Blog
Proffer Agreement vs. Immunity Agreement: Key Differences
The proffer agreement is the most misunderstood document in federal criminal practice. Not because the language is complex, though it often is, but because the name suggests a protection that the document does not, in its operative terms, actually confer. Clients arrive at our office using the words “proffer” and “immunity” as though they refer to the same instrument. They do not. The distance between the two is the distance between a handshake and a court order.
A proffer is a contract. An immunity grant is a constitutional mechanism, authorized by statute and imposed by judicial authority. One is negotiated over a conference table in an Assistant United States Attorney’s office; the other is compelled through 18 U.S.C. § 6002 and, in many cases, authorized by a federal judge. The protections differ not in degree but in kind, and the consequences of confusing them are severe enough that the confusion itself constitutes a form of harm.
The Proffer Session
The colloquial name for the proffer agreement carries an irony most clients do not appreciate until afterward. “Queen for a Day” implies a position of authority. The reality is closer to an audition.
In a proffer session, the defendant or subject of an investigation sits in a room with their attorney, the assigned AUSA, and one or more federal agents. The defendant provides information. All of it. Partial disclosure is treated as a breach. The session operates under a letter, signed by both parties, that provides limited use immunity: the government agrees not to introduce the defendant’s own statements in its case in chief. That limitation sounds broad. It is not.
The proffer letter does not prevent the government from pursuing investigative leads generated by the defendant’s statements. If you describe a transaction in your proffer and the government then subpoenas records of that transaction from a third party, those records are admissible. Your words are protected. The consequences of your words are not. The Supreme Court in Kastigar v. United States drew this distinction in 1972, and every standard proffer letter since has incorporated what practitioners call the Kastigar waiver: a clause permitting derivative use of the proffered information.
There is a further risk that most proffer letters now contain, one that receives less attention than it deserves. If any element of the defendant’s trial defense is inconsistent with what was said during the proffer, the government may introduce the proffer statements for purposes of impeachment. The scope of “inconsistent” is determined, in practice, by the prosecution. The proffer session was supposed to be a preliminary conversation. The letter makes it something closer to a deposition with consequences that arrive on delay.
The cross examination of government witnesses, the defendant’s own testimony, even the theory of the case, all become subject to a comparison with what was said in that room. A defendant who proffers and then proceeds to trial has constrained the range of available defenses, sometimes without realizing the constraint existed until preparation is underway.
We have taken the position, in cases where the proffer letter’s impeachment clause is drafted broadly enough to encompass cross examination, that the letter should be renegotiated before the session occurs. Not every AUSA will agree to narrower terms, but the request itself signals that defense counsel has read the operative language rather than the heading.
Statutory Immunity Under Federal Law
Formal immunity operates on entirely different authority. Under 18 U.S.C. § 6002, the government may compel a witness who has invoked the Fifth Amendment to testify, provided that the testimony and any evidence derived from it cannot be used against the witness in a subsequent criminal proceeding. The immunity is not negotiated. It is ordered, and the witness who refuses to comply faces contempt.
The federal system provides use and derivative use immunity, not transactional immunity. The distinction matters considerably. Use and derivative use immunity prevents the government from employing the witness’s compelled statements, or any investigative fruits of those statements, in a prosecution of that witness. But the government remains free to prosecute if it can demonstrate, in what is called a Kastigar hearing, that all of its evidence was derived from sources independent of the compelled testimony. The burden rests on the prosecution, and it is, as the Court noted, a heavy one.
Transactional immunity, by contrast, bars prosecution for any offense related to the compelled testimony regardless of the source of the evidence. It is broader than the Fifth Amendment requires. Some states still offer it. The federal system, since the Organized Crime Control Act of 1970, does not. Whether this represents a rational accommodation of competing interests or an insufficient substitute for the constitutional privilege is a question the Court addressed in Kastigar and has not revisited with any enthusiasm since.
The Distance Between a Letter and a Statute
The difference between a proffer agreement and a formal immunity grant is not procedural. It is structural, and the structure reveals the asymmetry.
A proffer agreement is a contract. Its terms are whatever the parties negotiate, which in practice means whatever the government drafts and the defendant agrees to sign. Proffer letters vary across districts and across individual prosecutors. Some are protective. Some contain provisions that, if the cooperation fails, leave the defendant in a position worse than silence would have. There is no statutory floor for a proffer letter’s protections, no judicial oversight of its terms, and no requirement that the letter be approved by anyone other than the signing AUSA. The letter is, if we are being precise, not a grant of immunity at all. It is a contractual limitation on the use of specific statements, governed by the principles of contract law, subject to breach determinations that the government itself often controls.
Formal immunity is authorized by statute, approved by senior officials within the Department of Justice, and imposed by court order. The witness does not negotiate the terms. The terms are the statute. And the statute’s protections, once imposed, carry the weight of the Fifth Amendment itself. The prosecution’s burden in a Kastigar hearing is not contractual. It is constitutional.
This distinction produces consequences that are predictable in the abstract and specific in practice. Consider the defendant who proffers, provides information that leads to the indictment of a coconspirator, and then finds that no plea agreement materializes. The defendant proceeds to trial. The government has generated new evidence (witnesses identified, documents located, accounts traced) through the derivative use of the proffer. The proffer letter permitted this.
Now consider the same defendant who had received formal use immunity under § 6002. The government would bear the burden of proving, for every piece of evidence at trial, that its source was independent of the immunized testimony. The prosecution’s entire case would need to survive a Kastigar hearing. The difference between these two scenarios is the difference between a defendant who volunteered information under a contract and a defendant whose testimony was compelled under constitutional protection.
I am less certain about whether this distinction has been appreciated by the circuits (the ones where the AUSA describes the process, where the word “immunity” is used loosely, where the client begins to believe the two instruments are one) though the outcomes in cases where cooperation collapses suggest it has not. The defendant who proffered under a letter and the defendant who testified under a § 6002 order face materially different risks, but the early conversations tend to obscure that difference until it is too late to reclaim it.
The proffer letter says your words will not be used against you. It does not say your words will not be used.
There is a particular silence that follows the reading of a proffer letter in a conference room. The client has heard the protections. The AUSA has described the process. And somewhere between what was promised and what was understood, the client discovers a difference that matters.
When the Proffer Collapses
Most proffers, in our experience, do lead to cooperation agreements. The government does not waste resources on sessions it does not intend to follow through on, though something like a quarter of the sessions we have participated in over the past several years have produced no agreement at all. That number is approximate, and the reasons vary.
When a proffer does not produce a cooperation agreement, the defendant holds a document whose protections are thinner than they appeared. The proffer statements cannot be used in the government’s case in chief. But the government has heard the defendant’s account. It has had time to identify inconsistencies. It has pursued derivative leads. The defendant has, in effect, provided the government with a map and then been told the map cannot be introduced at trial, but the destinations it revealed are admissible.
The exposure compounds in multi defendant cases. The decision to proffer follows the logic of the prisoner’s dilemma: if one defendant proffers while the other does not, the silent defendant faces the cooperator’s testimony at trial. Both have reason to proffer. But if the government declines the cooperation offer after the session, the defendant who spoke has weakened their position relative to the one who did not. The calculus is rational in advance.
The question a client poses, almost without exception, is whether they should cooperate. The answer depends on facts that are never available in full at the time the question is posed: the strength of the government’s case, the value of the client’s information, the reliability of the assigned AUSA, the jurisdiction’s sentencing practices, whether coconspirators have already begun cooperating. The decision to proffer is made under conditions of imperfect knowledge, and the proffer letter does not compensate for what is not known.
What the letter does, and all it does, is define the narrow terms under which speech will be exchanged for a limited restriction on use. It is a beginning, not a resolution. The formal immunity agreement, when it arrives, is a different instrument, carrying different authority, different protections, and different consequences for the government that granted it. One is an opening offer. The other is a constitutional event.
A consultation is where these distinctions become specific to a set of facts. A first conversation with our office costs nothing and assumes nothing; it is the point at which the general becomes particular, and the particular is where the law resides.