Blog
When a Proffer Agreement Backfires: Real Case Examples
The Cost of Candor
The proffer agreement is the most misunderstood document in federal criminal practice. It promises protection. It delivers constraint. A defendant walks into the United States Attorney’s office believing that candor will purchase goodwill, that the most damaging admissions will remain confined to that conference room, and that cooperation is the first movement toward resolution. In a significant number of cases, that belief is warranted. In the cases that concern this article, the belief proved to be a miscalculation, one whose consequences extended well beyond the proffer session itself.
Federal courts have, over three decades, constructed a framework in which the protections of a proffer letter are narrow by design and, in certain circuits, narrow to the point where mounting a defense at trial becomes an exercise in avoidance.
The Mezzanatto Framework
The architecture of the modern proffer agreement originates in a 1995 Supreme Court decision that most defendants have never read and every federal prosecutor has committed to memory. Gary Mezzanatto was arrested in San Diego on federal drug charges after selling methamphetamine to an undercover officer. Before the plea discussion commenced, the prosecutor imposed a condition: any statements Mezzanatto offered during the session could be used at trial to impeach inconsistent testimony. Mezzanatto consulted with his attorney. He agreed.
During the session, Mezzanatto conceded that he knew the package he sold contained drugs. He also conceded knowledge of a methamphetamine laboratory at a co-conspirator’s residence. At trial, he denied both. The prosecution cross-examined him with his own prior admissions. The jury convicted.
The Ninth Circuit reversed, reasoning that the exclusionary provisions of Federal Rule of Evidence 410 and Federal Rule of Criminal Procedure 11(e)(6) could not be waived by the defendant. The Supreme Court disagreed. Justice Thomas, writing for the majority, held that the exclusionary provisions were “presumptively waivable” absent some indication that the defendant entered the agreement unknowingly or involuntarily. The decision rested on a principle that sounds benign in the abstract: a defendant who agrees to speak with candor should not later be permitted to testify in contradiction without consequence.
Justice Souter, in dissent, identified the practical consequences. Standard prosecution forms already required these waivers before a prosecutor would consent to discuss cooperation at all. The waiver was not a negotiated term. It was the price of admission (which every circuit has since ratified, though the reasoning varies in ways that matter more than the holdings suggest). Defendants, he observed, were in no position to challenge the demand. The exception, he wrote, would consume the rule.
Most defendants who sign a proffer letter do not appreciate the scope of what they have waived until the proffer has already been given.
Derivative Use and the Kastigar Waiver
Before the proffer session, before the defendant’s attorney has finished reviewing the standard letter, the most consequential waiver has often already been embedded in the boilerplate.
The concept of derivative use, rooted in Kastigar v. United States (1972), defines the boundary between what the government may and may not accomplish with immunized testimony. Under Kastigar, the government must demonstrate that its evidence was obtained entirely independent of the immunized statements. Proffer agreements, however, include a clause waiving Kastigar protections. The defendant, in signing, surrenders the right to contend that evidence introduced at trial was tainted by the proffer.
The waiver operates as a one-way corridor. A defendant who mentions a bank account during the proffer session has furnished the government with an investigative lead the government may already possess. The defendant does not know this. But the proffer now supplies a plausible independent justification for introducing that evidence at trial, a justification that did not exist before the defendant spoke. The records become admissible not because the proffer was quoted but because the government can assert the records were obtained through independent means. The defendant’s own words constructed the argument against which the defendant can no longer object.
Whether the Kastigar waiver renders proffers categorically dangerous or merely situationally dangerous is a question I find myself answering differently depending on the district, the prosecutor, and the volume of evidence the government has already assembled.
When Cross-Examination Opens the Door
The decisions that followed Mezzanatto expanded the waiver’s reach in directions the original opinion left unresolved. The question was no longer whether a defendant’s own contradictory testimony could trigger the waiver. The question became whether any act of defense could accomplish the same.
United States v. Velez, decided by the Second Circuit in 2004, illustrates the constraint. José Velez was indicted in the Southern District of New York for possession of a firearm as a convicted felon. In his first proffer session, he maintained his innocence. He requested a second session with new counsel, and before that meeting began, he signed a proffer agreement containing a broad waiver provision: the government could use his statements to rebut any evidence or arguments offered by or on behalf of the defense, including arguments raised sua sponte by the court. In the second session, Velez admitted that he owned and possessed the firearm.
At trial, Velez confronted the consequence of that admission. He chose not to introduce testimony that would have supported his defense because doing so would have permitted the government to introduce his proffer statements. The jury convicted. The court sentenced him to 120 months.
At sentencing, Velez stated that he had been “trapped” into making his admissions. The practical consequence (that Velez could not present a defense without arming the prosecution with his own confession, that his attorney was constrained to strategies that did not contradict a statement made months earlier in a room where Velez believed he was bargaining, not testifying) was, in the court’s view, a fair result of a choice voluntarily made.
In the Seventh Circuit, United States v. Krilich (1998) extended the doctrine further. Robert Krilich, a real estate developer in suburban Chicago, admitted during a proffer to bribing the mayor of Oakbrook Terrace, Illinois, to obtain a favorable zoning change. The bribe was structured through a staged hole-in-one golf contest: Krilich palmed a ball, placed it in the cup, and declared the mayor’s son the winner, which triggered an insurance payout. The scheme coupled bribery with insurance fraud.
At trial, Krilich did not testify. His attorneys attempted to present an inconsistent position through the cross-examination of government witnesses, several of whom were friendly to the defense. The Seventh Circuit held that this was sufficient to trigger the waiver. Evidence is evidence, the court wrote, whether it emerges on direct examination or on cross. The conviction stood.
I have counseled clients through this calculus in the Southern District and in the Eastern District, and the honest assessment is that the line between permissible cross-examination and a waiver trigger is not always discernible before it has been crossed.
The proffer agreement does not prohibit a defense. It prices the defense in a currency the defendant has already spent.
The Krilich court did note that the prosecutor’s position was too extreme: the government had argued that mounting any defense at all would open the door to proffer statements. That reading was rejected. But the middle ground the court identified was thin enough that defense counsel, reviewing the proffer letter before the session, must consider the possibility that standard cross-examination will be treated as an inconsistent position. The filing sits in the record as a reminder of what happens when a defendant speaks before understanding the terms under which those words may return.
Proffer Statements at Sentencing
The protection that most defendants believe they have purchased applies, by its own terms, only to the government’s case-in-chief at trial. It extends no further.
Under the United States Sentencing Guidelines, judges consider all “relevant conduct,” a category that includes criminal activity forming part of the same course of conduct or common scheme, even if uncharged. Admissions made during a proffer session constitute relevant conduct. A defendant who enters the proffer room facing charges related to a single transaction and who, in the course of full disclosure, reveals a pattern of similar conduct spanning several years has expanded the court’s sentencing calculus. The loss amount increases, the offense level increases, and the Guidelines range shifts in a direction no proffer session can reverse.
This is, in our experience, the dimension of proffer risk that receives the least attention before the session and produces the most consequential results after it. We have observed, in three cases in the last eighteen months, sentencing exposure that exceeded what the government could have proved independently by a margin the client had not anticipated. A defendant charged with fraud involving a loss in the low hundreds of thousands may, through candor, disclose conduct that pushes the figure substantially higher. The proffer agreement’s trial protection becomes irrelevant if the defendant pleads guilty or is convicted. At sentencing, every admission returns.
We approach proffer preparation differently on this point. Before any session, we construct a detailed estimate of the sentencing exposure that would result from full disclosure, measured against the exposure the government can establish without the defendant’s cooperation. The distance between those two figures determines, in significant part, whether the proffer is advisable. If the gap is small, cooperation imposes little additional sentencing risk. If the gap is wide, the proffer may inflict more damage at sentencing than it prevents at trial.
The Decision to Cooperate
Nothing in the preceding sections should be read as an argument against cooperation. A proffer agreement, entered with full awareness of its limitations and its waiver provisions, remains the most common mechanism for resolving federal criminal exposure, and often the most effective one. The cases discussed above are cautionary not because cooperation failed but because the defendants’ understanding of the document did not match its operation.
The gap between expectation and operation is where the damage occurs.
Three questions govern the analysis. First, what can the government prove without the defendant’s assistance? If the answer is most of the case, the proffer adds relatively little to the government’s position and may purchase significant goodwill. Second, what will full disclosure reveal that the government does not already possess? If the answer involves conduct that expands the scope of the charges or the sentencing calculation, the cost must be weighed against the benefit with a precision that approximates, but cannot quite achieve, certainty. Third, is the defendant prepared to maintain consistency between the proffer and any subsequent testimony, argument, or defense strategy? If the answer is uncertain, the waiver provisions transform that uncertainty into a trial constraint that may prove irrecoverable.
A first consultation with this firm is without cost and carries no obligation; it is where this particular conversation begins, and where the decision is most productively examined.