White-Collar Showdown

I.     Introduction

Not many people would rank white-collar criminals among the downtrodden of the criminal justice system. Some courts and scholars rightly worry about the Department of Justice abusing its pre-trial leverage against individual white-collar suspects.1 But focusing on convicted white-collar criminals, the claim that they are punished too harshly would prompt reflexive skepticism in many. White-collar criminals abuse positions of trust, respect, and authority to their own advantage2—usually a very lucrative advantage. The FBI estimates that the annual cost of white-collar crime in the United States is around half a trillion dollars,3 roughly thirty times the cost for every other crime combined.4 It is against this backdrop that Judge Bennett and his co-authors offer the bold thesis that the Sentencing Guidelines ranges for white-collar fraudsters are overall too severe.5

It turns out that Judge Bennett is not the only judge who thinks this. Surveying an impressive cross-section of the judiciary, state and federal, trial and appellate, Judge Bennett and his co-authors found that most are inclined to sentence typical white-collar fraudsters at or below the minimum recommended by the Sentencing Guidelines.6 Without picking a side, we can all agree that this disconnect between judges’ sentencing practices and the Guidelines’ recommendations is troubling. Now we have to decide who has it right. Judge Bennett and his colleagues on the bench are insightful repeat players who see the facts on the ground and can look each white-collar defendant in the eye. Their perspective warrants serious attention. Yet, it is with the utmost respect that I suggest they may be dismissing the Sentencing Guidelines too lightly.                                                                                                           

II.     WWSCD?

One place to start is by asking: What would the Sentencing Commission do with Judge Bennett’s data? Are they likely, upon reading Judge Bennett’s article, to recognize their mistake and rush to fix it? Maybe. But I would suggest that a schoolyard “told-ya-so” is the more likely response. The Guidelines for fraud reflect the Sentencing Commission’s worry that judges were inclined to sentence white-collar criminals too lightly, a worry Judge Bennett’s data may be read to confirm.

Judge Bennett’s data shows that judges routinely discount the Sentencing Guidelines ranges for white-collar fraud.7 One thing worth emphasizing at the get-go is that there is no legal problem here with what any individual judge is doing. Ever since Booker, the Guidelines are not binding on judges.8 That basically means that, within pretty forgiving bounds, judges can sentence as they please. Those boundaries are mostly procedural. Judges cannot scrap the Sentencing Guidelines all together; they must still treat them as “advisory.”9 Courts have described what this means in various ways, but the gist is that judges should, at a minimum, start their sentencing deliberations with the Guidelines range.10 There is good evidence that, for the most part, judges do this, placing a heavy thumb on the scale in favor of the Guidelines: Roughly 80% of sentences for all offenses end up being within the recommended range (or deviating from it at the government’s request).11 In the remaining 20%, judges are well within their discretion to depart from the range after tipping their hats to it.

Even though Booker authorizes individual sentencing judges to depart from the Guidelines recommendation, Judge Bennett’s data raises concerns in the aggregate. If we look at the reasoning of any individual judge sentencing a white-collar fraudster, I am certain we are unlikely to see anything objectionable. But sometimes problems that are hidden at the individual level emerge when we shift focus to the collective level. If, as Judge Bennett’s data suggests, the judiciary as a whole sentences white-collar fraudsters at or below the Guidelines range, we might worry that the judiciary as a whole is not really taking advice from the Guidelines.12 If the judiciary were treating the Guidelines as advisory, we would expect that individual decisions would span the range and, under the law of large numbers,13 tend to cancel each other out. White-collar sentences would, on the whole, fall within the Guidelines ranges, as sentences for other crimes do. But when the trend dips low in a statistically significant way, any sociologist would say some other force must be at work.

As it turns out, the Sentencing Commission, which designed and revises the Sentencing Guidelines, has a theory about what that force is—good old cognitive bias.14 “Bias” sounds like an evil word, but the kind of cognitive bias at issue here is the sort of bias that we all exhibit;15 it is unconscious and out of our control.16 Still, we should counteract it to the extent possible.17

That is what the Sentencing Commission was trying to do by raising the ranges for white-collar fraud to their present level.18 Before the reforms, sentences for white-collar criminals were too light, especially when compared to street-crime sentences.19 One leading explanation was that judges are much more likely to overlap demographically with the sorts of upper-echelon folks who end up committing white-collar crimes.20 Beneath their robes and powdered wigs, judges are just humans subject to human foibles; like us all, they are more likely to empathize (even if unconsciously) with one of their own.21 The Sentencing Commission could not cure this suspected cognitive bias, but they could counteract it by raising the sentencing ranges for white-collar crimes. Which is what they did. The Sentencing Commission might look at Judge Bennett’s data and, rather than worry, see it as confirmation—“See, we told you that judges favor those white-collar folks! At least they are now sentencing at the bottom of a higher range.”22


But is that what the drafters of the Sentencing Guidelines should do? You might think that if judges are going to sentence white-collar fraudsters lightly, the Sentencing Commission should interpret the practice charitably and defer for the sake of harmony. Maybe the judges are on to something. And, in any case, does the Booker line of cases not tip the scales in favor of the judiciary when it comes to sentencing?

Booker certainly had the effect of vesting judges with a lot of discretionary power over sentencing. But it is important to recall that this was not because judges have a unique authority vis-à-vis Congress and the Sentencing Commission23 to impose sentences. Ironically, the reasoning in Booker is in many ways a strike against the institutional competency of the judiciary over sentencing. Booker was ultimately a case about the importance of the jury. Judges had been usurping the jury’s role by imposing sentences on the basis of facts found by the judge without the involvement of a jury.24 This, Booker tells us, judges cannot do.25Booker had the counter-intuitive result of giving judges more power to sentence criminals, without giving them more institutional authority to do so. The power to say what punishment criminals deserve lies, in the end, with Congress, which can still mandate minima and maxima for any number of crimes,26 thereby cabining judicial discretion to the extent it sees fit. Sentencing judges just get to act in the gaps Congress leaves behind.

Even if judges do not have any ultimate institutional authority over sentencing, their position may still give them unique insights into white-collar sentencing. If that is true, a charitable Sentencing Commission looking beyond the possibility of cognitive bias should take heed. But there is little reason, and Judge Bennett offers none, to think that judges have any privileged insight into the considerations relevant to white-collar sentencing policy.

Those who talk about these things generally recognize three broad purposes for criminal punishment: (1) deterrence; (2) rehabilitation; (3) and retribution.27 Different kinds of data are relevant to determining what sorts of punishment, and how severe, would best achieve these purposes. Economic and sociological data might tell us what sort of state response should or would deter prospective criminals. Psychological data might tell us when rehabilitation is possible, and how to bring it about. And moral truths about just deserts are the best guide to punishment’s retributive aims. None of this information is the sort that judges are particularly well-situated to gather. There is no reason to think that judges are drawing on a deeper reservoir of empirical data than the Sentencing Commission has available. There is also no reason to think that judges are any better at intuiting the moral seriousness of white-collar crimes. Judges are legal experts, but sentencing policy is something that precedes law.28

IV.     The Main Event: Judges vs. Sentencing Commission

Nothing I have said so far suggests that Judge Bennett and his co-authors are wrong that the Guidelines ranges for white-collar fraud are too high. At this point, I just hope to have instilled a reluctance to take judges’ say-so on the matter. That leaves us with a clean slate and a lineup between the judges and the Sentencing Commission. Let us look at the matter de novo. As Judge Bennett emphasizes, this should be an empirically informed standoff.29 There are three rounds up for grabs, one for each of criminal law’s purposes. Is the Sentencing Commission trying to smudge those well-starched collars, or are the judges’ intuitions just so much white-washing?


Round 1: Deterrence

White-collar crime exacts a heavy toll. As mentioned above, the costs of white-collar crime outpace the aggregate costs of every other crime thirty-to-one.30 This sounds like an area that calls for a heavy dose of deterrence. By threatening a sanction for misconduct, criminal law can offer exactly that.

Deterrence theorists assume that people are egoists and view criminal law as a mechanism for persuading them that misconduct is against their self-interest (once the threatened sanctions are factored in).31 Classic deterrence theory comes with a simple equation for figuring out how severely to punish any particular type of crime. The magnitude of the sanction multiplied by the chance of getting caught should be greater than the expected criminal gains.32 The theory works best with egoists who are capable of rationally assessing risks and rewards. It does not fare so well with irrational criminals, like those acting in the throes of some passion or who are incapable (because of youth, limited intelligence, etc.) of making and guiding their conduct by basic economic arithmetic.33 White-collar fraudsters seem like they should be the objects par excellence of deterrence—true rational egoists. They are often educated, financially savvy, and good at calculating risk.34

How can we tell whether we have hit the sweet spot where the expected sanction is just high enough? The best metric is to look at what people do.35 Observation is the surest way to see whether the expected sanction exceeds what criminals expect to gain. According to the usual deterrence-style thinking, if we still have unacceptable levels of some sort of crime, we need to punish it more harshly. That would be because too many criminals, calculating the risks and rewards of crime, see the rewards winning out. And that seems to be exactly what is happening with white-collar fraud. Incidence of white-collar fraud has only increased over the years.36 Based on the usual way deterrence theorists think about these things, that would confirm that Judge Bennett is right that the Guidelines ranges for white-collar fraud are defective. But for being too low!

At least, that is one explanation of why we are seeing elevated rates of white-collar fraud. It assumes that white-collar criminals are rational egoists. But the last thirty years of behavioral economics have taught us to be wary of arguments premised on individuals’ rationality.37 One striking circumstance of the increasing incidence of white-collar fraud is that the sentencing ranges were increasing at the same time.38 Regardless of whether the ranges ever hit the optimal level, we would expect higher sanctions to result in lower crime if the targets are rational egoists. Since that did not happen, we have some evidence that white-collar fraudsters do not respond rationally to the risk of sanction.

Here is some more evidence. The basic equation for deterrence theory—sanction times chance of getting caught must exceed expected gain—indicates that increasing sanctions and increasing enforcement, i.e. chance of capture, are interchangeable.39 This makes sense from the perspective of potential criminals balancing risks and rewards; either doubling the chance of capture or doubling the penalty would end up doubling the riskiness of the venture. At least, that is how rational criminals would view things. Recent data suggests, though, that the probability of getting caught looms larger in white-collar criminals’ calculus.40 In other words, doubling the chance of capture has a greater deterrent effect than doubling the sanction. This is just a seemingly irrational quirk of white-collar psychology, and one that should affect how we think about deterring white-collar crime.

What could explain this bug in white-collar risk calculus? It could be that white-collar criminals are literally irrational. But that is not a particularly charitable interpretation. Another possibility is that they are rational, but that there are other, extra-legal costs to getting caught, like loss of reputation and career, that white-collar fraudsters fear even more.41 Perhaps they fear these so much that the marginal effect of criminal sanctions fails to register.42 A final possibility is that white-collar fraudsters are rational, but not egoists. As discussed more fully in the next section, we know that institutions can affect how the individuals in them act. Perhaps white-collar criminals are committing fraud in institutional settings that drown out their self-interested egos and emphasize criminogenic norms that favor the institution (think of a corporation with aggressive performance quotas).43 In this case, the fraudsters could still be rational, but 9-to-5 when they commit their crimes,44 they are rationally pursuing institutional goals at the expense of their own.

Where does this leave us in terms of judges versus the Sentencing Commission? If criminal punishment fails to deter white-collar fraud because the punishments are too low, that clearly tells against the judges. But if there are counter-rational influences on the fraudsters that prevent the sanctions from having deterrent effect (whatever the correct explanation of that is), that suggests deterrence is not the most effective framework for thinking about how to punish white-collar fraudsters. To prevent white-collar fraud, we should look beyond individual criminal punishment and perhaps emphasize enhanced enforcement or institutional reform instead. Deterrence starts to seem orthogonal to the question of the Guidelines ranges. Judges: 0 | Sentencing Commission: 0.

Round 2: Rehabilitation

If deterrence is not a helpful way to gauge how severely to punish white-collar fraud, perhaps rehabilitation is. Though it has fallen out of favor in recent decades,45 many still list rehabilitation as one of the basic purposes of punishment.46 To fulfill this purpose, punishment should reform away criminals’ dispositions to commit crime.47 So for theorists focused on rehabilitation, recidivism rates are the key measure of success.48 If convicted criminals reoffend after being punished, the system failed to reform them.

In the context of white-collar crime, scholars rarely talk about rehabilitation.49 This may be because, as one scholar has suggested, rehabilitation is less of a concern for gainfully employed criminals (as white-collar fraudsters typically are) who can more easily reintegrate into society on their own.50 Ultimately, we should not rely on such armchair rationales. Data about recidivism rates and whether punishment can do anything about them should make or break our interest in rehabilitation for white-collar fraudsters.

Here is what we know about recidivism rates—they are quite high. One study found that as many as 39.2% of white-collar criminals were re-convicted within three years of their first offense.51 Data from another study indicates that one-in-four white-collar regulatory offenders had prior convictions; that number rises to one-in-three if we zero-in on fraudsters.52 These stats should prompt more interest in white-collar rehabilitation, and the failure of the Sentencing Guidelines to achieve it.

So again, the judges are right that the Sentencing Guidelines are off base. But in what way? Three possibilities suggest themselves. First, a classic rehabilitation theorist would conclude that the Guidelines ranges are too low. Perhaps the sentences are insufficient to convey to white-collar fraudsters that their offenses are serious breaches of trust and good order. Stewing in jail a bit longer might help them find religion, reflect on the suffering of their victims, and “learn their lesson.”53

This diagnosis is unattractive because its underlying premise—that jail time can reform—has proven false. Philosophers have long questioned the possibility of reform through punishment.54 And the history of the American criminal justice system backs them up. The larger part of twentieth-century penology was a sustained experiment in using punishment to rehabilitate.55 That experiment failed; decades of recidivism rates that proved impervious to punishment persuaded lawmakers to shift course in the 1970s.56 Rehabilitation is now off the table as a matter of law for fixing prison terms.57

A more promising explanation of what is wrong with the Sentencing Guidelines has nothing to do with their severity, but with the type of punishment they encourage. The lesson of the twentieth-century experiment in rehabilitation was not that rehabilitation is impossible but that jail time is a poor way to achieve it.58 The Federal Defenders are hot on the trail and have already proposed that white-collar sentencing should include alternatives to incarceration.59 It is possible that something like community service, shaming sanctions, or restrictions on computer access would be more successful. Some scholars argue that specific types of “reintegrative” shaming sanctions could be particularly effective in the white-collar context.60 And data reflecting very low white-collar re-offense rates while on probation or parole suggest that more extensive use of these mechanisms could improve things.61

A third possibility, again having nothing to do with punishment severity, is that the Guidelines are focused on punishing the wrong defendant. It may be that the person most in need of reform is not the individual white-collar criminal but the institution within which the individual operates.62 Some popular approaches to thinking about corporate criminality focus on the criminogenic characteristics that can arise in organizations.63 These approaches point to what scholars call a corporation’s ethos, which can deteriorate in the presence of lax compliance culture, aggressive performance goals, up-or-out promotion metrics, and the like.64 The thought is that institutions with bad ethos can make criminals of individuals who would otherwise have no criminal inclinations. In such cases, the individuals are the wrong place to focus if we are concerned about recidivism. We should be paying much more attention to using our criminal justice system to reform the institutions that are making white-collar employees into white-collar criminals.65

So, once again, the round is a draw. The Sentencing Commission got it wrong, but the judges were not right about why. The Guidelines ranges seem to be ineffective at rehabilitating white-collar fraudsters, but because they focus on the wrong types of punishments and the wrong defendants, not because they are too harsh. Judges: 0 | Sentencing Commission: 0.

Round 3: Retribution

That leaves us with retribution as a possible tie-breaker. Retributivists think that punishment should give criminals their just deserts,66 and that the justice deserved is proportional to the seriousness of the offense.67 At one end of the spectrum are the commonplace, relatively petty crimes like traffic violations, vandalism, and public intoxication. These deserve quite light punishment. At the other end of the spectrum and deserving much greater punishment are more serious crimes, like rape, murder, and armed robbery. The question for the retributivist is where white-collar fraud fits along the spectrum. Likely there is no single answer, since the seriousness of the fraud probably varies with the magnitude of the harm caused.68 The Sentencing Guidelines reflect this intuition by increasing the offense level for fraud (and hence the recommended range) in proportion to the loss69 and the number of victims suffering substantial financial hardship.70 Still, it is possible that the Guidelines ranges, even if they scale with harm, scale too fast, too slow, or are overall too harsh or lenient.

We know what Judge Bennett and his co-authors think. By their lights, the Guidelines ranges for white-collar fraud are harsher than is warranted by its relatively low moral seriousness.71 Relative to what? The closest street-crime analogue is probably something like robbery.72 Like fraud, robbery can be more or less serious depending on the specifics of the case, such as whether anyone was hurt in the course of the robbery73 and how much was stolen.74 But is Judge Bennett right that, as a general rule, white-collar fraud is less morally serious than robbery?75 It is hard to say. We would need a substantive moral theory to answer the question, and Judge Bennett provides none. Nor would I presume to offer one. Probably Judge Bennett is drawing on his no-doubt refined sense of moral intuition, and, I must confess, I feel its pull. Even if white-collar fraud can cause more economic harm, it rarely directly results in physical injury.76 And that fact alone holds significant sway in my moral economy. However, out of a spirit of moral humility and a fear of my own cognitive biases, I am inclined to turn to the intuitions of the American people for guidance.

On this topic, there is some good data. Some of the earliest studies contained data that would have been available to the Sentencing Commission when it was initially drafting the Sentencing Guidelines. In 1985, the Bureau of Justice Statistics released The National Survey of Crime Severity, which polled Americans on the “seriousness” of various sorts of crimes.77 The relative mean seriousness of various crimes is telling. Just to get a sense of the scale, stealing property worth $10 from outside a building (think pink flamingo) was ranked pretty low, with a mean seriousness score of 1.5772.78 Stabbing a victim to death was rated very high, 2.8929.79 Check fraud and embezzling $1,000 from an employer ranked in the middle, at 2.1982 and 2.1336, respectively.80 They ranked about the same as stealing property worth $1,000 from outside a building (2.1767), and just below robbing a victim of $1,000 at gunpoint (2.3277).81 These numbers indicate that, at least in the eyes of the average American, fraud and robbery are roughly on a moral par. That data is forty years old, but more recent studies suggest it is still good. People today think that white-collar crime is morally blameworthy,82 and, on average, people think that robbery and white-collar fraud deserve equivalent punishment.83 One of the most recent studies even found that respondents “viewed white collar crime as slightly more serious than traditional crime types.”84

There is no straightforward way to tell whether the Sentencing Guidelines reflect a moral equivalence between fraud and robbery.85 Each crime has different factors that can raise the offense level (e.g., causing substantial financial hardship versus using a gun)86 and the offense level for fraud scales up faster with the magnitude of the economic loss than does robbery.87 One sensible test could be to compare the Sentencing Guidelines offense levels for the average fraud and the average robbery. The average fraud causes a loss of roughly $5.5M.88 Assuming no other aggravating offense characteristics, the average fraud has an offense level of 24 (6 for the base offense level plus 18 for the loss).89 The average robbery is for $1,167 and involves some kind of dangerous weapon (gun, knife, or “other”).90 Again assuming no other additional aggravating offense characteristics, the average robbery has an offense level of . . . exactly the same, 24 (20 for the base offense level plus 4 for using a dangerous weapon).91

Finally, something the Sentencing Commission got right! So far as retributivism and general perceptions are concerned, the Sentencing Guidelines appropriately reflect the relative moral seriousness of white-collar fraud. Retributivists rejoice! From this perspective, the judges in Judge Bennett’s survey are mistaken if they apply some sort of discount when they sentence white-collar fraudsters. Judges: 0 | Sentencing Commission: 1.

V.     Conclusion: The Sentencing Commission Ekes Out a Win

In the final tally, the Sentencing Commission manages to pull ahead. Judge Bennett and other judges think that the Sentencing Guidelines are too harsh when it comes to white-collar fraudsters.92 The claim can only make sense in the context of a theory about what we are trying to do when we punish fraud. So far as deterrence and rehabilitation are concerned, some evidence suggests punishment severity may be pretty irrelevant. When it comes to retribution and giving fraudsters what they deserve, however, popular moral intuitions support the Guidelines ranges.

This is not a resounding victory for the Sentencing Commission. It turns on a conflict between two moral intuitions: the judges’ and the public’s. Conflicts of intuitions are infamously hard to resolve. On the one side, there is the common intuition that fraud and robbery are equally serious; on the other is the expert judicial opinion that fraud is less serious.93 I opted for the common intuition primarily because of worries about cognitive bias favoring defendants who share a socio-economic background with judges (and with me). But if the history of moral thought has taught us anything, it is that common moral intuitions are commonly wrong. It is probably best to incorporate this uncertainty into the final score and dock half a point from the Sentencing Commission. Final tally—Judges: 0 | Sentencing Commission: ½. That split is within an easy margin of error. A bit more data or a new argument could easily swing things the other direction.

Best two of three?


  1. [1]. See United States v. Stein, 541 F.3d 130, 157–58 (2d Cir. 2008) (holding that prosecutors breached white-collar defendants’ Sixth Amendment right to counsel by coercing their employer to advise them to meet with government agents without counsel); Julie R. O’Sullivan, The Last Straw: The Department of Justice’s Privilege Waiver Policy and the Death of Adversarial Justice in Criminal Investigations of Corporations, 57 DePaul L. Rev. 329, 329–30 (2008) (discussing the outrage and concern of defense bars about the Department of Justice’s practice of insisting that corporations waive attorney-client privilege to receive cooperation credit). As a parenthesis, these procedural abuses cannot hold a candle to the problems that often-destitute street criminals face. See, e.g., Wilkins v. May, 872 F.2d 190, 195 (7th Cir. 1989) (holding that federal agents violated the right to due process of law by holding a pistol to the defendant’s head during interrogation).

  2. [2]. See Edwin H. Sutherland, White Collar Crime: The Uncut Version 5–10 (1983).

  3. [3]. See Rodney Huff et al., The 2010 National Public Survey on White Collar Crime 12 (2010),

  4. [4]. See generally Kathryn E. McCollister et al., The Cost of Crime to Society: New Crime-Specific Estimates for Policy and Program Evaluation, 108 Drug & Alcohol Dependence 98 (Apr. 2010), (indicating that the cost of all other crimes is around $15 billion).

  5. [5]. Bennett and his co-authors have a much more sophisticated set of theses, including specific reforms they propose to the Sentencing Guidelines. See Mark W. Bennett et al., Judging Federal White-Collar Fraud Sentencing: An Empirical Study Revealing the Need for Further Reform, 102 Iowa L. Rev. 939, 980–88 (2017) (discussing proposed reforms). But the search for reform is prompted by the impression, supported by judicial survey data, that the Guidelines are generally too harsh on fraud. Whether directly or through approving quotation, Bennett and his co-authors say, with respect to white-collar defendants, that the Guidelines “tend to overstate the seriousness of many offenses,” id. at 979; “result in the imposition of excessively harsh prison sentences,” id. at 981; “result[] in unusually long sentences,” id. at 982; and produce “dissatisfaction with the[ir] harshness,” id. at 984.

  6. [6]. Id. at 972 (“[N]early three out of four federal district
    judges imposed the precise minimum sentence [to a hypothetical defendant described in the survey] ... .”).

  7. [7]. Id. at 944 (indicating that “judges regularly sentence economic criminals well below the minimum guideline in all but the smallest of loss cases” and that “three in four federal district judges sentenced [a hypothetical] defendant to the exact minimum sentence possible”).

  8. [8]. See United States v. Booker, 543 U.S. 220, 245 (2005) (holding that mandatory sentencing guidelines are unconstitutional).

  9. [9]. See id. (striking provisions that previously made the Guidelines ranges mandatory).

  10. [10]. See Gall v. United States, 552 U.S. 38, 49 (2007) (“[T]he Guidelines should be the starting point and the initial benchmark.”).

  11. [11]. See Julie R. O’Sullivan, Federal White Collar Crime: Cases and Materials 111 (4th ed. 2009).

  12. [12]. See Bennett et al., supra note 5, at 960 (“[W]hile sentenc[ing ranges] became stiffer, judges have chosen not to keep up.”).

  13. [13]. See Douglas D. Heckathorn, Respondent-Driven Sampling: A New Approach to the Study of Hidden Populations, 44 Soc. Probs. 174, 186 (1997) (“[T]he sum of proportional distributions of population must sum to one.”).

  14. [14]. See Andrew Weissmann & Joshua A. Block, White-Collar Defendants and White-Collar Crimes, 116 Yale L.J. Pocket Part 286, 289 (2007) (“One of the laudatory goals in promulgating the Sentencing Guidelines was to remedy the potential for hidden—or unhidden—bias in favor of ‘white collar’ defendants.”).

  15. [15]. See Robert J. Smith et al., Implicit White Favoritism in the Criminal Justice System, 66 Ala. L. Rev. 871, 891–904 (2015) (discussing how implicit racial biases form).

  16. [16]. See Jerry Kang et al., Implicit Bias in the Courtroom, 59 UCLA L. Rev. 1124, 1126 (2012) (discussing the nature of implicit bias).

  17. [17]. See generally id. (describing how the legal profession can begin to combat implicit bias).

  18. [18]. See Weissmann & Block, supra note 14 (noting the Sentencing Committee’s motivation).

  19. [19]. See Samuel W. Buell, Is the White Collar Offender Privileged?, 63 Duke L.J. 823, 833–34 (2014) (discussing how “lenient” and “notoriously lighter” sentencing for white-collar criminals motivated Congress to reform sentencing).

  20. [20]. See Kenneth Mann et al., Sentencing the White-Collar Offender, 17 Am. Crim. L. Rev. 479, 500 (1980) (discussing how judges feel empathy and sympathy for white-collar criminals, “whose position in society may be very much like their own”). There is data now suggesting that white-collar crime is no longer exclusively for white-collar types. See O’Sullivan, supra note 11, at 4–7 (providing an overview of white-collar crime statistics). But, as Judge Bennett et al. note, judges are less likely to deviate from Sentencing Guidelines recommendations in the lower-stakes fraud of middle-class, blue-collar folks. Bennett, supra note 5, at 963–64 (“The percent of sentences within the guideline range for offenders in the lowest loss category ($5,000 or less) was 84.1%. Between $7 and 20 million, it dropped to 26.8%, while for the nine offenders in the more than $400,000,000 loss category, the drop was 0.0%.” (citation omitted)).

  21. [21]. See Douglas O. Linder, Juror Empathy and Race, 63 Tenn. L. Rev. 887, 900 (1996) (“Vicariously identifying with another’s feelings, volitions, or ideas is most often accomplished with respect to a person whose experiences, values, and appearance are similar to one’s own.”).

  22. [22]. In all fairness, it should be noted that Judge Bennett is not always particularly sympathetic to “those white-collar folks.” See, e.g., United States v. VandeBrake, 771 F. Supp. 2d 961, 1012–13 (N.D. Iowa 2011) (sentencing antitrust defendant to four times the Guidelines maximum).

  23. [23]. It bears noting that the Sentencing Commission is a judicial agency. About the Commission, U.S. Sentencing Comm’n, (last visited Apr. 26, 2017) (“The U.S. Sentencing Commission, a bipartisan, independent agency located in the judicial branch of government ... .”).

  24. [24]. See Apprendi v. New Jersey, 530 U.S. 466, 490 (2000) (“Other than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury ... .”).

  25. [25]. United States v. Booker, 543 U.S. 220, 244 (2005).

  26. [26]. Whalen v. United States, 445 U.S. 684, 689 (1980) (“[T]he power to define criminal offenses and to prescribe the punishments to be imposed upon those found guilty of them, resides wholly with the Congress.”); see Beckles v. United States, 137 S. Ct. 886, 895 (2017) (“[T]he Guidelines advise sentencing courts how to exercise their discretion within the bounds established by Congress.”).

  27. [27]. See 18 U.S.C. § 3553(a)(2) (2012) (detailing factors the courts may consider, such as “the need for the sentence imposed—(A) to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense; (B) to afford adequate deterrence to criminal conduct; (C) to protect the public from further crimes of the defendant; and (D) to provide the defendant with needed educational or vocational training, medical care, or other correctional treatment in the most effective manner”); Albert W. Alschuler, The Changing Purposes of Criminal Punishment: A Retrospective on the Past Century and Some Thoughts About the Next, 70 U. Chi. L. Rev. 1, 1 (2003) (listing “retribution, deterrence, incapacitation, and rehabilitation” as the “textbook purposes of criminal punishment”).

  28. [28]. Judge Bennett and his co-authors are not the first to say we should give special deference to how judges think about white-collar sentencing. For another tribute, also without a clear statement of why, see Mann et al., supra note 20, at 480 (arguing that “sentencing policy [should] reflect the real world concerns of those who impose sentences”).

  29. [29]. See Bennett, supra note 5, at 943 (discussing results of an empirical study that interprets federal judges’ approaches to sentencing.).

  30. [30]. See generally McCollister, supra note 4 (calculating the economic costs of various criminal activities); Huff et al., supra note 3.

  31. [31]. See Donald C. Langevoort, Ego, Human Behavior, and Law, 81 Va. L. Rev. 853, 870 (1995) (discussing “ego-related explanations” within deterrence theory).

  32. [32]. Cindy R. Alexander & Mark A. Cohen, The Causes of Corporate Crime: An Economic Perspective, in Prosecutors in the Boardroom 11, 14–15 (Anthony S. Barkow & Rachel E. Barkow eds., 2011) (“Within this rational-choice ‘deterrence’ framework, individuals weigh the costs and benefits of crime-related activity against the expected sanction to maximize their private utility under the constraints of the organization in which they find themselves ... .”).

  33. [33]. See Richard H. McAdams & Thomas S. Ulen, Behavioral Criminal Law and Economics 4–5 (Univ. of Ill. Law & Econ. Research, Paper No. LE0-8-035, 2008) (discussing how behavioral studies revealing systematic problems in the average person’s decision-making process undermine the basic rational-choice presupposition of general-deterrence theory).

  34. [34]. See Stephanos Bibas, White-Collar Plea Bargaining and Sentencing after Booker, 47 Wm. & Mary L. Rev. 721, 724 (2005) (“[W]hite-collar crime is more rational, cool, and calculated than sudden crimes of passion or opportunity, so it should be a prime candidate for general deterrence.”).

  35. [35]. See generally P. A. Samuelson, A Note on the Pure Theory of Consumer’s Behaviour, 5Economica 61 (1938) (discussing the concept of revealed preferences).

  36. [36]. U.S. Sentencing Comm’n, Sentencing and Guideline Application Information for §2B1.1 Offenders 1, 1 fig.1 (2013).

  37. [37]. See Andrew J. Oswald, Irrational Actors, 348 Sci. 1099, 1099 (2015) (discussing the belief that one of the major failings in economics is the misplaced idea that people are rational actors).

  38. [38]. See Bennett et al., supra note 5, at 945–57 (recounting the history of the fraud guidelines).

  39. [39]. See Alexander & Cohen, supra note 32, at 20 (“Detection and sanctions are substitutes in the production of deterrence.”).

  40. [40]. See Daniel S. Nagin, Deterrence in the Twenty-First Century, 42 Crime & Just., 199–202 (showing that probability of detection affects deterrence calculus more than severity of sanction); see also A. Mitchell Polinsky & Steven Shavell, On the Disutility and Discounting of Imprisonment and the Theory of Deterrence, 28 J. Legal Stud. 1, 12 (1999) (“[White-collar criminals] are likely to be risk preferring in imprisonment, which suggests
    that less-than-maximal sanctions, combined with relatively high probabilities of apprehension, may be optimal.”).

  41. [41]. See Mann et al., supra note 20, at 483–84 (“Most judges share a widespread belief that the suffering experienced by the white-collar person as a result of apprehension, public indictment and conviction, and the collateral disabilities incident to conviction—loss of job, professional licenses, and status in the community—completely satisfies the [deterrence-based] need to punish the individual.”).

  42. [42]. See Peter J. Henning, Is Deterrence Relevant in Sentencing White-Collar Criminals?, 61 Wayne L. Rev. 27, 49 n.96 (2015) (“[A] short prison stay ... may not provide more than a marginal impact [for white-collar criminals] beyond the experience of prosecution, conviction, and sentencing.” (quoting David Weisburd et al., Specific Deterrence in a Sample of Offenders Convicted of White-Collar Crimes, 33 Criminology 587, 599 (1995)).

  43. [43]. See Pamela H. Bucy, Corporate Ethos: A Standard for Imposing Corporate Criminal Liability, 75 Minn. L. Rev. 1095, 1133–34 (1991) (discussing examples of institutional goals that encourage employees to commit fraud).

  44. [44]. The typical fraud takes place in the accounting department of a small private company. See Ass’n of Certified Fraud Exam’rs, Inc., Report to the Nations on Occupational Fraud and Abuse 32, 55 (2016) (indicating that more fraud originates from accounting departments than from any other business unit and that the average fraud takes place in a private company under 100 people).

  45. [45]. See Albert W. Alschuler, supra note 27 (indicating that a “new penology” emerged to end rehabilitation as a primary focus).

  46. [46]. 1 Wayne R. LaFave, Substantive Criminal Law § 1.5 (2d ed. 2016) (discussing rehabilitation for the purpose of re-entry into society as a goal of criminal law); John Kaplan et al., Criminal Law: Cases and Materials 46 (7th ed. 2012) (discussing the acceptance of rehabilitation as a purpose of punishment in criminal law).

  47. [47]. See RA Duff, Virtue, Vice and the Criminal Law – A Response to Huigens and Yankah, in Law, Virtue and Justice 195, 196 (Amalia Amaya & Ho Hock Lai eds.,
    2013) (“We could ... use criminal law and punishment as ways of directly fostering virtue and preventing vice ... .”).

  48. [48]. See Benjamin B. Sendor, The Relevance of Conduct and Character to Guilt and Punishment, 10 Notre Dame J.L. Ethics & Pub. Pol’y 99, 100 (1996) (“‘[B]ad character’ in this context means a ‘settled disposition’ ... to commit acts that violate the law.” (citation omitted)).

  49. [49]. Judges do not seem to be concerned with rehabilitation for white-collar criminals either. Mann, supra note 20, at 482.

  50. [50]. Patricia M. Jones, Sentencing, 24 Am. Crim. L. Rev. 879, 885 (1987) (indicating that it is commonly argued “that the defendant’s respectable community position and gainful employment make rehabilitation unnecessary”).

  51. [51]. See Katie A. Fredericks et al., White Collar Crime: Recidivism, Deterrence, and Social Impact, 2 Forensic Res. & Criminology Int’l J. 1, 7 (2016). We should take this number with a grain of salt, since the authors used an unusually broad definition of “white-collar crime,” which seems to include many non-violent property offenses, like motor vehicle theft, that would not appear on standard lists.

  52. [52]. Elizabeth Szockyj, Imprisoning White-Collar Criminals?, 23 S. Ill. U. L.J. 485, 494 (1999) (indicating that 72% of regulatory offenders and 64% of fraudulent offender have no prior adult convictions).

  53. [53]. John Irwin, Prisons in Turmoil 2 (1980) (indicating that early American rehabilitation theorists hoped to “ke[ep felons] in quiet solitude, reflecting penitently on their sins in order that they might cleanse and transform themselves”).

  54. [54]. See Gianluca Di Muzio, Aristotle on Improving One’s Character, 45 Phronesis 205, 214 (2000) (interpreting Aristotle to say that reform must be self-initiated and “can [only] be attained through a process that does not rely on means such as persuasion or punishment”).

  55. [55]. See Alschuler, supra note 27, at 2–6.

  56. [56]. Id. at 9 (“The demise of rehabilitation was attributable less to jurisprudential reflection than to apparent empirical failure.”). For a somewhat different and more detailed account of the decline of the rehabilitative ideal in punishment, see Richard C. Boldt, Rehabilitative Punishment and the Drug Treatment Court Movement, 76 Wash. U. L.Q. 1205, 1218­–45 (1998) (discussing the rise and decline of rehabilitation).

  57. [57]. See 18 U.S.C. § 3582(a) (2012) (“[I]mprisonment is not an appropriate means of promoting correction and rehabilitation.”); 28 U.S.C § 994(k) (2012) (“The Commission shall insure that the guidelines reflect the inappropriateness of imposing a sentence to a term of imprisonment for the purpose of rehabilitating the defendant ... .”).

  58. [58]. S. Rep. No. 98-225, at 37–38 (1983), as reprinted in 1983 U.S.C.C.A.N. 3182, 3220­–21 (1983).

  59. [59]. Michael Caruso, Fed. Pub. Def., S. Dist. Fla., Statement Before the United States Sentencing Commission Public Hearing on Economic Crime and Inflationary Adjustments 5 (Mar. 12, 2015),

  60. [60]. Kristina Murphy & Nathan Harris, Shaming, Shame and Recidivism: A Test of Reintegrative Shaming Theory in the White-Collar Crime Context, 47 Brit. J. Criminology. 900, 910 (2007) (“[T]hese results provide some support for the central prediction of reintegrative shaming theory—that reintegration of offenders leads to lower reoffending ... .”).

  61. [61]. Mark Motivans, U.S. Dep’t of Justice, Bureau of Justice Statistics, Federal Justice Statistics 2006—Statistical Tables tbl.7.3 (indicating that fraudsters are roughly half as likely, as compared to violent offenders, to reoffend while on probation or parole).

  62. [62]. See Ass’n of Certified Fraud Exam’rs, Inc., supra note 44.

  63. [63]. See generally Martin L. Needleman & Carolyn Needleman, Organizational Crime: Two Models of Criminogenesis, 20 Soc. Q. 517 (1979).

  64. [64]. Bucy, supra note 43, at 1101.

  65. [65]. See Mihailis E. Diamantis, Clockwork Corporations: A Character Theory of Corporate Punishment, 103 Iowa L. Rev. (forthcoming Jan. 2018) (arguing that criminal punishment of corporations should focus exclusively on such reforms).

  66. [66]. Michael S. Moore, The Moral Worth of Retribution, in Responsibility, Character, and the Emotions: New Essays in Moral Psychology 179 (Ferdinand Schoeman ed., 1987) (“Retributivism is the view that punishment is justified by the moral culpability of those who receive it.” (emphasis omitted)).

  67. [67]. Andrew von Hirsch, Censure and Sanctions 6–19 (1993).

  68. [68]. Though Judge Bennett and his co-authors vigorously dispute this point, see Bennett et al., supra note 5, at 957 (quoting Kate Stith & José A. Cabranes, Fear
    of Judging: Sentencing Guidelines in the Federal Courts 69–70 (1998)) and id. at 979 (quoting with approval United States v. Emmenegger, 329 F. Supp. 2d 416, 427 (S.D.N.Y. 2004) (“[T]he amount stolen is a relatively weak indicator of the moral seriousness of the offense or the need for deterrence.”)), they offer no empirical support for their resistance. I suspect that the data are not in their favor. Psychologists have long suspected that harm caused is one of two primary factors we use to judge how much a person deserves to be punished (the other is the mental state of the offender). See Morris B. Hoffman, The Punisher’s Brain: The Evolution of Judge and Jury 68 (2014) (“Our adult brains . . . have dedicated circuits devoted to the assessment of intentionality and harm, and to the calculation of blame based on those two assessments ... .”).

  69. [69]. U.S. Sentencing Guidelines Manual § 2B1.1(b)(1) (U.S. Sentencing Comm’n 2016).

  70. [70]. Id. § 2B1.1(b)(2).

  71. [71]. See Bennett et al., supra note 5, at 983 (stating that the Guidelines “seriously overstate criminal culpability” of white-collar fraud).

  72. [72]. It is not for nothing that the sentencing rules for fraud and burglary are situated right next to each other. U.S. Sentencing Guidelines Manual § 2B1–2B2.

  73. [73]. Id. § 2B3.1(b)(3).

  74. [74]. Id. § 2B3.1(b)(7).

  75. [75]. Bennett et al., supra note 5, at 979 (opining that the fraud Guidelines “tend to overstate the seriousness of many offenses” (quoting James E. Felman, Chairman, Am. Bar Ass’n Criminal Justice Section, Testimony on Behalf of the American Bar Association Before the United States Sentencing Commission for the Hearing on Proposed Amendments to the Federal Sentencing Guidelines Regarding Economic Crimes 2 (Mar. 12, 2015), ussceconcrimetestimony.authcheckdam.pdf)).

  76. [76]. I say “directly” because economic harm can cause physical injury indirectly by leaving victims with fewer resources to care for their physical well-being. Think medical bills.

  77. [77]. U.S. Dep’t of Justice, Bureau of Justice Statistics, The National Survey of Crime Severity 153–65 (1985).

  78. [78]. Id. at 155.

  79. [79]. Id.

  80. [80]. Id. at 156. One interesting feature of the measures of the seriousness of white-collar crime is that the standard deviation is greater than it is for the measures of street crimes. Id. at iii. This means that, regardless of what the average seriousness judgment is, there is a wider range of on white-collar crime than on street crime. So, when Judge Bennett and I have intuitions about fraud that differ from the mean, we just show ourselves to fall to one end of a pattern of deviation.

  81. [81]. Id. at 155.

  82. [82]. Stuart P. Green & Matthew B. Kugler, Public Perceptions of White Collar Crime Culpability: Bribery, Perjury, and Fraud, 75 L. & Contemp. Probs. 33, 52­–58 (2012) (presenting data on the fine-grained moral judgments people make about when fraud deserves to be punished).

  83. [83]. See Kristy Holtfreter et al., Public Perceptions of White-Collar Crime and Punishment, 36 J. Crim. Just. 50, 53 (2008) (finding that people think street criminals and fraudsters should receive the same punishment); Andrea Schoepfer et al., Do Perceptions of Punishment Vary Between White-Collar and Street Crimes?, 35 J. Crim. Just. 151, 157 (2007) (“Specifically, 31 percent of the respondents believed that the robbery should receive a more severe punishment and 31 percent believed that fraud should receive a more severe punishment, while 38 percent of the respondents believed that the two crimes should be equally punished.”).

  84. [84]. Huff et al., supra note 3, at 8–9 (emphasis added).

  85. [85]. This certainly seems to have been part of the purpose behind the fraud guidelines. See Alan Ellis et al., At a “Loss” for Justice: Federal Sentencing for Economic Offenses, 25 Crim. Just. 34, 36 (2011) (“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).”).

  86. [86]. U.S. Sentencing Guidelines Manual § 2B1.1(b)(2), § 2B3.1(b)(2) (U.S. Sentencing Comm’n 2016).

  87. [87]. Id. §§ 2B1.1(b)(1), 2B3.1(b)(7). This last feature actually seems to mirror people’s intuitions—people judge embezzling $1,000 to be much more serious than embezzling $10, but judge robbing someone at gunpoint of $1,000 as pretty comparable to robbing them at gunpoint of just $10. U.S. Dep’t of Justice, supra note 77, at 155–56.

  88. [88]. U.S. Sentencing Comm’n, Overview of Federal Criminal Cases Fiscal Year 2014 8 (2015).

  89. [89]. U.S. Sentencing Guidelines Manual § 2B1.1(a)–(b).

  90. [90]. Robbery, Fed. Bureau of Investigation: Uniform Crime Reporting, (last visited May 21, 2017).

  91. [91]. U.S. Sentencing Guidelines Manual § 2B3.1(a)–(b).

  92. [92]. See Bennett et al., supra note 5, at 984 (indicating “dissatisfaction with the harshness of guidelines sentences”).

  93. [93]. See id. at 972 (“[I]t is plausible that federal district judges’ ‘short’ sentences were not entirely due to empathy, but rather experience with (or rejection of) overly harsh sentencing guidelines.”).


Associate Professor of Law, University of Iowa, College of Law.

Many thanks to my research assistant, Curtis Underwood.