UNITED STATES of America, Plaintiff-Appellee, v. Armando NAVARRETE, Defendant-Appellant, 667 F.3d 886

POSNER, Circuit Judge. A jury convicted the defendant of defrauding LaSalle Bank, a large bank in Chicago (later acquired by Bank of America), in violation of federal bank-fraud law, and of related federal offenses. The district judge sentenced him to 96 months in prison and ordered him to forfeit the money that he had obtained from the fraud, which the judge determined to be $16,241,202, plus property that he had bought with proceeds of the fraud, and to pay restitution to the bank also in the amount of $16,241,202. The only challenge by the defendant’s lawyer on appeal is to the order of restitution.

We permitted the defendant to file his own “supplementary” briefs, which raise additional issues — but they are frivolous. A confusing feature of the case is that the government tells us that it’s planning to convey any forfeited assets that turn up to the bank (which is to say to its successor, Bank of America). One might think that if the government does that, the order of restitution will be moot — the victim of the fraud will have been made whole, especially since the amount of the forfeiture exceeds the bank’s loss because of the property that was ordered forfeited.

The cases, moreover, interpreting vague statutory language, see 18 U.S.C. §§ 981(a)(1)(C), 982(a)(2)(A); 28 U.S.C. § 2461(c), allow the sentencing judge to make the forfeiture order in personam rather than in rem, so that it is a personal judgment against the defendant rather than a claim to specified assets. E.g., United States v. Baker, 227 F.3d 955, 970 (7th Cir.2000); United States v. Newman, 659 F.3d 1235, 1242 (9th Cir.2011); United States v. McGinty, 610 F.3d 1242, 1246 (10th Cir.2010).

The judge did that in this case, and so the order will remain in force when the defendant is released from prison, just like the order of restitution. He appears to be a capable businessman, so, while he has no money now (he was permitted to appeal in forma pauperis), he may eventually be able to pay both debts in full. If so, the government will not be permitted to pay the bank twice the bank’s loss, because restitution (as we’ll see) cannot exceed the victim’s loss; instead the government will pocket the proceeds of the forfeiture.

So the defendant if able will be paying twice the victim’s loss, though not to the victim. And so a successful attack on the restitution order will reduce the defendant’s total liability by almost half (“almost” because of the property forfeited along with the defendant’s monetary gain from the fraud — but remember that he is not attacking the forfeiture order). It might seem that a defendant should never have to pay more than his victim’s loss, and therefore that a forfeiture order and a restitution order should be considered alternative rather than cumulative punishments. But the cases hold otherwise.

United States v. Emerson, 128 F.3d 557, 566-67 (7th Cir.1997); United States v. Neuman, supra, 659 F.3d at 1240-42; United States v. McGinty, supra, 610 F.3d at 1246-48. That’s not much of a paradox. It means, very appropriately in a fraud case since fraud is a concealable offense, that the combination of a forfeiture order and a restitution order results in a form of punitive damages piled on top of the other penalties for the defendant’s crime, such as imprisonment.

The defendant owned a company called Illinois National Safe (we’ll call it “the Company”) that was in the business of installing and maintaining security devices, such as surveillance cameras, closed-circuit television, locks, and fire and burglar alarms. One of its customers was LaSalle Bank. In 1998 George Konjuch, the bank’s vice president in charge of security, told the defendant that he would authorize the Company to provide additional services for the bank in exchange for bribes — and that he would cut off the Company if the defendant refused his offer. The defendant accepted the offer. Years passed.

The Company’s sales to the bank soared — it went from servicing 30 branches of the bank to servicing almost 150 — and Konjuch received larger and larger bribes until he was receiving more than $40,000 a month from the defendant. Eventually LaSalle’s management became suspicious of the large amounts of money being paid to the Company because the bank was being billed by it in hundreds of invoices each for less than $5,000.

Konjuch kept them below that level and thus within the range in which he could authorize payments without a supervisor’s approval; in addition, a large bank like LaSalle makes many small payments on a daily basis, and he could hope that his superiors wouldn’t add them up and learn how much the bank was paying the Company. The tactic failing, in 2006 the bank stopped doing business with the Company, having paid it more than $45 million since 2001. The defendant had paid Konjuch $1.3 million in bribes.

For purposes of calculating a prison sentence, if the victim’s loss can’t be estimated the offender’s gain can be used to approximate the loss. U.S.S.G. § 2B1.1, Application Note 3(B); United States v. Vrdolyak, 593 F.3d 676, 680 (7th Cir.2010); United States v. Chatterji, 46 F.3d 1336, 1340 (4th Cir.1995).

But an order of restitution, unlike either a prison sentence or an order of forfeiture (including the order of forfeiture in this case, which added to the defendant’s monetary gain the property that he bought with the proceeds of the fraud), can be based only on the victim’s loss, 18 U.S.C. § 3664(f)(1)(A); United States v. George, 403 F.3d 470, 474 (7th Cir.2005); United States v. Galloway, 509 F.3d 1246, 1253 (10th Cir.2007); see also 18 U.S.C. §§ 3663A(c)(3)(B), 3664(e), even though disgorgement of an ill-gotten gain is a standard example of restitution in civil cases.

It is not clear whether the district judge understood the relation between forfeiture and (criminal) restitution. He did not discuss restitution separately in sentencing the defendant except to say that the restitution would be equal in amount to the forfeiture. After severing its relationship with the Company, the bank hired Ingersoll Rand to provide security services that the Company had provided.

The government’s experts (forensic accountants) compared Ingersoll Rand’s prices to the Company’s prices, found that the former were lower, and used the difference to calculate the loss that the fraud had caused the bank between 2001 and 2006. The figure they came up with was $29.6 million. But the judge, in calculating the amount of the bank’s loss for purposes of determining the length of the defendant’s prison sentence, rejected the government’s loss estimate.

He noted that the services provided by the Company were not identical to those provided by Ingersoll Rand when it replaced the Company as the bank’s vendor and that had the Company charged the same prices that Ingersoll Rand charged it would have been operating at a loss— which could mean that the Company’s services had been better or more extensive than Ingersoll Rand’s and that was why they cost the bank more.

The judge didn’t think there was persuasive evidence that the Company’s cost structure had been significantly “inflated” as a result of the bribes, though the bribes had gotten the Company more business with the bank.

He was sure the bank had suffered a loss, but threw up his hands at quantifying it: “I am persuaded that LaSalle Bank paid [the Company] significantly more than it would have paid Ingersoll Rand or another similar company for the same services, but I just can’t reasonably determine exactly how much more.” Despairing of determining the loss, the judge ordered the defendant to pay restitution equal to the defendant’s gain from the fraud.

Though conceding as it must that the amount awarded as restitution must be based on the victim’s loss rather than on the defendant’s gain, the government defends the award (it thinks the amount should have been higher, but didn’t cross-appeal). It argues that since the bank would not have paid the Company a dime had it known the defendant was bribing Konjuch, the bank “lost” the money that ended up in the defendant’s pocket; his gain was the bank’s loss. That is too hasty. Suppose the Company’s (and so the defendant’

s) gain from the bribes was based on a combination of charging the bank more than it would have had to pay a competitor of the Company for identical services and selling more services to the bank — that is, edging out competitors.

Imagine that instead of selling the bank 1000 burglar alarms at a price of x that would yield a profit of $50 for each alarm (which let us say was the pre-bribe quantity and price sold by the Company to the bank), the defendant had been able by bribing Konjuch to induce the bank to buy 1500 burglar alarms from the Company at a price, $r + $10, that would yield a $60 profit per alarm, displacing a competitor who had been selling the bank 500 alarms each year also at $r apiece.

The Company’s ill-gotten gain would have been $40,000 (1000 X ($60 — $50 = $10) + 500 x $60), consisting of the entire profit on the additional 500 alarms the Company sold to the bank and the inflated profit (the extra $10 per alarm) on the original 1000 alarms that it had sold to the bank. But the bank’s loss would have been only $15,000 (1500 x ($60 — $50 = $10)), consisting of the $10 excess profit obtained by the Company on all 1500 alarms that it sold to the bank. There is, it is true, another victim of the fraud — the competitor who was edged out as a result of the bribes.

But the only victim for whom restitution is being sought in the present case is the bank. According to the government’s own figures, the Company’s sales to the bank rose from $2,404,622 in 2001 to $14,797,364 in 2004 — an increase of 515 percent. Is it conceivable that the Company increased its prices by 515 percent in three years? Or that Konjuch could have concealed a 515 percent excess markup from his superiors for so long a period without being caught? There is no evidence of that. The Company must therefore have been selling more to the bank.

Maybe the bank was paying more for the extra services than it would have had to pay a competitor edged out by the fraud, as in our example; or maybe they were services that the bank didn’t want and would not have bought had it not been for Konjueh’s treachery. There is little if any evidence to support either conjecture. Remember that in 1998 the Company was servicing 30 branches and in 2006 150. We don’t know exactly how much of this growth took place between 2001 and 2004.

We know only that the Company added an average of 15 branches each year over the period of the fraud (which had begun in 1998), which suggests (no stronger word is possible, because we don’t know how much variance there was), that in 2001 the Company was servicing 75 branches (30 + (3 x 15)) and by 2004 120 (75 + (3 x 15)).

This is only a 60 percent increase, consistent with the hypothesis that the Company kept raising its prices dramatically, but as we said we don’t know how much variance in the growth rate there was from year to year, and, more important, we do not know whether the Company was providing the same quality and quantity of services to each branch. In addition, in 2004 the Company and the bank signed a $7.8 million maintenance contract, which doubled the Company’s revenues from selling to the bank — from $7.16 million in 2003 to $14.8 million the following year.

It’s true that the bank paid less to Ingersoll Rand, the Company’s honest replacement, for comparable services than it had paid the Company. But the bank may have been effecting economies that it thought would make it more attractive to potential buyers, because it was looking to sell itself — and 18 months later it did. And the bank’s demand for the kind of services that the Company had provided to it may have shrunk for reasons unrelated to overpricing or gold plating by the Company or even wanting to save money.

Or perhaps the services that the Company had provided and Ingersoll Rand did not provide were still being bought by the bank, only from some other company or companies. Maybe the government’s experts explored all these possibilities and their $29.6 million figure is an accurate estimate. But the thrust of the government’s response to the appeal is not to defend its experts’ loss assessment against the judge’s criticisms or the defendant’s criticisms but instead to insist that in computing the defendant’s illicit gain the judge was- — -without knowing it — computing the victim’s loss.

The government bases this argument not on evidence, however, but on the proposition that any payment pursuant to contracts obtained by fraud must enrich the seller by the precise amount paid by the purchaser, net of the seller’s costs. That is not correct, as we have said. Indeed, in an extreme case, where the only effect of the fraud is to transfer business between two sellers whose prices and quality are the same, there is no loss to the purchaser. The price of the Ingersoll Rand replacement contract could, however, be made the starting point for computing the loss to the bank.

If that contract price was x percent lower than the Company’s price, then * percent of the quantity of services (in dollars) that the Company sold to the bank could be treated as the bank’s prima facie loss. The defendant could rebut by showing that the quality or quantity of services rendered under the Company’s contracts exceeded the quality or quantity of services rendered under Ingersoll Rand’s contract; if so, the bank’s loss would be less, and maybe zero (or even negative). Much of the relevant information may be buried in the worksheets of the government’s forensic accountants.

So maybe the judge threw up his hands too soon, and should try to estimate the bank’s loss along the lines just suggested, though alternatively he may decide to invoke 18 U.S.C. § 3663A(c)(3)(B), which allows the sentencing judge to decline to award restitution if he “finds, from facts on the record, that ... determining complex issues of fact related to the cause or amount of the victim’s losses would complicate or prolong the sentencing process to a degree that the need to provide restitution to any victim is outweighed by the burden on the sentencing process.” That is a choice for the district judge to make in the first instance; we note that the need to determine the bank’s loss is limited in this case because of the government’s decision to convey forfeited assets to the victim up to the limit of the victim’s loss.

All that is clear is that the order of restitution cannot stand. The rest of the judgment is affirmed. Affirmed in Part, Reversed in Part, and Remanded.