Recent protests, catalyzed by the murder of George Floyd in Minneapolis, call for an end to racist police violence. With their actions, the protesters have also moved beyond many of the stale policing debates of the recent past. Defund, disband, abolish—people who would never have even heard these words in discussions about the police are now seriously considering them.

The breakthroughs in the police debate would not have been possible without the protesters, who have remained steadfast despite being beaten and abused by police everywhere in the United States. But this is not about making a breakthrough in the debate. This is about life and death. To stop police from killing people, 1,000 a year, year after year, changes will have to be made to the system. The protesters will be vindicated only if the changes made are the right ones.

Reform programs will only be successful if they start from the premise that the policing institution has lost its social legitimacy, which it never deserved. Reforms that assume police legitimacy—whether they involve more body cameras, better oversight, a more diverse force, or more prosecution of killers among the police—do not do the job.

Once the police are viewed as an illegitimate institution, we are well on our way. As Mariame Kaba argues in the New York Times, we could do worse than to make a 50 percent cut to police budgets and let the logic of austerity get the job done, as it has with the rest of the public sector. But 50 percent can be bargained down to 10 percent, and 10 percent to 2 percent, as long as police and their advocates can continue to link public safety with policing. The backlash against abolishing police as “politically unrealistic” in light of public safety has begun at the local level where the issue is being debated.

The goal has to be to abolish the class of people who have the legal right to end lives (and to lie to you while you must tell the truth).

Do police currently have the right to kill? Absolutely. Using conservative estimates and public data, writer Lee Camp calculated that police killed an average of 900 people per year—in other words, at least 12,600 people from 2005 to 2019. In this period, Camp writes, a total of three police officers were convicted of murder and had those convictions stand up to appeal. That is less than one-tenth of one percent, but it rounds off cleanly to zero.

The license to kill, above all, must be taken away from police. It survives because of a mystique—helped by ubiquitous cop shows, books, and movies—which is based on three notions: the idea that they are courageous because their job is dangerous, the idea that they keep society safe, and the fact that you can call them in an emergency.

Courage? Yes, policing is the 16th most dangerous job in America, behind logging workers, fishing workers, pilots, roofers, refuse collectors, truck drivers, farmers, steelworkers, construction workers, landscapers, power-line workers, grounds maintenance workers, agricultural workers, construction trade helpers, and the first-line supervisors of mechanics, installers, and repairers. But no worker in any of the top 15 most dangerous jobs has the option of killing when they subjectively feel unsafe—police do.

Safety? Police have no special function keeping society safe. In Alex Vitale’s book The End of Policing, he cites criminologist David Bayley’s earlier book Police for the Future, in which Bayley calls this one of the “best kept secrets of modern life. Experts know it, the police know it, but the public does not know it.” We have known for 50 years that police don’t help public safety. French anthropologist Didier Fassin, in his 2013 book Enforcing Order, cites the Kansas City experiment of the 1970s:

“This unprecedented study, unique at the time, compared three zones of the city: in the first, ‘reactive,’ crews limited their activity to responding to residents’ calls; in the second, ‘proactive,’ they at least doubled the time they spent on patrol; in the third, serving as a ‘control’ zone, they continued their previous mixture of activity. The results of a full year of operations and measurement appeared identical: neither attacks on persons, whether assault and battery, sexual assaults or muggings, nor attacks on property, whether burglary or damage to vehicles, varied significantly as a result of the different systems implemented; similarly, the experience of crime and the feeling of insecurity as reported by residents and business owners showed no variation between the zones, nor did the level of satisfaction with the police; and it turned out that in all three cases, 60 percent of the officers’ time was spent on activities not directly related to law enforcement, including a quarter bearing no relation at all to police work… Ultimately, it was evident that the patrols used preventatively had no effect on crime, either in terms of offenses recorded by law enforcement or from the point of view of residents’ perception of risk.”

The results were ignored: police kept patrolling for the next five decades. Fassin, who hung out with Paris police as part of his study, made his own calculations of how they spent their time:

“In my experience, time spent in response to calls often amounted to approximately 10 percent of the shift time; it was rare that it rose to 20 percent (five calls per team per night shift was a maximum that was rarely reached), with the rest of the time being devoted to random patrols, and to the administrative recording of actions taken.”

Think Paris is anomalous? Think again:

“A number of studies conducted in the United States reveal that officers on patrol spend between 30 and 40 percent of their time responding to calls (on average five calls per team per hour in cities), only 7–10 percent of which are related in some way to offenses or crimes, and between 40 and 50 percent of their working day on random patrol and paperwork, with the rest of the time devoted to various tasks.”

Here’s how Fassin describes the daily work of the police he observed:

“Cruising around quiet streets and peaceful neighborhoods, the police wait for occasional calls that almost always turn out to be pointless, either because they relate to mistakes or hoaxes, or because the teams arrive too late or bungle the case due to their clumsiness, or because there is no cause for any official questioning or arrest.”

Fassin cites a criminologist from Ontario, Richard Ericson, who in 1982 found that police spent 76 minutes out of an eight-hour shift responding to calls, arguing that “the presence of police officers has become an end in itself.”

So, police have the 16th most dangerous job, and they are irrelevant for public safety—but society needs someone to call in an emergency. This role can be filled by trained civilian workers who will have to learn to solve daily social problems without a license to kill, which could be the direction Minneapolis goes given city councilors’ vow to disband the police in their city. Last year Canada’s Globe and Mail reported about a police force in Yukon that carried no weapons and could lay no charges. Some cities have child protection workers that operate to protect children with greater or lesser degrees of intrusiveness. Social workers can be trained to intervene in domestic disputes and in active mental health crises in the field. They can be deployed in teams that ensure one another’s safety, like in other professions. There are detailed proposals for taking responsibility for safety into the hands of the community—Olúfẹ́mi O. Táíwò describes one in Dissent Magazine; Zach Norris reframes the issue in his new book We Keep Us Safe; and Ejeris Dixon and Leah Lakshmi Piepzna-Samarasinha describe community approaches to safety in their edited volume Beyond Survival: Strategies and Stories from the Transformative Justice Movement.

There should be some cultural reforms too. Boots Riley suggests removing police and military consultants, who act like state censors, from movie and TV production. The #MeToo movement led to the creation of an intimacy coordinator role in film production, to ensure that sex scenes could be filmed without sexist exploitation. Studios could be responsive to this movement by drastically reducing their cop show production while removing censors for the shows that remained. This would go some way to reduce police mystique and police worship.

Police advocates may argue there might be some financial losses from abolition. Some police forces “eat what they kill” through civil asset forfeitures, fines, and tickets, keeping taxes low while making poor people’s lives miserable. Overall, however, money will be saved.

Initially, much of the money saved by defunding police would have to go to easing the transition of the people currently in police roles into other jobs. Pensions are a mechanism for taking police off the line for whatever reason, which police organizations use generously indeed. But pensioning every police officer indefinitely, while it would save lives, would not make any resources available for public safety. Instead, governments can provide retirement and retraining packages (the courageous police might look at retraining for one of the 15 more dangerous jobs) for them, as they do with other workers who are laid off.

For the duration of current union contracts, police could be paid to prepare for other jobs or simply to stay home—expensive in the short term, but it would save thousands of lives. After that initial period, the hundreds of billions of dollars that are spent on policing could be redirected to create and expand public services. The possibilities would be limited only by the number of billions that could be moved from police. Social workers, certainly, are a strong candidate to redirect funds toward, as well as free transit and other free basic services (especially health care in the United States).

Criminological data has told us for decades that police are irrelevant for public safety. Other data tells us a lot about what does influence safety. British researchers Richard Wilkinson and Kate Pickett in their classic 2009 book The Spirit Level show that a large number of social problems, including violence, correlate strongly with inequality. Their work also shows different options for achieving equality: high wages by private employers (as in Japan) or high taxes and redistribution (as in Northern Europe). In the United States, every option for increased equality has been blocked by the wealthy who have—as Martin Gilens and Benjamin Page make clear in their important 2014 study—captured politics. A real Green New Deal would do more for public safety than any conceivable police reform short of abolition.