Blog
Can the Government Break a Proffer Agreement?
The government can break a proffer agreement. It does so with regularity, through mechanisms that are not concealed in the agreement but printed in the same typeface as the protections, on the same page, in clauses that most defendants do not pause over until the session has ended and the room has been vacated.
Federal Rule of Evidence 410 was intended to prevent this. The rule shields statements made during plea negotiations from admission at trial, on the theory that candor in negotiation serves the system more than the advantage of using those statements against the person who made them. The Supreme Court affirmed this principle in Kastigar v. United States, holding that use and derivative use immunity is coextensive with the Fifth Amendment privilege. But the proffer agreement is, if we are being precise, not an immunity grant at all. It is a contract. And the government drafts the contract.
The Structure of the Proffer Letter
A proffer letter is a document of seven or eight pages, signed before a meeting at the United States Attorney’s office, in which the defendant agrees to provide truthful information about criminal activity in exchange for a limited promise: that the government will not use the defendant’s own words against the defendant at trial. That promise is real. It is also narrower than it appears.
The letter contains a second clause, one that most defense attorneys recognize but that many clients do not absorb on the day of signing. This clause permits the government to make derivative use of the information provided during the session. If the defendant mentions a co-conspirator’s address, the government may obtain a warrant for that address. If the defendant references a bank account, the government may subpoena its records. The evidence recovered through those derivative leads is admissible at trial, because the letter says it is, and because the defendant agreed to a waiver of Kastigar protections as a condition of the meeting.
A third clause, standard in most districts, provides that if the defendant later offers testimony or evidence at trial that contradicts statements made during the proffer, the government may introduce the proffer statements in rebuttal. After United States v. Mezzanatto, the Supreme Court confirmed that this waiver is enforceable so long as it was entered knowingly and voluntarily. The Second Circuit, in United States v. Velez, extended the principle to cover not merely the defendant’s own testimony but any evidence or argument offered by the defense. The practical effect is that a defendant who proffers and then proceeds to trial must construct an entire defense around the contents of a conversation that was supposed to be protected.
What the Waiver Provisions Permit
The waiver clause transforms the proffer from a protected conversation into something closer to a recorded deposition in which only one side controls the transcript. A defendant who proffers and then proceeds to trial cannot present a defense inconsistent with what was said in that room without triggering the government’s right to introduce the proffer statements themselves. Defense counsel, aware of this constraint, must shape the trial strategy around the contents of the proffer. The constraint is not theoretical. It forecloses arguments, limits cross-examination, and reduces the available defenses to those that do not contradict a single sentence spoken on the day of the session.
In our experience, the constraint operates most forcefully in cases where cooperation fails. The defendant proffers in good faith. The government concludes the information is insufficient, or that the defendant was not forthcoming on every point, or that the case does not warrant a plea offer. No agreement is reached. The session ends. The defendant returns to the posture of an adversary, except that the adversary now possesses the defendant’s own account of events. But the waiver survives.
The asymmetry is structural. The defendant’s statements are shielded from direct use but exposed to derivative use and conditional rebuttal. The government’s obligations amount to a promise not to introduce the defendant’s words in its case in chief, a promise that dissolves the moment the defendant presents evidence the government considers inconsistent. One federal district judge in the Eastern District of New York, writing in United States v. Duffy, concluded that permitting the government to extract these waivers as a precondition to negotiation would compromise the fairness of plea bargaining and federal trials. That opinion did not carry the day.
Whether the court intended this outcome or merely failed to prevent it is a question worth considering.
We approach proffer preparation differently than the standard guidance suggests. Before a client enters that room, we conduct a detailed assessment of what the government already knows from independent sources, what derivative leads the proffer might generate, and whether the potential benefit of cooperation outweighs the evidentiary exposure that follows. In something like three out of five consultations, the conclusion is that the proffer should not occur at all. The client is better served by declining the invitation. The clerks in the Southern District will sometimes accept a letter of declination without further inquiry; others will not, and the government’s response to a declined proffer varies by office and by the individual prosecutor involved.
The government does not need to violate the letter of the agreement to break it in substance. The derivative use clause is, in most circuits, sufficient to permit the government to construct a case out of leads generated by the defendant’s cooperation. In a First Circuit mortgage fraud prosecution, the defendant’s proffer statements were used to establish the factual basis for a search warrant targeting a co-conspirator’s property. The records obtained through that search were introduced at the defendant’s own trial. The court found no breach. The agreement permitted derivative use, and the court saw no reason to look further.
The Moment the Agreement Fails
The proffer agreement fails most visibly when it should have succeeded: when the defendant cooperated in good faith, provided truthful information, and expected a plea offer that never materialized. The letter does not promise a resolution. It promises a conversation. The government retains sole discretion (a phrase that appears in nearly every proffer letter and that means precisely what it says) over whether to file a 5K1.1 motion for a downward departure, and even cooperation that produces indictments of co-defendants may result in no benefit if the prosecution determines the information fell short of its expectations.
In 2019, before the wave of circuit opinions clarifying the scope of Mezzanatto waivers, there was a period of genuine uncertainty about whether derivative use clauses could reach the government’s case in chief. That uncertainty has resolved in the government’s favor, though a few district courts have continued to resist.
The defendant who has proffered and received nothing in return has lost ground. The government possesses the defendant’s admissions, the leads those admissions generated, and in many cases a map to evidence it would not have obtained without the session. The defendant possesses a letter that promises not to quote the defendant’s words during trial, subject to exceptions that encompass most of the ways a defense can be conducted. The letter is signed by both parties. It was entered into voluntarily.
Remedies and Their Limits
When the government does breach the agreement, as distinct from exercising its permitted rights under the agreement, the remedy depends on the nature of the breach and the circuit. Santobello v. New York established that a prosecutor’s promise, once made as consideration for a guilty plea, must be honored, and that even an inadvertent breach does not diminish the obligation. The remedies the Court identified were specific performance of the agreement or withdrawal of the plea.
The difficulty with proffer agreements is that the defendant has not yet entered a plea. The proffer is a pre-plea conversation. If the government uses information in violation of the agreement’s express terms, the defendant’s recourse is a motion to suppress, or a request for a Kastigar hearing at which the government must demonstrate that its trial evidence derives from sources independent of the proffer. But most proffer letters contain an explicit waiver of the right to a Kastigar hearing. The defendant agreed, on the day of the session, that no such hearing would occur. Courts have enforced these waivers with little hesitation.
There is a category of breach that the case law has not addressed with precision: the government that enters a proffer session without genuine interest in reaching an agreement, using the session to gather information under the appearance of negotiation. The Third Circuit acknowledged this concern in passing, and at least one law review article has argued that such conduct constitutes bad faith sufficient to void the letter entirely. That argument has not prevailed in a published opinion, and I am less certain about its prospects than some of the academic literature suggests.
The Contract Written in One Direction
A proffer agreement is a contract in which one party drafts every term, the other party signs under conditions that are coercive by definition, and the interpretation of breach runs in a single direction. The government decides whether the defendant was truthful. The government decides whether the information was valuable. The government decides whether to file the motion for a reduced sentence. The defendant decides only whether to sign.
None of this means the proffer is never worth entering. In cases where the government’s evidence is already substantial and the trial exposure is severe, cooperation through a proffer may represent the most rational course. The question is not whether to cooperate but whether to cooperate with a full understanding of what the agreement permits, what it withholds, and where its protections thin out.
The waiver clause is the provision most clients overlook on the day of signing. It should be the first provision discussed with counsel, not the last.
A first consultation costs nothing and commits the client to nothing. It is the beginning of an assessment, not an agreement.