Before there was the outrage of George Floyd’s death during his arrest for allegedly passing a counterfeit $20 bill, there was Freddie Gray. Five years ago, Gray died after being taken into custody by the Baltimore police for being in possession of a knife. This, in a city that failed to clear 65 percent of its homicides between 2007 and 2017.

Both cases, and too many more we could all name, are a symptom of what the late criminologist Mark Kleiman called the “randomized draconianism” of a criminal justice system that simultaneously under- and over-polices its citizens.

It’s a concept that’s key to understanding why, if we want to break the vicious, hopeless cycle in which we are trapped, the answer isn’t less policing; indeed, we might need more, but certainly much smarter, policing.

One inevitable result of radically uneven law enforcement is suspicion of cops in communities subject to the heaviest policing. Perhaps less predictably, ever more aggressive penalties aren’t even a very good way to lower crime. As Kleiman pointed out in his book, “When Brute Force Fails: How to Have Less Crime and Less Punishment,” unlikely but extra-severe punishment doesn’t much alter criminal behavior because perpetrators don’t expect to get caught. The decades of attempting to solve soaring crime rates with escalating harshness were not just a humanitarian disaster but also a policy failure; we could have gotten better results with milder punishments applied more consistently.

Keep that in mind as we push forward with police reform, because to control police brutality, we do need to control crime. Too often, such reforms follow a vicious cycle: a justifiably outraged public demands that police be brought to heel, new policies and restrictions are put in place, cops pull back from active policing out of either anger or a genuine fear that any mistake could cost them their jobs, and crime rises, so a frightened public demands greater law enforcement.

Unquestionably, we do need better policing. A system that simply lashes out at a subset of lawbreakers, often in ways that are wildly disproportionate, produces deep and lasting harms. In its place, we need a highly trained, professional police force and “swift, certain and fair” punishment that can actually change the incentives lawbreakers face.

Fortunately, over the past 60 years we’ve learned a lot, from Kleiman and others, about making criminal justice simultaneously more effective and less harsh by making punishment more likely, though often much milder. But to make such reforms work, we’d need police reform that embodies some of those very same principles: a system that offers swift, certain punishment for all violators, based on clear guidelines for community interactions, especially the use of force, and longer training in putting those guidelines into practice — instead of occasionally descending on a bad cop who happens to get caught on video.

That twin-pronged approach is our best chance at meaningful long-term reform. Police unions, however, have successfully resisted greater accountability for decades, simply because cops are a powerful single-issue constituency on protecting police privileges and compensation, while the public has many diverse concerns. If police wait long enough, public opinion will turn to some other crisis, and they’ll regain the upper hand.

Reform is thus more likely to stick if we co-opt the unions rather than trying to break them. Instead of “defund the police,” what if we offloaded the nonjudicial parts of their work, like dealing with the homeless and the mentally ill, to social workers, and then “stuffed their mouths with gold” to reform the policing part? We could offer a significant salary boost in exchange for accepting stricter standards and oversight, which wouldn’t just ease the political obstacles, but possibly attract higher-quality candidates to the police force.

But here we will run into a second, larger barrier: At the moment, no one wants to be seen doing or saying anything that sounds even a little bit like rewarding the police for being abusive.

This is not, of course, what I’m suggesting. It’s about producing greater professionalism and much tighter accountability, which must include firing and, where warranted, prosecuting the abusive cops. Yes, the police unions might take the money, resist the professional reforms and stall the accountability, but if it’s really impossible to design a reform they can’t undermine, then we should give up now.

I’m not ready to do that. And currently, we have a narrow window of opportunity before the public turns its attention to something else, which means we need to move quickly as well as boldly.

At this critical juncture, we can’t afford to mirror the great mistake of the 20th-century “law and order” types who thought they could fix the soaring crime rates simply by treating violators worse and worse. Since then, we’ve learned that we get further, faster and at much lower human cost if we focus not on our anger, but on changing the context and incentive structure in which criminals flourish — on both sides of the law.

Read more: