At a “Loss” for Justice: Federal Sentencing For Economics Offenses Alan Ellis, John R. Steer, and Mark H. Allenbaugh

By: Alan Ellis, John R. Steer, and Mark H. Allenbaugh

According to the United States Sentencing Commission data, economic offenses—which include larceny, fraud, and nonfraud white-collar offenses—now constitute the third largest portion of the federal criminal docket, with drug offenses holding second place and immigration first. (See US Sentencing Comm’n, Sourcebook of Federal Sentencing Statistics tbl. A (2009).) Such offenses can “rang[e] from largescale corporate malfeasance, to small-scale embezzlements, to simple thefts.” (See US Sentencing Comm’n, Fifteen Years of Guidelines Sentencing: An Assessment of How Well the Federal Criminal Justice System Is Achieving the Goals of Sentencing Reform, 55 (Nov. 2004).)

Such a wide array of disparate economic offenses both in kind and degree primarily are sentenced under USSG § 2B1.1, which arguably (and perhaps necessarily) is the most complex of all the sentencing guidelines with more than 16 specific offense characteristics and crossreferences, 19 application notes, and more amendments than any other guideline—40 to date, with more on the way. Currently, more than 300 federal criminal statutes are covered by this single guideline, far more than any other guideline.

ALAN ELLIS, past-president of the NACDL and former Fulbright Visiting Professor of Law, Shanghai, Jiaotong University, School of Law, practices in the area of federal sentencing, prison matters, postconviction remedies, and international criminal law with offices in San Francisco, California; Philadelphia, Pennsylvania; and Shanghai, China. He is a contributing editor to Criminal Justice magazine. Contact him at or go to

JOHN R. STEER is a partner at Allenbaugh Samini Ghosheh LLP in its Washington, D.C., office with a focus on federal white-collar practice, due diligence, and government relations. Prior to entering private practice, he served as deputy general counsel to the U.S. Sentencing Commission from 1986 to 1987 and as general counsel from 1987 to 1999, after which he was appointed a member and vice chair of the commission. MARK H. ALLENBAUGH is a founding partner of Allenbaugh Samini Ghosheh LLP with offices in Orange County, California; Washington, D.C.; and Guangzhou, P.R.C. Prior to entering private practice, he served as a staff attorney for the U.S. Sentencing Commission. He currently serves as cochair of the Sentencing Committee for the National Association of Criminal Defense Lawyers, and as a member of the ABA’s Corrections and Sentencing Committee. He is a coeditor of Sentencing, Sanctions, and Corrections: Federal and State Law, Policy, and Practice (2d ed., Foundation Press, 2002). Contact him at

This article argues that the lengthily named guideline “Theft, Embezzlement, Receipt of Stolen Property, Property Destruction, and Offenses Involving Fraud or Deceit” at USSG § 2B1.1, which, for purposes of brevity, we shall refer to simply as the “fraud guideline,” frequently relies too heavily on the ambiguous concept of “loss” (or “gain” in some cases). Although the calculation of “loss” is a “critical determinant” of a defendant’s sentence, see United States v. Rutkoske, 506 F.3d 170, 179 (2d Cir. 2007), and is often “the single most important factor in the application of the Sentencing Guidelines,” according to Peter J. Henning, in White Collar Crime Sentences After Booker: Was the Sentencing of Bernie Ebbers Too Harsh? (37 McGeorge L. Rev. 757, 767 (2006)), loss, as measured by the guideline, nevertheless, does not adequately account for extrinsic factors such as market conditions, inflation, credits from insurance payments, or the scienter of the offender. As applied by federal judges around the country, the fraud guideline often results in widely unwarranted sentencing disparity and a lack of certainty in sentencing, and produces sentences grossly disproportional to the actual seriousness of the offense. (See United States v. Parris, 573 F. Supp. 2d 744, 754 (E.D.N.Y. 2008) (noting that “the Sentencing Guidelines for white-collar crimes [can produce] a black stain on common sense”); United States v. Adelson, 441 F. Supp. 2d 506, 512 (S.D.N.Y 2006) (lamenting “the utter travesty of justice that sometimes results from the guidelines’ fetish with absolute arithmetic, as well as the harm that guideline calculations can visit on human beings if not cabined by common sense”).)

“Loss” simply needs to be given less weight under the fraud guideline relative to other factors germane to economic offenses, particularly in large-loss cases. Furthermore, in order to bring loss back to reality, the intrinsically speculative concept of “intended loss” should be substantially discounted if used as an alternative to “actual loss.”

In short, the increasingly complex fraud guideline is rapidly becoming a mess. While the Sentencing Commission valiantly attempted to make the economic crime guidelines more coherent a decade ago, those efforts have all but disappeared among the avalanche of congressional directives. Commission efforts to reform section 2B1.1 to incorporate new offenses add to the overall complexity of the guideline, and the now-advisory status of the guidelines necessarily has led to an increasing number of departures and variances.

As the commission continues several ongoing studies regarding the effects of Booker (543 U.S. 220 (2005)) and its progeny on federal sentencing law, policy, and practice, a substantive reevaluation of the role of loss in calculating guideline sentences for economic offenses, and, indeed, section 2B1.1 overall, needs to be incorporated into these studies.

A Brief History of the Fraud Guideline

The fraud guideline originally was located at USSG § 2F1.1 and the theft guideline at USSG § 2B1.1. As with the guidelines generally, the commission was tasked by Congress to collect data on past sentencing practices. Based on Sentencing Reform Act directives, the commission’s reading of legislative history, and its overall policy objectives, the commission modified past practices to roughly equalize sentences for “white collar” fraud and embezzlement offenses and “blue collar” theft offenses (in contrast to past practices in which theft offenders generally received more severe sentences). The commission also made all sentences for economic offenses somewhat more severe than what a review of past practice data had revealed in order to better achieve deterrence and just punishment sentencing goals.

Nevertheless, initial sentences for even the most severe fraud offenses were far less severe than the guideline sentences for economic offenses today. For example, the loss table topped out at a mere $5 million, and the fraud guideline had only two specific offense characteristics.

The commission took steps to ensure that fraud sentences were proportionally less onerous than for other offenses considered more serious (e.g., bribery, serious drug trafficking, and offenses involving violence).

Between November 1, 1987, when the initial guidelines took effect, and today, there have been three significant amendments that increased the severity of the loss table itself and extended the loss brackets to larger dollar amounts.

The Savings and Loan Scandal Prompts Change.

First, the 1989 amendments increased the severity for moderate- and large-loss offenses, as they were defined at the time (offenses greater than $140,000 in loss received an increase), and the amendments extended the loss table by four loss bracket increments from losses greater than $5 million to losses greater than $80 million. This early reconsideration by the commission in part responded to external developments, principally the savings and loan fraud crisis of the late 1980s.

Concurrent with these changes, the commission began adding additional specific offense characteristics (SOCs) that sought to address particular factors sometimes present in particular types of fraud offenses. Nearly all of these SOCs were of an aggravating nature. This practice began a trend that, over the years, has added some 16 SOCs, several of which are multipronged, to the fraud guideline. (One SOC for more than minimal planning was deleted by incorporating it into the loss tables in 2001.)

Many, if not most, of these new SOCs responded to congressional directives of a general or sometimes very specific nature. Approximately 21, or half the total number of amendments, were of this nature. For example, the savings and loan financial crisis of the late 1980s led to legislation that increased maximum penalties for financial fraud offenses and concurrently directed the Sentencing Commission to add specific aggravating factors to the fraud guideline.

As a result, the commission added major, four-level increases equating to an average 50 percent penalty increase for conduct that “substantially jeopardized the safety and soundness” of a financial institution, and for so-called “fat cats” who derived more than $1 million from the offense. (This SOC was later reduced two levels because of overlap/”double-counting” concerns.)

The Economic Crimes Package Overhauls Sentencing for Economic Offenses.

The second major severity increase in the fraud guideline loss table occurred in 2001 as part of the commission’s “Economic Crimes Package.” This comprehensive amendment merged three guidelines, 2Fl.l-fraud, 2B1.1-theft/embezzlement, and 2B3.1-property destruction, into one guideline, 2B1.1, with a new and more severe loss table, and combined specific offense characteristics.

This new loss table used two-offense level increments instead of the former one-level increment. It assumed the applicability of and incorporated the previously separate enhancement for more than minimal planning by gradually phasing it into the loss table. The amendment additionally provided severity increases ranging from two levels for losses greater than $5 million, to four levels (approximately a 50 percent increase) for losses greater than $200 million.

The amendment also provided a comprehensive and far more sophisticated new definition of “loss,” changing it from a theft-based definition to one of legal causation more akin to how financial losses are understood in the civil context. The new definition of “reasonably foreseeable pecuniary harm” had the effect of increasing the countable loss in many cases, because it included some previously excluded consequential and other harms.

Finally, the amendment also introduced a new, multipronged enhancement for multiple victims that henceforth would frequently interact with the loss table enhancement.

Attack of the Sarbanes-Oxley Act.

A third severity increase for many fraud offenses occurred a little more than a year later in response to the Sarbanes-Oxley Act and legislative directives in it. The commission increased the base offense level from level 6 to level 7 for fraud offenses carrying a statutory maximum penalty of 20 years or more, which now covered the frequently prosecuted wire and mail fraud offenses. And the loss table was again extended by two additional brackets such that the highest loss amounts went from offenses greater than $100 million to offenses greater than $400 million.

Additional enhancements were added for circumstances in which: (i) the offender was an officer or director of an organization (+4 levels or a 50 percent increase in the sentence), (ii) company insolvency resulted from the offense, and (iii) more than 250 victims were involved.

These multiple amendments to the loss table in and of themselves have dramatically increased sentencing severity for fraud offenses having substantial monetary losses. For example, without considering any other guideline enhancements, the adjusted total offense level for an offense causing just over $20 million in loss has been increased from level 19, which equated to a guideline sentencing range of 30-37 months, to level 29, or 87-108 months. In other words, the three amendments to the loss table in 1989, 2001, and 2003 effectively tripled sentences for large-scale fraud offenses.

At the same time, the commission has added a host of SOCs to the fraud guideline, most in response to congressional legislation creating new offenses, increasing statutory penalties, and frequent directives to the commission for higher penalties. Although these enhancements do not apply in all cases, their cumulative effect has been to further increase severity in some types of fraud cases, sometimes dramatically.

Structural Criticisms of the Fraud Guideline

For the reasons stated above, the loss table often overstates the actual harm suffered by the victim. Multiple, overlapping enhancements also have the effect of “double counting” in some cases. Furthermore, the guidelines fail to take into account important mitigating offense and offender characteristics.

With respect to the loss table specifically, while the commission has made multiple aggravating amendments over the years, it has failed to make any adjustments for the effects of inflation, which itself has effectively increased penalties.

The commission also has arbitrarily selected the loss brackets in the table rather than using a uniform multiplier. A lower multiplier at the high-dollar end of the loss table effectively escalates penalties more rapidly, especially when used in conjunction with the sentencing table, where every increase in offense level causes a progressively greater increase in nominal months of imprisonment. The net effect of this steadily escalating mathematical structure contrasts with an arguably more reasonable and rational approach exhibited in the multiple count guidelines (in chapter three, part D). The latter structure increases guideline penalties as additional offenses occur, but it does so at a “progressively decreasing rate.”

Furthermore, in response to legislative directives, the commission has added a host of SOCs and some additional chapter three enhancements. The net effect has been to heighten penalty severity, sometimes dramatically. For example, an executive officer of a publicly traded company who causes a large fraud loss can face a high guideline penalty derived from the loss table, further accentuated by a +6 level increase for 250 or more victims, +2 levels for receiving more than $1 million from the offense, +4 levels for being a corporate officer, up to +4 levels for an aggravating role in the offense, and possibly +2 levels for abuse of trust. The net guideline score can easily equate to a life sentence, which generally under the guidelines has been reserved for dangerous and/or violent offenses such as murder, terrorism, or high-level drug traffickers.

In early versions of the guidelines, penalty levels and enhancements were derived primarily from empirical analyses of past sentencing practices. The commission was careful to maintain proportionality among offenses, and it had a good sense of the effects of adding enhancements within a guideline. However, as legislatively directed enhancements have proliferated, particularly within the fraud guideline, the objective of proportionality among offenses has gone by the wayside. Guideline sentences for large-dollar fraud offenses can easily equal those of violent crimes such as murder, and far exceed those of other white-collar offenses such as bribery, which the commission initially considered more serious.

While the fraud guideline focuses primarily on aggregate monetary loss and victimization, it fails to measure a host of other factors that may be important, and may be a basis for mitigating punishment, in a particular case. Among these considerations are:
1. the scope and duration of the offense;
2. the extent to which the offender did or did not personally profit from the offense;
3. the motivation for the offense;
4. the extent to which the offense was exacerbated by factors beyond the offender’s control.

[Table 1] $20 Million + Frauds Sub $20 Million Frauds
Number 71 7,879
% of All Frauds 0.89% 99.11%
Total Loss $6,530,000,000 $2,268,970,000
% of Total Loss 74.21% 25.79%
Average Loss $91,971,831 $287,977


[Table 2] Offender Loss Sentence $ per mo.
Richard Harkless $39,000,000 1,200 (mos.) 32,500
James R. Nichols $21,123,830 292 72,342
Sholom Rubashkin $27,000,000 324 83,333
John Miller $21,000,000 159 132,075
Gary Vanwaeyenberghe $25,521,966 168 151,916
Calude LeFebrve $64,850,000 240 270,208
Jeff Skilling $80,000,000 288 277,778
Joseph Nacchio $28,000,000 72 388,889
Kevin S. Jackson $20,000,000 51 392,157
Morad Abu Sliman $26,000,000 57 456,140
John Rigas $102,708,142 144 713,251
Louis Pearlman $300,000,000 300 1,000,000
Mark Turckan $25,000,000 12 2,083,333
Marc Dreier $700,000,000 240 2,916,667
Lance Poulsen $2,000,000,000 360 5,555,556
Tom Peters $3,670,000,000 600 6,116,667
Ronald Ferguson $500,000,000 24 20,833,333
Bernard Madoff $65,000,000,000 1,800 36,111,111
Bernard Ebbers $11,000,000,000 300 36,666,667

Historically, the guidelines also have failed to measure offender characteristics other than criminal history. In large scale fraud offenses, the likelihood of important mitigating offender characteristics is greater than in many other types of offenses. Among these are:

1. age;
2. physical and/or emotional health;
3. positive contributions to the community, public service, charitable contributions; and,
4. family and community responsibilities.

Published in Criminal Justice, Volume 25, Number 4, Winter 2011. © 2011 by the American Bar Association. Reproduced with permission. All rights reserved. This information or any portion thereof may not be copied or disseminated in any form or by any means or stored in an electronic database or retrieval system without the express written consent of the American Bar Association.

Continued Criticism from the Judiciary

As discussed in detail below, especially as loss amount grows, the correlation between loss and the sentence imposed becomes increasingly arbitrary. Loss simply is not always an appropriate proxy for the seriousness of the offense, as even the guidelines themselves concede. (See USSG § 2B1.1, cmt. n.19(C) (noting a downward departure may be warranted in cases where the guideline sentence “substantially overstates the seriousness of the offense”).)

[S]ince Booker, virtually every judge faced with a top-level corporate fraud defendant in a very large fraud has concluded that sentences called for by the Guidelines were too high. This near unanimity suggests that the judiciary sees a consistent disjunction between the sentences prescribed by the Guidelines [in economic offense cases] and the fundamental requirement of Section 3553(a) that judges impose sentences “sufficient, but not greater than necessary” to comply with its objectives.

(Frank O. Bowman, III, Sentencing High-Loss Corporate Insider Frauds After Booker, 20 Fed. Sent’g Rep. 167, 169 (Feb. 2008).)

comparison_of_loss_to_sentenceIn fact, in a recent letter to the US District Court for the Northern District of Iowa regarding a high-profile sentencing case before that court (in which the authors of this article were part of the legal team), Brett Tolman, former US attorney for the District of Utah, and Paul G. Cassell, former US district court judge for the District of Utah, wrote that “[r]ather than resting on evidence of past, national sentencing practices, the white collar Guidelines are a product of the political environment in which they were promulgated, the Commission’s desire that the Guidelines reflect perceived congressional policy, and the Commission’s own independent policy determinations concerning the severity of a particular class of conduct.” (Letter to Chief US District Judge Linda Reade at 5 (Apr. 19, 2010) in United States v. Rubashkin, __ F. Supp. 2d __, 2010 WL 2471877 (N.D. Iowa 2010).) In this letter, the authors concluded that “[i]mposing massively lengthy sentences in white collar cases is not only wasteful of taxpayer dollars, but it is insulting to the victims of violent crimes. It is hard for victims of truly violent crimes to understand why defendants who have violently harmed them should receive a far shorter prison sentence than” a white-collar offender. (Id. at 7.)

A Review of Sentencing Data for Economic Offenses

In fiscal year (FY) 2009, 7,951 individuals were sentenced under the fraud guideline constituting 10.5 percent of all 81,732 guidelines cases. Only those sentenced under the drug guideline at USSG § 2D1.1 (24,901 or 32.9 percent) and the unlawful entry guidelines at USSG § 2L1.2 (17,310 or 23.6 percent) constituted larger groups of offenders.

The total loss amount aggregated over all 7,951 cases totaled nearly $9 billion (and that is assuming the loss was at the low end of the guideline range). (See US Sent’g Comm’n, Use of Guidelines and Specific Offender Characteristics 9 (2009).)

Again, using commission data (see Table 1), of the 7,951 fraud sentences in FY 2009, 7,879 or 99.11 percent involved loss amounts of less than $20 million. Adding these losses together (using the lowest loss level in the guideline range), these frauds resulted in a combined loss of just over $2.2 billion or 25.79 percent of the total overall loss amount. The average loss amount, however, was just over $287,000. In fact, over three-quarters of the frauds (76.34 percent) in the sub-$20 million loss class involved losses of less than $20,000.

So, in sum, more than 99 percent of frauds involve amounts far less than $20 million and total only a quarter of total fraud losses. However, the 1 percent (71 cases) of large frauds constitutes three-quarters of the aggregate fraud loss, and therefore as a group provide an excellent model for testing the correlation between loss and the resultant sentence.

Given the complexity of these cases, a straight comparison of loss amount to the sentence received is not very telling. Accordingly, we have made the comparison between loss amount and ultimate sentence received in these cases by dividing the amount of loss attributed to the defendant to the sentence received by the defendant in months. In short, we wanted to know how much a month in prison is worth under the fraud guideline, which is intended to promote greater uniformity and certainty in sentencing. One would expect, therefore, that a month in prison is worth approximately the same amount regardless of the amount of total loss attributed to the offender.

We found some startling, counterintuitive results as Table 2 illustrates from noteworthy, published cases.

Perhaps the most infamous white-collar criminal over the past decade, Bernard Madoff, received one of the longest sentences ever—150 years or 1,800 months—for a white-collar offender. However, when compared to the incomprehensible magnitude of the loss—$65 billion— Madoff’s sentence may not have been as incredible as it may seem at first blush. As Table 2 illustrates, Madoff received one month’s imprisonment for every $36.1 million of fraud he committed. Under this methodology, Marc Dreier should have received a sentence of approximately 19 months instead of 240 months.

Or, taking Marc Dreier as an example, Jeff Skilling should have only received 2.21 months, given the ratio of loss to months’ imprisonment for Dreier. Mark Turkcan, president of First Bank Mortgage based in St. Louis, who misapplied $35 million in loans that resulted in a loss of approximately $25 million, received a sentence of only a year and a day. But Gary Vanwaeyenberghe, with a loss amount approximately the same as Turkcan, received a sentence of 168 months or 14 times Turkcan’s sentence.

And to take examples from the opposing ends of Table 2, if Richard Harkless had been sentenced according to the same loss/months ratio as Bernard Ebbers, he should have only received just over one month in prison as opposed to 100 years.

Upon reviewing the data sample, there simply is no consistent correlation between loss and offense seriousness with respect to the ultimate sentence imposed. While the list of offenders in this table obviously is neither exhaustive nor a scientific sampling, it nevertheless strongly suggests that loss is a poor proxy for offense seriousness, at least as seen by federal judges under the now advisory guideline system. Table 3 makes a graphic comparison of the data in Table 2.


As the Supreme Court explicitly recognized, “in the ordinary case, the Commission’s recommendation of a sentencing range will reflect a rough approximation of sentences that might achieve § 3553(a)’s objectives.” (Kimbrough v. United States, 552 U.S. 85, 89 (2007) (emphasis added).) But the sentences resulting from the use of loss under the fraud guideline can hardly be considered even a “rough” approximation in light of the data reviewed, especially in large dollar-loss cases.

As it stands, the fraud guideline constitutes a series of ad hoc amendments covering a vast array of distinctly dissimilar conduct applying to offenders from the Gordon Gecko variety to the well-intentioned but desperate business owners. There simply is no way the sentences that result from them can be considered principled or even reasonable, especially because loss plays such a central role in determining the ultimate sentence.

It is time that the fraud guideline is revisited by the commission. First, the commission should conduct a serious, scholarly analysis of how well the guideline achieves the purposes of sentences and the goals of the Sentencing Reform Act. Second, the commission needs to conduct a comprehensive proportionality comparison of sentences in relation to offense seriousness within the fraud guideline, in comparison to other white-collar offenses, and then in comparison to other offenses generally. These analyses should inform a thorough revision of the fraud guideline. One would hope that revision would result in (i) less reliance on loss and the use of a loss table with fewer brackets and a progressively decreasing scale; (ii) less overlap between loss and other SOCs; and (iii) more weight given to the nature of the offense and the offender.

Until such a comprehensive review and revision is undertaken by the commission, defense counsel will have to challenge and work around the fraud guideline on a case-by-case basis, and judges will have to do the commission’s work by fashioning sentences as best they can per the mandates of 18 U.S.C. § 3553(a) to achieve justice.


Published in Criminal Justice, Volume 25, Number 4, Winter 2011. © 2011 by the American Bar Association. Reproduced with permission. All rights reserved. This information or any portion thereof may not be copied or disseminated in any form or by any means or stored in an electronic database or retrieval system without the express written consent of the American Bar Association.

Menu Menu