In Kumar, the government charged the former CEO of Computer Associates, a publicly traded company, along with the company’s head of North American sales, with various crimes relating to a securities fraud scheme. The case arose out of a fraudulent accounting practice known as the “35-day month.” The company had backdated contracts executed in the first few days of a financial quarter to recognize revenue from those contracts in the prior quarter. The illegal purpose of the practice was to make investors believe that the company had met or exceeded its quarterly earnings estimates. Eventually, the defendants pleaded guilty to fraud and other offenses. They were sentenced in November 2006.

Ordinarily, a sentencing court is required to apply the guidelines in effect at the time of sentencing. See U.S.S.G. § 1B1.11(a). If the application of that rule creates an ex post facto problem, however, then a sentencing court is required to apply the guidelines in effect at the time of the offense conduct. The binary choice required under the guidelines is commonly known as the “one-book rule.” See U.S.S.G. § 1B1.11(b)(1); Miller v. Florida, 482 U.S. 423, 430 (1987) (Ex Post Facto clause implicated if a law applies to events that occurred before its enactment and disadvantages the defendant affected by the law);  Weaver v. Graham, 450 U.S. 24, 30 (1981) (“the lack of fair notice and governmental restraint when the legislature increases punishment beyond what was prescribed when the crime was consummated” is central to the ex post facto prohibition).

The conduct underlying the charged fraud offenses in Kumar ended in 2000. Certain uncharged obstructive conduct relating to the fraud offenses had continued until 2005. The district court applied the 2005 guidelines (the version in effect at the time of sentencing) instead of the 1998 guidelines (the version in effect at the time of the offense conduct). The 2005 version of the guidelines contained several sentencing enhancements that did not exist in the 1998 version, including an enhancement for obstructive conduct. The district court, after deciding that defendants had obstructed the government’s investigation of the underlying conduct, applied the enhancement.

On appeal, the defendants argued that they were unfairly punished by application of the 2005 guidelines under the one-book rule because the enhancement did not exist before 2000 at the time of the offense conduct. The panel majority decided that the application of the 2005 guidelines clearly disadvantaged the defendants under the Miller standard because their sentencing ranges under the guidelines were much higher than they would have been if the 1998 guidelines had been applied. But the panel majority concluded that no ex post facto problem existed under Miller because the application of the 2005 guidelines was not retrospective. According to the Kumar majority, the defendants had fair notice of the consequences of committing obstructive conduct postdating the offense conduct because the one-book rule had been adopted before their offense conduct occurred.

Judge Sack, in dissent, stated that “the majority give[s] insufficient attention to the quality of notice that ex post facto jurisprudence requires. It is not notice simpliciter, but notice that is ‘fair.’” Slip op. at 19. In other words, Judge Sack stated a view that any notice created by the existence of the one-book rule existed at a level of generality too high to permit an enhancement based on an amendment to the guidelines that had been promulgated after the offense conduct. Ultimately, the disagreement between the panel majority and Judge Sack involves the types of line-drawing and contextual issues concerning the scope and meaning of notice that are often present in litigation under the Due Process Clause.

All of this invites a more basic question: What practical difference does ex post facto analysis make when the sentencing guidelines are simply one non-binding factor that a court must consider under 18 U.S.C. § 3553(a)? The Seventh Circuit, in an opinion by Judge Posner, has held squarely that the Ex Post Facto clause does not even apply to the post-Booker advisory sentencing guidelines. See United States v. Demaree, 459 F.3d 791, 795 (7th Cir. 2006) (“We conclude that the ex post facto clause should apply only to laws and regulations that bind rather than advise[.]”). According to Demaree, a sentencing judge who agrees with a guideline enhancement that was promulgated after the offense conduct can reach the same result through another means under 18 U.S.C. § 3553(a) as if he or she were directly applying the guideline enhancement. Id. (“[A] rule that a guidelines change cannot be applied retroactively if it would be adverse to the defendant would have in the long run a purely semantic effect”). (Even before United States v. Booker, when the sentencing guidelines were mandatory, Judge Easterbrook took the position that “sentencing guidelines are not ‘laws’ within the scope of [the Ex Post Facto] clause.” See United States v. Vivit, 214 F.3d 908, 924 (7th Cir. 2000) (concurring opinion).) Other circuits have disagreed with Demaree.  See, e.g., United States v. Turner, 548 F.3d 1094, 1099-1100 (D.C. Cir. 2008).

In Kumar, the district court had held (consistent with Demaree) that “because the Guidelines [are] advisory, the Ex Post Facto clause [is] not implicated by sentencing decisions.” Slip op. at 8. The court of appeals, however, chose to sidestep that bedrock question entirely. See slip op. at 25 n.12 (majority opinion); id. at 11 n.7 (dissenting opinion). Presumably, the Second Circuit proceeded in that manner because the government on appeal disclaimed any reliance on the district court’s position. Slip op. at 25 n.12. In Demaree, the government similarly had argued on appeal that the Ex Post Facto clause applies to the advisory guidelines. Not mincing words, Judge Posner characterized the government’s position as a “rearguard action against Booker” and an effort to make the guidelines “bind as tightly as possible.” 459 F.3d at 795.  (Now, the government has taken a position at odds with its stance in Demaree and Kumar. See United States v. Lanham, et al., Nos. 08-6504/6506, 09-5094/5095, slip op. at 20 (6th Cir., Aug. 24, 2010) (rejecting government’s argument that harsher, post-conduct version of the guidelines applied and stating that advisory nature of guidelines does not eliminate ex post facto concerns).  This is either a change of course or the Acting Solicitor General will need to impose order here.)

Whether one chooses to accept Judge Posner’s brand of legal realism, there is no question that a district judge has broad discretion to reach results under Section 3553(a) that vary substantially from the guidelines. In Kumar itself, the district court varied downward from life sentences recommended by the guidelines and imposed sentences of 7 years and 12 years. In the end, the type of intense legal skirmishing exemplified by Kumar may demonstrate, more than anything else, that the advisory guidelines continue to have a tethering effect despite their status under Section 3553(a) as a single factor among many that must be considered by courts in federal sentencing proceedings.

The logic behind Dura has enormous importance in federal criminal cases involving publicly traded securities. Under the advisory United States Sentencing Guidelines, the calculation of a sentence in a fraud case is largely driven by the amount of “loss” caused by the fraud. U.S.S.G. Section 2B1.1.  “Actual loss” is defined as the “reasonably foreseeable pecuniary harm that resulted from the offense.” Id., Commentary, 3(A)(i). Under the guidelines, a district court is required only to make a “reasonable estimate of loss.” Id., Commentary, 3(C). If a corporate executive – Enron’s Jeff Skilling and WorldCom’s Bernie Ebbers are two prominent examples – is convicted of fraud involving hundreds of millions of dollars in decreased market capitalization, he or she may be subject to decades in prison under the guidelines (or what amounts to a life sentence).

Given the liberty interests involved in a criminal case, one might think that the standards would be at least as strict as those that apply in a civil case. Not so. Recently, the Ninth Circuit rejected the proposition that Dura should apply to sentencing in a federal criminal case. See United States v. Berger, No. 08-50171 (9th Cir., Nov. 30, 2009). According to the Ninth Circuit, the criminal sentencing context is different because “a court gauges the amount of loss caused, i.e., the harm that society as a whole suffered from the defendant’s fraud.” Slip op. at 15629 (emphasis in original). But that reasoning is utterly circular. Whether “society as a whole” suffered any loss depends on the facts involved. If a fraud is committed but is never disclosed to the market, it is possible that the fraud did not actually affect share price at all. And even if a fraud is disclosed, the actual harm to any particular purchaser can’t simply be determined by subtracting the share price on the date of purchase from the price on the date the fraud was revealed. The effect of the fraud needs to be isolated from the many other variables that affect share price over time.

Other courts of appeals, as the Ninth Circuit recognized in Berger, have embraced the logic of Dura in the criminal context. See United States v. Rutkoske, 506 F.3d 170, 179 (2d Cir. 2007) (“we see no reason why considerations relevant to loss causation in a civil fraud case should not apply, at least as strongly, to a sentencing regime in which the amount of loss caused by a fraud is a critical determinant of the length of a defendant’s sentence.”); United States v. Olis, 429 F.3d 540, 546 (5th Cir. 2005) (“The civil damage measure should be the backdrop for criminal responsibility both because it furnishes the standard of compensable injury for securities fraud victims and because it is attuned to stock market complexities.”). Still, the practical reach of those cases may be limited in light of the requirement in the guidelines that a court need only make a “reasonable estimate of loss.” Defendants will argue that the government must prove actual loss with sophisticated analysis; the government will argue that no such exactitude is required.  See United States v. Ferguson, 584 F. Supp.2d 447, 451-52 (D. Conn. 2008).

Time will tell whether the Ninth Circuit’s view of Dura in the criminal context will stand. Even in the Ninth Circuit, however, Berger cannot be allowed to end the story. In every federal criminal case involving alleged shareholder loss, defense counsel must be prepared to advocate Dura’s plain logic. To the extent that a sentencing court views Dura as inconsistent with the sentencing guidelines, defense counsel must be prepared to convince the court that the guidelines don’t deserve to be followed. They are merely advisory.  A district court is free to disagree with a policy of the Sentencing Commission that has not been mandated by Congress. See Kimbrough v. United States, 128 S. Ct. 558, 564 (2008).

Congress has not mandated that criminal defendants faced with potentially long prison sentences be subject to standards less exacting than those applicable to defendants in civil cases. To the contrary, Congress has mandated that a sentencing court must consider the “nature and circumstances of the offense and the history and characteristics of the defendant” and that a court “court shall impose a sentence sufficient, but not greater than necessary,” to “reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense,” and to “afford adequate deterrence to criminal conduct” and “protect the public from further crimes of the defendant.” See 18 U.S.C. Section 3553(a). Those principles, and proof of actual harm, must be guiding lights in all cases involving alleged shareholder loss.  Anything less exacting serves to elevate administrative convenience and excessive retribution over fundamental fairness.