Subjects and lessons to explore.

Law

Criminal Justice & Criminology · Courts

Plea Bargaining

13 min read 5 practice questions
10 key termsSourcesReport an error
Want it in plain words first? Jump to Eli explains — the same idea, no jargon.
On this page 9 sections
  1. In 30 seconds
  2. Why this matters
  3. The college version
  4. Eli explains
  5. Worked example
  6. Key takeaway
  7. Quick check
  8. Study tools
  9. Sources & references

In 30 seconds

Nearly every criminal conviction in the United States is a guilty plea, and most of those pleas are negotiated. The prosecutor offers to drop charges, drop counts, or take a position on sentence; the defendant gives up trial, confrontation, silence, and often appeal. Federal Rule of Criminal Procedure 11 supplies the form, and a line of Supreme Court cases sets the constitutional floor. This lesson covers the mechanics, the pressure behind them, and the argument over that pressure. It is educational material, not legal advice.

Why this matters

If you study the courts by studying trials, you are studying the exception. The negotiation that precedes a plea is where charges are chosen, sentences are effectively set, and most defendants meet the system's real decision points, yet almost none of it happens on the record. Understanding lets you read a conviction statistic correctly, see why a mandatory minimum matters even in cases that never invoke it, and follow the debate that runs through American criminal-justice policy: whether a system that resolves ninety-plus percent of its cases by agreement is efficient, coercive, or both. It also explains why wrongful convictions are not confined to trials.

The college version

What actually gets traded

A plea bargain is an agreement in which a defendant pleads guilty, or in some jurisdictions no contest, in exchange for something the prosecution controls. Practitioners sort the trades three ways. substitutes a less serious offense for the one charged, as when first-degree murder becomes second-degree murder. leaves the offense alone but reduces how many times it is charged, which matters when counts carry consecutive terms or separate mandatory minimums. leaves both alone and buys the prosecutor's position at sentencing: a recommendation, an agreement not to oppose a defense request, or a stipulation about how a sentencing rule applies. The Department of Justice's own manual uses a close version of this taxonomy, calling them charge, sentence, and mixed agreements. Real agreements usually combine them, and often add terms with nothing to do with the count sheet, such as forfeiture, restitution, cooperation against others, or a waiver of the right to appeal.

The three federal agreement types

Federal Rule of Criminal Procedure 11(c)(1) is the cleanest statement of the mechanics in American law, and it is worth learning by its letters. The parties may agree that the government will: (A) not bring, or will move to dismiss, other charges; (B) recommend, or agree not to oppose, a particular sentence, range, or guideline application; or (C) agree that a specific sentence or range is the appropriate disposition. The difference between (B) and (C) is the difference between a suggestion and a deal. A (B) recommendation does not bind the court, and the rule requires the judge to tell the defendant that he has no right to withdraw the plea if the court declines to follow it. A (C) agreement binds the court once the court accepts it. For (A) and (C) agreements, the judge may accept, reject, or defer until after reading the presentence report; if the judge rejects one, Rule 11(c)(5) requires an on-the-record warning and an opportunity to withdraw the plea. One structural rule governs all of them: the court must not participate in plea discussions.

The colloquy

A guilty plea is a conviction without a trial, so Rule 11(b) makes the judge build a record before accepting it. The judge must address the defendant personally in open court and confirm the defendant understands the right to plead not guilty, to a jury trial, to counsel, to confront witnesses, to be free from compelled self-incrimination, and to testify and compel witnesses, and that pleading guilty waives all of them. The judge must also state the nature of each charge, any maximum penalty, any mandatory minimum, forfeiture and restitution exposure, the obligation to calculate and consider the guideline range, the possibility of removal for a non-citizen, and the terms of any provision waiving appeal or collateral attack. Then two separate findings: that the plea is voluntary and did not result from force, threats, or promises other than those in the agreement, and, before judgment, that a exists. State procedures differ in detail but nearly all require some equivalent inquiry.

Voluntariness: Brady v. United States (1970)

Brady pleaded guilty to kidnapping under a statute that allowed a death sentence only on a jury's recommendation, and later argued that the statute had coerced him. The Supreme Court disagreed. A plea entered to avoid the possibility of execution is not for that reason involuntary. The Court adopted a standard from the Fifth Circuit: a plea by a defendant fully aware of the direct consequences must stand unless it was induced by threats, by misrepresentation including unfulfilled or unfulfillable promises, or by promises improper as having no proper relationship to the prosecutor's business. Because a plea also waives constitutional rights, it must be knowing and intelligent as well as voluntary. Brady also described what it called the "mutuality of advantage" that explains the practice, and noted that even in 1970 well over three-fourths of convictions rested on guilty pleas. The Court was careful, though, saying it would have serious doubts if offers of leniency substantially increased the likelihood that defendants advised by competent counsel would falsely condemn themselves.

Promises must be kept: Santobello v. New York (1971)

Santobello pleaded guilty to a lesser offense on the prosecutor's promise to make no sentencing recommendation. Months later a different prosecutor, apparently unaware of the commitment, recommended the maximum, and the judge imposed it while stating that the recommendation had not influenced him. The Court vacated anyway. When a plea rests in any significant degree on a promise or agreement of the prosecutor, so that the promise can be said to be part of the inducement or consideration, that promise must be fulfilled. Notice what the case does not do. It did not hold that the defendant picks the remedy: the Court remanded so the state court could decide between specific performance, with resentencing before a different judge, and letting Santobello withdraw the plea. Santobello is also the source of the Court's most-quoted defense of the practice, calling negotiated disposition an essential component of the administration of justice that, properly administered, is to be encouraged.

How far charging leverage may go: Bordenkircher v. Hayes (1978)

Paul Hayes was charged with uttering a forged check for $88.30. The prosecutor offered five years and said plainly that if Hayes refused he would seek a new indictment under Kentucky's Habitual Criminal Act, which carried a mandatory life sentence given Hayes's two priors. Hayes refused, the prosecutor followed through, a jury convicted, and Hayes was sentenced to life. By five votes to four, the Court held the Due Process Clause was not violated. The Court reasoned that the escalation had been disclosed before Hayes chose, that the recidivist charge was fully justified by the evidence, and that a prosecutor with probable cause to believe a statutory offense was committed generally has entire discretion over whether to prosecute and what to charge, subject to the limit that selection may not be deliberately based on an unjustifiable standard such as race or religion. Blackmun, Brennan, Marshall, and Powell dissented. Note what the holding is and is not: a constitutional floor, not an endorsement. Department of Justice policy tells federal prosecutors that charges should not be filed simply to exert leverage to induce a plea.

Pleading guilty without admitting guilt: North Carolina v. Alford (1970)

Facing a capital first-degree murder charge and strong evidence, Alford pleaded guilty to second-degree murder while telling the court he had not committed the killing. The Supreme Court held the plea constitutionally acceptable. An express admission of guilt is not a constitutional requisite to criminal punishment; an accused may knowingly consent to a sentence while unwilling or unable to admit the acts, and a plea containing a protestation of innocence may be taken where the defendant intelligently concludes his interests require it and the record contains strong evidence of actual guilt. Two limits are easy to miss. A defendant has no absolute constitutional right to have any guilty plea accepted, and states may bar their own courts from accepting pleas from defendants who assert innocence. Some do. Federal policy also discourages them: the Justice Manual directs that Alford pleas be avoided except in the most unusual circumstances and requires supervisory approval before a prosecutor agrees to one.

Counsel at the bargaining table: Frye and Lafler (2012)

Two cases decided the same day extended the Sixth Amendment into the negotiation itself. In Missouri v. Frye, defense counsel let two written plea offers expire without telling his client, who later pleaded guilty with no agreement and received three years instead of the offered ninety days. The Court held that the right to effective assistance reaches offers that lapse or are rejected, and that counsel generally has a duty to communicate formal offers on terms favorable to the accused. In Lafler v. Cooper, counsel advised rejecting an offer of fifty-one to eighty-five months on the mistaken theory that intent to murder could not be proven because the victim was shot below the waist; after trial the defendant received a mandatory minimum of 185 to 360 months. Both cases require real proof of prejudice: the defendant must show a reasonable probability that the offer would have been presented to the court, accepted by it, and adhered to by the prosecution, and that the result would have been less severe. Remedies must neutralize the taint without handing the defendant a windfall.

How many convictions are pleas, and what each number counts

Three correct figures point in slightly different directions because they count different things. The U.S. Sentencing Commission reports that of the 66,662 individuals sentenced in federal court in fiscal year 2025, 65,099 (97.7 percent) pleaded guilty and 1,563 (2.3 percent) were convicted at trial; that denominator is people sentenced, so it excludes everyone whose case was dismissed. The Bureau of Justice Statistics, counting all 71,866 defendants adjudicated in U.S. district court in fiscal year 2023, reports 88.9 percent convicted by plea, 2.2 percent convicted at trial, and 8.9 percent not convicted, mostly by dismissal. For state courts there is no comparable current annual national series; the best-known figure is the Supreme Court's own statement in Frye and Lafler in 2012, drawing on Bureau of Justice Statistics sources, that ninety-seven percent of federal and ninety-four percent of state convictions result from guilty pleas. The plea share is also not uniform: in fiscal year 2025 it ran from 99.7 percent in federal immigration cases down to 74.8 percent in murder cases.

What makes the offer hard to refuse

Four levers do most of the work. First, charging discretion: because the prosecutor chooses the counts, the prosecutor largely chooses the exposure the defendant is bargaining against. Second, statutory exposure. Under 21 U.S.C. 841(b)(1)(A) a qualifying drug quantity carries a ten-year minimum, rising to fifteen years after one prior serious drug or violent felony and twenty-five after two; under 21 U.S.C. 851 that enhancement applies only if the U.S. attorney files an information listing the priors before trial or before the plea, which puts a decision worth years squarely inside the negotiation. Third, the guidelines build in a discount: USSG 3E1.1 gives a two-level reduction for accepting responsibility, plus a third level on the government's motion when the offense level is sixteen or higher and the defendant gave timely notice of intent to plead. Among the 61,245 fiscal year 2025 federal cases with complete guideline data, 60.4 percent got the three-level reduction and 35.7 percent the two-level; only 3.9 percent got none. Fourth, detention: one study using randomly assigned bail judges estimated that pretrial detention raises the probability of conviction largely by increasing guilty pleas.

The trial penalty and the innocence problem

The gap between the post-plea and post-trial sentence is usually called the , and it is easier to describe than to measure. The National Association of Criminal Defense Lawyers, an advocacy organization for the defense bar, compared average sentences in Sentencing Commission data and reported large within-offense gaps, roughly threefold for fraud in 2015. That report itself cautions that the comparison does not control for case-level differences and that, because plea negotiations happen off the record, the full extent cannot be calculated. The mechanisms above are documented in primary law regardless of the aggregate gap's size. So is the consequence the Court worried about in Brady. Counting the National Registry of Exonerations' public case table as published in August 2026, 927 of the 3,848 exonerations recorded since 1989, about twenty-four percent, involved a defendant who pleaded guilty; the Registry codes conservatively, so that is a floor. In Harris County, Texas, nearly 200 people were cleared of drug possession convictions after the district attorney's conviction integrity unit found they had pleaded guilty before laboratory results showed the seized material was not a controlled substance.

Efficiency or coercion

The argument is old and unresolved, and both sides have institutional voices. On one side, Santobello reasoned that if every charge were subjected to a full trial the states and the federal government would need to multiply their judges and courtrooms many times over, and Brady described a mutuality of advantage: the defendant limits exposure and starts serving sooner, the government conserves scarce resources for genuinely contested cases. On the other, the Bordenkircher dissenters saw the Court retreating from its own protections against vindictive charging, and defense-bar and academic critics argue the differential is now large enough to function as a price on a constitutional right rather than a discount for cooperation. Nothing here resolves that. What it should let you do is separate three questions that are often blended: what the law permits, what agencies say their own lawyers should do, and what people argue the rules ought to be. This is educational material about United States law, it describes federal procedure with state practice varying substantially, and it is not legal advice.

Eli, the EliExplains learning guide

Eli explains

The same idea, in plain words

Explain it like I’m 10

Imagine a school where almost no disagreement ever reaches the principal. Instead, the person who writes up the incident and the student's advisor work out the consequence in the hallway, and the principal's job shrinks to checking that the student understood the deal and meant to take it. That is roughly how American criminal courts work. The trial is the official procedure everyone learns about, but it is not the usual one. Instead, the prosecutor decides what to accuse the person of, then offers to accuse them of less, or of fewer things, or to ask the judge for a lighter sentence, if the person admits guilt now. The person almost always has to decide without knowing exactly how a trial would come out. The judge is not allowed to join in the bargaining, but before accepting the deal must ask the person, out loud and on the record, whether they understand what they are giving up.

Picture it like this

It works like settling a dispute out of court rather than going before a judge: both sides trade the chance of a very good outcome for the certainty of a middling one, because a hearing costs time and could go badly.

Where the picture stops working

The analogy breaks in the place that matters most. In a civil settlement the two sides are formally equals and the worst case is money. Here one side is the government, it decides on its own how much the other side stands to lose by picking the charges, and the worst case is years of a person's life. A defendant may also be sitting in jail while deciding, which a civil litigant is not. And a settlement does not brand anyone guilty of a crime.

Worked example

Work through a federal case. A defendant is charged with one count of drug trafficking carrying a ten-year mandatory minimum and one firearm count. She has a prior qualifying drug felony, so the prosecutor could file a 21 U.S.C. 851 information and raise the floor to fifteen years, but only if he files it before she pleads. The offer: plead to the drug count, the firearm count is dismissed, no 851 information, and the government moves for the third acceptance-of-responsibility level. Identify each piece. Dismissing the firearm count is count bargaining. Declining to file the 851 is charge-related leverage exercised by not exercising it. The acceptance motion is a guideline reduction the government controls. If the agreement only says the prosecutor will recommend a number, it is Rule 11(c)(1)(B) and the judge is free to ignore it; if it fixes the sentence, it is 11(c)(1)(C) and binds the judge once accepted. Her lawyer must convey all of this to her, and the judge must confirm on the record that she understands the mandatory minimum and any appeal waiver before the plea is taken.

Key takeaway

Plea bargaining is not a shortcut around the American criminal process; it is the process, and the constitutional rules governing it are thin by design: the plea must be voluntary and informed, promises that induced it must be kept, and counsel must be competent, but the prosecutor's charging leverage that produces the plea is largely unreviewed. This lesson is educational material about United States law, not legal advice, and state practice varies.

Quick check

3 questions here, of 5 in this lesson’s practice set. Answers stay hidden until you check.

Question 1 of 3foundational

Under Federal Rule of Criminal Procedure 11(c)(1), which type of plea agreement binds the sentencing court once the court accepts the agreement?

Choose an answer, then check it.
Question 2 of 3intermediate

What did the Supreme Court actually hold in Bordenkircher v. Hayes (1978)?

Choose an answer, then check it.
Question 3 of 3advanced

A student writes that "97.7 percent of federal criminal defendants plead guilty, according to the U.S. Sentencing Commission." What is wrong with the sentence?

Choose an answer, then check it.
Practice all 5

Keep learning

Ready to build on this? Continue to the next lesson.

Practice this lesson
Study tools & related lessonsYou’ll learn to · Common mistakes · Easily confused · Key vocabulary · Related

You’ll learn to

  • Define plea bargaining and distinguish charge, count, and sentence bargaining.
  • Explain what Federal Rule of Criminal Procedure 11 requires of a plea colloquy and how the three agreement types under Rule 11(c)(1) differ in what they bind.
  • Summarize the holdings of Brady, Santobello, Bordenkircher, Alford, Frye, and Lafler, and state what each adds to the constitutional framework.
  • Interpret plea-share statistics correctly by identifying the denominator, the date, and the source of each figure.
  • Analyze how charging discretion, mandatory minimums, guideline reductions, and pretrial detention shape a defendant's decision to plead.
  • Evaluate the competing positions in the efficiency-versus-coercion debate without adopting either.

Common mistakes

  • Treating Bordenkircher v. Hayes as the Supreme Court approving of hard bargaining.

    It held only that due process was not violated when a disclosed threat to bring plainly provable charges was carried out. That is a constitutional floor. Department of Justice policy separately instructs federal prosecutors that charges should not be filed simply to exert leverage to induce a plea, and it was a five-to-four decision with two dissents.

  • Assuming a plea agreement binds the judge.

    Only a Rule 11(c)(1)(C) agreement binds the court, and only once the court accepts it. An 11(c)(1)(B) recommendation does not bind the judge at all, and the judge must warn the defendant that there is no right to withdraw the plea if the recommendation is not followed.

  • Quoting one plea-rate percentage as the plea rate.

    Ask what the denominator is and what year it comes from. The 97.7 percent federal figure for fiscal year 2025 counts people sentenced; the 88.9 percent BJS figure for fiscal year 2023 counts all defendants adjudicated, including the roughly nine percent whose cases ended without conviction; and the ninety-four percent state figure is a 2012 statement in Frye and Lafler.

  • Believing that pleading guilty means the person did it.

    A plea is a conviction, not an established fact about the world. Alford pleas are guilty pleas entered while asserting innocence, and roughly a quarter of the exonerations recorded since 1989 in the National Registry of Exonerations involved someone who had pleaded guilty.

  • Thinking bad plea advice is unreviewable because the trial that followed was fair.

    Frye and Lafler rejected exactly that argument. But relief still requires showing prejudice: that the offer would have been presented, accepted by the court, adhered to by the prosecution, and would have produced a less severe result.

Easily confused

Rule 11(c)(1)(B) sentence agreement vs. Rule 11(c)(1)(C) sentence agreement

A (B) agreement is a recommendation or a promise not to oppose; the judge may disregard it and the defendant cannot withdraw the plea if she does. A (C) agreement fixes the disposition and binds the judge once accepted, which is why the judge may defer accepting it until after reading the presentence report.

Alford plea vs. Nolo contendere plea

Both avoid an admission of the conduct. An Alford plea is formally a guilty plea entered while the defendant asserts innocence, so it is admissible like any conviction. A nolo plea neither admits nor disputes the charge and requires the court's consent, and under Federal Rule of Evidence 410 evidence of a nolo plea is not admissible against the defendant who made it in a later civil or criminal case.

Brady v. United States (1970) vs. Brady v. Maryland (1963)

Different cases, different subjects, same name. The 1970 case sets the voluntariness standard for guilty pleas. The 1963 case requires the prosecution to disclose material exculpatory evidence. Students routinely cite one for the other.

Missouri v. Frye vs. Lafler v. Cooper

Decided the same day on the same principle. Frye concerns an offer counsel never communicated, so it lapsed. Lafler concerns an offer the defendant rejected on incompetent advice and then lost at trial. Frye is about the duty to convey; Lafler is about the remedy after trial.

Key vocabulary

Plea bargaining
Negotiation in which an accused person agrees to plead guilty or no contest in exchange for something the prosecution controls, such as dropped charges or a favorable sentencing position.
Charge bargaining
Substituting a less serious offense for the one originally charged, so the conviction itself carries lower exposure.
Count bargaining
Reducing how many separate offenses a defendant is convicted of while leaving the offense definition unchanged, which matters when counts stack.
Sentence bargaining
Buying the prosecutor's position at sentencing rather than a change to the charges, through a recommendation, an agreement not to oppose, or a stipulation.
Plea colloquy
The on-the-record exchange in which a judge personally questions a defendant before accepting a guilty plea, covering rights waived, penalties faced, and whether the decision was freely made.
Factual basis
The requirement that a court satisfy itself, before entering judgment on a guilty plea, that the conduct admitted actually amounts to the offense charged.
Alford plea
A guilty plea entered by a defendant who continues to assert innocence, permitted where the record contains strong evidence of guilt and the defendant intelligently concludes the plea serves his interests.
Nolo contendere
A plea of no contest, accepted only with the court's consent, in which the defendant neither admits nor disputes the charge but accepts punishment.
Appeal waiver
A plea-agreement provision giving up the right to challenge the conviction or sentence on direct review or by collateral attack; a federal judge must state its terms before accepting the plea.
Trial penalty
The observed gap between sentences imposed after conviction at trial and sentences imposed after a guilty plea in comparable cases.

Sources & references

  1. Federal Rules of Criminal Procedure (effective March 21, 1946, as amended to December 1, 2025) — Administrative Office of the U.S. Courts (uscourts.gov)
  2. Federal Rules of Evidence (as amended to December 1, 2025) — Administrative Office of the U.S. Courts (uscourts.gov)
  3. Brady v. United States, 397 U.S. 742 (1970) — Supreme Court of the United States (United States Reports, via govinfo)
  4. Santobello v. New York, 404 U.S. 257 (1971) — Supreme Court of the United States (United States Reports, via govinfo)
  5. North Carolina v. Alford, 400 U.S. 25 (1970) — Supreme Court of the United States (United States Reports, via govinfo)
  6. Bordenkircher v. Hayes, 434 U.S. 357 (1978) — Supreme Court of the United States (United States Reports, via the Library of Congress)
  7. Missouri v. Frye, 566 U.S. 134 (2012) — Supreme Court of the United States (United States Reports, via the Library of Congress)
  8. Lafler v. Cooper, 566 U.S. 156 (2012) — Supreme Court of the United States (United States Reports, via govinfo)
  9. 2025 Guidelines Manual (Annotated), Chapter Three, USSG 3E1.1: Acceptance of Responsibility — United States Sentencing Commission
  10. Sourcebook of Federal Sentencing Statistics, Fiscal Year 2025, Table 12: Guilty Pleas and Trials by Type of Crime — United States Sentencing Commission
  11. Sourcebook of Federal Sentencing Statistics, Fiscal Year 2025, Table 22: Acceptance of Responsibility Reductions by Type of Crime — United States Sentencing Commission
  12. Federal Justice Statistics, 2023 (NCJ 309946) — Bureau of Justice Statistics (BJS), U.S. Department of Justice
  13. Justice Manual 9-27.000 - Principles of Federal Prosecution (including 9-27.300, Selecting Charges) — U.S. Department of Justice
  14. 21 U.S.C. 841 - Prohibited acts A — Office of the Law Revision Counsel, U.S. House of Representatives
  15. 21 U.S.C. 851 - Proceedings to establish prior convictions — Office of the Law Revision Counsel, U.S. House of Representatives
  16. The Effects of Pretrial Detention on Conviction, Future Crime, and Employment: Evidence from Randomly Assigned Judges — American Economic Review (American Economic Association)
  17. The Trial Penalty: The Sixth Amendment Right to Trial on the Verge of Extinction and How to Save It (2018) — National Association of Criminal Defense Lawyers
  18. Explore Exonerations (National Registry of Exonerations data table) and Understanding the Registry (code glossary) — National Registry of Exonerations, a project of the University of California Irvine Newkirk Center for Science and Society, University of Michigan Law School, and Michigan State University College of Law
  19. Exonerations in 2024 (annual report) — National Registry of Exonerations

EliExplains lessons are original prose written from the open, credible references above. See Copyright & Licensing.

Researched 2026-08-19

Educational content only. It is not medical, legal or professional advice. Found an error? Tell us.