No Gain, Yes Pain

United States v. Hsu, No. 09-4152-cr (2d Cir. February 17, 2012) (Winter, Lynch, Carney, CJJ)

Norman Hsu, a prominent, if corrupt, political fundraiser, used the connections he made in politics to run a giant Ponzi scheme. He pled guilty to mail and wire fraud, and was convicted by a jury of campaign finance fraud. In all, the district court imposed a 292-month guideline sentence.

The main, but not only, issue on his appeal concerned an interesting sentencing issue. The district court found that the Ponzi scheme caused a loss of between $50 million and $100 million, but in doing so included earnings that the victims reinvested in the scheme - even though those earnings were invented as part of the scheme - in the intended loss. The circuit agreed that this was permissible.

Normally, in fraud cases, the guidelines measure the amount of principal the victims lost, and not the amount of lost principle plus the promised profit that never materialized. But the “situation is different” where the investor is “told not simply that his investment will grow, but that it has grown, and that the total of his original investment and the accrued interest or other gain is now available to be withdrawn or reinvested in the scheme, depending on the investor’s preference.” For Ponzi-scheme participants who can choose either to withdraw or reinvest, the choice to reinvest, which is usually necessary to keep the scheme alive, “transforms promised interest into realized gain that can be used in the computation of loss.” Only the “most recent promised or reported” gains are excluded from the guidelines calculation as interest.

The alternative, calculating losses only by looking at the money actually invested by victims, would “fail to take into account” the very nature of Ponzi schemes: the victims’ changing behavior in the face of the fraudulent report of the success of their investments and the fact that the purported proceeds of this “success” are maintained in accounts controlled by the defendant, not withdrawn by the investors. When the defendant gives the victims the option of withdrawing the proceeds, but then induces them to instead reinvest the money and again put it at risk, the “victims have suffered further loss.”

The court recognized that the task of defining this “reinvestment” will in many cases be difficult and will have to be “pursue[d] with care,” because “what constitutes interest precluded from consideration during sentencing in one context may be the very loss intended in another.” But here, the case was “straightforward. Hsu’s victims frequently returned post-dated checks to him for reinvestment, thereby relinquishing the opportunity to cash those checks and withdraw from the scheme. When this occurred, the reinvested checks - including the previously promised returns - became part of their principle investment, and therefore constitute the very losses that Hus intended to inflict.” Hsu’s contrary argument - “that the ‘gains’ did not exist ... reduces to the claim that the victims losses do not count because he was unable to pay them back.”

The court concluded by noting that this method is not “the only way to measure loss in a Ponzi scheme,” but is one way to develop a “reasonable estimate” of loss for the guidelines. In the end, all the court really held is that the exclusion of interest from loss calculations did not apply, and that the district court’s calculation was otherwise reasonable and appropriately measured the harm to the victims.

No Gain, Yes Pain

United States v. Woolf Turk, No 09-5091-cr (2d Cir. November 30, 2010) (Katzmann, Hall CJJ, Jones, DJ)

Ivy Woolf Turk was a principal in a real estate development company. Between 2003 and 2007 she and her partner persuaded investors to lend them $27 million, primarily to renovate apartment buildings in upper Manhattan. They induced the loans by promising that the investors would hold recorded first mortgages on the buildings as collateral. This was a lie - they never recorded the mortgages, so the investors were merely unsecured creditors. At the same time, the developers obtained loans from banks, and those liens were recorded.

Eventually Woolf Turk began defaulting on the victims’ loans. The victims became suspicious and discovered that, despite Woolf Turks’ representations, their mortgages had never been recorded. In May of 2007, the investors sued; only then did they learn that, not only were their mortgages unrecorded, but that the bank loans were recorded, and thus that their interests in the properties, if any, were secondary to the banks’.

Eventually, the development company went bankrupt and its assets were liquidated. Most went to the banks - the secured creditors - and the rest to pay various court and regulatory fees. The bankruptcy trustee was only able to distribute half a million dollars to the victims, who thus lost nearly all of the $27 million they lent to Woolf Turk.

Woolf Turk pled guilty to conspiracy to commit mail and wire fraud. At sentencing, the district court agreed with the government that the loss calculation for Guidelines purposes was more than $20 million, and calculated the Guidelines accordingly. With a resulting range of 121 to 151 months, the court sentenced Woolf Turk to 60 months’ imprisonment.

On appeal, Woolf Turk disputed the district court’s loss calculation, pursuing the same argument she made in the district court - that the victims’ loss was caused by the housing-market crash and not by Woolf Turks’ fraud. The circuit disagreed that “the loss amount should have been treated as zero because the properties in which her victims thought they were investing arguably had some market value at the time her fraud was discovered.”

This argument was based on the faulty premise that the loss was measured by the decline in value of what was promised as collateral. Rather, the victims’ loss “is the principal value of the loans they made to Woolf Turk which were never repaid and which the buildings were supposed to have collateralized but never did.”

Here, the buildings were “arguably” not collateral at all because the victims’ mortgages were never recorded, and under New York law, an unrecorded mortgage is void as against a lien on the same property recorded in good faith. It did not matter that, had the value of the properties had gone up, instead of down, the victims might have recovered all or part of their losses. Here, the purported collateral had no meaningful value at the time of sentencing.

The Guideline holds defendants accountable for any “reasonably foreseeable pecuniary harm,” which it defines as harm that the defendant knew or should have known could result from the offense. The potential for loss that Woolf Turk’s victims faced by not having recorded mortgages met this standard. Thus, the loss amount was the full principal of the loans that Woolf Turk fraudulently obtained. And it is irrelevant that some value might have remained in the collateral at the time the fraud was discovered, because the victims had no interest in the collateral, obtained no value from its sale, and no value remained at the time of sentencing.

Cash and Quarry

United States v. Byors, No. 08-4811-cr (2d Cir. October 29, 2009) (Cabranes, Livingston, CJJ, Korman, DJ)

Defendant, while ostensibly raising money for a Vermont marble quarry, made material misrepresentations to his investors. He also converted substantial amounts of their money to pay for his personal expenses, including vacation homes, cars and horses. He pled guilty to multiple fraud and money laundering offenses and was sentenced to 135 months’ imprisonment. On appeal, he raised two unsuccessful challenges to his Guidelines calculations.

He first argued that the district court should have deducted from the loss calculation - about $9 million - the “legitimate business expenditures” that went into his efforts to “capitalize” the quarry business. The circuit disagreed. Under the “plain language” of Application Note 3(E) to the fraud guideline, the loss amount is only offset by any “value” that the victims receive, and not by legitimate expenditures. Byors' expenditures conferred nothing of value and no benefit to his victims. He rendered no “services” to them and did not deliver any return on their “investment.” Even accepting his claim that he used the money for the purposes he promised his victims, there was no error here. Byors' victims were left with nothing of value when the fraud was uncovered.

Byor, who tampered with a witness during the investigation into his fraud, also raised an issue about the interaction between the general obstruction of justice guideline, § 3C1.1, and the specific provision dealing with obstruction in the money laundering Guideline, § 2S1.1, comment. n.2(C). These two provisions seemingly conflict in cases where the defendant has obstructed a predicate offense, but not the subsequent money laundering offense itself.

The money laundering guideline provides that the application of § 3C1.1 “shall be determined based on the [laundering of criminally derived funds] ... and not on the underlying offense from which the laundered funds were derived.” Byors argued that under this provision the Chapter 3 adjustment could only apply if the obstruction related to the money laundering offense, and not the underlying fraud. The circuit disagreed. The Chapter 3 obstruction enhancement covers the offense of conviction, “any relevant conduct,” or “a closely related offense.” The fraud that underlay the money laundering offense, during which Byors obstructed justice, was either relevant conduct or “closely related” to the money laundering offense.

The court refused to conclude that an application note to a separate offense conduct guideline “creates an exception” to § 3C1.1, since that would be contrary to its practice of seeking to “harmonize” commentary with the Guidelines.

Nothing In Store

United States v. Uddin, No. 07-3121-cr (2d Cir. January 6, 2009) (Kearse, Sack, Katzmann, CJJ)

Mohammed Uddin owned a small grocery store in Manhattan, and used it to commit food stamp fraud between 2003 and 2006 by dispensing cash in exchange for food stamps. He pled guilty but admitted in his allocution only that the amount of fraud exceeded $5,000 - the jurisdictional amount. After a Fatico hearing, the district court concluded that the loss amount was $377,799, and sentenced Uddin accordingly. On appeal, Uddin challenged the loss calculation.

The District Court Proceedings

The government had been seeking a loss in excess of $1.2 million, arguing that all of Uddin’s food stamp redemptions exceeding $50 during the relevant time period were fraudulent. Uddin argued instead that the loss should be limited to $5,000, the amount he admitted in his plea.

The evidence at the Fatico hearing showed that his store redeemed several times more in food stamp benefits than did two comparably sized stores nearby, and that he exchanged food stamps for cash with a CI on fourteen occasions in 2006 alone.

An agent also testified that, based on his experience, any food stamp redemptions of more than $50 at Uddin’s store were fraudulent. He characterized this as a “conservative” estimate based on the fact that the store had a very limited supply of eligible food items for sale, and those that it did stock were “dusty” and “outdated.” In addition, the store was small, lacked baskets or carts, and did not offer delivery service. Video surveillance in 2006 showed no customers leaving the store carrying groceries worth $50 or more; some appeared to be counting cash as they exited.

A second agent testified that, while the store had been well stocked in 2002 - when it first obtained its license to redeem food stamps - by 2006 the store sold very little eligible food. Moreover, the average food stamp transaction in New York City was about $12, thus the activity in Uddin’s store was quite unusual.

After hearing this testimony, the district court concluded that the government’s estimate of loss was too high, while Uddin’s was too low. The court first assumed that the store’s stock declined steadily and gradually between 2002 and 2006. This led it to discount the government’s proposed loss amount by half, which led to a loss of $629,665. The court also disagreed with the government that every transaction of $50 or more was completely fraudulent, and discounted that assumption by forty per cent. Sixty per cent of $629,665 is $377,799, and that was the amount the court settled on.

The Circuit’s Decision

The court of appeal affirmed, noting that the Sentencing Guidelines require only a “reasonable estimate” of loss in financial crimes. A court can make such an estimate by “extrapolating the average amount of loss from known data.” Here, the district court’s estimate was reasonable and was supported by a preponderance of the evidence, given the evidence of the decline in the store’s stock during the relevant time period. The use of the $50 transaction figure as a “general point of reference for likely fraudulent transactions” was likewise reasonable. Even if “not based on precise data,” it was based on “known” data such as the average dollar amount of food stamp redemptions at similar stores in New York City and the witnesses’ observations of Uddin’s own store.

The court then added - not particularly helpfully - that while “there will undoubtedly be situations in which a district court’s estimate of a loss amount falls outside the boundaries of reasonableness, we need not define precisely what those boundaries are. It is enough that the district court here did not exceed them.”

In Search of Lost Time

United States v. Abiodun, No. 06-5335-cr (2d Cir. July 30, 2008) (Cardamone, Cabranes, Katzmann, CJJ)

Emmanuel Abiodun was one of a group of people who ran a large credit card and identity fraud scheme in which credit reports were illegally downloaded and used to obtain credit cards in the victims’ names. Abiodun himself purchased between 300 and 400 reports and, the district court found, was responsible for a loss of between $1.6 and 2.0 million.

The court also increased his offense level by six levels based on its finding that Abiodun’s conduct involved more than 250 victims. The court included in this number individuals who suffered no actual financial loss, but who spent time securing reimbursement from banks and credit card companies.

On appeal, the circuit agreed that this was appropriate. The fraud guideline defines a victim as anyone “who sustained any part of the actual loss” for which the defendant is accountable. But this can include monetary harm, or any other type of harm that is “readily measurable in money.” On appeal, Abiodun argued that under these definitions it was error for the court to include individuals whose identities were stolen but who were fully reimbursed for their financial losses.

The circuit disagreed. Fully reimbursed individuals can still be victims under this guideline if “as a practical matter,” they suffered “(1) an adverse effect (2) as a result of the defendant’s conduct that (3) can be measured in monetary terms.” Thus, the district court correctly concluded that, if someone whose credit information was stolen spent “an appreciable amount of time securing reimbursement,” this loss of time could be measured monetarily and thus that such a person would be a "victim."

Despite this holding, which is in essence, an affirmance, the court vacated the sentence. The district court counted such “time losers” as victims, but did not add in the value of this time when calculating the overall loss amount. But the definition of “victim” is limited to persons whose loss is included in that total. Thus, on remand, the district court either has to (1) recalculate the loss amount to include the value of the time lost by these individuals or (2) keep the loss amount the same, exclude these people from the victim count, and determine whether Abiodun’s crime still affected more than 250 victims.

Comment

There are a couple of oddities about this decision. The first is that it could result in a higher sentence on remand, since the loss amount might go up. Usually when this is the case, the court offers the defendant the option of withdrawing his appeal instead of pursuing the remand, but it did not do so here. See, e.g., United States v. Harrington, 354 F.3d 178, 186 n.5 (2d Cir. 2004). The second is that the court also vacated the sentence of Abiodun’s co-defendant, even though there is nothing in the opinion to suggest that his case presented this same issue.

Hollywood Accounting

United States v. Leonard, No. 05-5523-cr (2d Cir. June 11, 2008) (Kearse, Calabresi, Katzmann, CJJ)

In this case, the court concludes that interests in film production companies were “investment contracts,” and hence securities, under federal securities law. It also holds, however, that the district court erred in treating the entire cost of the securities as the loss amount under the guidelines.

Facts

The defendants ran sales offices that peddled interests in LLC’s formed to finance the production and distribution of motion pictures. Potential investors were solicited over the phone and, if they expressed an interest, would be sent offering materials, including brochures, operating agreements, and other such documents. Investors could purchase $10,000 “units” by completing and mailing back a subscription agreement.

The defendants’ sales offices would receive a commission of around 45% for each unit sold. This was the fraud - although the offering materials indicated that a commission would be paid, they did not accurately disclose the size. Read generously to the defendants, they seemed to indicate that no more than 20% of the unit price would go toward sales commissions.

The Sufficiency of the Evidence

The defendants’ first claim was that there was insufficient evidence that the investment units were “securities.” The definition of “security” covers many types of instruments. Here, the specific question was whether the units were “investment contracts.” An investment contract is one where the investor “is led to expect profits solely from the efforts of the promoter or a third party.” Here, the defendants argued that the investors were supposed to help manage the LLC’s, and hence never expected to profit “solely” from the efforts of others.

There is a difference between companies that seek “passive investors,” which fall under the securities laws, and those in which there is a reasonable expectation of significant investor control, which do not. But here, this distinction was complicated by the fact that the investment units were shares in LLC’s. According to the circuit, LLC’s, due to their hybrid nature, require a case-by-case analysis of the “economic realities” of the underlying transaction.

Here, the organizational documents describe an investment that, on its face, was not a security. Those documents were intended to create the impression that subscribers would play an active role in the management of the companies. In reality, however, the members of the LLC’s “played an extremely passive role in the management and operation of the companies.” They voted rarely, and only a small number of them served on any committees. Thus, according to the circuit, “the vast majority of investors in both companies did not actively participate in the venture, exercising almost no control.” For example, the trial evidence showed that the managerial rights mentioned in the subscription documents were illusory, because others made almost every significant decision about the making of the films before the fundraising was even complete. Moreover, the investors themselves had no experience in film production, and played no role in shaping the organizational agreements themselves, casting further doubt as to whether the members were truly expected to have significant control over the enterprise.

And that was enough for the jury to properly conclude, considering “substance over form,” that, “from the start,” there was no reasonable exception of investor control.

The Loss Calculation

At sentencing, the district court held the defendants accountable under for the entire cost of the securities they sold, on the theory that the investors would not have participated if they knew the true size of the sales commissions. The circuit deferred to the district court’s finding about the investors’ decision to participate, but held that this did not “in and of itself” mean that the securities they received were “entirely without value.”

Since the investors actually obtained an interest in a company engaged in producing and distributing a motion picture, the district court should have deducted from the purchase price the actual value of the instruments. The court notes that these were “illiquid securities for which there is no public market,” and thus that it will be quite difficult to value them. Nevertheless, the district court must make a “reasonable estimate” of this figure.

The Loan Arranger

United States v. Confredo, No. 06-3201-cr (2d Cir. June 10, 2008) (Newman, Winter, Parker, CJJ)

This case takes on the difficult question of fixing the loss amount under the sentencing guidelines when the case involves fraudulently obtained that loans have been partially repaid. It also addresses an interesting Apprendi claim.

1. The Loss Amount

Defendant Confredo and his associates coordinated the submission of more than 200 fraudulent loan applications to New York banks. The borrowers were small businesses, which paid Confredo a fee, and knew that the applications were false, in most instances because the businesses were not credit worthy. Most of the applications were cosigned by second parties with good credit, but none were secured by real collateral. In total, more than $24 million was sought, and more than $12 million was actually lent, mostly from Citibank.

At sentencing, the probation department recommended that the full $24 million be treated as intended loss under the guidelines, although the available evidence suggested that the banks had actually lost significantly less than that, because some of the $12 million in loans had been repaid. When Confredo was originally sentenced, in 1997, the district court used the $24 million figure, rejecting Confredo’s argument that the guidelines should be based on the actual loss, which he conceded was between $10 million and $20 million.

He won his first appeal, which included an unpreserved claim that the loss amount was incorrect. Because the government had conceded on other issues, the court directed that the district court revisit loss amount on remand.

At resentencing, the government urged the district court to stick to its original ruling. Confredo argued that the intended loss was less than $20 million because he expected that (1) the banks would reject some of the applications and (2) some of his customers would repay their loans, at least in part. The district court sided with the government, but the circuit disagreed.

The court first dealt with the “uncertainty” as to whether the district court’s loss calculation was a fact-finding, reviewed only for clear error, or an interpretation of the guidelines, which would be reviewed de novo. Here, the court treated the ruling here a legal question - as a matter of law, does the presenter of loan applications intend a loss equal to the aggregate amount of the loans when the presenter is not the borrower?

Then the court then surveyed the law on this point. Until 1991, it had held that the proper measure of intended loss was always the value of the loan obtained or sought, even if the defendant intended to repay it. A 1991 guideline amendment, however, permitted greater flexibility on this issue, giving a defendant credit for actions - such as loan repayments or assets pledged to secure the loan - that might reduce the intended loss amount.

The court viewed Confredo’s case as more difficult, however, because he was not the borrower himself, and no assets had been pledged to secure the loans. Nevertheless, the court concluded that the 1991 amendment means that a defendant “should have an opportunity to persuade the sentencing judge that the loss he intended was less than the face amount of the loans.” The court remanded the case for this purpose, directing that the district court “determine the extent, if any, to which Confredo has proven a subjective intent to cause a loss of less than the aggregate amount of the loans.”

2. The Apprendi Issue

Confredo also received to a 3-level enhancement for committing some of the offenses to which he had pled guilty while on bail for others. He argued that this enhancement violated his right to a jury trial, under Apprendi. Interestingly, the court held that Apprendi does apply in this situation because, even though Confredo did not receive a sentence above the unenhanced statutory maximum, the enhancement “expose[d him] to the risk of a sentence that exceed[ed] the statutory maximum.” But the court also held that Apprendi’s jury fact-finding requirement was not violated, because Confredo “sufficiently” admitted that he committed offenses while on release, by admitting to conduct that “the public record indisputably establishe[d]” had occurred after his release on bail.

The court also recognized that there was a second Apprendi violation here - the absence from the indictment of an allegation that Confredo committed the offenses while on release. But it held that such an omission is harmless error “where the evidence is overwhelming that the grand jury would have found the fact at issue.”

Score: Form 1; Substance 0

United States v. Rutkoske, No. 06-4067-cr (2d Cir. October 25, 2007) (Newman, Winter, Katzmann, CJJ).

This stock fraud decision deals primarily with the timeliness of a superseding indictment.

An initial indictment not naming Rutkoske was filed on December 11, 2003; S1, the first superseder, was filed on April 6, 2004. It named Rutkoske, and described a single overt act within the five-year statute of limitations. Suspiciously, that act occurred “on or about April 9, 1999,” making the indictment timely by only about three days. After repeatedly being pressed by the defendant to pin down the details of the April 9 act, the government superseded again, in July of 2005. S2 charged Rutkoske with the same offenses as S1, but the government dropped the April 9 overt act and instead alleged two others, on April 15 and April 16, 1999. When Rutkoske moved to dismiss S2 as untimely under the five-year statute of limitations, the government conceded that the April 9 act had not occurred on that date, rendering S1 retroactively untimely. Nevertheless, the district court denied the motion, and the court of appeals affirmed.

The court began with the two-part test for relation back of a superseder: the original indictment must have been be validly pending and the superseder must not “materially broaden or substantially amend the charges.” The question raised here, one of first impression in this circuit, was “whether an indictment that is facially valid only because of one alleged overt act within the limitations period should be considered . . . validly pending . . . when the Government concedes that [that] overt act did not occur within the limitations period.” In answering this question in the affirmative, held that since S1 was “facially timely” when it was returned, it did not matter than it was, in actuality, untimely. The court noted that if the case had gone to trial on S1, the government could have satisfied the statute of limitations by proving a different, timely overt act. The court also noted that the government’s concession that S1 was untimely did not occur until after the return of S2. Thus, S1 was “facially timely and validly pending” at the time that S2 was returned.

The court also held that S2 did not broaden or amend the charges, since it merely extended the dates of the conspiracy by one week.

Comment: This is a disturbing decision. S1 was, as a factual matter, untimely both when it was filed and when S2 was filed. Why on earth should the case turn on the fact that S1 erroneously appeared to be timely? Is form really more important than substance? While the court carefully notes that there is no evidence that the government “deliberately withheld” its concession that the April 9 overt act had not occurred until after S2 was filed, the court’s confidence in the government would seem to be a bit naive. The facts here surely support a strong inference of deliberate withholding. The government must have known that there was a problem with the April 9 act when it decided to supersede; the defense had been pressing for an explanation of that act for more than a year. Given this, why else would the government have superseded unless it knew it had a timing problem? The government saved its case by not disclosing the defect until after S2 was filed. This decision would thus seem to give a free pass to all ethically challenged prosecutors - as long as they successfully hide their misconduct until it is cured, the defendant has no remedy.


On the brighter front, this case has a nice discussion of one of the most vexing Guidelines issues - the instruction that loss calculations under § 2B1.1 need not calculated with precision and that a “reasonable estimate” is sufficient.

Calculating loss can be particularly difficult in stock fraud cases because so many factors contribute to the decline of share prices. Surprisingly, this is not an area where the circuit has given much guidance. Here it does, turning, unusually, to civil law - the “principles governing recovery of damages in civil securities fraud cases” - for assistance. These principles a triggered remand for resentencing because the district court relied exclusively on the testimony of a NASD expert that, in essence, attributed the total decline in the stock price to the defendant’s conduct. The court's failure to “even consider[] other factors” relevant to the decline was error.