Blog
Proffer Sessions With the FBI vs. the U.S. Attorney's Office
The proffer session at the United States Attorney’s Office and the interview conducted by FBI agents are not the same event, though a surprising number of defense attorneys permit their clients to treat them as though they were. One occurs under a written agreement that confers limited use immunity. The other occurs under no agreement at all, in a room where every sentence becomes evidence the moment it leaves the speaker’s mouth.
The call from the government typically arrives in one of two forms. An Assistant United States Attorney contacts defense counsel and proposes a formal proffer session, accompanied by a letter. Or an FBI agent contacts the client directly, sometimes by telephone, sometimes at the front door of a residence or place of business, and requests an interview. The first scenario involves a document. The second involves a handshake, a badge, and a request that sounds more like an invitation than a legal proceeding. The client who does not perceive the difference between these two encounters has already conceded ground that may not be recoverable.
The letter is where the distinction begins.
The Proffer Agreement and Its Terms
A proffer agreement is a written contract between a federal prosecutor and the individual under investigation or indictment. The agreement, sometimes called a “Queen for a Day” letter, establishes that the individual will provide information to the government and that the government will not use the individual’s own statements as direct evidence in its case in chief at trial. The agreement is signed before the session commences, typically by the client, defense counsel, and the AUSA assigned to the matter.
What a proffer agreement does not confer is transactional immunity, which would prohibit prosecution for the underlying conduct entirely. The Supreme Court in Kastigar v. United States drew this line in 1972: use and derivative use immunity satisfies the Fifth Amendment, but it is not a pardon. The distinction matters because it defines the outer boundary of what the letter actually protects. The agreement describes its protections in terms that can appear more generous than the protections actually are. The client hears “your statements cannot be used against you” and perceives a shield. The agreement means something narrower, if we are being precise about what the agreement actually prohibits: your statements cannot appear in the government’s case in chief, unless you breach the agreement, commit perjury, or later take a position at trial inconsistent with what you said in the session.
The proffer agreement offers a form of protection. That protection is thinner than most clients perceive it to be.
Derivative use is where the real exposure resides, and it is the provision that most clients do not fully comprehend until the consequences have already materialized. If a client mentions a business partner, a bank account, or a transaction during the proffer session, the government can pursue that lead independently, and anything the government discovers through that independent pursuit is admissible. The government can and will follow those leads, subpoena those records, and interview those individuals. A standard proffer letter contains what practitioners refer to as a Kastigar waiver, by which the individual agrees to forgo the right to a hearing at which the government would otherwise need to demonstrate that its evidence derives from sources independent of the proffer statements; this waiver, which most clients sign without understanding its full procedural implications, effectively removes one of the few remaining checks on how the government uses what it learns in the room, because the hearing it waives is the very mechanism that Kastigar established as the safeguard against indirect use of immunized testimony. Whether the court intended this as a safeguard or merely failed to address the gap is a question that remains, at this writing, unresolved. Most proffer letters in circulation contain that waiver, though the Second Circuit and the Southern District have produced conflicting guidance on how far it extends.
The Room Without a Letter
FBI agents do not require a proffer agreement to conduct an interview. They do not need one. What they need is a willing speaker, and the architecture of the encounter is designed to produce one.
Two agents arrive, usually at a place and time not of the client’s choosing. The tone is conversational. The agents introduce themselves, explain that they are conducting an investigation, and ask whether the individual would be willing to answer a few questions. There is no letter. There is no written immunity of any kind. There is no prosecutor present. There is, in a formal sense, nothing resembling the structured proceeding that occurs at the U.S. Attorney’s Office. The interview sits in a kind of procedural wilderness: the individual speaks freely, and every word is governed by the same federal criminal statute that governs sworn testimony, but none of the protections that attach to formal cooperation apply.
The agents write their own account of the conversation on a form called an FD-302, sometimes hours or days after the meeting, in language that reflects what they believe was said rather than what was.
You sign the letter and then you discover what the letter permits. But in an FBI interview, there is no letter at all. There is no document defining the terms of the exchange. The individual speaks, the agent (who, it should be noted, has in most instances already reviewed financial records, communications, and prior witness statements before arriving at the door, and who is therefore not seeking information so much as testing whether the individual’s account conforms to what the government already possesses) listens and takes notes, and the FD-302 becomes the official record.
Three of the last four clients who came to us after an FBI interview had already spoken for over an hour without counsel present. In each instance, the client believed the conversation was preliminary, exploratory, something less than a formal proceeding. In each instance, the government treated the conversation as a completed interview.
The agents do not record the session.
Section 1001 and the Shared Exposure
Under 18 U.S.C. § 1001, a materially false statement to any federal agent constitutes a felony carrying a sentence of up to five years. The statute operates identically whether the conversation occurs under a proffer agreement at the U.S. Attorney’s Office or on a Thursday afternoon in the lobby of a Marriott.
This is the provision that collapses the distinction between the two settings into something more unsettling. The proffer agreement may protect the individual from the direct use of truthful statements. It does not protect the individual from the consequences of an inaccurate one. If the government concludes that a statement made during the proffer session was false, the agreement is voided, and the individual faces prosecution for the false statement itself under Section 1001 in addition to whatever charges the government was already contemplating. The same exposure exists in the FBI interview, with the additional problem that the interview is not recorded, no written agreement defines the scope of the conversation, and the sole record of what the individual said is the FD-302 composed by the agents after the fact.
The cases that illustrate this pattern are familiar. Martha Stewart was not convicted of securities fraud. She was convicted of making false statements to investigators about the conduct the government was examining. The conviction arose from the interview, not from the trade. General Flynn entered a guilty plea for statements made to FBI agents, not for the underlying conduct that prompted the investigation. In both cases, the interview produced the charge. The underlying investigation produced nothing, or nothing sufficient.
A misremembered date, an approximate figure offered in good faith, a denial of attendance at a meeting that cell tower records later contradict: each of these has formed the basis for a separate federal prosecution.
The question is whether the person sitting across from the agents understood the stakes before the first answer was given.
Deciding Whether to Sit Down
The invitation to proffer arrives, in most of the cases we have handled, before the client has retained counsel at all. The agent calls. The AUSA sends a letter. The client, who may have been aware of the investigation for weeks or months, perceives the invitation as an opportunity to explain. The instinct to cooperate is not irrational. It is, however, premature.
Before we permit a client to enter any proffer session, we conduct an independent investigation of the matter that is, in some respects, as thorough as the government’s own. We interview the client at length. We obtain and review the documents the government is likely to possess. We identify the witnesses the government has probably already contacted. The purpose of this process is not to prepare the client to perform well in the session, though preparation matters. The purpose is to determine whether a proffer session serves the client’s interest at all, or whether it merely provides the government with information it could not otherwise obtain. The standard advice in the field is to cooperate early and cooperate fully. We do not follow that approach in every case, because we have observed instances in which early cooperation provided the government with leads it would not have developed on its own timeline, and in which the client received no meaningful consideration in return.
A reverse proffer, in which the government presents its evidence to the defense rather than the other way around, can sometimes reveal whether the government’s interest in cooperation is genuine or performative. Most prosecutors who conduct reverse proffers do so because they believe the case is strong. The reverse proffer is designed to demonstrate that strength, not to solicit information. We have, on occasion, requested a reverse proffer before committing a client to a formal session, because the government’s willingness to show its evidence is itself a form of information.
Whether a client should accept the government’s invitation depends on factors that cannot be reduced to a general rule, which is itself part of the problem with most of the advice circulating on the subject.
The client who has already spoken to agents without counsel has a different set of problems than the client who has received a target letter and is weighing cooperation for the first time. The client whose co-defendants have already proffered faces a different calculus than the client who would be the government’s first cooperator. The client who is a subject faces different exposure than the client who is a target. Each of these distinctions alters the analysis. I am less certain than the preceding paragraphs might suggest about how often early cooperation produces outcomes that justify the risk, though the cases in which it did tend to share a common feature: the client’s information was both truthful and valuable to an ongoing investigation that extended well beyond the client’s own conduct.
A first consultation assumes nothing and costs nothing; it is the point at which strategy begins to replace speculation.