The report on prosecutorial misconduct in U.S. v. Stevens is noteworthy in many respects. It is thorough, candid and well documented. The special counsel, Henry F. Schuelke, had unprecedented access to internal communications among members of the prosecution team and testimony from the prosecutors themselves about their conduct. This kind of direct evidence about the preparation and conduct of a trial is usually shielded from public view. For that reason in particular, the report is valuable. Its conclusion that the prosecution was “permeated by the systematic concealment of exculpatory evidence” is amply supported by facts painstakingly drawn from the participants and carefully and objectively reported.

What makes the report more valuable is the force with which the facts prove the need for discovery reform in criminal cases. As the latest and best documented example of concealment by experienced prosecutors in a high-stakes case, its rare behind-the-scenes view of experienced prosecutors debating about evidence that was clearly exculpatory demonstrates why nothing less than changes in the discovery rules will ensure compliance with the demands of Brady v. Maryland (1963).

As the report discusses, shortly before midnight on Oct. 1, 2008, in the middle of the trial, the prosecution disclosed to counsel for then-Senator Ted Stevens and to the court a redacted FBI report. That FBI report revealed that the main cooperating witness against Stevens, Bill Allen, had stated that he thought Stevens would have paid his bill if Allen had sent him an invoice. The exculpatory quality of this information was not, to any reasonably objective observer, debatable. After all, the senator’s defense was that he did not believe his ethics filings needed to report the cost of renovations to his Alaskan residence as gifts.

The internal discussions surrounding that disclosure, which are not fully discussed in the report, highlight the crux of the Brady problem. The lead prosecutor, who was the principal deputy chief of the public integrity section, and one other did not think that Allen’s statement was exculpatory. Another disagreed. The chief of the section, who had by that time been largely removed from review of discovery decisions in the case, was called into the meeting for advice. He was surprised at the discussion. He did not think the question was close, and he ordered the team to produce the information. Yet, even after that decision, the principal deputy chief maintained, in her letter disclosing the FBI report to the court and counsel for Stevens, that the information disclosed was only “arguably” Brady material.

The next day, the presiding judge rebuked the team for the delayed production, found the information to be unquestionably exculpatory, and ordered the prosecutors to produce every interview report and grand jury transcript of any witness associated with the Stevens investigation. Several days later, the then acting assistant attorney general and acting principal deputy assistant attorney general of the Criminal Division opined that the judge was wrong and that the information was not exculpatory.

This vignette illustrates the intractability of the Brady problem. How could four federal prosecutors with a combined 50 years of prosecutorial experience in some of the Department of Justice’s most prestigious offices get it so wrong? The answer is quite simple. All prosecutors have a conflict of interest where exculpatory information is concerned. Brady determinations require interpretation and judgment, and for some prosecutors their desire to win and their strongly held belief in their cases, their witnesses and their evidence cloud that judgment. No matter how many policy pronouncements the Department of Justice makes and how much discovery training it may provide, this conflict will never change.

We believe a few simple modifications of the discovery rules, not tinkering with the Brady obligation itself, will go far to eliminate the impact of this conflict on prosecutors’ selection of material to be provided to the defense.


In 1958, the state of Maryland prosecuted two men for murder. One of them, John Brady, received the death penalty. The state had argued at sentencing that Brady strangled the victim to death. After trial, the defense learned that the prosecutor had information that Brady’s accomplice had confessed to strangling the victim, and the prosecutor had not disclosed the evidence to Brady’s counsel. The U.S. Supreme Court reviewed the case and announced a constitutionally based requirement that prosecutors turn over to the defense evidence favorable to the accused on guilt or punishment.

The Supreme Court did not find remarkable a constitutional requirement that prosecutors disclose evidence favorable to the accused. The essence of the holding is that our adversary system is not so purely adversarial that a prosecutor can conceal evidence from a defendant that might help him avoid conviction or reduce punishment. Every prosecutor in the country, state and federal, acknowledges the obligation to produce evidence favorable to the accused and asserts that he unfailingly complies with it.

In the 50 years since the Supreme Court’s Brady decision, however, the objective truth is that, although prosecutors acknowledge the obligation, and most abide by it, numerous failures to produce exculpatory evidence have come to light, almost always after conviction. A disturbing number result in release or new trials for men and women who have been incarcerated for long periods of time. Statistics are not available, but whatever the actual number, one is too many, and there certainly have been more than one. The kind of misconduct documented in the Stevens case is regrettably not uncommon. See, e.g., Maynard v. Government of the Virgin Islands, 392 Fed. Appx. 105 (3d Cir. 2010) (convictions reversed for prosecutor’s neglect in providing exculpatory information to defendant); U.S. v. Chapman, 524 F.3d 1073, 1086 (9th Cir. 2008) (indictment dismissed for Brady violations that rose “to the level of flagrant prosecutorial misconduct”); U.S. v. Fitzgerald, 615 F. Supp. 2d 1156 (S.D. Cal. 2009) (indictment dismissed for “flagrant” misconduct, which may be shown by either intentional or reckless disregard of disclosure obligations).

Some prosecutors’ offices have become notorious for systematic noncompliance. New Orleans is one. The Supreme Court is this term reviewing for the third time concealment of exculpatory information in that office. One of its alumni described the policy as “When in doubt, don’t give it up.”

In 2009, Chief Judge Mark Wolf of the District of Massachusetts commented on the long-standing problem in the U.S. attorney’s office there, observing that “[i]n the District of Massachusetts, the government has had enduring difficulty in discharging its duty to disclose exculpatory material in a timely manner.”

Recently, Judge James Turk of the Western District of Virginia ordered the release of a Virginia state prisoner convicted of murder, finding that “[h]ere the Prosecutor’s Office engaged in a series of lies and failures to disclose exculpatory evidence.” The county prosecutor had concealed and lied about promises to an informant who claimed that the defendant had confessed to him. The informant had testified repeatedly as a government witness, and the prosecutor had caused the defendant to be moved into the cell with the informant for purposes of creating another “confession.” The prosecutor lied about the reason for the move. Turk concluded that the trial was an example of “an extreme malfunction in the state criminal justice system.”


On April 1, 2009, Attorney General Eric Holder Jr. moved to vacate the conviction of then former Senator Stevens and to dismiss the prosecution for its failures to disclose exculpatory evidence to the defense. At about the same time, an understandably irate Judge Emmet Sullivan, who had presided over the case, appointed a distinguished and experienced attorney, Henry F. Schuelke, to investigate and decide whether any members of the prosecution team should be held in contempt or prosecuted for withholding the evidence. A former federal prosecutor himself, Schuelke understood the challenges for prosecutors who conduct a complex lengthy investigation followed by a trial of a well-represented public figure. His lengthy report detailed stunning failures to disclose exculpatory evidence by the trial prosecutors. It left no room for doubt that the evidence was exculpatory, that the trial prosecutors knew about it, and that they withheld it. It rejects the more nonsensical justifications for the failures and simply lets the facts speak for themselves.

In addition to nondisclosure of the Allen statement described above, the lengthy work details how the prosecution team, led by the principal deputy of the section, decided, with the approval of the assistant attorney general and his deputy, not to produce interview reports of witnesses because they were not literally covered by the Jencks Act, how the overworked prosecution team let FBI agents conduct the Brady review, and how some members of the team instructed agents to write interview reports that minimized the exculpatory nature of a witness’s testimony, and in one case at least acquiesced in an agent not making any report at all of a vital exculpatory remark by a key witness.

In summary, we have an objective written history of how in a high-profile case experienced prosecutors can lose their way so badly that the attorney general himself found the case unsalvageable.

The Stevens case has spawned calls for reform from many quarters. One of them is, not surprisingly, the U.S. Senate. There is already a proposal for legislation, the goal of which would be to prevent more Stevens-type lapses. Legislation may be helpful, but in our view the legislation will not create duties that do not already exist.

Instead, it is time to explore honestly why the Brady obligation seems so difficult to follow. Reform will not have any meaning until the relevant policymakers examine and understand the fundamental problem. The Brady obligation creates a deep and difficult, if constitutionally required, conflict between the obligation to convict the guilty and the duty to reveal information that makes conviction less likely. For the men and women who prosecute in state and federal courts, making a Brady determination requires them to step out of their prosecutorial shoes and view their evidence through the eyes of defense counsel. For some, this is an impossible undertaking.

Prosecutors have strong and sincerely held beliefs in the merits of their cases. They believe the defendant is guilty. They want to win. No prosecutor advances a career with a string of not guilty verdicts. They have often lived with their cases for years. They have met with and believe in their witnesses and they have bonded with the victims. They rely on investigators with whom they have developed close and personal relationships. Those same investigators believe just as strongly in their cases, and in some instances forge close and personal relationships with the witnesses and victims.

It is unrealistic to expect that every prosecutor can set aside all these personal, sometimes emotional commitments, and, in the heat of a controversy, identify and hand to his adversary material that makes it more likely that the prosecutor will lose. This is precisely what Brady requires and precisely why compliance founders.

Legislative reforms, like the proposed Murkowski bill, that simply change the standard for exculpatory from “material” to something like “more favorable” to the defendant, while they may be helpful, do nothing to alter the inherent conflicts of interest: They simply shift the subjective decision-making from one standard to another.


Unlike civil discovery, discovery in criminal cases is not based primarily on relevance. Under Federal Rule of Criminal Procedure 16, the governing rule, prosecutors are obliged to produce documents or things that are “material,” but they have no obligation to disclose the identity of witnesses with knowledge or the statement of any trial witness unless it is verbatim (or adopted by the witness or exculpatory). Some prosecutors go beyond the narrow definitions, but not all, and there is no enforceable requirement that they do so. Indeed, the practice of producing interview reports varies widely within the Department of Justice itself. Some offices produce interview reports as witness statements without hesitation. At least half do not. Some prosecutors do not produce interview reports of witnesses they do not intend to call, even if the witnesses have relevant things to say.

Thus, today’s prosecutor may view discovery, understandably, as an obligation not to provide his adversary with relevant testimony, inculpatory or exculpatory, but only with transcripts, if they exist, of the testimony of witnesses the prosecutor has called or intends to call. If some testimony of a witness they intend to call may be exculpatory, the obligation to disclose it may literally be satisfied by producing the transcript after the witness testifies, or shortly before. And whether or not there are transcripts, some prosecutors, as the Stevens team did, resolve all Brady duties by summarizing exculpatory evidence rather than producing the underlying information.

To make matters worse, current Department of Justice policy allows all of these variations. The broad language of the Department of Justice’s memorandum on criminal discovery, issued by Deputy Attorney General David W. Ogden on Jan. 4, 2010, would sanction a prosecutor’s decision not to disclose interview reports and permit an agent’s review of interview reports for Brady even in a routine case. In other words, the errors in Stevens are replicable today under current Department of Justice policy, thus leaving the issue of “poor management” and “poor supervision” over discovery obligations uncorrected at the senior levels of the department.

In our view, the time has come to acknowledge that discovery for prosecutors in criminal cases should be broadly relevance-based. At a minimum, the rule must require prosecutors to disclose all interviews of witnesses, whether they are to be called or not, and all grand jury testimony, well in advance of trial. When witness safety, national security or other factors make disclosure dangerous, courts can adjust the timing of discovery, or its distribution, to meet those concerns. This reform would assure that, unless prosecutors affirmatively hide evidence, they will not fail to disclose exculpatory testimony because they do not think it is sufficiently exculpatory. It should not replace the fundamental and independent duty to disclose exculpatory evidence if not otherwise produced, but it will go far to eliminate the nonproduction of testimony that the defense could use simply because the rules do not require its production in the form in which it exists and when the exculpatory information, or the material nature of it, is debatable in the mind of the diligent prosecutor.

It will also eliminate the largely pretextual argument that the Jencks Act justifies not producing exculpatory information in transcripts until after the witness testifies, which is what the Jencks Act’s language explicitly requires. And it should make clear that all witness statements, whether verbatim or not, must be turned over.

The Department of Justice will undoubtedly resist the reform we suggest and insist that the answer to the problem is more training and not mandated disclosures. We think those arguments are shortsighted, particularly given that current department policy would allow the errors in Stevens to reoccur. There is no reason why discovery in criminal cases should not be relevance-based as it is in civil cases. And there is no reason, in most cases, why defense counsel should not obtain from prosecutors the identity of witnesses who have relevant information. The Brady obligation is basically subjective. It requires, as we noted, an adversary to conclude that some bit of evidence would be useful to his adversary. Our view is that all participants in the system will benefit from unmistakable objective standards for discovery.

Like it or not, the department would ultimately benefit from such a rule. It would afford prosecutors more protection in those cases in which unintentional or inadvertent Brady mistakes do happen. The department will be in a better position to argue that any nondisclosed information in fact was cumulative, or the inadvertent mistake did not rise to the level of constitutional error.

Requiring prosecutors to produce all interview reports and grand jury testimony subject to certain exceptions will certainly ensure more robust and complete discovery to the defense. It will reduce discovery issues and allow for more surgical, narrowly tailored discovery requests germane to the issues at hand. It will also allow defense counsel to have earlier, more complete and meaningful discussions with their clients about proposed plea agreements.

Finally, requiring prosecutors to produce all interview reports and grand jury testimony shortly after charges will generate greater efficiencies for the courts. Discovery disputes will be reduced. More cases would resolve in earlier, rather than last-minute, plea agreements as defense counsel and defendants have more information to evaluate the merits of a prosecution.

For discovery reform to be true reform, it must correct the fundamental issues at play and have uniform and consistent application across the board. Otherwise, justice will remain individualized and idiosyncratic, and that is not justice for anyone.

William M. Welch II was formerly the chief of the public integrity section in the U.S. Department of Justice. He served in that position at the time of the Stevens prosecution. He was a federal prosecutor for 22 1⁄2 years. William W. Taylor III is a partner at Zuckerman Spaeder in Washington. He was one of the attorneys who represented Welch during the Schuelke investigation. Steven N. Herman, an associate at the firm, provided assistance in the preparation of this piece.