Concept
Plea Agreements
Guilty pleas
Plea Agreements
Most criminal cases end in a bargain between prosecutor and accused, not in a verdict. In Brady v. United States the Court observed that well over three-fourths of the criminal convictions in this country rest on guilty pleas, a great many of them motivated at least in part by the hope or assurance of a lesser penalty than a trial might bring.1 Justice Harlan's dissent in Duncan v. Louisiana relied on estimates by two experts that about 75 percent of prosecutions for crimes triable to a jury are settled by guilty plea.2 Santobello v. New York drew the practical conclusion. Disposing of charges by agreement between the prosecutor and the accused is an essential component of the administration of justice, because if every charge were tried in full, the States and the Federal Government would need to multiply their judges and court facilities many times over.3
The constitutional justification for the practice is a mutuality of advantage. A defendant who sees slight chance of acquittal gains a reduced exposure, an immediate start to the correctional process, and relief from the burdens of trial, while the State gains a prompter punishment and saves scarce judicial and prosecutorial resources for cases in which guilt is genuinely in doubt.4 Because each side gives up something and receives something, the Court could not hold it unconstitutional for the State to extend a benefit to a defendant who extends a substantial benefit in return.5 The same opinion cautions that the prevalence of pleas, though explainable, does not by itself validate them or the system that produces them.6 The bargain therefore enters the validity analysis from the previous sections as one more source of inducement, and the question becomes which inducements leave the plea voluntary and intelligent.
Santobello answers by attaching a condition to its endorsement of plea bargaining: the benefits of the practice presuppose fairness in securing the agreement between the accused and the prosecutor.7 The plea must be voluntary and knowing, and if promises induced it, the essence of those promises must in some way be made known.8 Once the promise exists, the prosecutor is bound by it, since a promise that rests a plea in any significant degree on the prosecutor's word, so that it forms part of the inducement or consideration, must be fulfilled.9 The case itself arose from the State's failure to keep a commitment concerning the sentence recommendation on a guilty plea.10 Brady supplies the matching test for validity. A plea entered by a defendant fully aware of the direct consequences, including the actual value of any commitments made by the court, the prosecutor, or the defendant's own counsel, must stand unless it was induced by threats, by misrepresentation (which includes unfulfilled or unfulfillable promises), or by promises improper by their nature, such as bribes.11 That standard has two halves. The first asks what produced the plea and so speaks to voluntariness. The second asks what the defendant understood about the deal's real worth and so speaks to intelligence.
Rule 11(c) turns this concern for fairness into procedure.12 The government's attorney and the defendant's attorney, or the defendant when proceeding without a lawyer, may discuss and reach an agreement, but the court must not participate in those discussions.13 Excluding the judge removes a structural temptation. The judge will later decide whether to accept the plea and will impose the sentence, so a judge who helped negotiate would be pressing one side toward a particular deal while holding power over its outcome, and a defendant would find it hard to refuse an offer from the person who will pronounce the sentence. Keeping the judge out also leaves the court free to review, as an independent decision-maker, a bargain it did not make. Disclosure of the agreement to the court does the complementary work, because a promise written into an agreement the judge can examine is one the record preserves and the colloquy can test, while a promise made privately is the hidden inducement that the voluntariness inquiry exists to expose. The subsections below follow the agreement in the order the rule does: who negotiates and how the agreement reaches the court, which promises bind the court and which do not, and how the court accepts or rejects what the parties present.
The two versions differ in one respect, the character of the promise. In the first, the promise was within the prosecutor's power, appeared on the record, and was kept, so the defendant chose with an accurate picture of the choice. In the second, the promise could not be delivered and was concealed from the court, so the defendant's picture of the choice was false, and the court had no chance to correct it in the colloquy.14
Brady and Santobello address different points in the life of a bargain, and keeping them apart prevents a common error. Brady asks whether the State's role in motivating the plea made it involuntary, and it holds that pleas are valid despite the State's responsibility for some of the factors that prompted them, so long as no actual or threatened physical harm or mental coercion overbore the defendant's will and the defendant acted with awareness of the consequences.15 Santobello begins after a legitimate bargain exists and asks whether the State kept the promise that induced the plea.16 Students lose points by treating the State's benefit from a plea as coercion, when the benefit to both sides is what makes the bargain permissible, and by forgetting that a legitimate inducement binds the prosecutor once the defendant has relied on it.17
Negotiation and Disclosure
Rule 11(c)(1) fixes who sits at the bargaining table. An attorney for the government and the defendant's attorney may discuss and reach a plea agreement, and a defendant who is proceeding pro se, meaning without a lawyer, negotiates in place of defense counsel, while the court must not participate in these discussions.20 As explained above, a judge who bargains would hold both the power to press a defendant and the power to sentence. The procedural consequence is a division of labor: the parties produce the proposal, and the court first meets its terms when the plea is offered, in a position to evaluate a bargain it had no hand in making.
Keeping the court out of the room does not leave the table unprotected. The Sixth Amendment guarantees counsel at the critical stages of a prosecution, and the Court has treated the negotiation of a plea bargain as one of those stages for purposes of the right to effective assistance.21 In Missouri v. Frye, a prosecution offer more lenient than the plea the defendant later entered lapsed after counsel's alleged ineffectiveness, and the Court held that the right to counsel reaches the negotiation and consideration of offers that expire or are rejected, not only the plea finally entered.22 The lawyer's role in the negotiation is therefore a constitutional one, and it is the lawyer, or the pro se defendant, who carries the exchange with the prosecutor.
What the parties may say across the table is broad, because plea bargaining is a real exchange in which both sides apply pressure. In Bordenkircher v. Hayes the prosecutor told the defendant at the outset of the negotiations that he would seek a recidivist indictment if the defendant would not plead guilty, and the Court held that following through on an offer stated that openly did not violate due process, since the defendant knew the true terms of the choice when he decided to plead not guilty.23 In United States v. Mezzanatto a defendant agreed, as the price of entering plea discussions, that his statements could be used to impeach him at trial, and the Court enforced that waiver because nothing indicated it was made unknowingly or involuntarily.24 Even sentencing facts are on the table: defendants can either stipulate to the facts that raise a sentence or contest them, so bargaining over those facts already goes on.25 The permissive reading of what may be said explains why disclosure carries so much weight. Brady's validity standard asks whether the defendant understood the actual value of every commitment made by the prosecutor or counsel, so each commitment that moved the defendant has to reach the judge who tests the plea.26
Rule 11(c)(2) supplies that mechanism. The parties must disclose the plea agreement in open court when the plea is offered, unless the court for good cause allows them to disclose it in camera, which means privately to the judge rather than in public session.27 The timing and the default each do work. Because disclosure comes when the plea is offered, the judge has every term before deciding whether to accept the plea, and because the default is open court, the terms enter the record that a reviewing court will later read. A promise that appears in that record can be tested in the colloquy, while a side promise that never reaches the judge is an inducement the voluntariness inquiry has no chance to examine.
The exception is narrow by design. In camera disclosure still gives the court the entire agreement, so the only thing it changes is who else learns the terms, and the rule leaves the decision to the court rather than to the parties. A bare preference for privacy gives the court no cause to find, and an agreement the parties simply decide to keep off the record has not been disclosed at all.
The two scenarios differ in what the judge did before the plea was offered. In the first, the court's first contact with the bargain was the disclosure in open court, so it could review the terms as a neutral decision-maker. In the second, the court had already signaled a sentence, which makes refusing the deal feel risky and undercuts the free choice the plea must reflect.28 Students lose points by treating the parties' agreement to keep the terms private as enough to justify in camera disclosure, when the rule makes open court the default and requires the court itself to allow the exception for good cause.29
Binding and Non-Binding Terms
Rule 11(c)(1) names three kinds of promise that a government attorney may make in a plea agreement, and the most useful way to sort them is by how far each promise reaches. Two of them concern the sentence, and they sit at opposite poles: one leaves the court entirely free, while the other controls the court once the court accepts the agreement. The third concerns charges. The rule describes the forms this way:
- Type (A): the government will not bring, or will move to dismiss, other charges.32
- Type (B): the government will recommend, or agree not to oppose the defendant's request, that a particular sentence or sentencing range is appropriate or that a particular Guidelines provision, policy statement, or sentencing factor does or does not apply, and such a recommendation or request does not bind the court.33
- Type (C): the government agrees that a specific sentence or sentencing range is the appropriate disposition, or that a particular Guidelines provision, policy statement, or sentencing factor does or does not apply, and such an agreement binds the court once the court accepts the plea agreement.34
The distinction shows most clearly at the two sentencing poles, so the discussion takes (B) and (C) first and then places (A) beside them.
Type (B) is a promise about advocacy. The prosecutor will say, or will refrain from disputing, that a certain sentence or Guidelines position is right, and because the judge alone decides the sentence, a promise by one party to the case cannot hand that decision to that party. What the defendant obtains is therefore a favorable argument from the government, worth only as much as the judge finds it persuasive. A limited promise is still a promise, though. In Santobello v. New York the commitment at issue concerned the sentence recommendation, and the Court held that when a plea rests in any significant degree on a prosecutor's promise, so that the promise is part of the inducement or consideration, the promise must be fulfilled.35 Fulfilling a (B) promise means doing what was promised, making the recommendation or abstaining from opposition as agreed. A judge who sentences above the recommendation has therefore caused no breach, while a prosecutor who fails to make the recommendation has.
Type (C) works differently because its force reaches the court. The government agrees that a specific sentence or range is the appropriate disposition, and the rule provides that this agreement binds the court once the court accepts the plea agreement.36 The timing is part of the rule. Until the court accepts, the agreement is a proposal, and the judge, who was kept out of the negotiations, remains an independent decision-maker free to review a bargain the judge did not make. Acceptance is the moment at which that independent judgment is exercised, and after it the agreed term controls. The structure gives the defendant something no other type provides, an assurance about the sentence itself, and it places the court's check at the point where the court has the whole bargain before it.
Type (A) concerns charges rather than sentences. The government agrees to not bring other charges, or to move to dismiss them, so the promise is about the prosecutor's own conduct in shaping the case.37 Charge bargaining of this kind has constitutional footing, because the Court observed that holding plea benefits unconstitutional would require prosecutors and judges to refuse guilty pleas to selected counts, to lesser included offenses, or to reduced charges, and that the Fifth Amendment does not reach so far.38 Types (A) and (C) also share a procedural feature: they are the two types for which the court may accept the agreement, reject it, or defer its decision until it has reviewed the presentence report, a point the next subsection develops.
The three types catalogue what the government may promise, but an agreement also records what the defendant gives. In United States v. Hubbell the defendant pleaded guilty to mail fraud and tax evasion, and in the plea agreement he promised to provide the Independent Counsel with “full, complete, accurate, and truthful information” about matters relating to the Whitewater investigation.39 A cooperation term of that kind belongs on the defendant's side of the exchange and is not one of the Rule 11(c)(1) types. It matters here because the value of the whole bargain includes what each side commits to, and a defendant choosing intelligently has to understand both halves.
The reason the type matters to validity comes from Brady v. United States. A plea entered by a defendant fully aware of the direct consequences, including the actual value of any commitments made by the court, the prosecutor, or the defendant's own counsel, must stand unless it was induced by threats, by misrepresentation (which includes unfulfilled or unfulfillable promises), or by promises improper by their nature.40 The actual value of a sentencing promise depends on its type. A (B) recommendation is a hope, because the judge may sentence higher, and a (C) term that the court has accepted is a guarantee, so a defendant who treats the first as the second has misjudged what the bargain is worth. If counsel or the prosecutor encourages that belief, the expected result is one that no one can deliver, which is the kind of misrepresentation, including unfulfillable promises, that the standard treats as a defect. Requiring the agreement to name its type lets the court explain accurately what follows if it accepts the agreement or declines to follow a recommendation, and it keeps the defendant's choice informed.
The two versions share the same defendant, the same count, and the same five-year figure. What differs is the verb: in the first the government promises to recommend the figure, and in the second it agrees that the figure is the appropriate disposition. That single difference decides whether the number is advice the court may decline or a term the court must follow once it accepts the agreement.41
On an exam stem that recites a promised sentence, the first task is to identify which promise the government actually made, and the answer sits in the verb attached to the sentence figure: recommending or not opposing is type (B), while agreeing that the figure is the appropriate disposition is type (C).42 A second error treats a type (C) term as binding the court when the parties sign, although the rule attaches the binding effect to the court's acceptance of the plea agreement.43 A charge-dismissal promise under type (A) carries no sentence figure at all, so it cannot be read as a promise about the sentence.44
Accepting or Rejecting Agreements
Disclosure under Rule 11(c)(2) puts the entire agreement before a judge who took no part in the bargaining, and the rest of Rule 11(c) says what that judge may do with it. For an agreement of the type in Rule 11(c)(1)(A) or (C), the court has three choices: it may accept the agreement, reject it, or defer a decision until it has reviewed the presentence report.47 The rule ties these choices to those two types, and the structure explains why. A charge dismissal under type (A) changes what the defendant will be sentenced for, and a type (C) term fixes the sentence itself once the court accepts. A type (B) recommendation never binds the court, so it leaves nothing for the court to accept or refuse.
The power to refuse comes from the judge's independent role in sentencing. In Santobello v. New York the Court observed that there is no absolute right to have a guilty plea accepted and that a court may reject a plea in the exercise of sound judicial discretion.48 Missouri v. Frye makes the same point from the defendant's side: a defendant has no right to be offered a plea and no federal right that the judge accept it, and Rule 11(c)(3) gives trial courts leeway to accept or reject agreements.49 The reason follows from who decides the sentence. A type (C) agreement binds the court once accepted, so acceptance is the one moment at which the court can ask whether the agreed term fits the offense and the offender. Without that check, two parties could settle a sentence that neither the judge nor anyone else had examined.
Deferral is the middle option, and it exists because a judge often lacks the information needed to judge a specific sentence on the day the plea is offered. The presentence report gathers information about the defendant and the offense that the parties may not have put before the court. By deferring, the court neither commits itself nor refuses: the agreement stays a proposal, and the judge can measure an agreed sentence against the report before the agreement binds anyone. A judge who defers has rejected nothing, so the rejection procedure described below has not yet been triggered.
If the court accepts the agreement, Rule 11(c)(4) adds a duty of explanation. The court must inform the defendant that, to the extent the agreement is of type (A) or (C), the agreed disposition will be included in the judgment.50 The duty arrives when it matters most. Acceptance is the point at which a type (C) term begins to control the court, so the defendant hears from the judge, and not only from counsel, that the bargained result will appear in the judgment. A defendant who knows the court has committed itself can see exactly what the plea has produced.
Rejection carries the strictest procedure, because it tells the defendant that the bargain on which the plea was offered will not be honored. When the court rejects an agreement containing type (A) or (C) provisions, Rule 11(c)(5) requires it to act on the record and in open court, or, for good cause, in camera.51 The rule goes on to list further steps the court must take after announcing the rejection, and the full text of Rule 11(c)(5) should be consulted for them. The on-the-record requirement does two jobs. It gives the defendant notice that the bargain has failed while choices about the case are still open, and it leaves a record from which a reviewing court can see what the judge did. The in camera alternative mirrors the disclosure rule discussed earlier: open court is the default, and privacy needs good cause.
The judge held the same power in both versions, since the court may decline to be bound by a type (C) agreement.52 In the first version, the judge used the deferral option and then took the acceptance step, ending with the required statement that the agreed disposition will be in the judgment.53 In the second, the judge exercised the power to refuse without the announcement the rule demands, so the defendant is left to discover the failed bargain from the sentencing notice.54 Students lose this issue when they treat a judge's silent refusal to follow the agreement as one more sentencing choice, when Rule 11(c)(5) requires a rejection to be made on the record and in open court, or in camera for good cause.55