Blog
Is It Dangerous to Cooperate in a Federal Case?
The Asymmetry at the Center
Cooperation in a federal case is not a negotiation. It is a sequence in which the defendant acts and the government evaluates.
Under USSG Section 5K1.1, a federal court may impose a sentence below the guideline range only if the government files a motion attesting that the defendant provided substantial assistance in the investigation or prosecution of another person. The court cannot act on its own initiative. The defendant cannot file the motion. Defense counsel cannot file it. The authority rests with the prosecution alone, and the Supreme Court confirmed as much in Wade v. United States in 1992: the government has a power, not a duty, to file when a defendant has substantially assisted.
The practical consequence of this structure is that a defendant who agrees to cooperate performs every obligation first: the debriefings, the testimony, the disclosure of information that may incriminate friends, family, or business partners. The government observes, records, and then determines (the determination itself being subject to almost no judicial review, save for claims of unconstitutional motive such as race or religion) whether the assistance rises to the level of “substantial.” That word carries no fixed definition in the guidelines. The government supplies its content after the cooperation has already been performed, and most defendants who enter cooperation agreements do not perceive this asymmetry until sentencing. The evaluation belongs to the prosecution, and the defendant learns the result at sentencing.
The Proffer Session and Its Architecture
Before any cooperation agreement is signed, the government will request a proffer session.
The proffer, sometimes called a “Queen for a Day” agreement, is a meeting in which the defendant sits with prosecutors and federal agents, typically with counsel present, and provides information about the criminal conduct under investigation. The agreement typically states that the defendant’s statements will not be used directly against them. But it permits the government to pursue leads obtained from the session and to use the statements for impeachment if they later prove inconsistent.
The statute is not entirely clear on whether this protection extends to all derivative evidence, which is part of the difficulty. What the proffer agreement does not protect against is the government’s ability to follow the threads of what was said and construct a case from what those threads reveal. In one case from the Eastern District, a defendant who described a billing arrangement during a proffer identified, without intending to, a co-conspirator the government had not yet contacted; that co-conspirator’s testimony became the centerpiece of the trial. The defendant’s own words were never introduced. The witness those words produced was introduced, and the result was a conviction.
There is a second risk that alters the calculus of every cooperation decision. Once a defendant has proffered, going to trial becomes a different proposition. Under Federal Rule of Evidence 410, proffer statements cannot be used at trial, with one exception: if the defendant testifies and says anything inconsistent with the proffer, those statements become admissible for impeachment. A defendant who has sat through hours of questioning about events that may have occurred years earlier has, in effect, locked in a version of the facts. Any deviation at trial, even one produced by the ordinary degradation of memory, can be used to destroy credibility. Three hours of questioning can produce sentences that, years later, foreclose an entire defense.
Our practice begins the cooperation analysis before the proffer, not during it. We assess the strength of the government’s existing case, the value of what the client can provide, and the specific terms of the proffer letter before a word is spoken in that room. We begin there because everything that follows depends on what was said in that room.
Timing and the Problem of Information
In multi-defendant cases, the first person to cooperate receives the most favorable consideration. The person who acts second provides testimony the government may already possess.
I am less certain about the reliability of timing as a strategy than most writing on this subject suggests. The general principle holds: prosecutors value information they do not yet have, and the first defendant to provide it occupies a position of advantage that diminishes with each subsequent cooperator. But the calculation is not as clean as game theory implies. A defendant who cooperates early with incomplete knowledge of the government’s case may provide less than expected. A defendant who waits may discover that the government’s interest has shifted to a different target entirely.
The information asymmetry runs in both directions. The defendant does not know what the co-defendants are doing. The co-defendants do not know what the defendant is doing. The government knows what everyone is doing and calibrates its offers accordingly. Whether the court intended this incentive structure or merely failed to prevent it is a question worth considering.
What Happens When Cooperation Fails
In 2018, before the second sentencing, Paul Manafort’s cooperation agreement with the Special Counsel’s office collapsed. He had agreed to cooperate truthfully and completely. The government’s position was that he had not. Judge Amy Berman Jackson found that he had made intentionally false statements and ruled the Special Counsel’s office was no longer bound by its obligations under the plea agreement.
The consequence was not simply the loss of a sentencing benefit. By entering the cooperation agreement, Manafort had already pleaded guilty. He had already provided statements in debriefing sessions. The plea stood. The admissions stood. The cooperation credit did not. Partial honesty, in this context, is treated as dishonesty.
The breach did not merely eliminate the benefit of cooperation; it supplied the prosecution with admissions it could use at sentencing without any corresponding obligation. This is the risk that cooperation agreements carry within them, the way a contract reads differently once you have already signed it. The defendant who cooperates and fails to satisfy the government is in a worse position than the defendant who never cooperated at all, because the cooperating defendant has spoken and cannot unspeak.
It requires, if we are being precise, not just honesty but a particular quality of honesty that most people do not practice in ordinary life: the willingness to say things that cause harm to oneself without qualification, omission, or self-protective shading. Most people, under pressure, minimize. Prosecutors know this. Their questioning is designed to surface minimization, and any minimization, once identified, can be characterized as a false statement under 18 U.S.C. Section 1001. The agreement does not say so, but that is what it means.
The Costs That Do Not Appear in the Agreement
Physical safety is the concern most defendants raise first, and it is not an unreasonable one. In cases involving organized crime or drug trafficking organizations, the label of cooperator carries real danger, and the government’s capacity to provide protection is not unlimited. The U.S. Marshals Service operates the Witness Security Program, and its track record with participants who follow program guidelines is strong. But WITSEC requires relocation, a new identity, and separation from everyone who is not enrolled. For most defendants, that trade is not one they are prepared to consider.
Courts do recognize the danger a cooperator faces, and it is one of the five factors a judge considers when determining the extent of a sentencing departure.
The strain on marriages, business partnerships, and family relationships, which begins at the moment of arrest and intensifies the moment cooperation becomes known, is considerable. Nobody includes that in the sentencing memorandum.
When Cooperation Is the Correct Decision
There are cases in which cooperation is not merely advisable but, given the weight of the evidence, the only rational path. When the government possesses recordings, documents, and corroborating testimony sufficient to secure a conviction at trial, and when the defendant holds information about other participants that the government does not yet have, cooperation can reduce a sentence by years. In drug cases carrying mandatory minimums, a Section 5K1.1 motion combined with a Section 3553(e) motion is often the only mechanism that permits a sentence below the statutory floor.
What matters is whether the decision to cooperate or to decline was reached with complete information about what each path entails, which in our experience is rarely the case by the time a defendant first contacts an attorney. The question is not whether cooperation is dangerous. The question is whether, in a given case, the danger of cooperating exceeds the danger of not cooperating, and that determination requires a specific assessment of the evidence, the co-defendants, the jurisdiction, and the terms being offered.
We evaluate the cooperation question before the first meeting with prosecutors, not after. We assess what the government likely knows, what the client can provide, and what the proffer letter permits and prohibits, and we do this before the client is in a room with federal agents. The sequence matters because the conversation with prosecutors, once it begins, cannot be taken back.
- Assess the strength of the government’s evidence independently.
- Identify what information the client possesses that the government does not.
- Review the proffer letter for derivative use and impeachment provisions.
- Evaluate physical safety and relational consequences.
The federal system does reward cooperation in many cases. A first conversation assumes nothing and costs nothing; it is the point at which a diagnosis becomes possible.