Federal prosecutors wield enormous power. They have the authority to make charging decisions, enter cooperation agreements, accept pleas, and often dictate sentences or sentencing ranges. There are currently no effective legal checks in place to police the manner in which prosecutors exercise their discretion. As a result, in the current era dominated by pleas instead of trials, federal prosecutors are not merely law enforcers. They are the final adjudicators in the 95% of cases that are not tried before a federal judge or jury. In a government whose hallmark is supposed to be the separation of powers, federal prosecutors are a glaring and dangerous exception. They have the authority to take away liberty, yet they are often the final judges in their own cases. One need not be an expert in separation-of-powers theory to know that combining these powers in a single actor can lead to gross abuses. Indeed, the combination of law enforcement and adjudicative power in a single prosecutor is the most significant design flaw in the federal criminal system. Although scholars have made persuasive cases for greater external controls on prosecutors, these calls for reform are unrealistic in the current political climate. The solution must be sought elsewhere.

This Article looks within the prosecutor's office itself to identify a viable corrective on prosecutorial overreaching. In particular, by heeding lessons of institutional design from administrative law, this Article considers how federal prosecutors' offices could be designed to curb abuses of power through separation-of-functions requirements and greater attention to supervision. The problems posed by federal prosecutors' combination of adjudicative and enforcement functions are the very same issues raised by the administrative state—and the solutions fit equally well in both settings. In both instances, individuals who make investigative and advocacy decisions should be separated from those who make adjudicative decisions, the latter of which should be defined to include some of the most important prosecutorial decisions today, including charging, the acceptance of pleas, and the decision whether or not to file substantial assistance motions. Using this model from administrative law would not only be effective, it would also be more politically viable than the leading alternative proposals for curbing prosecutorial discretion.