“Today,” Stuntz explains, “our cities are considerably more violent than before the great crime wave of the twentieth century’s second half, yet the nation’s imprisonment rate is quintuple the rate before that crime wave began. If punishment deters crime, we seem to be getting much less deterrent bang for the imprisonment buck than we once did. Add it all up, and the picture is quite different than the conventional wisdom allows.”

Stuntz’s thesis is that the misrule of politics has replaced the rule of law, with a ratchet of ever-expanding criminal laws giving boundless discretion to police and prosecutors, leading to a system that wrongly punishes too many poor young black men. When the law gives that much discretion, he writes, it stops functioning as law and instead becomes an assertion of power. The recent decline in crime is less a sign of success than of pathology. The encouraging numbers are misleading. They conceal devastating failure.

Stuntz writes, “Discretion and discrimination travel together.” The percentage of adults who are black, white, and Latino using illegal drugs is roughly the same (10 percent, 9 percent, and 8 percent, respectively), but blacks are three times more likely than Latinos to do prison time for drug crimes and nine times more likely than whites. Why? The misrule of politics, according to Stuntz. Specifically, the misrule results from suburban voters in counties having a lot of say in who gets elected as prosecutors in the urban areas where serious crime is concentrated. As Stuntz writes, prosecutors “are usually elected at the county level” and “counties that include major cities have a much higher percentage of suburban voters than in the past.” Think here, for example, of Fulton County, Georgia, or of Wayne County, Michigan, both so much larger than Atlanta and Detroit, respectively, that they even include some rural stretches. In other words, it is voters for whom crime is largely an abstract problem who exercise sway, while residents for whom the problem is real have less power.

The disappearance of the jury trial symbolizes this shift. Almost all felony criminal convictions today—96 percent—come from guilty pleas obtained by prosecutors elected with the support of suburban voters, not from verdicts reached by juries drawn from residents in areas where crime is concentrated. The system, in Stuntz’s words, has become an “arbitrary, discriminatory, and punitive beast,” which is undemocratic in vesting decisions about punishment in those who aren’t part of the community where those being punished live. Stuntz’s main remedies for this include putting more cops on the street, making more lawyers available to represent criminal defendants, letting local rules about sentencing prevail, and shifting responsibility between local and state governments for who pays for local police and state prisons.

More cops would mean fewer prisoners and more robust local democracy. More lawyers for criminal defendants would mean better-prepared cases, fewer coerced pleas, and more reliable outcomes. Letting local rules about sentencing prevail would reduce the severity and the racial disparity in sentencing, and, with judges presiding over this phase, reduce the power of prosecutors. Shifting responsibility for payment, by having local governments pay a larger share of prison costs and a smaller share of local police costs, would give them an incentive to sentence fewer prisoners—and remove a disincentive from hiring more cops.

Stuntz was troubled by “institutional design and incentives” in criminal law and politics that push toward ever harsher rules and sentences. Power over criminal law is allocated to the three branches of government—the legislature makes it, the executive branch enforces it, and the judiciary interprets it—but they are not checks on one another in this sphere. In fact, legislators and the executive branch’s prosecutors both benefit from “more and broader crimes”: Legislators get more power when they define crimes more broadly because they reduce the role of judges in deciding who is guilty; and prosecutors have more power because they have more discretion about what and how to prosecute. As a result, legislators and prosecutors tacitly cooperate with each other, leading to both more law and less: more on the books, and less on the street, in the sense that the laws are so broad the police and prosecutors get to decide whom to go after and find guilty. Those decisions are about power. In the “rule of too much law,” Stuntz advises, “too much law amounts to no law at all.”

His solution to this set of problems is to replace the vicious cycle that creates them with a virtuous cycle based on cultivating a relationship between those who break the law (or are tempted to) and those who enforce it. For most of the twentieth century in the Northeast and Midwest, the ratio of police officers to prison inmates was two to one. Today, it is less than one to two. “More than any other statistic,” Stuntz writes, “that one captures what is most wrong with American criminal justice.” More cops mean more deterrence. More deterrence means fewer arrests and fewer convictions. In the 1990s, New York City had the biggest drop in urban crime during the decade. It also had the biggest increase in its police force.

Another important component would be fewer prisoners. This would require reducing the severity of sentencing, which is now “more punitive than Russia’s,” reducing the discrimination that contributes to blacks outnumbering whites among prisoners, and reducing “excessive prosecutorial power”—which is “unchecked by law and, given its invisibility, barely checked by politics.” And too much power for prosecutors doesn’t mean there are enough of them: Stuntz calls for many more, so there are more lawyers to litigate cases and the pressure on them to obtain plea bargains is alleviated. That would also require more money for public defenders to represent defendants in court.

A more drastic aspect of his reform vision would be sweeping changes in criminal laws—defining more crimes vaguely so courts would need to resort to jury trials to decide who was guilty. This would excuse from liability for the most serious offenses the least guilty members of a group of criminals and would even allow some guilty defendants to claim that, though their conduct fit the definition of a crime, it wasn’t so “wrongful” that it merited punishment. This would mean “constitutionaliz[ing]” much of basic criminal law, by asking courts to define its boundaries instead of legislators and prosecutors—and giving courts more power when many perceive them to have too much power already.