In response to:

Why Innocent People Plead Guilty from the November 20, 2014 issue

‘The Two Lawyers’; painting by Honoré Daumier

To the Editors:

Judge Jed S. Rakoff’s article “Why Innocent People Plead Guilty” [NYR, November 20, 2014] is spot on, but doesn’t go far enough. True, we have a federal plea system, not a trial system. True, to call the process “plea bargaining” is a cruel misnomer. There is nothing here remotely like fair bargaining between equal parties with equal resources or equal information. The prosecutors’ power—as Judge Rakoff describes—is extraordinary, far surpassing that of prosecutors of years past, and in most cases, far surpassing the judge’s. Judge John Gleeson, a federal judge of the Eastern District of New York, made this clear during a case involving a charge for which there is a mandatory minimum sentence.1 As a result of the prosecutor’s decision to charge the defendant with an offense for which there is a mandatory minimum sentence, no judging was going on about the sentence. The prosecutor sentenced the defendant, not the judge, with far less transparency and no appeal.

Indeed, there were times during my seventeen-year tenure on the federal bench in Massachusetts that inquiring of a defendant as to the voluntariness of his guilty plea felt like a Kabuki ritual. “Has anyone coerced you to plead guilty,” I would ask, and I felt like adding, “like thumbscrews or waterboarding? Anything less than that—a threatened tripling of your sentence should you go to trial, for example—doesn’t count.”

It is true too, in view of such threats of long terms in prison, that there is a strong possibility that the innocent may plead guilty. It may well be a rational calculation, given the penalty of going to trial, for there is clearly such a penalty. The prosecutor typically induces a plea by offering a “carrot,” the lesser charge, and at the same time a gigantic “stick.” It is not simply that he may well tack on additional charges enabling mandatory or even consecutive punishments, should the defendant go to trial. He also can threaten that he will introduce evidence of uncharged conduct at the sentencing, or even evidence of counts for which the defendant was acquitted, so long as the defendant is convicted of something. No other common law country in the world enables the prosecutor to seek a sentence based on criminal conduct never charged, never subject to adversary process, never vetted by a grand jury or a jury, or worse, charges for which the defendant was acquitted.

Under the Federal Sentencing Guidelines introduced in 1984, such alleged conduct by the defendant can “count” toward a higher sentence, as evidence of additional “loss” in a white-collar case or quantity in a drug case. It can substantially increase the sentence, an effect unheard of in other, nonfederal jurisdictions, and indeed, unheard of before the passage of the guidelines. (Prior to the guidelines a judge could have considered uncharged or even acquitted conduct generally, but it did not have the same serious consequences in the post–mandatory minimum, post-guideline era in which a specific quantity of drugs or a specific amount of criminally obtained money, or both, almost exclusively drive the sentence.)

Whatever we call them, the prosecutor’s tactics impose very serious penalties for going to trial, for choosing to have a jury decide your fate, or even for challenging unconstitutional conduct. In fact, the guidelines encourage rushing to the prosecutor to plea bargain as soon as possible, before your codefendants do.

I retired from the bench in September 2011, after seventeen years. Prior to that, I had been a criminal defense lawyer. I am back in practice (among other activities like teaching and writing). I can see how the defense of federal criminal cases has changed. Do you risk challenging this unconstitutional system when it may mean that the more compliant—perhaps more culpable—defendants get a better deal? Do you risk litigating the case in the face of the trial penalty that could well double or triple a defendant’s sentence?

Remember Aaron Swartz, the twenty-six-year-old Internet prodigy accused of wire fraud and violations of the Computer Fraud and Abuse Act who committed suicide after being offered a choice between a plea of guilty with a six-month sentence or a trial in which he risked a seven-year sentence under the guidelines if found guilty. Then there is the case of Kevin Ring, a lobbyist for Jack Abramoff, who was convicted at trial. Abramoff, the conceded ringleader, pled guilty and got four years. Offered a deal with no prison time if he cooperated, Ring refused, taking the case to trial. After a trial finding him guilty, the prosecutor urged a seventeen-to-twenty-two-year sentence, which the sentencing judge acknowledged could well have a “chilling effect” on the exercise of the right to a jury trial. Indeed, it is not just the innocent who may well plead guilty. It is also possible that claims of prosecutorial or police misconduct will not be litigated because of a rush to plead guilty.


The remedy Judge Rakoff proposes would help, namely, a review process before a magistrate judge. But the magistrate judge is not appointed for life; his tenure is limited to a term and dependent upon the decision of the federal district court he serves. How free will he be to probe the plea deal? How much information would he have? Often he is a former prosecutor—although as Judge Rakoff’s judicial record suggests, that background does not predict a tilt toward the prosecutor.

Still, I am skeptical that anything will change unless there are changes in the structure of punishment, including the previously mentioned mandatory minimums and guidelines, as well as the ability of prosecutors to refer to uncharged or acquitted conduct at sentencing. I am skeptical that anything will meaningfully change until we reduce the weapons that the prosecutor brings to the bargaining table. I am skeptical unless the courts frankly characterize enhanced trial sentences for what they are, an improper penalty for going to trial, a substantial burden on the rights the Constitution provides. Without these changes, I fear that even review by the magistrate judge will become a Kabuki ritual.

Judge Nancy Gertner
Harvard Law School
Cambridge, Massachusetts

To the Editors:

Jed S. Rakoff does an excellent job, first, of differentiating between criminal justice as conceived by most citizens and the plea system that dominates criminal courts throughout the US, and second, of describing how the current system unfairly leads innocent defendants to plead guilty. But his suggested solutions, of involving judges in the plea-bargaining process, or of using magistrates to oversee the process before the results are presented in court, are halfway measures that face two problems. The fundamental difference between criminal justice as the Founders intended it and criminal justice as currently practiced is that in the first, two opposing sides present a case before an independent decision maker (judge or jury) who operates with no bargaining interest, whereas in the current system, there is still bargaining instead of an independent judgment regarding guilt. Judge Rakoff’s solution also maintains the bargaining, and while it may make the procedures fair by equalizing the power of prosecutors and defense attorneys, it would still lead innocent yet rational defendants to plead guilty.

That is the first problem. The second problem is purely economic. For a number of years, I led workshops in ethics and philosophy of law for local and state judges around the country. I always raised ethical problems with plea bargaining and I always received the same response: “I have too many cases to avoid plea bargaining.” Of the hundred or so judges with whom I discussed the issue, only a couple had dockets of less than seven hundred, while many had a backlog of more than a thousand cases. We now try less than 10 percent of criminal cases. If we were to try every criminal case, we would need six or seven times the number of criminal judges we now have, along with new courtrooms and staff. Judge Rakoff’s solution would not require as much, but involving judges or magistrates in every plea bargain would surely require two or more likely three times the number of judges or magistrates now on the bench. The public is unlikely to approve a doubling or tripling of court expenses.

I have no perfect solution, but it is very clear that if we eliminated excessively long sentences, legalized drug use, and decriminalized other drug offenses (along with some other crimes, such as shoplifting), we would give judges more time to try the important cases and allow plea bargaining only for cases involving short sentences or no felony record. The money we would save from reduced incarceration could then be used to increase the number of judges and courtrooms.

Bruce Brower
Associate Professor and
Chair of Philosophy
Tulane University
New Orleans, Louisiana

To the Editors:

Judge Jed S. Rakoff’s article “Why Innocent People Plead Guilty” is accurate in its basic description of the failings of our criminal justice system: mandatory minimum sentences (indeed harsh sentencing laws generally) are used to coerce guilty pleas; defense lawyers are often at a distinct informational disadvantage; prosecutors, not judges, decide sentences in a high proportion of cases; and trials have all but vanished.

Judge Rakoff’s prescription is more judicial involvement in the plea-bargaining process. After arrest or indictment, a magistrate judge would meet separately with the parties, assess the strengths and weaknesses of the case, and make a recommendation—“dismiss the case…proceed to trial…or…enter into a plea bargain along lines the magistrate might suggest.” The proceedings before the magistrate would be sealed, and “no party would be required to follow the magistrate’s suggestions.”


This dog won’t hunt. Consider this simple case. A defendant is stopped at the airport after his conduct raises suspicion. A lawful search of the suitcase he is carrying reveals a pound of heroin hidden in a secret compartment. The defendant tells the arresting officer that an acquaintance asked him to bring the suitcase to New York and that he had no knowledge of the hidden compartment or its contents.

What happens before the magistrate in Judge Rakoff’s proposed procedure? Does the defendant testify? Does the magistrate cross-examine him? Is the prosecutor obliged to provide the magistrate with materials that might impeach the defendant’s credibility? Why would a prosecutor give the defendant a preview of her case? If the defendant exercises his constitutional right and declines to testify, how can the magistrate assess the strength of the case? Does “sealed” mean sealed? What if the magistrate learns that the defendant has changed his story at trial? Can the defendant (and prosecution witnesses) lie to the magistrate with impunity?

These questions—and one could raise many more—show that the proposal is almost certainly unworkable. There is, however, a more fundamental problem. As many criminal procedure scholars have observed, one reason that trials are so rare is that they have become so expensive. With the proliferation of defendants’ rights, even a misdemeanor trial that once took a day can now take a week. Which is why in New York City, where I practice, there are none.

Sadly, simplifying our criminal justice system so that trials are fair and efficient is on no one’s agenda. (As Judge Rakoff correctly notes, we also lack the political will to eliminate harsh sentences that make going to trial so perilous for so many defendants.) But adding more procedure is not the solution to our problems. Nonbinding reviews by magistrates would be a leap in the wrong direction.

Paul Shechtman
Zuckerman Spaeder, LLP
Adjunct Professor, Columbia Law School
New York City

Jed S. Rakoff replies:

It is gratifying that people as diverse as a former federal judge (Nancy Gertner), a former New York state director of criminal justice (Paul Shechtman), and a chair of a major university’s philosophy department (Bruce Brower) all agree with my assessment of why the current criminal justice system unfairly coerces guilty pleas and effectively transfers sentencing power from judges to prosecutors. While they also seem to agree with me that eliminating mandatory minimum sentences and other harsh sentencing laws is the most direct solution, the writer with the most familiarity with the political process, Mr. Shechtman, acknowledges that, as I argued, this is unlikely to occur in the foreseeable future. A politician perceived as “soft on crime” is a politician who had better start looking for another job.

The letter writers, however, are more skeptical that my fallback proposal—allowing judges to participate in the plea-bargaining process—is likely to be effective. I share their concerns that there are potential pitfalls in my proposal, which is why I suggested that it first be tried as a modest pilot project. Clearly, like most reforms in our criminal justice system, it would place some additional (though I think modest) demands on judicial resources. Nonetheless, I think that there are good reasons to think it would work.

First, an essentially analogous program that has proven very effective already exists in civil cases in the form of nonbinding mediation. In my own judicial district, for example, civil litigants regularly meet with magistrate judges or court-appointed mediators shortly after a case is filed and, in separate, confidential presentations to the mediator, describe their respective evidence and positions. The mediator then meets again with the parties separately and, based on what the mediator now knows about the underlying factual and legal positions, points out to the respective parties the pitfalls they each face.

As every mediator will tell you, it is not that the parties have been wholly unaware of these pitfalls in the past that makes mediation effective, but rather that the parties have never taken these shortcomings seriously because they were never previously described with the forcefulness and objectivity that a mediator brings. Duly shaken, the parties are now open to a settlement proposal from the mediator that they would have previously rejected out of hand. The result is that, in the overwhelming majority of cases, the mediator succeeds in settling the case voluntarily. In my proposed program, a judge or magistrate would, in effect, conduct a criminal case mediation. I do not agree with Judge Gertner that magistrate judges, appointed by the tenured judges in their district, would lack the will or independence necessary to serve as successful criminal mediators, and I think Mr. Shechtman’s concern that the parties would be unwilling to share their proof with their adversaries is eliminated, as in all mediations, by having the parties meet separately with the mediator and present their proof on a confidential basis.

Prosecutors, in particular, would find their factual and legal contentions subject to a scrutiny wholly absent from current pretrial practice. In many cases, this would lead them to accept a recommended plea bargain far less onerous than those now obtained through their raw exercise of power. In some cases, they might even come to question whether they have charged the real culprit and whether, instead, the case should be dismissed. And in all cases, at a minimum, some modest degree of judicial scrutiny would be exercised over the only part of our criminal justice system that currently matters: the plea bargaining process.

Second, a program in some respects similar to what I recommend has, in fact, been in place in Connecticut for some years, where a judge is permitted to engage in fact-based plea bargaining as long as he or she does not preside over the subsequent trial if the plea bargaining fails. Contrary to Mr. Shechtman’s assertion, it appears that these “dogs” do “hunt,” at least in the wilds of Bridgeport and New Haven, without excessive demands on judicial resources and with considerable success. Thus, a 2006 survey of Connecticut prosecutors, defense lawyers, and judges found a broad consensus that the enhanced judicial role in the plea-bargaining process had increased the fairness and legitimacy of the process and provided a much-needed check on prosecutorial overreach.2

Third, in any event, I respectfully submit that the only real way to find out if my proposal would work effectively in a jurisdiction like the federal courts is to try it out on a pilot basis. It is all too easy to conjure up supposed flaws in an experiment that has not been tried: But how can a problem as deep-seated as this one ever be solved without some modest attempt at innovation?