Blog
Cooperation in Multi-Defendant Federal Cases: Strategy Considerations
The decision to cooperate is not, in any meaningful sense, a legal question. It is a question of sequence: who moves first, what they reveal, and whether the information retains its value by the time the government receives it. In a multi-defendant federal case, the architecture of cooperation resembles an auction more than a negotiation, and the commodity being traded (the defendant’s knowledge of others’ conduct) depreciates with each passing week.
Most defendants arrive at the cooperation question too late. They arrive after the initial shock of indictment has subsided, after the discovery materials have been reviewed, after counsel has assessed the government’s evidence and delivered a preliminary opinion. By that point, in cases involving three or more defendants, at least one co-defendant has already entered a proffer session. The information that person carried into the room has reshaped the government’s understanding of the case, and every subsequent cooperator offers something the prosecution already possesses, in whole or in part.
Whether the court intended this asymmetry or the system merely produced it is a question worth considering.
The Proffer Session
The proffer session takes place in a government office with a structure that is deceptively informal. The defendant sits with counsel. Across the table are the assigned AUSA and one or more investigating agents. The agents ask questions. The defendant answers. The session is documented but not recorded in most districts, a fact that becomes relevant if the cooperation fails and the government’s characterization of what was said diverges from the defendant’s recollection.
The written proffer agreement, which must be signed before the session begins, contains a provision that is both the session’s primary protection and its principal limitation. The government agrees not to use the defendant’s statements in its case in chief. That phrase, “case in chief,” occupies a narrower territory than most defendants perceive. It does not encompass rebuttal testimony. It does not extend to sentencing proceedings. It does not prevent the government from using the statements to impeach the defendant if the defendant later testifies in a manner the government considers inconsistent with the proffer.
The agreement permits derivative use. The government cannot introduce the defendant’s words at trial, but it can follow every lead those words provide, locate witnesses the defendant identified, obtain documents the defendant described, and construct a case from the information the defendant chose to disclose. A proffer that fails to produce a cooperation agreement leaves the defendant exposed to prosecution informed by everything the defendant shared, assembled through sources the government might never have located without the defendant’s assistance.
We approach proffer preparation with attention to the sequencing of disclosure rather than only the substance of the information itself. A proffer session is not a debriefing. It is, if we are being precise, a controlled exchange in which the order and framing of each disclosure determines whether the government perceives the defendant as a valuable witness or an unreliable narrator.
Derivative Use and Its Consequences
The distinction between direct use and derivative use is where most cooperation agreements encounter their most serious problems. Direct use immunity prevents the government from playing the defendant’s own words back to a jury. Derivative use is the permission to treat those words as a map.
In practical terms, this means the government can take a defendant’s description of a meeting that occurred on a specific date, subpoena records from that date, interview attendees the defendant named, and present those records and that testimony at trial without reference to the proffer. The evidence appears to have been gathered through independent investigation. The defendant’s role in its discovery is invisible.
Six months after signing a proffer agreement that produced no cooperation deal, a defendant may discover that the case against them has grown stronger. Not because new criminal conduct came to light, but because the government now possesses evidence it would not have found without the proffer. The agreement did what it promised. It did not do what the defendant needed.
I am less certain about the uniformity of this risk across districts than the preceding paragraph might suggest. Some U.S. Attorney’s offices exercise restraint in derivative use; others treat the proffer as an open license to investigate. The written agreement does not vary much, but the culture of the office determines how far it extends the permission.
The Cooperator Queue
In a case with five defendants, the first to cooperate carries information the government does not yet possess. The second carries some of the same information, supplemented by details the first cooperator either did not know or did not disclose. By the third, the government’s need has diminished considerably. By the fourth, the cooperation is ceremonial in character: the defendant admits guilt in exchange for a Section 5K1.1 motion whose magnitude reflects the marginal value of cumulative testimony.
The Sentencing Commission’s data on substantial assistance departures confirms what practitioners observe in individual cases. The volume of departures is not trivial, but the distribution of benefit is uneven. Early cooperators receive the most substantial reductions.
The difficulty is that no defendant knows their position in the queue. The government does not announce that a co-defendant has entered a proffer session. Defense counsel for co-defendants do not share this information, and joint defense agreements, to the extent they exist, do not require disclosure of cooperation discussions. A defendant weighing the decision to cooperate is making a calculation with a critical variable absent: whether the information they intend to offer has already been offered by someone else.
Three cases in the past eighteen months, all in the Southern District, involved defendants who entered proffer sessions believing they were first in line. They were not. The government had already secured cooperation from a co-defendant (who, it should be noted, entered that proffer session without knowledge that counsel for the remaining defendants had been operating under a joint defense agreement for several weeks and had already shared considerable strategic information across defense teams). The proffer sessions proceeded, and the information those defendants provided was received as corroboration rather than as original intelligence. The 5K1.1 motions filed on their behalf reflected that diminished position.
Joint Defense Agreements and the Moment They Collapse
A joint defense agreement permits co-defendants and their counsel to share information, coordinate strategy, and extend the attorney client privilege to communications between their respective lawyers. The agreement functions when all parties share a common interest in contesting the government’s case. It ceases to function, and begins to inflict damage, the moment one party’s interest diverges.
In 2000, the Ninth Circuit vacated the convictions in United States v. Henke on conflict of interest grounds, but the decision’s significance extends beyond the relief it granted. In Henke, Silicon Valley executives charged with conspiracy and securities fraud entered an oral joint defense agreement and shared confidential information across defense teams. When one defendant, Surendra Gupta, accepted a plea agreement and agreed to testify for the government, defense counsel for the remaining defendants found themselves unable to cross-examine Gupta. His trial testimony contradicted statements he had made during joint defense meetings, but the duty of confidentiality the agreement imposed prevented effective impeachment. The defendants were convicted. The structural problem the case exposed has not been resolved by the passage of time.
The lesson of Henke is not that joint defense agreements are inadvisable. The lesson is that any such agreement drafted without provisions addressing the withdrawal and cooperation of a member is a document that protects the defendant only until the moment protection is needed. We include specific clauses addressing this contingency:
- What happens when a member cooperates with the government
- What information can be used for cross-examination of a cooperating former member
- Whether remaining members waive the right to seek disqualification of counsel
- How the transition from allied party to adverse witness will be managed
Most agreements we review from other firms do not contain these provisions.
There is a particular silence in the conference room when counsel announces that a co-defendant has signed a cooperation agreement. The joint defense calls, which may have occurred weekly for months, cease without explanation. Documents that were shared become potential exhibits. The strategy that was constructed collectively now belongs, in part, to the government’s witness.
The mathematics of joint defense exposure are straightforward. An agreement among two defendants means one potential cooperator who possesses knowledge of your defense strategy. An agreement among seven means six. The larger the joint defense group, the more information circulates, and the more devastating the defection of any single member becomes.
Whether a joint defense agreement should be entered at all depends on the relative exposure of the parties, the strength of the government’s case against each defendant individually, and the likelihood that any member of the group will receive a cooperation offer that exceeds in value what the joint defense provides. In cases where the disparity in exposure is significant, the defendant with the least to lose has the most to gain from cooperating. That defendant is also the one the government approached first.
And the question that governs all of this, the one that ought to be asked before the joint defense agreement is even drafted, is whether the case is one in which collective resistance serves every defendant’s interest or only the interest of the most exposed. Most of the time, the answer to that question is visible in the indictment itself, if one reads it with attention to the distribution of culpability rather than the distribution of charges.
We do not enter joint defense agreements in cases where the exposure differential among defendants exceeds a certain threshold, a practice that has, in something like forty percent of the cases where we have declined, proved to be the correct decision within six months. The standard advice is to join and monitor for signs of defection. Our experience suggests that monitoring is less effective than selection.
When Cooperation Is Not the Answer
The presumption in federal practice is that cooperation benefits the defendant. The presumption is often correct. It is not always correct.
Cooperation requires an admission of guilt. A defendant who cooperates pleads guilty, typically to the most serious charge, and provides a complete account of their own criminal conduct before offering information about others. If the evidence against the defendant is weak, or if the charges involve conduct where the line between aggressive business practice and fraud is not a line at all but a region, cooperation trades away a defense that might have succeeded.
A defendant who cooperates and whose cooperation the government later determines was not substantial receives the worst possible outcome: a guilty plea without a 5K1.1 motion. The plea stands. The admission stands. The sentence is calculated without the benefit of a departure.
The evidence regarding how often this occurs is not precise, but practitioners observe it with regularity. A meaningful number of defendants who enter cooperation agreements do not receive the benefits they expected. Some because their information was less valuable than anticipated. Some because the government’s assessment of “substantial” changed between the signing of the agreement and the filing, or non-filing, of the motion. For defendants whose information concerns targets the government has already indicted, cooperation may reduce a sentence by months rather than years. Whether that reduction justifies the risks is a calculation that can only be conducted with full knowledge of the specific circumstances.
Timing and the Shape of the Decision
The single variable that most influences the outcome of a cooperation decision is one that receives insufficient attention in most initial consultations: when. The distinction between a defendant who cooperates in the second month after indictment and one who cooperates in the eighth month is not merely procedural. It determines the government’s valuation of the information, the scope of the cooperation agreement, and the magnitude of any eventual departure.
The calculation is complicated by the fact that the decision to cooperate is, in most cases, irreversible. A defendant who proffers and fails to reach a cooperation agreement cannot un-proffer. The information has been disclosed. The leads have been provided. What remains is a defendant whose position is weaker than it was before the proffer session began.
What the system produces, in aggregate, is a set of incentives that reward speed over deliberation. The defendant who cooperates early receives the greatest benefit. The defendant who considers the decision with appropriate care may find, by the time consideration has concluded, that the benefit has been claimed by someone who moved first. The tension between analysis and action is, in this firm’s view, the central strategic difficulty of multi-defendant federal cases. We do not resolve it by defaulting to speed. We resolve it by beginning the cooperation analysis before the indictment is returned, where that is possible, so that the decision can be both informed and timely.
Consultation is where this conversation begins. A first call requires nothing beyond the willingness to describe the situation, and it initiates the one analysis that cannot be conducted without counsel: whether the facts of the case, the identity of the co-defendants, and the posture of the government favor cooperation, resistance, or a sequence of considerations that has not yet been assembled.