This case presents a question of first impression under the Sentencing Guidelines: whether, in a fraud case, the Government prosecutor may estimate an amount of fraud loss requiring the District Court to raise the offense level under U.S.S.G. § 2F1.1 when the fraudulent activity constituting the offense could have resulted in no actual loss. We conclude that the Sentencing Code does not provide for an automatic increase in offense level based on an estimate of fraud loss when there is an impossibility of any actual loss.

* * * * * *

Muhammad Khan was convicted of eight counts of mail fraud, in violation of 18 U.S.C. § 1341, and aiding and abetting, 18 U.S.C. § 2, for carrying out a scheme to defraud two insurance companies of a total of $450,000. Muhammad Khan and his wife, Shehla Khan, natives of Pakistan, traveled from the United States to Pakistan where they got documents to show Muhammad's death. On her return to the United States, Shehla Khan filed claims for benefits under the two life insurance policies. As Muhammad Khan had arranged, Shehla Khan acted under the guidance of a friend, Sultan Wahla, also a native of Pakistan and a lawyer. The insurance companies declined to pay the claims, Ms. Khan sued for payment, and the fraud was discovered. In addition, Ms. Khan, with the help of, or at the instigation of, Wahla filed an application with the Social Security Administration for survivor's benefits for herself and the Khans' five children. The Social Security Administration denied the application for benefits because Muhammad Khan had not worked a sufficient number of quarters for his family to be eligible for survivor's benefits.

Following his conviction Muhammad Khan was sentenced under the Sentencing Guidelines, § 2F1.1, Fraud and Deceit. This section of the Guidelines, in the June 1988 version applicable to Khan's offense, carries a base offense level of 6 simply for conviction of the substantive offense. In addition, if the Government establishes a fraud loss greater than $2,000, the base offense level is raised in increments, depending on the amount of the fraud loss.1 The Government can establish a fraud loss amount based on the offense of conviction as well as on relevant conduct. Under the Sentencing Code, Khan's base offense level was automatically increased seven levels on the basis of the insurance policy amounts, $450,000, and one additional level on the basis of the fraudulent social security claim Shehla Khan also made, the latter brought in as relevant conduct under § 1B1.3(a)(2) and § 3D1.2(d) and estimated as $69,000. Khan did not challenge the fraud loss amount of the insurance policies but objected to the inclusion of $69,000 as an estimated amount of fraud loss on the social security claim. Thus, the question before us involves only the estimate of fraud loss brought in as relevant conduct.2

* * * * * *

We have before us the rare case where, in the face of complete success, the fraud generated no loss. The Khans completed their fraud. The Social Security Administration did not discover the fraud. Thus, the deception succeeded. Nevertheless, the Social Security Administration did not and could not suffer any dollar loss based on payments to the Khans. It refused payment to the Khans because Muhammad Khan had not worked a sufficient number of quarters for his family to qualify for survivor's benefits. The Social Security Administration could, and perhaps did, believe every detail of the deception, the whole story of Muhammad Khan's death, but did not award any money because none was available to the Khans.

In such a case as this, where no dollar loss is possible for reasons entirely unrelated to the fraud or its discovery, the court does not have available to it the increases in sentencing level based on fraud loss. The Government prosecutor may not validly insist that the District Court raise the sentence based on such information disclosed by the prosecutor. Instead, as explained below, it has available the base offense level, the increase in offense level for such offense characteristics as a misrepresentation that the defendant was acting in behalf of a charity or concealed the fraud in a foreign bank account, and an upward departure based on features of the offense which show special gravity but which are not addressed in the other offense level calculations.

Our decision that an offense level may not be increased on the basis of an estimated fraud loss when no actual loss is possible is grounded in an examination of the following: (1) the overall design of the Guideline provision on Fraud and Deceit; (2) the context of the cases in which this provision has been used to arrive at a proper sentence; and (3) the alternative ways by which a court may adjust a defendant's sentence so that it reflects the seriousness of the offense.

Fraud statutes cover a broad spectrum of illegal acts and kinds of harm "with extreme variation in severity." U.S.S.G. § 2F1.1, Background (June 15, 1988). The Guideline framework is therefore designed to allow a range of punishments for defendants convicted of fraud to address "potential harmfulness and the dangerousness of the offender, independent of the actual harm." Id. The provisions of § 2F1.1, read together with the application notes, allow fine tuning so that the punishment meted out to a particular defendant can specifically address the harm to the victim and the gain to the defendant.

Fraud, even without regard to specific harm to a victim, presents a grave problem. Thus, the Guidelines provide for a base offense level of 6 without regard to whether the fraud could or did create any specific harm and without regard to any other circumstance involved in the fraud. See § 2F1.1(a). The defendant is punished at this offense level simply on the basis that the Government proved the elements of the fraud offense.

Because the base offense level often does not identify the seriousness of the offense, the Guidelines offer a series of ways to enhance a sentence for fraud. First, the Guidelines allow what the Seventh Circuit in United States v. Schneider, 930 F.2d 555, 559 (7th Cir.1991), has called "bonus punishment points for different levels of proven loss beginning with $2,000." See § 2F1.1(b)(1). This loss is "the value of the property taken, damaged, or destroyed." § 2B1.1, Application Note 2 (June 15, 1988) (cross reference from 2F1.1, Application Note 7). In some cases, the actual loss may not reflect the gravity of the loss that would have occurred but for discovery. In such a case, the court is not limited to the actual loss figure. If the "probable or intended loss ... can be determined," the court should use that figure "if it is larger than the actual loss." § 2F1.1, Application Note 7. Although the fraud loss estimate must reflect "economic reality," see Schneider, 930 F.2d at 559, the loss figure "need not be precise" but may be a "reasonable estimate." § 2F1.1, Application Note 8.

These application notes describing the automatic, mandated, incremental increases for fraud loss amounts assume a fraudulent scheme that would have created some actual loss but for interruption of the scheme by detection or failure to carry out all the steps necessary to succeed. None of these notes suggests that a fraud loss must be estimated and punished in the case where a fraudulent scheme could result in no loss. That is, the Guidelines do not provide for incremental increases in offense level based on fraud loss in the case where, even if the fraud were never detected, there would be no actual loss because the scheme fails on other grounds.

Cases decided under fraud statutes with sentences under the Fraud and Deceit guideline are consistent with this proposition: These cases all involve a context of fraudulent schemes that could result in actual loss, and they have underlying them a presumption that actual loss was possible if the deception entailed in the fraud succeeded. See United States v. Jones, 933 F.2d 353 (6th Cir.1991) (in credit card fraud case, fraud loss included interest charge as well as cost of items charged); United States v. Sloman, 909 F.2d 176 (6th Cir.1990) (in insurance fraud case, Court approved a fraud loss calculation of the amount insurance company paid on falsely obtained claim insuring boat, less salvage value of boat). Cases in other circuits illustrate similar contexts. See United States v. Koenig, 952 F.2d 267 (9th Cir.1991) (fraud loss calculated for scheme to use counterfeit automatic teller machine cards); United States v. Kopp, 951 F.2d 521 (3d Cir.1991) (fraud loss calculated in context of fraudulently obtained loans); United States v. Brach, 942 F.2d 141 (2d Cir.1991) (calculation of loss for fraudulently obtained loan); United States v. Schneider, 930 F.2d 555 (7th Cir.1991) (fraud loss calculation for illegally obtained contract); United States v. Davis, 922 F.2d 1385 (9th Cir.1991) (fraud loss calculation based on goods defendant ordered in scheme whereby he persuaded merchants he had wired money for goods to their accounts).

Our understanding of § 2F1.1(b)(1) as requiring an automatic incremental increase in offense level only when actual loss is possible is supported by the alternative measures available to increase a sentence if the sentence does not otherwise reflect the seriousness of the defendant's fraud. The system of bonus punishment points outlined in § 2F1.1(b)(1) is only one of the ways the Guidelines mandate increasing a sentence under a fraud statute. In the version of the Guidelines under which Khan was sentenced, effective June 15, 1988, a court must enhance the defendant's sentence if the offense involved "more than minimal planning," a scheme to defraud more than one person, a misrepresentation that the defendant was acting on behalf of a charitable institution, or a violation of a judicial order. See § 2F1.1(b)(2). Under this provision the court must add 2 to the offense level or, if the offense level was less than 10, must increase the level to 10. Id. The Application Notes add an element of restraint here, however, by calling attention to the fact that these enhancement factors are written as alternative rather than as cumulative factors. A court could not, on the basis of these factors, add more than two levels unless the offense level was less than 10 after the enhancement. See also 2F1.1(b)(3) (addressing another serious factor in fraud cases, the problem that occurs when the defendant has used "foreign bank accounts or transactions to conceal the true nature or extent of the fraudulent conduct"; in such a case the court is to increase the offense level to 12 if it is not already at 12).

The Guidelines provide even further for fine tuning the defendant's sentence under the Fraud and Deceit section by calling attention to the broad departure provisions. Although the enhancements available in § 2F1.1(b)(2) are alternative rather than cumulative, so that the court may not stack a series of offense characteristics, adding 2 for each one, the court may make an upward departure if several of these features are present. See 2F1.1, Application Note 1. Because dollar loss may not illustrate the seriousness of the fraudulent conduct, the Guidelines encourage departure so that the offense level may reflect the true nature of the offense. See Application Note 9 (upward departure) and Application Note 11 (downward departure). Because the bonus punishment points do not increase for amounts of loss greater than $5,000,000, the court may depart upward based on amount of fraud loss if the loss "substantially exceeds that amount." Application Note 10. Finally, enhancements available for all offenses are available for fraud offenses. See, e.g., Application Note 13 (calling attention to the § 3A1.1 enhancement for vulnerable victim). This examination of the alternatives available to design a defendant's sentence within the Guidelines framework for Fraud and Deceit, § 2F1.1, makes clear that the increased offense level based on monetary loss is only one of the ways the Guidelines allow a court to adjust a sentence for fraud to make it fit the seriousness of the defendant's conduct and the degree of harm created.

In the case before us Khan was not eligible for the benefits he fraudulently applied for. Thus the Social Security Administration risked no loss from its benefits program as a result of his fraudulent application. Therefore the Government may not seek an automatic mandated increase in offense level based on an amount of benefits he might have received over a ten year period if he had been eligible for benefits. The Government could, however, have sought an upward departure if the sentence based on the insurance loss amount did not reflect the seriousness of the harm caused by the Khans' fraudulent application for benefits.

Khan also challenged as an abuse of discretion the trial court's admission at trial evidence concerning the fraudulent Social Security application. We find that objection to be without merit.

This Court reviews a trial court's admission of Rule 404(b) evidence of "other acts" under a deferential abuse of discretion standard. United States v. Acosta-Cazares, 878 F.2d 945 (6th Cir.), cert. denied, 493 U.S. 899, 110 S. Ct. 255, 107 L. Ed. 2d 204 (1989). This Court " 'look[s] at the evidence in a light most favorable to its proponent, maximizing its probative value and minimizing its prejudicial effect.' " United States v. Pollard, 778 F.2d 1177, 1170 (6th Cir.1985) (citations omitted). When the District Court does a Rule 403 balance of the probative value compared to the prejudicial effect of evidence, it should exclude evidence "only where the probative value of the relevant evidence is substantially outweighed by the danger of unfair prejudice." Id. (citation omitted). In denying the defendant's motion in limine to exclude this evidence, the District Court noted that in a mail fraud case the Government had to prove " 'scheme or artifice to defraud' " and that "other fraudulent acts which are substantially similar and reasonably near in time to the acts which form the basis for the indictment are admissible under Fed.R.Evid. 404(b) to prove the required scheme." In addition, the court considered the submission of false claims under Social Security to be evidence relevant to show that the defendants knew what they were doing when they submitted false insurance claims. The court thought the evidence particularly relevant in a case in which the defendants were likely to "assert a defense of ignorance or lack of familiarity with the English language or American customs." In a footnote, the Court said the defendants do not speak English and that an interpreter would be necessary for trial. The Court considered the balance under Rule 403 and held that the relevant nature of the evidence outweighed the prejudicial impact. This evidence of a multi-prong scheme makes the Khans' defense--that they did not know what they were doing--difficult to sustain. We do not believe the District Court abused its discretion in admitting this evidence.

We thus affirm the defendant's conviction and remand for resentencing in light of this opinion.