what the proponent claims them to be and that the summary accurately summarizes the source materials." Milkiewicz, 470 F.3d at 396. Maldonado argues that several of the government's exhibits flunked these requirements because they were impermissibly based on hearsay and embellished with inferences and conclusions.

Page 11Although the record is not crystal clear on this point, we agree with Maldonado that inadmissible hearsay may have played an improper role in the development of the summary exhibits. Domínguez testified that at least in some instances, she relied on government interviews to create exhibits:

[Government:] So, you go [through] all 10,000 transactions, how do you categorize each transaction? What documents did you review to satisfy yourself that you're going to categorize each transaction as A instead of B?

[Domínguez:] Okay, so it's multiple transactions. So if it's a deposit a bank will provide you with a deposit slip, a check. Also, when I'm categorizing some transactions, for example related to investors, I would review if there's [an] interview, if there is [a] contract, any other supporting documentation.

Thus, it is not apparent from Domínguez's testimony whether she relied on the interviews merely to confirm what the admissible documentary evidence already suggested, or whether she used the interviews as an independent basis for categorizing some transactions as involving investors. Nor did Maldonado elicit this testimony on cross-examination. But if the out-of-court interviews informed Domínguez's categorizations, then these categorizations were inadmissible because they amounted to either single or double hearsay.⁶ See id.

Page 12Regardless, any error in admitting Domínguez's categorizations into evidence was harmless. "An error will be treated as harmless only if it is highly probable that the error did not contribute to the verdict." Abdelaziz, 68 F.4th at 70 (quoting United States v. Kilmartin, 944 F.3d 315, 338 (1st Cir. 2019)). "Practically speaking, a reviewing court may find an error harmless when the properly admitted evidence, in and of itself, furnished overwhelming proof of the defendant's guilt." Kilmartin, 944 F.3d at 338.

On our review of the record, "we are confident that the guilty verdict[]" on Maldonado's securities fraud conviction is fully sustained "by the properly admitted evidence." Id. at 339. To make this determination, we look at the evidence that Maldonado does not challenge. Then, we consider whether this evidence "overwhelmingly" supported the government's theory that Maldonado recycled money from one client to another instead of engaging in profitable investment activity. Id. at 338.

Crucially, Maldonado does not challenge the admissibility of the underlying bank records. Nor did Maldonado Page 13challenge during trial the majority of Exhibit 327. To be sure, Maldonado objected to the "category" columns of the spreadsheet as "not based on the bank records." But he expressly approved the remaining columns based on those same underlying records. These columns included information like the "Payor" and "Payee" for each transaction, along with a "Description." This data alone supported Maldonado's conviction because it showed whether each transaction involved payment to or from an individual or business or investment activity. For example, one transaction in the spreadsheet was a wire payment listing "Scottrade Inc." -- a brokerage company -- as the "Payor." It was clear from the face of the transaction, then, that it represented an inflow from an investment rather than a payment to or from a client. Other transactions in the spreadsheet, by contrast, listed the names of individuals as "Payor" or "Payee." Still others listed neither Payor nor Payee but included a "Description" (also not a challenged column) that clearly evidenced non-business activity -- for example, one $23.54 outflow to "Burger King 3987 Rio Piedra."

The government charged Maldonado with operating a Ponzi scheme. What mattered to its case was showing that returns to earlier investors were paid from contributions by later investors, rather than from legitimate business profits. The government relied on the properly admitted portions of Exhibit 327 to demonstrate that Maldonado, while touting that he would create Page 14profits through various forms of business activity, overwhelmingly recycled funds from initial investors to pay off subsequent investors. And it also relied on this evidence to show that hardly any transactions indicated income-generating business activity.

Even if Exhibit 327 were excluded, then, other unchallenged summary exhibits would overwhelmingly support the jury's verdict. For one, Maldonado does not argue that Exhibit 328 was inadmissible. That exhibit excerpted portions of Exhibit 327 and did not include a "Category" or a "General Category" column. It presented snapshots of the times immediately preceding and following deposits from the testifying witnesses. In its closing argument, the government walked the jury through Exhibit 328. And it pointed out that, for example, on May 21, 2009, BPRIC's bank account had a negative balance of $3,084.09. Then, the exhibit documented that Maldonado accepted a $15,000 check from Sonia Delgado, bringing the balance to $11,915.91, even though BPRIC had yet to turn a profit. The exhibit also showed that, over the next month, Maldonado spent money at several retail stores and on transfers to various individuals until his remaining balance was just $11.99. None of those transactions involved "invest[ment] in different companies," which is what Delgado testified Maldonado sold to her.

Additionally, the unchallenged evidence showed that Maldonado did not engage in meaningful business activity with Page 15Datavos, Glorimar Fashions, or Pet Card Systems. Take Exhibit 313, which Maldonado does not challenge, and which summarized the activity in those accounts until they closed. It indicated that BPRIC deposited $50,000 in the Datavos account but withdrew $49,000 less than one month later. Similarly, it indicated a cash deposit of $4,000 into that account in September 2009, followed by a $3,500 check to BPRIC less than two weeks later. Other than administrative charges and fees "and a few debits to restaurants and gasoline," there was no other significant activity in the Datavos account. Similarly, Exhibit 313 showed no significant transactions in the Glorimar Fashions account, other than a $299.60 deposit in September 2010, a $229.60 check paid from the account in April 2011, and bank charges. And in the Pet Card Systems account, Exhibit 313 showed just five small deposits totaling $950, one cash deposit of $4,000, one check to BPRIC in the amount of $3,500, and administrative fees.

Several witnesses testified that Maldonado told them their money would be spent developing these businesses, and Díaz testified that promoters were instructed to make that exact pitch. But this unchallenged summary evidence demonstrated that the so-called business activity was minimal at best. In its closing argument, the government made much of Exhibit 313, pointing out that the limited transactions in these accounts showed that when Maldonado received money from clients, "it [went] out to pay other Page 16investors" rather than "to Datavos," for example. Even without Exhibit 327, then, we conclude that it is highly probable that the jury would have convicted Maldonado of securities fraud based on this unchallenged evidence.

Maldonado also contends that the "investor" label that Domínguez assigned and used in her spreadsheet invaded the province of the jury to decide ultimate issues in the case, and the district court should have excluded the exhibit for that reason. According to Maldonado, Domínguez should not have been permitted to label any category of individuals as "investors" rather than "creditors" because the parties disputed whether the contracts at issue were investment contracts or merely credit agreements. But, as we have explained, any evidentiary error related to this spreadsheet would be harmless. The jury had ample evidence in other exhibits and in witness testimony to support its verdict that Maldonado touted his business ventures to potential clients as investments. And the properly admitted evidence overwhelmingly showed that Maldonado shuffled funds from one client to another without generating a profit, thus supporting the government's theory that he had operated a Ponzi scheme.

Next, Maldonado argues that the government failed to prove beyond a reasonable doubt that the Productive Development Contracts were securities. We review a preserved challenge to the Page 17sufficiency of the evidence de novo, viewing the evidence in the light most favorable to the verdict. See United States v. Falcón-Nieves, 79 F.4th 116, 123-24 (1st Cir. 2023). "[T]he relevant question is whether . . . any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt." Id. at 123 (quoting United States v. Woodward, 149 F.3d 46, 56 (1st Cir. 1998)). Based on the record as a whole, we reject Maldonado's sufficiency challenge.

We begin by setting out the applicable legal framework. The Securities Act of 1933 defines a "security" as "any note, stock, . . . investment contract, . . . or any certificate of interest or participation in . . . any of the foregoing."¹⁵ U.S.C. § 77b(a)(1) (emphasis added). In turn, the U.S. Supreme Court has held that "an investment contract for purposes of the Securities Act means a contract, transaction or scheme whereby a person invests his money in a common enterprise and is led to expect profits solely from the efforts of the promoter or a third party . . . ." SEC v. W.J. Howey Co., 328 U.S. 293, 298-99 (1946). Thus, an investment contract comprises (1) an investment of money (2) in a common enterprise (3) with an expectation of profits to be derived solely from the efforts of others. See SEC v. SG Ltd., 265 F.3d 42, 46 (1st Cir. 2001). The Supreme Court went on to emphasize in Howey that the definition of an investment contract "embodies a flexible rather than a static principle, one that is Page 18capable of adaptation to meet the countless and variable schemes devised by those who seek the use of the money of others on the promise of profits." 328 U.S. at 299.

Elaborating on the "common enterprise" prong, we have held that a common enterprise can be shown through horizontal commonality, defined as "the pooling of assets from multiple investors in such a manner that all share in the profits and risks of the enterprise." SG Ltd., 265 F.3d at 50. Contrary to Maldonado's assumption, however, we have not held that horizontal commonality is necessary to establish the existence of an "investment contract."⁷ But, because the government chose to prosecute this case under the horizontal commonality theory, our analysis focuses on that theory of liability.

Maldonado contends that the evidence was insufficient to show that the clients' money was pooled in common or put towards a single business enterprise. He also notes that the Productive Development Contracts did not use the words "security" or "investment," and he points out that the clients did not have proportional shares in BPRIC. Finally, he insists that, even if the Productive Development Contracts were securities, the Page 19government did not meet its burden to show that Maldonado knew that the agreements were securities.

We conclude that each of Maldonado's arguments is either unsupported by the record or legally incorrect. To begin, we have held that "Ponzi schemes typically satisfy the horizontal commonality standard," and Maldonado was convicted on a Ponzi scheme theory. Id. at 51. Thus, Maldonado faces an uphill battle to show that the government failed to prove a common enterprise.

As to Maldonado's first argument, the record belies his claim that no evidence demonstrated that the clients' money was "pooled in common." Maldonado's counsel specifically asked Domínguez if deposits from each client were "mixed up with the moneys" already in Maldonado's accounts, and Domínguez confirmed that they were. And Maldonado's bank records corroborated this testimony.⁸ As a result, Maldonado's failure to turn a profit affected his clients collectively. Thus, the government did present evidence of pooling.

Next, Maldonado asserts that he did not put clients' money toward "a single business and common enterprise." Instead, he contends, "a variety of businesses were discussed with no specific representation that any one person's money would be used Page 20in a specific way." We reject Maldonado's assumption that under the common pooling theory, there must be only a single common pool. But regardless, his argument fails because the government presented evidence that each client gave money to BPRIC. The fact that this enterprise promised to make money through various businesses and investments does not mean that the clients' investments in BPRIC were not investments in a common venture.

Maldonado then turns to the text of the Productive Development Contracts, noting that they identified participants as "creditors" and disclaimed any "promise of return by way of investment." But we have emphasized that the test for determining whether an agreement is an investment contract "must be administered without regard to nomenclature." SG Ltd., 265 F.3d at 47. Thus, the labels in the agreements do not control.

With his last argument on this issue, Maldonado contends that the record does not show "that payments received were made according to any pro rata or uniform system." But we have never held that a pro rata share or uniform system of profits is required to demonstrate a common enterprise.⁹ Instead, even if each Page 21instrument has a different value, we have observed that when "most of the potential gain" depends on the "development" of the instruments collectively, a security can exist.¹⁰ Rodriguez v. Banco Cent. Corp., 990 F.2d 7, 11 (1st Cir. 1993).

Maldonado argues in the alternative that even if the BPRIC contracts qualify as securities, that was news to him. Thus, Maldonado contends, the government failed to prove that he knew that the contracts were securities, meaning that he could not have acted "willfully." See 15 U.S.C. § 78ff(a) (requiring securities violations to be "willful[]" to subject the violator to fines or imprisonment). The government responds that it was not required to prove that Maldonado knew the contracts were securities, and in any case, the evidence demonstrated that he was in the know.

We see no basis for Maldonado's alternative argument. To act "willfully" means to act "voluntarily and intentionally, and with the specific intent to do something which the law Page 22forbids . . . ; that is to say, with a bad purpose either to disobey or to disregard the law." United States v. Bailey, 405 F.3d 102, 111 n.4 (1st Cir. 2005); see also Rehaif v. United States, 588 U.S. 225, 232 (2019) (stating that the "purpose of scienter" in criminal statutes is "to separate wrongful from innocent acts"). But Maldonado cites no authority interpreting § 78ff(a) to require that a defendant know that the agreements at issue were securities in order to have a "bad purpose" to "disobey or disregard the law." For example, he does not cite to any precedent suggesting that specific intent to defraud his clients would not be enough to qualify as willful conduct under § 78ff(a).

In any event, the evidence at trial demonstrated that Maldonado knew the contracts were securities. As explained above, Maldonado attacks the "common enterprise" prong of the test for investment contracts. But the evidence at trial fully supported a finding that he knew that BPRIC was a common enterprise. Recall, a common enterprise may involve "the pooling of assets from multiple investors in such a manner that all share in the profits and risks of the enterprise." SG Ltd., 265 F.3d at 50. The evidence unequivocally demonstrated that Maldonado managed BPRIC's money, and that he ran the enterprise through his own accounts. Thus, a rational jury could conclude that Maldonado was well aware that the money was pooled and that BPRIC's profits and risks were Page 23shared among clients. As a result, we reject Maldonado's sufficiency challenge.

Maldonado also argues that the district court made mathematical and factual errors in calculating his sentence, but we disagree.¹¹ We generally review preserved procedural sentencing challenges for abuse of discretion. See United States v. Viloria-Sepulveda, 921 F.3d 5, 8 (1st Cir. 2019). When a defendant "challenges the factual predicate supporting the district court's application of a sentencing enhancement, 'we ask only whether the court clearly erred in finding that the government proved the disputed fact by a preponderance of the evidence.'" United States v. Cannon, 589 F.3d 514, 517 (1st Cir. 2009) (quoting United States v. Luciano, 414 F.3d 174, 180 (1st Cir. 2005)).

As we have explained, the district court applied an 18-level enhancement because it concluded that Maldonado caused over $3.5 million of losses and a six-level enhancement because Page 24the offense resulted in substantial financial hardship to twenty-five or more victims. "In a fraud case resulting in financial loss, the defendant's guideline sentencing range is determined in part by calculating the greater of either the intended loss or the actual loss." United States v. Naphaeng, 906 F.3d 173, 179 (1st Cir. 2018); see also U.S. Sent'g Guidelines Manual § 2B1.1, cmt. n.3(A)(i)-(ii) [hereinafter "U.S.S.G."]. The court may consider "both losses stemming directly from the conduct underlying the offenses of conviction and losses stemming from any relevant conduct (including uncharged conduct)." United States v. Flete-Garcia, 925 F.3d 17, 28 (1st Cir. 2019). Further, the sentencing court has "considerable discretion in determining what evidence" it regards as reliable for the loss calculation, and the "loss calculation need not be precise." Id.

Maldonado presents two challenges to the district court's calculations of loss and the number of victims, which in turn determined his sentence. First, he argues that the district court made mathematical errors in calculating the loss amount. Second, he argues that the district court lacked sufficient evidence to consider any purported losses by individuals who did not testify at trial. We reject both arguments as inconsistent with our case law and the record.

As to the loss calculation itself, Maldonado contends that the district court's methodology did not support the number Page 25it reached. The court stated that it would "consider loss amount[s] based on both victims who testified during trial as well as those others mentioned in the PSR." Because the sum of the loss amounts for the named victims in the PSR is just over $2.1 million, Maldonado insists that the district court committed a reversible mathematical error by concluding that the loss amount exceeded $3.5 million.

We discern no mathematical error by the district court. In addition to listing the restitution victims, the PSR stated that "[a]s evidenced by financial analysis and documentation in support for the number of victims and restitution provided by the [g]overnment, restitution was conservatively estimated at $3,786,270. According to the [g]overnment, this represents the amount of financial loss to the victims, including others not listed in the [i]ndictment."¹² When the court explained that it relied on the PSR, then, we understand it to have relied on this $3.7 million figure.

Maldonado next points to a separate alleged mathematical error. He argues that, even assuming that the court relied on the PSR's $3.7 million number, which was based on the government's financial analysis that Domínguez presented at trial, this Page 26evidence does not add up to a loss of over $3.5 million. Maldonado notes that Exhibit 306 computes a net loss for all confirmed investors of only $3.2 million. But Maldonado misunderstands the court's loss calculation. As the government points out, the U.S. Sentencing Guidelines instruct how courts should calculate loss resulting from a Ponzi scheme. See U.S.S.G. § 2B1.1, cmt. n.3(F)(iv). As the relevant application note explains, in a Ponzi scheme, "the gain to an individual investor in the scheme shall not be used to offset the loss to another individual investor in the scheme." Id. When the investors who received payments from the recycled funds of other investors are removed from the exhibit, the resulting net loss is over $3.5 million.¹³

We turn next to Maldonado's argument that the district court was not permitted to calculate loss or the number of victims Page 27based on individuals who did not testify at trial. Maldonado contends that the court lacked evidence that any of these individuals were victims of criminal conduct. He insists that the record "did not unambiguously establish fraudulent representations" to non-testifying individuals, nor did it show Maldonado's "criminal fraudulent intent" as to non-testifying individuals. According to Maldonado, even if he entered into contracts with these individuals and then failed to repay them, these facts do not show that he criminally defrauded them.

We conclude that the district court did not err in considering the loss suffered by non-testifying witnesses as set forth in Exhibit 306. Again, sentencing courts have considerable discretion in determining what sources are reliable, and they may even consider uncharged conduct. See Flete-Garcia, 925 F.3d at 28. In this case, the government conducted an extensive financial analysis, spending more than a year examining thousands of pages of transactions from Maldonado's bank accounts. We see no clear error in the district court's decision to credit this analysis. And the fact that Maldonado objected at trial to the admissibility of Exhibit 306 under Federal Rule of Evidence 1006 does not change our conclusion. At sentencing, "the court can consider all kinds of relevant information regardless of admissibility at trial (including hearsay that has never been tested by cross-examination), provided it has 'sufficient indicia of Page 28reliability to support its probable accuracy.'" United States v. Mills, 710 F.3d 5, 15 (1st Cir. 2013) (citations omitted) (quoting U.S.S.G. § 6A1.3(a)). Finally, Maldonado presented only general objections to these factual findings during his sentencing proceedings. He argued that the government's assessment "d[id] not encompass all the information that would be relevant to make a proper and complete forensic accounting analysis" and that it did not include at least one of Maldonado's accounts. But he failed to introduce contrary evidence. See United States v. Flores-Seda, 423 F.3d 17, 21 (1st Cir. 2005) (concluding that the district court's loss calculation was reasonable in part because the defendant offered no "evidence to the contrary").

Further, we have approved district courts' decisions to consider losses suffered by non-testifying witnesses when they have found that those losses were part of the defendant's "scheme." See, e.g., United States v. Curran, 525 F.3d 74, 78 (1st Cir. 2008). Here, Maldonado contends that "while the government presented evidence of what witness[es] 'believed' . . . about their money when they handed over checks and at other points, it proffered no commensurate evidence as to the non-testifying" individuals. But the evidence showed that Maldonado systematically used the same deceptive tactics to draw in clients. For example, Díaz explained that Maldonado instructed promoters that BPRIC invested in several specific companies and provided Page 29materials for them to distribute to prospective clients describing these companies. In turn, Domínguez's unchallenged testimony showed that BPRIC did not provide sufficient funds to those entities to make them viable businesses. Thus, BPRIC failed to turn a profit, all while it continued to accept checks from more and more unsuspecting people. In light of this evidence, it was reasonable for the district court to consider the losses of the non-testifying victims in its calculation. See id. at 81 (holding that, in light of patterns evidencing defendant's "scam," it was reasonable for the district court to conclude that payments to the defendant "were actual losses resulting from the offense").

Finally, Maldonado brings two challenges to the district court's restitution award. First, he contends that the district court was "limited to the specific transactions listed in Count [One] of the indictment" and therefore could not order restitution to the other victims included in the PSR. Second, he argues that the district court lacked sufficient evidence to conclude that the non-testifying individuals identified in the PSR were victims of criminal conduct. But for the reasons explained below, Maldonado has at least forfeited these objections. Because the district court did not plainly err in entering its restitution order, we affirm it. See United States v. Sansone, 90 F.4th 1, 6 (1st Cir.

Page 302024) ("Unpreserved claims of error, if not deemed waived, are reviewed only for plain error.").

The district court ordered restitution for Maldonado's securities fraud conviction under 18 U.S.C. § 3583(d) as a condition of his supervised release.¹⁴ Section 3583(d) allows a district court to impose as part of a defendant's supervised release sentence "any condition set forth as a discretionary condition of probation in section 3563(b) and any other condition it considers to be appropriate."¹⁸ U.S.C. § 3583(d)(3). Section 3563(b), in turn, permits a court to require the defendant to "make restitution to a victim of the offense . . . ." Id. § 3563(b)(2); see also United States v. Bok, 156 F.3d 157, 166 (2d Cir. 1998). In this context, a "victim" includes:

[I]n the case of an offense that involves as an element a scheme, conspiracy, or pattern of criminal activity, any person directly harmed by the defendant's criminal conduct in the course of the scheme, conspiracy, or pattern.

18 U.S.C. § 3663(a)(2) (emphasis added).¹⁵

Page 31The district court concluded that Maldonado's securities fraud charge, as set forth in the indictment, constituted a "scheme." The court therefore reasoned that, as part of Maldonado's conditions of supervised release, it could order restitution to individuals "directly harmed" by his criminal activity, including those not identified in the indictment.

Although Maldonado now contends that the individuals listed in the PSR but not the indictment were not "victim[s] of the offense" within the meaning of § 3563(b), he failed to properly raise this objection below. Indeed, at Maldonado's sentencing, the district court asked whether there were "[a]ny objection[s] on the terms of supervised release," and defense counsel replied "[n]o."

By failing to object to the terms of his supervised release, Maldonado has forfeited his challenges to the restitution order. See United States v. Walker, 538 F.3d 21, 23 (1st Cir. 2008) (noting that forfeiture occurs when "a party fails to make a timely assertion of a right" (internal quotation marks omitted)). We therefore review Maldonado's forfeited arguments only for plain error.¹⁶ See Sansone, 90 F.4th at 6.

Page 32To prevail under the plain-error standard, an appellant must show that "(1) the district court committed an error; (2) that error was 'plain -- that is to say, clear or obvious,' (3) the error affected his substantial rights, and (4) leaving the error uncorrected would 'seriously affect[] the fairness, integrity or public reputation of judicial proceedings.'" United States v. Langston, 110 F.4th 408, 419 (1st Cir. 2024), cert. denied, 145 S. Ct. 581 (2024) (quoting United States v. Ortíz-Mercado, 919 F.3d 686, 689 (1st Cir. 2019)). The four prongs may be evaluated in any order. See id.

We start with the second prong, which requires us to decide whether the error -- to the extent there was one -- was "plain." See id. "[A] plain error must be 'indisputable.'" Id. (quoting United States v. Correa-Osorio, 784 F.3d 11, 22 (1st Cir. 2015)). Thus, Maldonado must show that the district court "obviously" exceeded its authority under §§ 3583 and 3563(b) to Page 33enter the restitution order at issue here. See id. He can do so by "point[ing] either to binding on-point precedent or show[ing] that [his] theory is compelled by constitutional law, statute, regulation, or other legal mandate." Id. (internal quotation marks omitted).

Maldonado cannot demonstrate clear or obvious error because he has cited neither binding on-point precedent nor compelling authority that would foreclose the district court's ruling. Specifically, he has not pointed to any controlling or compelling authority that would have prohibited the court from determining that the victims identified in the PSR were "person[s] directly harmed" by the scheme charged in the indictment.¹⁷ See 18 U.S.C. § 3663(a)(2).

To challenge the scope of the restitution order, Maldonado principally relies on the Supreme Court's decision in Hughey v. United States, 495 U.S. 411 (1990). But that case does not require a different result. In Hughey, the Supreme Court held that restitution orders entered under § 3663(a), which defines the "victims" eligible for restitution, are limited to "the loss caused by the specific conduct that is the basis of the offense of conviction." Id. at 413. However, subsequent amendments toPage 34 § 3663(a) created an exception to Hughey when the crime of conviction includes as an element a scheme, conspiracy, or pattern of criminal activity. See United States v. Acosta, 303 F.3d 78, 87 (1st Cir. 2002) ("Under 18 U.S.C. § 3663(a)(2), any conduct that is in the course of the scheme, conspiracy, or pattern may be considered in calculating restitution."); see also United States v. Batson, 608 F.3d 630, 636-37 (9th Cir. 2010) (noting that Hughey applies "so long as th[e] offense does not involve an element of scheme, conspiracy[,] or pattern of criminal activity" (emphasis added)). At trial, Maldonado himself asserted that "securities fraud is the willful participation in a scheme to defraud." (Emphasis added.) Thus, although we do not resolve today whether 15 U.S.C. §§ 78j(b) and 78ff(a), as a categorical matter, include as an element a "scheme" -- or whether such a categorical interpretation of the statutes is necessary to permit restitution for all "person[s] directly harmed" by securities fraud that amounts to a scheme -- the district court did not "obviously" err in concluding that the offense of conviction here involved a scheme.¹⁸ See Langston, 110 F.4th at 419. For that reason, its decision to order restitution to victims directly harmed by Page 35Maldonado's conduct, but not identified in the indictment, is not obviously wrong.

Further, Maldonado has not cited any authority to suggest that the district court's reliance on the PSR to determine the number of victims constituted "indisputable" error. See id. Indeed, "[a] PSR generally bears 'sufficient indicia of reliability to permit the district court to rely on it at sentencing.'" United States v. Prochner, 417 F.3d 54, 65-66 (1st Cir. 2005) (quoting United States v. Cyr, 337 F.3d 96, 100 (1st Cir. 2003)). As a result, the district court did not plainly err in concluding that the individuals listed in the PSR were, by a preponderance of the evidence, victims of Maldonado's securities fraud crime. See United States v. Vaknin, 112 F.3d 579, 582-83 (1st Cir. 1997). Accordingly, we affirm the restitution order.

For all these reasons, we vacate Maldonado's bank fraud convictions and affirm his securities fraud conviction and sentence, including the district court's restitution order.