Appeals from the United States District Court for the Central District of Illinois. Nos. 3:10-cr-30058-RM-BGC et al.-Richard Mills, Judge.
The opinion of the court was delivered by: Easterbrook, Chief Judge
Before EASTERBROOK, Chief Judge, and POSNER, FLAUM, KANNE, ROVNER, WOOD, WILLIAMS, SYKES, TINDER, and HAMILTON, Circuit Judges.
A member of this court called for a vote on the question whether these four appeals should be heard en banc on the court's own initiative. A majority of the active judges did not vote in favor of rehearing en banc, and the proposal therefore fails. Petitions for rehearing or rehearing en banc will not be accepted; this decision is the court's final judgment. Three members of the court have written opinions explaining their votes.
EASTERBROOK, Chief Judge, with whom FLAUM, KANNE, SYKES, and TINDER, Circuit Judges, join.
These four appeals were filed by the United States with the Solicitor General's authorization. Eight days after the United States prevailed, the prosecutor filed a document styled "Notice of Changed Position" announcing that the Attorney General disagrees with this court (and apparently with the Solicitor General too). The "Notice of Changed Position" does not ask us to do anything in particular, but some members of the court believe that we should grant rehearing en banc and overrule United States v. Fisher, 635 F.3d 336 (7th Cir. 2011), which led our panel to decide these four appeals in the prosecutor's favor. I am content to leave Fisher undisturbed.
The Attorney General's "Memorandum for all Federal Prosecutors", dated July 15, 2011, directs United States Attorneys to argue that the Fair Sentencing Act of 2010, Pub. L. 111-220, 124 Stat. 2372 (2010), applies to all criminal prosecutions in which sentence was imposed on or after August 3, 2010, the day the President signed the bill. The Memorandum also directs United States Attorneys to argue that the 2010 Act does not apply to cases in which sentence was pronounced on August 2, 2010, or earlier, even if they were pending in the district court or appeal on August 3. In other words, the Attorney General has concluded that the 2010 Act is partially retroactive.
As far as I am aware, the Supreme Court has never held any change in a criminal penalty to be partially retroactive. The choice always has been binary: retroac- tive or prospective. And what makes application "retroactive" is a change in the legal consequences of activity that predates the new law's enactment. See generally Landgraf v. USI Film Products, 511 U.S. 244 (1994), which discusses what it means for application of a new statute to be retroactive, and the two exceptions to the presumption against retroactivity: new procedural rules and new jurisdictional requirements. The 2010 Act does not affect judicial procedure; it changes the penalty for criminal conduct. And it does not affect jurisdiction.
The common law distinguished increases in criminal punishments from reductions or repeals. Any law that repealed a criminal statute or reduced the defendant's punishment was fully retroactive, while in light of the Constitution's Ex Post Facto Clause a law creating a crime or increasing criminal punishment could apply only to conduct that occurred after the law changed. But in 1871 Congress enacted the General Saving Statute, now codified as 1 U.S.C. Â§109, which makes all changes prospective unless the new statute provides otherwise. Warden v. Marrero, 417 U.S. 653, 659--61 (1974), discusses this history. Section 109 provides:
The repeal of any statute shall not have the effect to release or extinguish any penalty, forfeiture, or liability incurred under such statute, unless the repealing Act shall so expressly provide, and such statute shall be treated as still remaining in force for the purpose of sustaining any proper action or prosecution for the enforcement of such penalty, forfeiture, or liability.
Defendants argued that Â§109 is irrelevant to the 2010 Act, because it reduces rather than "repeals" the penalties for crack cocaine. They also contended that a criminal does not "incur" a punishment until sentenced. Every circuit has concluded, to the contrary, that a law reducing criminal punishment is a repeal of the old statute and the enactment of a new one for the purpose of Â§109, and that a punishment is incurred when the crime is committed. Marrero supports both of these conclusions. Our precedent is United States v. Bell, 624 F.3d 803, 814--15 (7th Cir. 2010), which holds that Â§109 makes the 2010 Act prospective, because it lacks an express declaration of retroactivity. A footnote collects other circuits' equiva-lent decisions.^
Bell and the other circuits rejected arguments for retro-activity made by defendants whose appeals were pending on August 3, 2010. That's why courts could decide the question so quickly. A second wave of defendants, those sentenced on or after August 3, 2010, asked for partial retroactivity. Our circuit was the first to con-
^United States v. Goncalves, 642 F.3d 245 (1st Cir. 2011); United States v. Acoff, 634 F.3d 200 (2d Cir. 2011); United States v. Reevey, 631 F.3d 110 (3d Cir. 2010); United States v. Rhodes, 2011 U.S. App. LEXIS 10238 (4th Cir. May 20, 2011) (one of at least five opinions in that circuit, all non-precedential); United States v. Doggins, 633 F.3d 379 (5th Cir. 2011); United States v. Carradine, 621 F.3d 575 (6th Cir. 2010); United States v. Brewer, 624 F.3d 900 (8th Cir. 2010); United States v. Baptist, 2011 U.S. App. LEXIS 11056 (9th Cir. June 2, 2011); United States v. Lewis, 625 F.3d 1224 (10th Cir. 2010); United States v. Gomes, 621 F.3d 1343 (11th Cir. 2010). The D.C. Circuit has yet to address this subject. sider that possibility. The panel in Fisher rejected the argument that the date of sentencing matters. If the 2010 Act is retroactive, then it applies to all pending cases no matter how far they have got in the judicial system; if it is not retroactive, then it applies only to crimes committed on or after August 3, 2010. Nothing depends on the sentencing date, which reflects how long it took to catch a criminal, and the state of the district judge's calendar, rather than principles of deterrence or desert. Section 109 says that the former law "shall be treated as still remaining in force" for pre-amendment conduct. If the old law is "treated as still remaining in force", the new law can't be applied to persons newly sentenced for pre-amendment crimes; Â§109 forecloses partial retroactivity.
United States v. Douglas, 644 F.3d 39 (1st Cir. 2011), reached a contrary conclusion, apparently unaware that it was creating a conflict with Fisher, which had been issued 20 days earlier. Douglas held that the new minimum and maximum sentences take effect for defendants sentenced on and after November 1, 2010. (I'll come back to the source of that date.) United States v. Rojas, 2011 U.S. App. LEXIS 13677 (11th Cir. July 6, 2011), then misread Douglas as holding that the new rules take effect with sentencing on and after August 3, 2010, and applied that transition date. (Rojas was sentenced on August 3 and would not have been eligible for a lower sentence under Douglas.) On July 7 our panel in Holcomb remanded four cases with instructions to apply Bell and Fisher. And on July 15 the Attorney General issued his memorandum agreeing with Rojas. That led to the United States Attorney's "Notice of Changed Position" in the four appeals this circuit had decided on July 7. Then United States v. Dixon, 2011 U.S. App. LEXIS 16374 (3d Cir. Aug. 9, 2011), followed Rojas without explaining why it chose August 3 rather than November 1 as the transition date. Most recently, United States v. Sidney, 2011 U.S. App. LEXIS 16421 (8th Cir. Aug. 10, 2011), agreed with Fisher and concluded that Â§109 does not permit partial retroactivity.
When the Executive Branch confesses error, this circuit gives respectful consideration to the rationale for the new position. A recent example is United States v. Corner, 598 F.3d 411 (7th Cir. 2010) (en banc), which overruled several of the circuit's decisions after the Solicitor General filed a brief carefully explaining where the circuit had gone wrong. That explanation carried the day; Corner was unanimous. Unfortunately, the Attorney General's "Memorandum for all Federal Prosecutors" lacks the sort of analysis that was so helpful in Corner. The Memorandum does not discuss Â§109 or the language of the 2010 Act. It does not explain why partial retroactivity is appropriate-or why the transition should depend on the date of sentencing rather than some other event, such as a guilty plea or appeal. The Attorney General does quote from the caption of S. 1789, which describes the proposal as "A bill to restore fairness to Federal criminal sentencing", but this language precedes the enacting clause and is not part of the United States Code. It also is unhelpful in evaluating a proposal for retroactive application. Every law lowering sentences expresses a legislative conclusion that sentences had been excessive. The common law responded by applying penalty reductions retroactively. But Â§109 provides otherwise. The observation that Congress, the President, and many federal judges think the former rules excessively severe does not distinguish the 2010 Act from any other law reducing sentences and does not justify disregarding the anti-retroactivity norm created by Â§109.
Douglas asked what reason there could be to continue imposing sentences that the 2010 Act condemns as excessive. The same question could be asked about every other law that reduces criminal sentences. The answer must be that Â§109 itself supplies the reason. It tells us that statutory lenience does not reduce the punishment for acts completed before the new law took effect. Perhaps the common law reflects greater wisdom than does Â§109, but Congress has displaced the common law.
Although Â§109 says that only an "express" provision in a later statute can support retroactivity, Congress is entitled to change that rule just as it is entitled to change the punishment for distributing crack cocaine. The legislature of 1871 can't tie the hands of the legislature sitting in 2010. This may be why the Court suggested in Marrero that a "fair implication" in a new law could allow retroactive application. 417 U.S. at 659 n.10. An earlier decision, Great Northern Ry. v. United States, 208 U.S. 452, 465 (1908), said that when a new law "by necessary implication" applies to pre-enactment crimes, the courts must follow the newer law rather than Â§109. A necessary (or fair) implication falls short of an express provision but could show that Congress has amended Â§109 to that extent. Still, unless superseded, Â§109 is as authoritative as other rules found in the Dictionary Act, 1 U.S.C. Â§Â§ 1--8, and in 18 U.S.C. Â§Â§ 5--27. Definitions, presumptions, and presets are essential to understanding legal texts. They are subject to revision, but a court should not lightly infer that Congress has tossed out all the framework laws that facilitate interpretation-and thus facilitate legislation too, by giving the legislature a formulary to use. See, e.g., Rowland v. California Men's Colony, 506 U.S. 194 (1993) (holding that the context clause in the Dictionary Act allows departure from the presumptive definitions only if there is no other plausible linguistic understanding of the new statute). See also Kentucky Association of Health Plans, Inc. v. Miller, 538 U.S. 329 (2003) (enforcing the rule in the McCarran-Ferguson Act that only federal laws expressly applying to insurance supersede state regulatory schemes).
Neither the Attorney General nor any member of this court believes that the 2010 Act is fully retroactive. To say that the 2010 Act is not fully retroactive is to say that Congress did not supersede Â§109, expressly or by implication. Section 109 forecloses partial retroactivity by providing that the former law "shall be treated as still remaining in force" for pre-amendment conduct. If Â§109 has not been superseded, what is the justification for partial retroactivity? The Attorney General, like the first, third, and eleventh circuits, is silent on that subject.
Some of my colleagues believe that support for partial retroactivity can be found in Â§8 of the 2010 Act, which provides:
The United States Sentencing Commission shall-
(1) promulgate the guidelines, policy statements, or amendments provided for in this Act as soon as practicable, and in any event not later than 90 days after the date of enactment of this Act, in accordance with the procedure set forth in section 21(a) of the Sentencing Act of 1987 (28 U.S.C. 994 note), as though the authority under that Act had not expired; and
(2) pursuant to the emergency authority provided under paragraph (1), make such conforming amendments to the Federal sentencing guidelines as the Commission determines necessary to achieve consistency with other guideline provisions and applicable law.
A requirement for a change in Guidelines within 90 days of the new law's enactment does not imply anything about minimum and maximum sentences on August 3, 2010. The new Guidelines came into effect on November 1, 2010. This is the source of the date that Douglas chose. The first circuit thought it would be incongruous if the new Guidelines, but not the new minimum and maximum sentences, applied to defendants sentenced on or after November 1, 2010. This does not support the Attorney General's view that August 3 marks the transition. Putting the statutory changes into effect while the Commission was still deliberating would be just as incongruous as putting the Guidelines but not the new minimum and maximum penalties into effect. Yet none of my colleagues concludes that the rules change for sentences on November 1, 2010, or for that matter November 1, 2011-when the 2010 Guidelines will be given retroactive effect. See Sentencing Commission Release of July 1, 2011, making Amendment 750, which implemented the 2010 Act, retroactive as of November 1, 2011. (The Commission's authority to apply new Guidelines to closed cases comes from 18 U.S.C. Â§3582(c)(2). See Dillon v. United States, 130 S. Ct. 2683 (2010).)
A reader might be inclined to ask why the 2010 Act's changes to minimum and maximum sentences should not take effect on November 1, 2010, the same date as the revised Guidelines-for revised Guidelines apply to new sentences even if the conduct took place years earlier. See 18 U.S.C. Â§3553(a)(4)(A)(ii); United States v. Demaree, 459 F.3d 791 (7th Cir. 2006). There is no inconsistency however, because the Guidelines and the 2010 Act are doing different things. The statute that provides penalties for cocaine and cocaine base, 21 U.S.C. Â§841(b), sets minimum and maximum punishments; the Guidelines then influence where within that range the judge imposes sentence. The 2010 Act amended Â§841(b). Judges are free to disagree with the Commission, see United States v. Booker, 543 U.S. 220 (2005); Kimbrough v. United States, 552 U.S. 85 (2007), but they are not free to disagree with Congress. Thus we have two retroactivity dates. One is when the new minimum and maximum penalties take effect; the other is when the revised Guidelines take effect. The Commission has, and has used, statutory authority to apply the lower Guidelines even to closed cases starting November 1, 2011. The Commission lacks any equivalent authority to make different statutory minimum and maximum sentences applicable to cases in which the criminal conduct predated August 3, 2010.
These four appeals are about the retroactivity of the changes to the statutory minimum and maximum sentences, not about the amended Guidelines. And this is why Â§8 of the 2010 Act does not affect these appeals. Changes to the Guidelines have nothing to do with minimum and maximum sentences. That was settled in Neal v. United States, 516 U.S. 284 (1996), which the Court reaffirmed in Kimbrough and DePierre v. United States, 131 S. Ct. 2225 (2011). When the Commission dramatically lowered the Guideline ranges for LSD, defendants argued that it would be preposterous to apply the new Guidelines (which do not count the weight of the carrier medium) while leaving unchanged the statutory minimum, which Chapman v. United States, 500 U.S. 453 (1991), holds does count the carrier's weight. Neal held, however, that the statute and the Guideline are unconnected and that arguments about incongruity do not justify modifying the statutory minimum and maximum sentences.
The ratio in the Guidelines has not been 100:1 since 2007. That year, the Sentencing Commission dropped most cocaine-base sentences by two levels. See Amendment 706, effective November 1, 2007 (and made retroactive by Amendment 713 as of March 3, 2008). The result of the 2007 change was a ratio that, depending on quantity,
could be as low as 25:1 or as high as 80:1. For many offenders the further change in 2010 does not matter. No one who distributes 8.4 kilograms or more of cocaine base received a lower Guideline range. Many common quantities have the same base offense level before and after the 2010 change to U.S.S.G. Â§2D1.1. For example, the level for distributing 1 kilo dropped from 36 to 34 in 2007 but stayed at 34 in 2010. Some defendants receive a benefit from both revisions: the level for 100 grams of crack fell from 30 to 28 in 2007 and to 26 in 2010. My point is not that the 2010 changes are slight for everyone-the benefit can be large for persons who distribute small quantities (the level for 5 grams drops from 24 to 16)-but that the Guidelines abandoned the 100:1 ratio in 2007, not 2010. Neither in 2007 nor in 2010 did Congress link the time of change in the Guidelines' ratio to the time of change in the minimum and maximum penalties.
Section 8, which tells the Commission to get a move on in revising its Guidelines, does not imply anything about when the new minimum and maximum sentences go into force. Because for some quantities the difference between the 2007 and 2010 quantity tables is small or nonexistent, one effect of the rapid revision is to increase penalties swiftly for the most serious offenders (and to increase the difference between the penalties for the worst offenders and the least serious ones)-as sections 5 and 6 of the 2010 Act call for higher Guide-lines when certain aggravating factors are present, while Â§7 directs the Sentencing Commission to reduce the punishment for offenders with minimal roles. Those changes are unrelated to the crack/powder ratio.
Section 10 of the 2010 Act tells the Sentencing Com-mission to study the effects of the new legislation and report within five years. Dixon observed that, unless the new legislation applies retroactively, the study will be limited to the law's effect on persons who distribute cocaine base after August 2, 2010. So? There will be plenty of people in that category. The point of such a study is to ascertain how lower penalties affect the volume of crime. People who distributed cocaine before the 2010 Act expected to be subject to the old penalty structure; their behavior cannot be changed by a later drop in sentences. A study of the 2010 Act's effects will produce meaningful results only if limited to persons whose criminal conduct occurs while the 2010 Act is in force. I do not think that Â§10 supplies much footing for an inference one way or the other, but, if Â§10 is relevant, Dixon got things backward.
Thoughtful people might wonder what sense it makes for Congress, having decided that a 100-to-1 ratio is excessive, to leave the minimum and maximum sentences alone for persons whose crimes predate August 3, 2010. It is a good question, to which there is no satis-factory answer other than the observation that legislation is an exercise in compromise. Some legislators sup-ported the existing 100-to-1 ratio between cocaine base and powder cocaine, while others thought that the two versions of this drug should be treated the same, as the Sentencing Commission once recommended. Some members of Congress wanted to reduce the disparity by raising the penalties for powder cocaine; others wanted to address it by reducing the penalties for crack. Members of Congress compromised at a ratio of 18 to 1, with most change coming through reductions in minimum and maximum terms of imprisonment.*fn1
There's no scientific basis for the 18-to-1 ratio, or for getting there by reducing crack sentences rather than increasing powder sentences, but it was the best that the advocates of parity could achieve (or, equivalently, the most that other legislators would concede). I don't mean by this that the 18:1 ratio is irrational, only that it is arbitrary, in the same sense that a statute of limitations is arbitrary. (Why 90 or 270 days for employment-discrimination suits, 2 years for claims under the Federal Tort Claims Act, 4 years for the residual statute in 28 U.S.C. Â§1658, and 5 years for most federal felonies?) Many Members of Congress who wanted parity also favored retroactivity, and Members who supported a higher ratio also favored no retroactivity. One way proponents of this law could achieve a lower ratio was to give up on retroactivity. The ratio, and retroactivity, are among the several dimensions of this compromise.
Most legislative deals are struck off the floor. I do not claim inside knowledge about this one. Perhaps I err in guessing about how this law came to have an 18:1 ratio and to allow the Sentencing Commission to implement retroactive Guidelines. I broach this subject only to say why I do not find persuasive an argument along the lines of: "The revised Guidelines were in place by November 1, 2010, so the new minimum and maximum penalties must apply to at least some persons whose crimes occurred before August 3, 2010." That theme disregards the compromise nature of legislation.
When Congress enacts a bill, a majority agrees on its text, not on grand principles. Neither side got everything it wanted in this statute, and judges disserve the legislative process by giving one side more than it secured at the bargaining table. Indeed, the tendency to provide one side with "just a little more in the right direction" can make legislation harder to accomplish by requiring Congress to take up, and resolve, all of the ways in which the judiciary might be tempted to tinker. "[N]o legislation pursues its purposes at all costs. Deciding what competing values will or will not be sacrificed to the achievement of a particular objective is the very essence of legislative choice-and it frustrates rather than effectuates legislative intent simplistically to assume that whatever furthers the statute's primary objective must be the law." Rodriguez v. United States, 480 U.S. 522, 525--26 (1987) (emphasis in original). For all we can know, a belief in Congress that the judiciary would make the law partially retroactive would have stiffened the opposition to the bill, and the 100:1 ratio would still be in force today.
Choosing an effective date for new legislation can be as arbitrary as deciding how many grams of cocaine hydrochloride receive the same treatment as one gram of cocaine base. The Attorney General relies heavily on the word "fair" in the title of the Fair Sentencing Act, but what's fair about condemning someone sentenced on August 2 to more time in prison than a person sentenced the next day, even though they committed their crimes on the same date (and may have been co-conspirators)? Suppose comrades in crime distribute cocaine in mid-2009 and are caught promptly. One confesses, pleads guilty, and testifies at the trial of the other, who fights tooth and nail and falsely denies culp-ability. The first is sentenced on August 1, 2010, the second on September 1. How would it be "fair" (or even conscionable) to give the lower sentence to the person who refused to accept responsibility for his crimes, just because by dragging out the process that person was sentenced after August 2?
It would be weird to conclude that, the longer it takes to issue an indictment, or the better the offender at evading capture, and hence the later the sentencing date, the lower the sentence. Why should the changes in the minimum and maximum terms take effect before the changes in the Guidelines (November 1, 2010)? Rojas, Dixon, and the Attorney General do not even try to explain why they chose August 3 rather than November 1 as the transition date. Why change the rules as of the date of sentencing rather than the date of arraignment, plea, or trial, the date the appeal is decided, or some other event? Any of those transition dates would produce incongruities. Only full retroactivity, or no retroactivity, treats equal criminal conduct equally.
If the President wants to apply the lower minimum and maximum penalties to all cases, pending and closed, he has only to issue a general commutation. The pardon power permits the President to achieve retroactive lenience if he is willing to pay the political price. By contrast, the judiciary must implement compromises faithfully, even when most judges wish that the political decision had been different. I have therefore voted not to hear these appeals en banc.
WILLIAMS, Circuit Judge, with whom POSNER, ROVNER,
WOOD, and HAMILTON, Circuit Judges, join, dissenting from the denial of rehearing en banc. The sentences Congress had originally mandated for crack cocaine offenses premised on drug quantities that were one hundred times lower than those for powder cocaine offenses are indefensible. There is no debate about that. Recognizing this, Congress wiped them out in the Fair Sentencing Act of 2010 to, in its own words, "restore fairness in Federal cocaine sentencing" by eliminating the 100:1 mandatory minimums. The only question in this case, odd as it might sound, is whether Congress wanted everyone sentenced after the Fair Sentencing Act became law to receive a "fair" sentence, or just some.
Our circuit should have heard this case en banc. Three other circuits have ruled that judges no longer must impose unfair sentences after the Fair Sentencing Act. This issue affects pending cases and many cases to come in light of the five-year statute of limitations on drug prosecutions. There were equal votes ...