Academic journal article Albany Law Review

A Nearly Perfect System for Convicting the Innocent

Academic journal article Albany Law Review

A Nearly Perfect System for Convicting the Innocent

Article excerpt


A law school casebook asks whether plea bargaining "convict[s] defendants who are in fact innocent (and would be acquitted [at trial])." (1) As the question indicates, the fact that plea bargaining may lead some innocent defendants to plead guilty is not a powerful criticism of this practice. The casebook asks whether plea bargaining increases the number of wrongful convictions. Because no one can know how many wrongful convictions are produced either by trials or by guilty pleas, the question may seem unanswerable.

But in fact the answer is easy. Convicting defendants who would be acquitted at trial is one of the principal goals of plea bargaining. "Half a loaf is better than none," prosecutors say. (2) "When we have a weak case for any reason, we'll reduce to almost anything rather than lose." (3) If the correlation between "weak cases" and actual innocence is better than random, plea bargaining surely "convict[s] defendants who are in fact innocent (and would be acquitted [at trial])."

Prosecutors engage in both "odds bargaining" and "costs bargaining." That is, they bargain both to ensure conviction in doubtful cases and to save the costs of trial. Were a prosecutor to engage in odds bargaining alone, he might estimate a defendant's chance of conviction at trial at 50% and this defendant's probable sentence if convicted at trial at ten years. Splitting the difference, the prosecutor then might offer to recommend a sentence of five years in exchange for a plea of guilty. Five years is what economists would call the defendant's "expected" sentence--his predicted post-trial sentence discounted by the possibility of acquittal. (4)

An offer of five years, however, would leave a risk-neutral defendant indifferent between pleading guilty and standing trial, and the prosecutor hopes to avoid a trial. He does not want the defendant to be indifferent. The prosecutor therefore engages in costs bargaining as well as odds bargaining. He tailors his final offer, not to balance, but to overbalance the defendant's chances of acquittal. (5) This prosecutor may offer four years in exchange for a plea--or two or three. (6) One can easily discover real-world cases in which prosecutors fearful of defeat at trial have struck bargains allowing defendants facing potential life sentences to plead guilty to misdemeanors. (7)

When a prosecutor has no chance of obtaining a conviction at trial, he may be unable to make an offer that will overbalance the defendant's chances of acquittal. (8) In every other case, however, the prosecutor can reduce the offered punishment to the point that it will become advantageous for the defendant to plead guilty whether he is guilty or innocent. Trials should occur only when defendants irrationally press their luck or when prosecutors and defendants disagree about probable trial outcomes and sentences. (9) In fact, trials are extremely rare in the American criminal justice system. The Supreme Court noted in 2012 that 97% of federal convictions and 94% of state convictions are the result of guilty pleas. (10)

Shawn Bushway, Allison Redlich, and Robert Norris recently provided evidence that real-world plea bargaining fits the economic model just described. They presented a hypothetical aggravated robbery case to 1585 prosecutors, defense attorneys, and judges, offering one of sixteen evidentiary variations on the case to each respondent. (11) They asked each respondent to estimate the likelihood of conviction at trial, the probable sentence following conviction at trial, and what sentence the respondent would accept as part of a plea agreement. For all but a few variations, the average acceptable sentence following a guilty plea was less that the predicted trial sentence discounted by the likelihood of acquittal. (12) Plea bargaining plainly makes it advantageous for innocent defendants with good prospects of acquittal to plead guilty. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.