This case involves an appeal from a 25-month sentence given to the defendant-appellant, Ronnie Lee Johnson. Johnson bases his appeal on two sections of the U.S. Sentencing Guidelines. He argues 1) that the District Court erred when it increased his offense level under § 3B1.3 for an abuse of trust, and 2) that the District Court incorrectly applied § 5G1.3, which governs sentencing when a defendant is already subject to an ongoing term of imprisonment. On the § 5G1.3 issue, we find that the District Court properly applied the Sentencing Guidelines. We have no occasion to pass on the abuse of trust issue because we conclude, based on our review of the record as a whole, that any error on that issue would not have affected Johnson's sentence. We therefore affirm the District Court's judgment.

Ronnie Lee Johnson operated an employee leasing and payroll service business in Hammond, Indiana. Trucking companies throughout the United States would arrange for their employees to become the employees of Johnson's business, thereby taking advantage of Indiana's low workers' compensation rates. Johnson, however, never paid over $1 million in federal taxes that had been withheld from employee paychecks. The Government charged Johnson in the U.S. District Court for the Northern District of Indiana for willful failure to collect and pay over taxes in violation of 26 U.S.C. § 7202. Johnson pleaded guilty pursuant to a plea agreement that provided that Johnson was to receive the minimum sentence under the Sentencing Guidelines.

The District Court's calculations put Johnson at offense level 20 (which included a two-level enhancement for abuse of trust) and Criminal History Category (CHC) II, suggesting a sentencing range of 37-46 months. When Johnson was sentenced in Indiana, however, he had already served 16 months of a 37-month sentence arising from a conviction in the U.S. District Court for the Middle District of Florida. That sentence was for wire and mail fraud and was also related to Johnson's employee leasing and payroll service business. Because Johnson was subject to an undischarged term of imprisonment, the Indiana federal court had to apply § 5G1.3 of the Sentencing Guidelines. Based on those calculations, the District Court departed downward, sentencing Johnson to 25 months in prison to run concurrently with the 21 months remaining on his existing sentence. The District Court also ordered Johnson to pay $1,135,484.81 in restitution to the Internal Revenue Service (IRS).

Under 18 U.S.C. § 3742(a), we ordinarily do not have jurisdiction to review a sentence in which a district court departed downward from the applicable guideline range. See United States v. Burnett, 66 F.3d 137, 139 (7th Cir.1995); Bischel v. United States, 32 F.3d 259, 265 n. 6 (7th Cir.1994). Unless the court violated the law or incorrectly applied the Sentencing Guidelines (as opposed to merely exercising the court's discretion), a defendant generally may not appeal from a downward departure. See Burnett, 66 F.3d at 139.

Johnson's first assertion of error, however, is that the District Court miscalculated his offense level--an application error that we certainly may review. Johnson argues that the District Court erred when it found that Johnson abused a position of trust and therefore enhanced Johnson's offense level from 18 to 20 under § 3B1.3 of the Sentencing Guidelines. As we explain below, however, we find that any such error was harmless because the District Court would have imposed the same sentence even under Johnson's proposed offense level. When the court went through the mechanics of applying § 5G1.3, the two-point enhancement for abuse of trust in effect dropped out of the court's sentencing calculations. Our inquiry, therefore, will focus on Johnson's second assertion of error, which involves the District Court's grouping of offenses when applying § 5G1.3.

Both Johnson and the Government agree that § 5G1.3(c) was the proper starting point for the District Court. In the 1995 version of the Sentencing Guidelines (which was the version that applied to Johnson), § 5G1.3(c) provides that "the sentence for the instant offense may be imposed to run concurrently, partially concurrently, or consecutively to the prior undischarged term of imprisonment to achieve a reasonable punishment for the instant offense." The commentary to the Sentencing Guidelines (which generally binds the federal courts like the Guidelines themselves, see Stinson v. United States, 508 U.S. 36, 44-46, 113 S. Ct. 1913, 1919, 123 L. Ed. 2d 598 (1993)) further provides that "[t]o achieve a reasonable punishment and avoid unwarranted disparity," courts applying § 5G1.3(c) should consider the standard sentencing criteria enumerated in 18 U.S.C. § 3553(a) as well as "be cognizant" of:

(a) the type ... and length of the prior undischarged sentence;

(b) the time served on the undischarged sentence and the time likely to be served before release;

(c) the fact that the prior undischarged sentence may have been imposed in state court rather than federal court, or at a different time before the same or different federal court; and

(d) any other circumstance relevant to the determination of an appropriate sentence for the instant offense.

U.S. Sentencing Guidelines Manual (USSG) § 5G1.3, comment. (n.3) (1995).

When sentencing Johnson, the District Court relied on our then-recent case, United States v. Goudy, 78 F.3d 309 (7th Cir.1996). Goudy involved the application of the 1994 version of § 5G1.3. That version and its commentary gave courts more definitive instructions than the 1995 version:

To the extent practicable, the court should consider a reasonable incremental penalty to be a sentence for the instant offense that results in a combined sentence of imprisonment that approximates the total punishment that would have been imposed under § 5G1.2 (Sentencing on Multiple Counts of Conviction) had all of the offenses been federal offenses for which sentences were being imposed at the same time.

USSG § 5G1.3, comment. (n.3) (1994). By counterfactually anchoring sentences around the punishment a defendant would have received had the sentences been imposed simultaneously, this approach attempts to prevent the Government from taking advantage of multiple prosecutions to get longer total punishments. See generally Witte v. United States, 515 U.S. 389, 403-07, 115 S. Ct. 2199, 2208-09, 132 L. Ed. 2d 351 (1995); United States v. Lechuga, 975 F.2d 397, 400 (7th Cir.1992). This approach also promotes one of the overall goals of federal sentencing, namely parity "among defendants with similar records who have been found guilty of similar conduct," 18 U.S.C. § 3553(a)(6).

The 1994 Guidelines accomplish this goal only imperfectly, however, because they did not require courts to depart downward to achieve parity. Rather, the 1994 Guidelines state that a court generally "may achieve an appropriate sentence through its determination of an appropriate point within the applicable guideline range for the instant federal offense, combined with its determination of whether that sentence will run concurrently or consecutively to the undischarged term of imprisonment." USSG § 5G1.3 comment. (n.3) (1994) (emphasis added); see also United States v. Brassell, 49 F.3d 274, 278 (7th Cir.1995). In other words, the 1994 Guidelines instruct a court to promote parity as best as it can within the guideline range for the instant offense. Indeed, the 1994 Guidelines specifically contemplate situations where a defendant's total punishment will not fall within the same guideline range that the defendant would have faced if sentenced simultaneously for both crimes. See USSG § 5G1.3(c) comment. (n.3) (1994). If a defendant, for example, has already served most of his undischarged sentence, the minimum sentence for the instant offense--even if ordered to run concurrently--may well result in a total term of imprisonment in excess of the term that simultaneous sentencing would have imposed. The 1994 Guidelines, moreover, do not instruct courts applying § 5G1.3(c) to give defendants credit for time already served on undischarged sentences.1

The district courts, therefore, arguably have more discretion when applying § 5G1.3(c) under the 1995 Guidelines than under earlier versions. The 1995 Guidelines have shifted to a more flexible approach under which courts need only "consider" certain factors, notably including "the time served on the undischarged sentence and the time likely to be served before release." One might argue that the 1995 Guidelines give district courts new freedom to depart downward based on time served on undischarged sentences. And because the District Court here ostensibly followed the approach suggested by Goudy and the 1994 Guidelines, the court might have misunderstood its discretion. Remand is inappropriate here, however, because the District Court actually went beyond the 1994 Guidelines approach and, in effect, gave Johnson credit for the time already served. As detailed below, the District Court correctly calculated the hypothetical sentence Johnson would have received and then used its discretion to depart downward based on time already served.

The District Court first followed § 5G1.2 and asked what Johnson's punishment would have been had he been sentenced for both counts at the same time. Section 5G1.2 referred the court to the "grouping" analysis set forth in §§ 3D1.1-3D1.5. The District Court concluded that the two offenses would not have been grouped together as "closely related counts" under § 3D1.2. Johnson challenges this conclusion, arguing that the counts should have been grouped. He argues that the victims in both the Florida and Indiana cases are the same because "they are all members of the trucking industry." Appellant's Br. at 16. The Sentencing Guidelines, of course, mandate the grouping of offenses that "involve substantially the same harm," which can mean that the "counts involve the same victim and the same act or transaction." See USSG § 3D1.2(a). Johnson's argument falls flat on its face, however, because the primary victims were different in the two cases. Johnson's conviction in the Florida case was based on his defrauding of truckers by providing non-existent insurance coverage. His conviction in Indiana, however, is based on his failure to pay taxes to the federal government. After all, Johnson has to pay over $1 million in restitution to the IRS, not to members of the trucking industry. Mail fraud and tax evasion would seem in most cases to involve different harms, see United States v. Seligsohn, 981 F.2d 1418, 1425 (3d Cir.1992), and in this case, the counts certainly do. The District Court therefore did not err when it refused to group the Florida and Indiana offenses.

The District Court then applied § 3D1.4 to determine Johnson's hypothetical combined offense level, and it is here where the abuse of trust enhancement drops out of the sentencing calculations and thus becomes moot. Johnson's Florida conviction involved an offense level of 21, and the District Court determined that the offense level for Johnson's Indiana conviction was 20. Applying § 3D1.4, the District Court therefore calculated that Johnson's combined offense level would have been 23. Prior to these two offenses, Johnson would have been in CHC I, so the District Court used the Sentencing Table to conclude that Johnson "should have" received a total punishment of between 46 and 57 months. Based on "the significant level of victim impact associated with the defendant's fraudulent insurance scheme," the District Court chose 50 months as the hypothetical combined sentence on which to base the Indiana sentence. The District Court, however, lowered this 50-month hypothetical sentence to 41 months in order to match the 9-month sentence reduction given to Johnson in the Middle District of Florida based on Johnson's substantial cooperation with the Government. See Fed.R.Crim.P. 35(b).

How could the District Court effectuate 41 months of total imprisonment? The Sentencing Guidelines prescribed a guideline range of 37-46 months for the Indiana offense (based on offense level 20 and CHC II). Without departing, therefore, the best the District Court could have done to promote parity would have been a 37-month sentence to run concurrently with Johnson's undischarged sentence. Because Johnson had already served 16 months, such a sentence would have kept Johnson in prison for a total of 53 months--one year more than the 41 months the District Court thought appropriate. The District Court therefore departed downward, giving Johnson a sentence of 25 months to run concurrently with his prior sentence. Thus, the District Court not only considered the hypothetical sentence but actually departed downward to achieve it, in effect giving Johnson credit for the time already served on his Florida sentence.

Remember, however, that Johnson argues that his Indiana offense level should have been 18 rather than 20. Even if Johnson is correct, the Sentencing Guidelines' grouping analysis yields the same combined offense level that the District Court actually used.2 See USSG § 3D1.4(a). So regardless of the abuse of trust enhancement, the District Court would have calibrated Johnson's sentence using the same hypothetical guideline range. And there is not the slightest hint in the record that the District Court's selection of a sentence within that range had anything to do with the abuse of trust enhancement. The District Court's decision that Johnson should be in prison for a total of 41 months was therefore independent of any error regarding the abuse of trust issue. Moreover, an offense level of 18 would not have given the District Court any new latitude to lighten up on Johnson without making a downward departure. An offense level 18 combined with CHC II would have meant a guideline range for the Indiana offense of 30-37 months. The District Court's 25-month sentence, in other words, would require a downward departure even under Johnson's assumptions.3

This case is therefore one of the few sentencing cases where we can conclude that any "error was harmless, i.e., that the error did not affect the district court's selection of the sentence imposed." Williams v. United States, 503 U.S. 193, 203, 112 S. Ct. 1112, 1121, 117 L. Ed. 2d 341 (1992). Throughout the sentencing process, the District Court's touchstone was the sentence Johnson would have received had he been sentenced at the same time for the Florida and Indiana offenses. Because the abuse of trust enhancement did not affect this touchstone, we are confident that any error regarding the enhancement was harmless. Cf. United States v. Hardy, 99 F.3d 1242, 1253 n. 13 (1st Cir.1996) (finding harmless error where district court used correct hypothetical sentencing range but for wrong reason).

The judgment of the District Court is AFFIRMED.