United States v. Spaulding

GORSUCH, Circuit Judge,

dissenting.

Say a criminal defendant enters an involuntary guilty plea. Maybe because of improper threats. Or maybe thanks to *1128unlawful inducements. After sentencing, he seeks to withdraw the involuntary plea without the necessity of a full appeal or a collateral lawsuit. Seeing the problem with the plea, the government agrees and joins the defendant’s request. Can the district court grant the uncontested motion? Or must it grind on and gift the parties additional months and maybe years of needless judicial process to arrive at a result everyone admits the law requires? All while, most likely, the defendant sits in prison?

As my colleagues see it, Rule 11(e) and § 3582(c) require the district court to grind on. In their view, these provisions forbid the district court from undoing a guilty plea after sentencing — no matter how unlawful it may be, no matter if everyone concedes its illegality, no matter the matter at all. These provisions are jurisdictional, they reason, and must be enforced always and everywhere: if the district court fails to enforce them this court has a duty to do so. Of course, my colleagues take no pleasure in reaching their result: it is simply one they believe the law compels. My colleagues, too, offer a very thoughtful and persuasive opinion in support of their conclusion.

Respectfully, though, I do not believe the law is quite as unforgiving as they suggest. In case after case over the last decade, the Supreme Court has cautioned us that not every rule or statute qualifies as a “jurisdictional” requirement that courts must enforce mu sponte; most instead impose only claim-processing procedures that are subject to traditional defenses like waiver and forfeiture. See, e.g., Kontrick, v. Ryan, 540 U.S. 443, 456, 124 S.Ct. 906, 157 L.Ed.2d 867 (2004). In my view, this decade of guidance shows why Rule 11(e) and § 3582(c) belong in the claim-processing camp and why there is no lawful impediment to the parties and district court agreeing to withdraw a guilty plea everyone knows to be unlawful. To be sure, the facts of Mr. Spaulding’s case aren’t quite as troubling as those I’ve posited above: it’s pretty apparent the district court in this case agreed to withdraw the guilty plea because of objections to the sentence it yielded, not problems in the plea itself, a result the government doesn’t appeal now and didn’t contest in writing in the district court. But make no mistake, the jurisdictional rule today’s opinion creates from the cloth of Rule 11(e) and § 3582(c) covers the one scenario as well as the other, for the court holds that a district court may never undo a guilty plea after sentencing — no matter what.

To understand the jurisdictional question in this case, you first have to understand 18 U.S.C. § 3231. There and long ago Congress gave the federal district courts “original jurisdiction, exclusive of the courts of the States, of all offenses against the laws of the United States.” Pub.L. No. 80-772, 62 Stat. 683, 826 (1948). (codified at 18 U.S.C. § 3231). This general grant of jurisdiction is unqualified: it does not end after sentencing or any other artificial period of time. Indeed, “[t]he jurisdiction of a court is not exhausted by the rendition of its judgment, but continues until that judgment is satisfied.” Field v. United States, 193 F.2d 86, 90 (2d Cir.1951) (Reed, J.) (quoting Wayman v. Southard, 23 U.S. (10 Wheat.) 1, 23, 6 L.Ed. 253 (1825)).1 Neither can this much come as a surprise: district courts every day exercise post-judgment jurisdiction. In the civil context, they afford relief from *1129a final judgment if someone can demonstrate it was procured by fraud or the like. See Fed.R.Civ.P. 60(b). In the criminal context, they may “vacate any judgment and grant a new trial” up to three years after the verdict for similar reasons — or even many years later if the parties decline to invoke the applicable rule’s (nonju-risdictional) time limits. See Fed. R.Crim.P. 33; Eberhart v. United States, 546 U.S. 12, 19, 126 S.Ct. 403, 163 L.Ed.2d 14 (2005) (per curiam) (holding that Rule 33’s time limits are not jurisdictional). All this is a matter of routine.

Neither, of course, do the rules of procedure that anticipate these post-judgment activities — Rule 60(b) in the civil law and Rule 33 in the criminal law — themselves confer jurisdiction. They are but rules adopted by the Supreme Court pursuant to the Rules Enabling Act, a statute that only allows the Court to “prescribe general rules of practice and procedure” for the district courts. 28 U.S.C. § 2072. Indeed, the Court has expressly recognized that the Rules Enabling Act does not bestow on it the power “to extend or restrict jurisdiction conferred by [another] statute.” Willy v. Coastal Corp., 503 U.S. 131, 135, 112 S.Ct. 1076, 117 L.Ed.2d 280 (1992) (citing Sibbach v. Wilson & Co., 312 U.S. 1, 10, 61 S.Ct. 422, 85 L.Ed. 479 (1941)). So it is that a district court considering motions under Rule 60(b) and Rule 33 must be exercising “some reservoir of jurisdiction-distinct from the rules of criminal [or civil] procedure.” United States v. Smith, 467 F.3d 785, 788 (D.C.Cir.2006). And when it comes to Rule 33 and the criminal law that reservoir is § 3231.

To suggest that the district court lacked jurisdiction to allow Mr. Spaulding’s plea withdrawal, then, my colleagues must find some source of law withdrawing § 3231’s otherwise general grant of jurisdiction. What reservoir Congress fills, it may of course drain away. But in searching for legislative authority that does so much work, we must be careful about mistaking even “important,” “mandatory,” and “emphatic” claim-processing rules for true jurisdictional limitations. United States v. Kwai Fun Wong, — U.S. -, 135 S.Ct. 1625, 1632, 191 L.Ed.2d 533 (2015). The Court has instructed that we should read legislation as' “cabin[ing] a court’s power” — its jurisdiction — “only if Congress has clearly stated as much. Absent such a clear statement, courts should treat the restriction as nonjurisdictional. That does not mean Congress must incant magic words. But traditional tools of statutory construction must plainly show that Congress imbued a procedural bar with jurisdictional consequences.” Id. (citations, alterations, internal quotation marks, and brackets omitted).

As its first candidate for a jurisdictional barrier to Mr. Spaulding’s plea withdrawal in this case, my colleagues point to Federal Rule of Criminal Procedure 11(e). The terms of that rule are clear enough: “After the court imposes sentence, the defendant may not withdraw a plea of guilty or nolo contendere, and the plea may be set aside only on direct appeal or collateral attack.” No one doubts that the rule is important or mandatory, and I don’t question that the district court must observe the rule when it’s invoked. Indeed, most guilty plea challenges arise on direct appeal or collateral review precisely because the rule is invoked or everyone is sure it will be. But just like Rule 60(b) and Rule 33, Rule 11(e) is a waivable claim-processing rule. Just like them it was promulgated pursuant to the Rules Enabling Act. And just like them it cannot withdraw the post-judgment jurisdiction of the district court conferred by another statute, here § 3231. See Willy, 503 U.S. at 135, 112 S.Ct. 1076.

*1130Of course, Congress sometimes chooses to adopt independent statutes that imbue an otherwise waivable claim-processing rule with jurisdictional force. So, for example, this court has recognized that a time limit in Rule 8002 of the Federal Rules of Bankruptcy Procedure is jurisdictional because a federal statute explicitly directs that certain appeals “be taken ... in the time provided by Rule 8002 of the Bankruptcy Rules.” 28 U.S.C. § 158(c)(2); see In re Latture, 605 F.3d 830, 836-37 (10th Cir.2010). But I can find no independent statute endowing Rule 11(e) with jurisdictional force — and my colleagues point to none.2

That leaves my colleagues’ second candidate for a possible jurisdictional barrier: 18 U.S.C. § 3582(c). Under that statute, a district court “may not modify a term of imprisonment once it has been imposed,” except in limited circumstances not applicable to Mr. Spaulding’s case. Here again I do not doubt the statutory command is important and mandatory and must be observed when asserted. But using traditional tools of statutory interpretation, it’s equally clear to me that § 3582(c) doesn’t strip the district court of any of its preexisting post-judgment jurisdiction and is instead and again a claim-processing rule.

The- trouble starts in the statute’s text. As its object § 3582(c) speaks only to the treatment of “sentences,” and even then only to the treatment of a certain class of sentences: those involving a “term of imprisonment.” Nowhere do the nouns “plea” or “guilty plea” appear. Yet we know Congress well understands the difference between sentences and pleas for it distinguishes between the two all the time. See, e.g., 18 U.S.C. § 3593(a) (discussing a “plea of guilty” in cases involving a potential “sentence of death” (emphasis added)); id. § 3771(d)(5) (listing circumstances in which a victim may move “to reopen a plea or sentence” (emphasis added)). Then there’s trouble with the verb too. The statute forbids district courts from “modify[ing]” sentences. And often enough that verb denotes a modest change or alteration, not a complete withdrawal or about-face. See, e.g., MCI Telecomm. Corp. v. Am. Tel. & Tel. Co., 512 U.S. 218, 225, 114 S.Ct. 2223, 129 L.Ed.2d 182 (1994); 9 Oxford English Dictionary 952 (2d ed.1991) (defining “modify” as “[t]o alter in the direction of moderation or lenity” or “[t]o make partial changes in; to change (an object) in respect of some of its qualities”). Now, I readily acknowledge a plausible counter-argument here. If § 3582(c) prevents courts from “modifying] a term of *1131imprisonment,” you could argue that a court shouldn’t be able to vacate a plea on which that sentence depends. In holding that § 3582(c) applies to Mr. Spaulding’s case, my colleagues at least implicitly adopt this view. But the plain terms of the statute do not include a clear statement stripping the district court of jurisdiction to accept a motion seeking to withdraw a guilty plea. And a clear statement is what we need. Kwai Fun Wong, 135 S.Ct. at 1632.

Neither is that the end of it, for the trouble that started with the text grows exuberantly when you recall Rule 33. In statutory interpretation it’s always our hope to fit the relevant statutes and rules together “into a harmonious whole” and render none nugatory. FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 133, 120 S.Ct. 1291, 146 L.Ed.2d 121 (2000). But the jurisdictional gloss my colleagues give § 3582(c) all but leaves Rule 33 for dead. It does because, if § 3582(c) really stretched beyond preventing sentence modifications to preclude district courts from taking any action with the knock-on effect of undoing a sentence, then district courts could never grant a Rule 33 motion undoing a criminal judgment if the defendant’s sentence included a term of imprisonment. And in federal court almost every criminal trial resulting in a conviction yields a term of imprisonment. See, e.g., United States Sentencing Commission, Statistical Information Packet, Fiscal Year 2011, Tenth Circuit tbl. 4 (around 90% of convicted federal defendants receive prison time). Coming at the point from the other direction, if § 3231’s grant of jurisdiction didn’t live on beyond a final judgment a district court decision granting a post-trial motion to vacate a judgment under Rule 33 would be almost always ultra vires. Far from creating a harmonious whole, reading § 3582(c) as jurisdictional introduces a seriously discordant note into our law by suggesting Rule 33 just doesn’t mean what it says.3

Broadening our view beyond subsection (c) and Rule 33 to take in § 3582’s other subsections further supports the conclusion that the statute isn’t jurisdictional. The section as a whole answers questions related to the claim-processing mechanics of imposing a sentence of imprisonment, not the jurisdiction of the court. Subsection (a), for example, sets forth the factors a court must consider when deciding whether to sentence a defendant to prison, while subsection (d) allows a judge to bar a defendant sentenced to prison for certain racketeering and drug charges from contacting criminal confederates. It’s hard to imagine that — tossed in among these provisions governing the finer points of sentencing- — -we would find one single subsection depriving the district court of jurisdiction.

An even broader view still — of Title 18 as a whole — continues to point in the same direction. Chapter 211 of Title 18 is helpfully entitled “Jurisdiction and Venue.” *1132That’s where you’ll find § 3231, the general grant of criminal jurisdiction discussed above. Meanwhile, § 3582 lives several doors down in Chapter 227, a chapter entitled “Sentences” (in Subchapter D, entitled “Imprisonment”). Neither was this arrangement manufactured by some bureaucratic commission compiling the U.S. Code: Congress itself enacted it. See Pub.L. No. 98-73, 98 Stat. 1837, 1998 (1984); Pub.L. No. 80-772, 62 Stat. 683, 826 (1948). And the Supreme Court has previously “found Congress’s decision to locate a statute ‘separate’ from jurisdictional provisions suggestive contextual evidence that the statute in question was non-jurisdictional.” Hobby Lobby Stores, Inc. v. Sebelius, 723 F.3d 1114, 1158 (10th Cir.2013) (Gorsuch, J., concurring), aff'd sub nom. Burwell v. Hobby Lobby Stores, Inc., - U.S. -, 134 S.Ct. 2751, 189 L.Ed.2d 675 (2014). Section 3582’s location in Title 18 suggests just that.

Yet one more structural point is worth taking in. Reading the Federal Rules and Title 18 together, we can see a comprehensive scheme regulating pleas, sentences, and judgments. Rule 11 governs pleas. Chapter 227(C) of Title 18 governs the imposition of fines. And Chapter 227(D), where § 3582 is found, governs sentences of imprisonment. Here again it seems unlikely that Congress would elevate only the very last little portion of this regime to the jurisdictional plane, leaving all others as claim-processing rules. And putting this interpretive clue together with all the rest, I just cannot see how I might fairly say Congress “plainly” endowed § 3582(c)’s procedural bar with jurisdictional consequences.4

Of course, precedent sometimes compels us to reach results that seem doubtful on first impression. But looking to the case law in this area I see nothing that commands us to treat Rule 11 or § 3582(c) as jurisdictional limits on the district court’s authority. Indeed, at least one other circuit has expressly held that § 3582(c) is not jurisdictional. See United States v. Taylor, 778 F.3d 667, 669-71 (7th Cir.2015). Another court, although reserving judgment on the issue, has similarly suggested that recent Supreme Court precedent “calls into question a jurisdictional reading of § 3582.” United States v. Smith, 467 F.3d 785, 788 (D.C.Cir.2006). True, and as the Seventh Circuit’s opinion in Taylor points out, some other courts have spoken of the statute in jurisdictional terms. See 778 F.3d at 670 (collecting cases). But these cases do not confront all of the relevant textual evidence I’ve noted here and no precedent in this circuit obligates this'court today to join their side of the circuit split. In response, my col-' leagues do claim some measure of support from United States v. McGaughy, 670 F.3d 1149 (10th Cir.2012). Yet that decision held jurisdictional only Rule 35(a)’s fourteen-day deadline, and no one before us claims that rule applies to this case. In the process of reaching its result, it’s true that *1133McGaughy did speak in passing about § 3582(c) and seemed to proceed on the premise that it’s jurisdictional, but not even my colleagues suggest it rendered a binding holding that the statute (or Rule 11(e)) qualifies as jurisdictional.5

After addressing text and precedent, my colleagues conclude their opinion with an appeal to policy. They contend that affording Rule 11(e) and § 3582(c) a jurisdictional reading and denying the ability of a district court to undo a guilty plea, even on the parties’ stipulated motion, “comports with the heightened interest in finality that attaches to criminal judgments founded upon guilty pleas.” Ante at 1126. An interest that stems, my colleagues seem to suggest, from the fact that plea bargaining and guilty pleas are both reliable and widespread. Id. (noting, among other things, that guilty pleas “rest, after all, on a defendant’s profession of guilt” (quoting United States v. Dominguez Benitez, 542 U.S. 74, 82-83, 124 S.Ct. 2333, 159 L.Ed.2d 157 (2004))).

This policy appeal too seems to me a doubtful substitute for a clear statutory statement withdrawing the district court’s longstanding post-judgment jurisdiction. In the first place, statutes and rules are the product, of many competing interests and compromised objectives and the best guide to the “policy” they seek to vindicate is their terms and structure, not our supplemental conjuring. If Congress wants to withdraw district court jurisdiction in this or any other area of course it may, but it is not our business to substitute its judgment for ours based on our own views of optimal policy. See, e.g., In re Dawes, 652 F.3d 1236, 1243 (10th Cir.2011) (“[Departing from a precise statutory text may do no more than disturb a carefully wrought legislative compromise.” (quoting John F. Manning, Textualism and the Equity of the Statute, 101 Colum. L.Rev. 1, 18 (2001))).

Neither am I persuaded by my colleagues’ claimed policy interest even taken on its own terms. My colleagues contend for a heightened finality interest in guilty pleas in part because of their supposed reliability. But by now it’s surely well known that structural features of the plea bargaining process can lead innocent defendants to take the same deals as guilty ones. And I can think of no reason to suggest a heightened confidence in plea bargaining when compared to its alternative: the trial. .It is at trial, after all, where the prosecution’s case is put to the test, the defendant is guaranteed so many procedural protections, and a jury must unanimously decide. When it comes to the confidence we bear in a conviction and the repose to which it is entitled, I would have thought -the trial guaranteed by the Constitution qualifies as at least the equal of the plea bargain that has evolved in its shadow. See Robert E. Scott & William J. Stuntz, Plea Bargaining As Contract, 101 Yale L.J. 1909, 1911 (1992); Stephanos Bibas, Plea Bargaining in the Shadow of Trial, 117 Harv. L.Rev. 2463 (2004); Wil*1134liam J. Stuntz, Plea Bargaining and Criminal Law’s Disappearing Shadow, 117 Harv. L.Rev. 2548 (2004).

Finally and relatedly, the majority’s worry that my reading of Rule 11(e) and § 3582(c) would “eliminate[] the concept of finality with regard to guilty pleas” seems to me much overblown. Ante at 1126. Under my reading of the relevant rules and statutes, the government need merely assert its rights under Rule 11(e) and the guilty plea will stand and finality will be preserved. Even if the government fails to assert a Rule 11(e) objection, a district court might still be able to enforce the rule on its own motion if it disagrees with the parties’ arguments. See United States v. Mitchell, 518 F.3d 740, 750 (10th Cir.2008) (suggesting that courts may raise certain claim-processing rules sua sponte). It is only in the (surely exceedingly) few cases where no one disputes the plea’s illegality that a plea may be undone without engaging the great grinding gears of an appeal or collateral lawsuit. Nothing about this arrangement should come as a surprise either for, after all, “[o]urs is an adversarial system of justice” in which the parties are usually “presum[ed] ... responsible for raising their own defenses.” Id. at 749. Congress may of course override this presumption anytime it wishes and elevate a claim-processing defense into a jurisdictional limitation. But that remains its job to do, not ours.

I respectfully dissent.

. The majority notes that "the power of a court over its judgments at common law expired with the term of Court.” Ante at 1125 (quoting United States v. Smith, 331 U.S. 469, 473, 67 S.Ct 1330, 91 L.Ed. 1610 (1947)). Bui Congress abolished formal terms for district courts over half a century ago. Pub.L. No. 88-139, 77 Stat. 248 (1963).

. My colleagues suggest that the Supreme Court’s "insistence that only those procedural rules implementing a statutory directive can be jurisdictional” might be disregarded because it is inconsistent with some of the Court's own precedents. Ante at 1120 n. 9. To prove their point, they cite Bank of Nova Scotia v. United States, 487 U.S. 250, 108 S.Ct. 2369, 101 L.Ed.2d 228 (1988). According to my colleagues, Nova Scotia determined that "Federal Rule of Criminal Procedure 52(a) obligates a federal court to disregard errors that do 'not affect substantial rights,' even if the government does not insist on the application of such a rule." Ante at 1120 n. 9 (emphasis added). But, respectfully, Nova Scotia says no such thing: it was not answering, and did not purport to answer, whether Rule 52 is a jurisdictional limitation on the district court’s authority that a court of appeals must observe sua sponte when no one's invoked it at any point in the proceedings in that court or this one. Instead, Nova Scotia held simply that Rule 52 is "as binding as any statute duly enacted by Congress, and federal courts have no more discretion to disregard the Rule's mandate than they do to disregard constitutional or statutory provisions.” 487 U.S. at 255, 108 S.Ct. 2369. If to say as much is to say Rule 52 is jurisdictional, then it is to say every statute is jurisdictional. And we know that’s not the case. See Kontrick, 540 U.S. at 455, 124 S.Ct. 906.

. The majority responds to this harmony point with one of its own, claiming a nonju-risdictional reading of Rule 11(e) undermines AEDPA’s procedural hurdles for "second or successive” motions for collateral relief. See ante at 1122 (discussing 28 U.S.C. § 2255(h)). But Mr. Spaulding moved to withdraw his plea the day after sentencing and before ever filing a § 2255 motion, so no one could plausibly argue that his motion was a second or successive anything. Despite my colleagues’ suggestion, then, this case simply does not implicate our practice of construing motions that follow an earlier collateral attack as arising under § 2255. Indeed, we have expressly warned district courts against following that practice when it comes to a first request for post-judgment relief. See United States v. Torres, 282 F.3d 1241, 1245 (10th Cir.2002). I am thus unable to see the majority’s perceived tension — and without that tension, there’s no need to adopt its creative jurisdictional solution.

. As my colleagues point out, § 3582(c) does direct its command to the court, not the parties: “The court may not modify a term of imprisonment once it has been imposed....” (emphasis added). And my colleagues are of course right that whether a statute's instructions are aimed at the court or the parties is a relevant consideration when weighing whether it has jurisdictional effect. See Hobby Lobby, 723 F.3d at 1157-58 (Gorsuch, J., concurring). But neither is this clue dispositive. In United States v. Dolan, 571 F.3d 1022 (10th Cir.2009), aff'd, 560 U.S. 605, 130 S.Ct. 2533, 177 L.Ed.2d 108 (2010), for example, this court and the Supreme Court recognized that a victims' restitution statute addressing its provisions to the court was not jurisdictional in light of the remaining interpretive evidence. The same is true here: while the statute undoubtedly addresses the courts, not the parties, too many other textual, structural, and contextual clues conspire to foreclose a conclusion that the statute is "plainly” jurisdictional.

. In Taylor, the Seventh Circuit suggested that this court's decision in United States v. Graham, 704 F.3d 1275 (10th Cir.2013), held § 3582(c) jurisdictional. But my colleagues mention Graham only in passing and for good reason. That ruling did not consider whether § 3582(c) withdraws § 3231's general grant of jurisdiction, instead asking whether that statute confers some power on the district court. To be sure, Graham suggested that "[f]ederal courts generally lack jurisdiction .to modify a term of imprisonment once it has been imposed.” 704 F.3d at 1277. But as my colleagues seem to agree, that wording is best understood not to implicate the court’s “jurisdiction” in the true and precise sense, as explained by Kontrick. After all, the case on which Graham relies for the proposition doesn’t discuss jurisdiction at all. See Dillon v. United States, 560 U.S. 817, 130 S.Ct. 2683, 177 L.Ed.2d 271 (2010).