Withdrawal and Finality

Withdrawal and Finality

The rules taught so far in this chapter govern how a plea comes into being: the arraignment, the colloquy, the factual basis, the plea agreement, and the defendant's presence. One question remains once all of that has happened, which is whether the defendant can take the plea back. Rule 11(d) answers by asking how far the case has progressed, and it gives a different answer before and after the court accepts the plea.¹ Sentencing marks a further line, drawn by Rule 11(e), so the defendant's freedom to withdraw narrows at two events: the court's acceptance of the plea and the court's imposition of sentence.

Before acceptance, a defendant may withdraw a guilty or nolo contendere plea for any reason or no reason.² That generous rule fits the stage, because until the court accepts the plea nothing has been adjudicated and no judgment exists for a withdrawal to disturb. After acceptance but before sentence, the defendant must make a showing under Rule 11(d)(2), and a change of mind alone no longer suffices.³ The tightening has an incentive explanation. If withdrawal stayed free all the way to sentencing, a defendant could plead, learn from the presentence report or the judge's comments what sentence to expect, and retreat whenever the outlook worsened. The colloquy and the court's acceptance would then amount to a rehearsal. Requiring a showing after acceptance gives the court's decision real weight, because the defendant can undo it only with a reason the court is willing to credit.

Sentencing closes the door. Once the court imposes sentence, Rule 11(e) bars withdrawal altogether, and the plea can be set aside only through direct appeal or collateral attack. A direct appeal is review by a higher court within the same case, while a collateral attack is a separate proceeding, such as a habeas corpus petition or a motion to vacate the sentence, that challenges the judgment from outside the original case. The plea decisions taught earlier in this chapter reached the Supreme Court through those two channels, which is why the final stage is best learned alongside them. The system treats a sentenced judgment as final except for these formal avenues, so the question after sentence shifts from whether the defendant still wants the plea to whether the plea was constitutionally or procedurally defective. Students lose this issue when they treat a motion to withdraw after sentence as still available, since the only routes left are appeal and collateral attack. The first sub-section below takes up the two stages before sentence, and the second takes up finality once sentence is imposed.

Withdrawing Before Sentence

Rule 11(d) governs when a defendant may take back a guilty or nolo contendere plea, and it divides the time before sentencing at a single event, the court's acceptance of the plea.⁴ The line falls there because acceptance is when the plea turns from a proposal awaiting the court's approval into the foundation for a conviction. Before that event the defendant's freedom to retreat is complete, and after it the freedom depends on a showing. The two windows are taught below in order, followed by the question of what use a withdrawn plea can later be put to.

Before the court accepts the plea, the defendant may withdraw it for any reason or no reason.⁵ Until acceptance there is no conviction, nothing has been adjudicated, and neither the court nor the government has acted on the plea, so a withdrawal disturbs nothing that anyone has relied on. The freedom also gives the Rule 11(b) colloquy practical force. A defendant who hears the court explain the trial rights being surrendered and the penalties that follow may decide in the middle of the hearing that the plea was a mistake, and the rule lets that decision stand without any explanation to the judge. The court therefore does not ask why the defendant has changed course, because at this stage no answer could change the result.

Once the court accepts the plea, the defendant may still withdraw before sentence, but only if the court rejects a plea agreement under Rule 11(c)(5) or the defendant can show a fair and just reason for requesting the withdrawal.⁶ The first situation follows from the structure of a bargain: the defendant pleaded in exchange for the agreement, so when the court declines to honor the agreement, the basis for the plea has disappeared. The second is the general standard, and it places the burden on the defendant. That allocation reflects what acceptance changed. The court has used the colloquy to find the plea voluntary, informed, and supported by a factual basis, and a request to undo a plea that has passed that scrutiny needs a reason weighty enough to outweigh it.

The rule does not define a fair and just reason, so its content comes from decisions about pleas that were obtained unfairly. In Santobello v. New York the prosecution made a commitment concerning the sentence recommendation on a guilty plea and then failed to keep it, and the Court stated the principle that when a plea rests in any significant degree on a prosecutor's promise, so that the promise is part of the inducement, the promise must be fulfilled.⁷ The opinions differed on the remedy, and that difference shows what a fair and just reason looks like. Justice Marshall, joined by two other Justices, would have let the defendant withdraw the plea, reasoning that when a defendant offers a reason for vacating a plea before sentence and the government has not relied on the plea to its disadvantage, the right to trial may be regained, and that a prosecutor who breaks the bargain undercuts the basis for the waiver of trial rights implicit in the plea.⁸ Justice Douglas, concurring, would have left the state court to choose between specific performance of the bargain and an option to go to trial on the original charges, giving the defendant's preference considerable weight.⁹ He also recalled that, since Kercheval v. United States, a court on timely application will vacate a plea "shown to have been unfairly obtained or given through ignorance, fear or inadvertence".¹⁰

Santobello came from a state conviction, so Rule 11 did not govern it, as Justice Douglas observed.¹¹ The opinions therefore do not interpret Rule 11(d)(2), but they show the kind of reason a court is likely to credit: a plea induced by a broken bargain, or by ignorance, fear, or inadvertence, as opposed to a simple change of mind. Justice Marshall's observation that a motion made before sentencing ordinarily meets a government that has taken no action in reliance on the plea also explains why this window is more forgiving than the period after sentence.¹²

The first two defendants differ only in whether the court had accepted the plea, and that single fact moves the answer from a right to withdraw to a request that needs support.¹³ The third differs from the second only in offering a reason, and the reason supplies what the fair-and-just standard demands. Students lose this issue when they carry the pre-acceptance freedom forward and treat a bare change of heart after acceptance as enough, since the any-reason rule ends when the court accepts the plea.¹⁴

Withdrawing a plea raises a second question, which is whether the withdrawn plea and the discussions around it can be used against the defendant later. Rule 11(f) does not answer that question inside Rule 11; it sends the admissibility of a plea, a plea discussion, and any related statement to Federal Rule of Evidence 410.¹⁵ The cross-reference matters because a right to withdraw means little if the withdrawn plea could be offered as an admission at trial, and the Rule 20(c) provision on a transfer statement, taught earlier, reflects the same concern for a defendant who changes course.

Finality After Sentence

Rule 11(e) marks the last stage in the life of a plea. 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.²⁴ The rule shuts one route and leaves two open, so a post-sentence problem starts with sorting the defendant's complaint into the route where it belongs. The reason for closing the first route is practical. Sentencing converts the plea into a judgment on which the court, the government, and the public have relied, and a defendant who could reopen that judgment by motion could undo a bargain after the other side had already performed its part. Finality also serves defendants as a group. A dissent in Blackledge v. Perry observed that most criminal cases end in pleas, so defendants share the prosecution's interest in pleas that stay settled, since pleas that are easily upset give prosecutors less reason to bargain.²⁵ The two remaining routes keep review available for the defects that matter most, such as a plea that was not voluntary or not intelligent.

Direct appeal is the route for a defect visible in the plea record itself. In Boykin v. Alabama the Supreme Court reversed on direct review, agreeing with three dissenting justices of the Alabama Supreme Court that the record did not disclose a plea entered voluntarily and with understanding.²⁶ That is the pattern of a direct-appeal claim: the reviewing court reads the transcript of the plea hearing and asks whether the required inquiry took place. The plea does not by itself close every direct-appeal argument, however. In Class v. United States the Court held that a guilty plea alone does not waive the right to challenge, on direct appeal, the constitutionality of the statute of conviction.²⁷ One court of appeals has likewise refused to enforce an appeal waiver attached to a plea when the defendant received no reduced charge or sentence in return.²⁸

Collateral attack is the route for a claim that depends on facts outside the transcript, or that arises after direct review has run its course. In North Carolina v. Alford the defendant first sought post-conviction relief in state court, where a hearing produced a finding that the plea was willing, knowing, and understanding and made on the advice of competent counsel, and he then petitioned for federal habeas corpus twice, with the Fourth Circuit eventually ruling the plea involuntary.²⁹ The Supreme Court vacated that ruling and restated the governing standard: a plea is valid if it represents a voluntary and intelligent choice among the alternative courses of action open to the defendant.³⁰ Hill v. Lockhart shows the same route on a narrower claim. More than two years after pleading guilty, Hill filed a federal habeas petition alleging that his plea was involuntary because his lawyer had misinformed him about when he would become eligible for parole.³¹ Where counsel advised the defendant, the voluntariness of the plea depends on whether counsel's advice fell within the range of competence demanded of attorneys in criminal cases.³² Hill still lost, because he failed to allege the kind of prejudice from the allegedly incompetent advice that would have entitled him to a hearing.³³

The contrast between these decisions shows why the routes differ. The claim in Boykin could be decided from the transcript, since the question was whether the judge had inquired into the plea at all. The claims in Alford and Hill turned on events away from the bench, namely fear of the death penalty and what counsel said in private about parole, so they moved through post-conviction proceedings and habeas petitions, where a court can take evidence.

The two defendants differ only in the route chosen, and that single choice moves the result from barred to reviewable.³⁴

Collateral attack after a plea is narrower than collateral attack after a trial. When a defendant pleads guilty voluntarily and intelligently on the advice of competent counsel, the plea generally forecloses later challenges to constitutional errors that came before it, even a coerced confession in McMann v. Richardson or racial discrimination in grand jury selection in Tollett v. Henderson.³⁵ The mechanism is that a plea admits factual guilt, which makes irrelevant those violations that do not contradict a valid establishment of factual guilt.³⁶ Tollett described the plea as breaking the chain of events that preceded it, which leaves the defendant free to attack only the voluntary and intelligent character of the plea.³⁷

One class of claim survives that bar. In Blackledge v. Perry the State had proceeded against the defendant on a misdemeanor charge, and the Court held that due process then barred it from requiring him to answer a more serious felony charge, so his guilty plea to the felony did not stop him from raising the claim on federal habeas.³⁸ What separates Blackledge from McMann and Tollett is whether the defect could be repaired. A tainted indictment could be cured by a new indictment from a properly selected grand jury, and a coerced confession could be left out of a new prosecution, but nothing could cure a charge the State had no power to bring.³⁹ The same reasoning supports collateral challenges based on double jeopardy in Menna v. New York and on the unconstitutionality of the statute of conviction in Class, but only where the State's lack of authority is apparent on the face of the record, which is why United States v. Broce refused to reopen a plea to two counts that facially described separate offenses.⁴⁰ Students lose this issue when they assume that a claim about what happened before the plea survives the plea merely because the claim is constitutional, since McMann and Tollett involved constitutional errors and still failed.⁴¹

A successful attack raises a further question, which is what the defendant risks by winning. In North Carolina v. Pearce, Rice had pleaded guilty to four burglary charges and received prison terms totaling ten years, but his judgments were set aside because he had no lawyer when he pleaded, and after retrial on three charges he was sentenced to 25 years with no credit for time already served.⁴² The Court held that punishment already exacted must be credited in imposing a new sentence for the same offense, and that a harsher sentence imposed on retrial to punish the defendant for exercising the right to appeal violates due process.⁴³ Bordenkircher v. Hayes later explained that the violation in Pearce lay in the danger that the State would retaliate against a defendant for lawfully attacking a conviction, and that the give-and-take of plea bargaining contains no comparable retaliation so long as the accused is free to accept or reject the offer.⁴⁴ The protection means a defendant can pursue the routes Rule 11(e) leaves open without risking a harsher sentence for succeeding.

Every later reviewer begins with the record made at the plea hearing. The appellate court reads it for the colloquy, and the habeas court reads it to decide whether the plea was voluntary and intelligent. For that reason Boykin demanded the utmost solicitude in canvassing the matter with the accused: a judge who does so leaves a record adequate for any review that may later be sought and forestalls collateral proceedings that seek to probe murky memories.⁴⁵ The colloquy therefore protects the defendant, whose plea is shown to rest on understanding, and it protects the court, whose judgment can be defended years afterward without relying on anyone's recollection of what was said.