Multi-million-dollar-fraud sentencing: How can you seek a reasonable sentence when the sentencing range will be driven by the high loss amount?

Your client has been convicted of a financial crime, and the high loss amounts—perhaps in the tens of millions of dollars—are driving the possible sentencing range to an extreme. How do you argue for a reasonable sentence?

I recently posted about a successful Sixth Circuit appeal in a $70 million credit-union collapse. After remand back to the sentencing judge, I filed this Sentencing Memorandum, contending that the fraud Guidelines under §2B1.1 are too severely influenced by the loss amount. Excerpts of that memorandum appear below, including (i) statistics from the U.S. Sentencing Commission regarding the average sentences imposed relative to the Guideline range, and (ii) a detailed chart compiled at the Federal Defender website showing sentences imposed for many of the most-famous financial crimes in the country—many of those sentences were below the Guidelines range. In this case, the initial sentence of 18 years was reduced to 10 after the successful appeal.

[Introduction, Background, etc., omitted]


I. Sentencing Principles Applicable Here.

A. General principles and fraud guidelines.

Every federal sentencing case involves application of 18 U.S.C. § 3553, and this Court has no doubt seen and recited the relevant portions of that statute hundreds of times.  But it is worth emphasizing the fundamental command of that statute: The sentence imposed on a person must be sufficient, but “no greater than necessary” to comply with the purposes of sentencing set forth in the statute.  18 U.S.C. § 3553(a).

To determine what sentence is sufficient, but no greater than necessary, the Court considers the seven factors of § 3553(a), summarized as follows:

(1) the offense and the person being sentenced;

(2) the need for the sentence to provide just punishment, adequate deterrence, protection of the public, and appropriate treatment for the defendant;

(3) the kinds of sentences available;

(4) the Guidelines range for the defendant;

(5) pertinent policy statements of the Guidelines;

(6) the need to avoid unwanted sentence disparities among defendants with similar records who have been found guilty of similar misconduct; and

(7) the need to provide restitution to any victims.

The applicable Guidelines range (the fourth factor) generally provides a starting point that drives the entire sentence because it methodically incorporates many facets of the individual defendant and the circumstances of the crime.  The range is built on the defendant’s criminal history and the offense level, which captures many details of the crime itself.  For example, when (as here) a fraud crime drives the guideline range, the Guidelines account for the amount of loss, the defendant’s role in the offense, whether the loss was from a financial institution, and similar factors. 

Typically, the sentences on grouped counts run concurrently, as the sentencing range generally falls below the statutory maximum for the crime.  Sometimes, however, that Guidelines range (which embodies what the Guidelines refer to as the “total punishment”) is greater than the statutory maximum of the count driving the range.  If that occurs, the Guidelines provide that the district court shall impose consecutive sentences for various counts “but only to the extent necessary” to reach the “total punishment” within that range.  U.S.S.G. § 5G1.2(d).  In other words, consecutive sentences are permissible when necessary for a district court to reach up to a sentence within the range.   In the more-common scenario where the Guidelines range falls within the statutory maximum—as in this case—the “sentences on all counts shall run concurrently” (except for situations not applicable here where a statute specifically requires consecutive sentences).  Id.; see, e.g., United States v. Alcorn, 27 F. App’x 317, 321 (6th Cir. 2001) (“The district court then has discretion to sentence consecutively up to the upper limit of the range but cannot sentence consecutively beyond the upper limit.” (pre-Booker decision)).  If the district court then wishes to imposes a sentence above the range, it can impose an upward variance from the top of the range based on the standard post-Booker principles governing such variances (e.g., no presumption of reasonableness applies to sentences outside the range, greater explanation is needed for such sentences, etc.).  See United States v. Erpenbeck, 532 F.3d 423, 437–38 (6th Cir. 2008).

As the Supreme Court has explained, there are two reasons why it is generally fair to assume that the Guidelines reflect a rough approximation of sentences that “might achieve § 3553(a)’s objectives.”  Rita v. United States, 551 U.S. 338, 348 (2007).  First, the original Sentencing Commission used an empirical approach, which began “with an empirical examination of 10,000 presentence reports setting forth what judges had done in the past.”  Id.  Second, the Commission can review and revise the guidelines based on judicial feedback through sentencing decisions, and consultations with experts.  Id.  This is why so many federal sentences are imposed within the guideline range and why such sentences are routinely affirmed on appeal as “reasonable.”  Sentences outside the range are more unusual.  The “farther the judge’s sentence departs from the Guidelines sentence . . . the more compelling the justification based on factors in section 3553(a) must be.”  United States v. Aleo, 681 F.3d 290, 299 (6th Cir. 2012) (holding that sentence for child pornography at 145% over the top end of the range was substantively unreasonable, noting that Guidelines accounted for much of the conduct).

And while a sentence within the Guidelines range for most crimes is often considered appropriate, courts have repeatedly concluded that sentences driven by the fraud guidelines range are too extreme because the single factor of loss amount magnifies the range to unreasonable heights.  Federal courts in the Southern District of New York have had extensive experience with financial crimes, and they have recognized that the loss amounts create fraud guidelines that are too severe:

  • “By making a Guidelines sentence turn on this single factor, the Sentencing Commission ignored [§ 3553(a)] and . . . effectively guaranteed that many such sentences would be irrational on their face.”  United States v. Gupta, 904 F. Supp. 2d 349, 351 (S.D.N.Y. 2012) (emphasis added).


  • “What drove the Government’s calculation in this case, more than any other single factor, was the inordinate emphasis that the Sentencing Guidelines place in fraud cases on the amount of actual or intended financial loss.”  United States v. Adelson, 441 F. Supp. 2d 506, 509 (S.D.N.Y. 2006).

This helps to explain why defendants in fraud cases are frequently sentenced below the Guideline range, and almost never sentenced above the range. The Sentencing Commission’s statistics bear this out, as noted in the charts on the next three pages.  Sentencing and Guideline Information for §2B1.1 Offenders, United States Sentencing Commission Symposium on Economic Crime (2013).



B. Three nationwide facts regarding fraud sentencing.



1. 97.7% of all fraud offenders are sentenced within or below the range.

In 2012, sentences for all 8,507 fraud cases were imposed as follows: Within the range: 50.6% (red).  Below the range: 47.1% (blue & green). Above the range: 2.3% (yellow).  (This is shown on right-most part of the graph below).


Sentencing Trends for §2B1.1 Offenders: Fiscal Years 2003 - 2012



2. For loss amounts between $2.5M and $7M, the sentence imposed is within or below the guideline range approximately 98.8% of the time (and mostly below).

When the loss amount is between $2.5M and $7M (as here), the sentence is imposed within the range 32.9% of the time (red) and below the range 65.9% of the time (blue & green)—a total of 98.8% of cases.  The sentence is above the range in 1.2% of cases.


Analysis of §2B1.1 Offenders in Each Loss Table Category: Fiscal Year 2012



3. The average sentence imposed for range used here (87–108) is 66 months (5.5 years).

The average sentence in fraud cases is almost always below the average Guidelines range.  Specifically, when the bottom of the Guideline range is 87 months (the range used for Nikolovski), the average sentence is approximately 66 months, as shown in the chart below. 


Analysis of §2B1.1 Offenders in Each Loss Table Category: Fiscal Year 2012


C. Even the most-notorious fraud offenders in the country are typically sentenced below the applicable range.

The following chart gives specific examples of a number of fraud cases involving large loss amounts, including some of the most infamous examples in the country—such as Bernard Ebbers, John Rigas, and Marc Drier.  As shown, every one of these defendants received a sentence below the guidelines range.  [*This table, along with citations to the source information in it, is available at the Federal Defender website at this link


II. Nikolovski Requests a Sentence of 108 Months, as the Government Recommended.

A. The mathematically correct Guidelines range here is 78–97 months.

At Nikolovski’s first sentencing, the agreed-upon Guidelines range was set at 87–108 months in prison.  As noted, Nikolovski argued on appeal that the mathematically correct range is actually 78–97 months in prison.  The Government never disputed the merits of this point, but it argued that Nikolovski had waived this argument, and the Sixth Circuit agreed.  Thus, Nikolovski acknowledges that the Sixth Circuit has held that he could not challenge the applicable Guideline range of 87–108 months as the starting point for his sentence.

But that does not mean that this Court should ignore that this range is mathematically incorrect as the Court considers all the various sentencing factors under § 3553(a).  The Sixth Circuit has requested that this Court state the extent to which Nikolovski’s sentences should run concurrently or consecutively— whether the Guideline range is in fact too high may affect this Court’s ultimate view of the proper sentence to be imposed.  Thus, what follows is how the correct Guideline calculation would have been reached.    

When a person is convicted of sufficiently related financial crimes that include money laundering, those crimes are grouped together and treated as a single offense for purposes of establishing the Guideline calculation.  U.S.S.G. § 2S1.1 n.6 (referring to grouping of counts embodying conduct in other counts under U.S.S.G. § 3D1.2(c)).  The offense level for the group is simply the highest offense level (after proper Guidelines adjustments for that individual offense) among all of the grouped offenses.  Thus, with the three offenses here, the first step is to determine which has the highest offense level.


The bank-fraud offense level comes to 28, as follows:

The bank-bribery offense level is lower, because the loss amount for the bribery (i.e., payments from Nikolovski to Raguz) is significantly less than the bank-fraud amount (i.e., total losses of the credit union in Raguz’s scheme).  See U.S.S.G. § 2B4.1. Thus, the remaining question is whether the third and final offense—money laundering—results in a higher offense level than the level 28 for the bank fraud.  Whichever is higher controls the sentencing range.

The offense level for money laundering was calculated as 29 at the first sentencing, but the proper level is only 27.  Money laundering begins with the adjusted offense level of the underlying offense (here, bank fraud), except that the enhancement for role in the offense must be the role in the money laundering—not the role in the underlying offense. U.S.S.G. § 2S1.1 n.2C.  See United States v. Anderson, 526 F.3d 319, 328 (6th Cir. 2008) (holding that money-laundering calculation starts with offense level and adjustments for underlying crime, but Note 2(C) to § 2S1.1 makes clear that Chapter 3 adjustments such as role in the offense are determined based on the money-laundering offense); United States v. Medina, No. 11-05542, 2012 U.S. App. LEXIS 13518, at *5 (6th Cir. June 29, 2012) (“[T]o apply the four-level aggravating role enhancement pursuant to USSG § 3B1.1(a), Medina must have been an organizer or leader with respect to the money-laundering conspiracy, as opposed to the drug conspiracy.”); United States v. Pass, 413 F. App’x 832, 835 (6th Cir. 2011) (“[W]e have read Note 2(C) to permit the application of chapter three adjustments where the adjustments are applied to the money laundering offense.”); United States v. Preldakaj, No. 10-4885, 2012 U.S. App. LEXIS 25160, at *10 (2d Cir. Dec. 10, 2012) (vacating and remanding under Application Note 2(c) where money-laundering calculation included role-in-the-offense enhancement for role in the underlying offense and not for role in the money-laundering offense).  The only way to enhance by 3 or 4 levels for a managerial or leadership role is to show that the activity involved five or more participants, which is not the case for Nikolovski’s money-laundering conviction (he simply spent the money or had it deposited)—the maximum enhancement, if any, would be 2 levels for that offense.  (This is the point that the Sixth Circuit deemed waived as to the applicable guideline range).

Once these amounts are properly calculated, an additional point is added for money laundering under 18 U.S.C. § 1957. 

Thus, the money laundering offense comes to 27, as shown on the next page:

At the first sentencing, everyone believed that the calculation for money-laundering included Nikolovski’s role in the offense for bank fraud (4 levels instead of 2 for money laundering), which improperly added 2 points, for a total offense level of 29.  And that offense level was higher than the 28 for bank fraud, so it became the governing offense level for the Guideline range. The highest offense level among the offenses was actually the bank-fraud offense level of 28.

Under a proper calculation—with an offense level 28 and zero criminal history—the mathematically correct guidelines range for Nikolovski is 78–97 months.

To be clear, Nikolovski is not arguing that this Court should do anything contrary to the Sixth Circuit’s conclusion that Nikolovski waived his argument on appeal with respect to the “applicable” guidelines range that serves as the starting point for sentencing.  That range, though incorrect, is 97–108 months.  Nikolovski is simply noting that this Court can—and should—recognize when imposing its ultimate sentence under 18 U.S.C. § 3553 that the correct range would have actually been 78–97 months.

B. The Government’s recommendation of 108 months (9 years) is appropriate here, and would generally eliminate any further appeal.

This Court should honor the Government’s position that a sentence exceeding the high end of the Guidelines range would be greater than necessary here.  As originally calculated, that high end is 108 months.  (As properly calculated, that high end is 97 months.) 

But why is it appropriate to remain within the Guidelines range in this case, even if at the high end?  The answer lies partly in a § 3553(a) factor that this Court did not address (and was admittedly not presented with) at the first sentencing: The national disparities in sentences for fraud cases—that is, the requirement to avoid unwanted sentence disparities among defendants with similar records who have been found guilty of similar misconduct. 18 U.S.C. § 3553(a)(6).  As shown above, because of the severe fraud Guideline calculations—driven, as here, by the loss amount—a sentence imposed just over the top of the range would be more severe than 98.8% of all defendants with similar records who have been found guilty of similar misconduct.  Supra, at 19.  Indeed, the average sentence imposed in fraud cases for the applicable range here is 66 months in prison—approximately 5.5 years.  Supra, at 20.  A sentence of 108 months would also be more severe, in relation to the range, than all defendants listed in the table above who have been sentenced for infamous fraud crimes.  Some examples: E. Kirk Shelton, $3.275 billion loss, sentenced to 79% of bottom of range; Bernard Ebbers, over $1 billion loss, sentenced to 83% of bottom of range; John Rigas, $2.3 billion loss, sentenced to 31% of bottom of range.  In short, a sentence at the top of the range (108 months) would be near the limits of conformity with sentencing in fraud cases across the country.

And what about the alleged previous threats and misconduct for which Nikolovski was never convicted but that the Government and this Court referred to during the first sentencing?  (Put aside for the moment that Nikolovksi disputes the accounts of these allegations).   How should this Court weigh those allegations at sentencing without acting arbitrarily or simply going by “feel”?  There are two guideposts.  One is the view of the United States, which has spent considerable time assessing this case and others involving defendants with similar allegations of misconduct.  Its view, of course, is that 108 months in prison adequately accounts for this conduct within the scope of Nikolovski’s offense and other characteristics.  The second guidepost is the Guidelines themselves.  We can simply look to what the Guidelines range would be if Nikolovski had indeed been convicted of assault charges in those incidents, thus erring even further against Nikolovski (and putting aside his disputes with the allegations).  If so, his criminal history points would not be 0 (as they actually are); instead, they would likely be closer to 3 points, making him a Category II offender.  See U.S.S.G. § 4A1.1 (noting points added for previous sentences of imprisonment).  That would change his applicable Guideline range from 87–108 to 97–121.  Of course, as noted, his Guidelines range is already incorrectly high, so the best, true mathematical measure of his range, even as a Category II offender, would move his range from 78–97 months to 87–108 months.  Thus a sentence of 108 months adequately accounts for these allegations of other misconduct.  [See sentencing table at this link]

The question facing this Court is whether the Government’s recommended sentence at the high end of the range, which amounts to a sentence more severe than 98% of fraud cases, should be rejected as too low.  That is, the Court must determine whether that sentence is sufficient, and no greater than necessary.  In light of Nikolovski’s good behavior since his first sentencing, along with the support that his family has shown and will continue to show upon his release, we respectfully suggest that a sentence exceeding 108 months (9 years) would indeed be “greater than necessary.”  Moreover, a sentence of 108 months would put this matter to an end, as the appellate waiver provision of the plea agreement here generally bars appeals of sentences within the Guideline range.  

Nikolovski’s mother has written that she “will never see him again.”  (Exhibit B.)  She may be correct.  Yet Nikolovski’s family is hopeful that her prayers for him “to join his family” will be fulfilled after he has served a sufficient amount of time in prison paying for and learning from his wrongdoing—their request is simply that this time served is no longer than necessary in light of the purposes of federal sentencing.


Koljo Nikolovski committed serious crimes, which he and his family acknowledge.  But his family has also shown that he has redeeming qualities.  Ultimately, Nikolovski requests the same sentence as recommended by the Government: a sentence at the high end of the applicable Guidelines range—108 months in prison.