Second-Best Criminal Justice


Criminal procedure reform can be understood as a “second-best” enterprise. The general theory of second best applies where an ingredient necessary for a “first-best” ordering is unattainable. That’s an apt description of the contemporary criminal process. Our normative ideals of criminal justice require fair and frequent trials to judge guilt or innocence, but the criminal trial rate has been falling for at least a century; today it is vanishingly close to zero. What may be even worse is how we’ve eliminated trials—by endowing prosecutors with enough leverage to coerce guilty pleas. Excessive prosecutorial leverage is the source of some of criminal procedure’s deepest pathologies.

This Article asks the reader to accept—as a thought experiment—that a negligible trial rate is a constraint on criminal procedure reform in the near term. From that starting point, the crucial question becomes whether there is a less destructive way to ensure a negligible trial rate. There is: inefficiency. The road to a more just, humane, and rational criminal process could begin with making formal criminal litigation more inefficient. In matters of institutional design, the general theory of the second best counsels using unseemly practices, like inefficient procedure, to offset fixed constraints, like the absence of criminal trials. If the formal process of criminal litigation could be made unreasonably expensive for both parties, both would want to settle to avoid it. Policymakers would then be free to dismantle the tools of prosecutorial leverage—overlapping offenses, draconian sentencing laws, punitive pre-trial detention, and more—without worrying about increasing the trial rate. The result would not achieve our criminal justice ideals—no second-best solution can—but it could be better than the status quo. Without more trials, it may be the best we can do.


Criminal justice, Plea bargaining, Prosecutorial leverage, Trial constraints, Plea reform, Litigation costs, Second best



William Ortman (Wayne State University)



Publication details



All rights reserved

File Checksums (MD5)

  • pdf: d39b4ecdb071b1f78ee91c0861cc99c9