In Minneapolis, police union president Bob Kroll made headlines for warning against a “rush to judgment” of the officers responsible for George Floyd’s death — even as video of Floyd’s death was splashed across social media. The comment was in-character for Kroll, who has opposed even the mildest reforms and praised Donald Trump for rolling back moderate Obama-era regulations and thereby letting “cops do their job, put[ting] the handcuffs on the criminals instead of us.”

More recently, despite New York mayor Bill DeBlasio denying the existence of patent police brutality caught on camera, the president of the second-largest police union in the city, Ed Mullins, has criticized him for allegedly muting the police response to protests. Mullins even posted DeBlasio’s daughter’s personal information on Twitter.

For many critics police unions are the chief obstacle to fighting police abuse, and Kroll and Mullins exemplify the worst of police unionism: a proud willingness to defend even the worst agents of police violence at any cost, and to gum up investigations and lawsuits and prevent penalties when abuse occurs. Critics’ solution: Abolish police unions.

Others, however, worry that union-busting the police invites broader attacks on public-sector unionism. Many critiques of police unions — that they are too powerful, that they bargain for compensation that costs cities too much and that their interests intrinsically contradict the public interest — have been leveled against other public-sector workers like teachers and nurses. How can we remedy the harms police unions have promoted and perpetrated for decades without undermining public-sector unionism as a whole?

The history of police unionism offers one solution: strict limits on when police can acceptably resort to violence.

For the first half-century after police began organizing around 1900, police unions were illegal in most of the United States, especially after the suppressed Boston police strike of 1919. City governments feared the negative political repercussions of permitting the independent organization of the agents of state violence. In some cities, these de jure bans did not prevent police associations from lobbying over compensation and working conditions. But overall, these associations were relatively weak.

This picture changed in the 1950s and ‘60s, when police unions first won widespread recognition and collective bargaining rights, organizing in response to political reformers who worked to rein in police corruption and the police abuse that fell heaviest on the backs of black communities. For instance, under Mayor Richardson Dilworth in Philadelphia, a hotbed of mid-century police reform, scores of officers were fired for corruption, and the city established the nation’s first civilian review board.

Mayors like Dilworth were not “anti-police” — they believed they were making departments better at fighting crime, and they often paired reforms with increased funding — but police unions aggressively opposed these measures as the meddling of overbearing reformer-managers who misunderstood or actively undermined their work. They bristled at civilian oversight, seeing it as uninformed, inappropriate and demeaning.

Like Kroll and Mullins today, postwar police unionists framed themselves as victims, “second-class citizens” “handcuffed” by politicians more interested in stripping them of their rights than in letting them fight crime. They claimed to have their hands tied during the urban uprisings of the 1960s, prevented by civilian authorities from restoring order.

Both feeding and riding the rising wave of law-and-order politics, police unionists used this framing to win contracts, legislation and court decisions that shielded their members from oversight and discipline. Some provisions limited the window to file a complaint and blocked public access to personnel files, preventing the reporting or investigation, let alone disciplining, of abuse.

Police unionists claimed to need the protection of these unions and these laws to effectively safeguard “law-abiding citizens” from the threat of crime. Crime, however, was always a nebulous and flexible concept. It could be manipulated to justify almost unlimited police authority.

In Philadelphia, the Fraternal Order of Police (FOP) turned these strategies into political advocacy that locked reformers out of the mayor’s office for 16 years and shut down the civilian review board.

With few reversals, in towns and cities across the country, this steady march toward police independence has continued from the 1960s to the present day. Its endpoint is a world in which, due to the virtually unlimited license for police violence, police may account for up to 10 percent of homicides in the U.S. each year.

There have been, however, a few instructive exceptions to this overall trend. While police abuse crested in 1970s Philadelphia under mayor and former police commissioner Frank Rizzo, black Philadelphians and a multiracial coalition around them helped push his successor, William Green III, to institute substantial police reforms beginning in 1980. One of the most important reforms was a more restrictive use of force policy.

Commissioner Morton Solomon enforced these limits strictly and consistently despite “low morale” from some of the rank and file and vigorous opposition by the FOP. Officers who violated the policy met swift and clear punishment, up to and including firing and criminal charges. The number of those killed by police annually fell from a high of 21 in 1978 to just four in 1982, and the number of those wounded fell from 67 to 22 in the same period. This outcome showed a powerful and determined social movement could fight successfully to curb police violence, but the policy in no way ended police abuse in Philadelphia. And the FOP has continued to erect barriers to accountability and discipline. Philadelphia’s current police union contract, for instance, guarantees the expunging of reprimands from an officer’s file if they aren’t disciplined for any other infractions for two years.

What lessons can we draw from this history?

First, in many ways police unions are unexceptional: They fight for better pay, job safety and limits to management’s power to discipline. But if police unions are typical in this way, police work is not, distinguished as it is by the legal right to use violence — often to address social concerns that such violence only perpetuates or exacerbates. This right should be subordinate to the public interest in regulating and limiting it. But police unions have rejected this principle for nearly a century, campaigning for unlimited latitude to use violence with little accountability to anyone outside the police department.

Organizers and legal scholars have proposed various solutions to this problem, but one of the clearest is for states and cities to set legal standards for the use of force and for reporting, investigating and punishing police abuse, both within departments and in the courts. Nurses’ and teachers’ unions are bound by legal standards that do not undermine them and that they generally embrace as integral to effective practice. As the case of Philadelphia in the early 1980s shows, such limits can work for police, too. And if passed by law, unions do not have to sign off on these limits

Right now, many organizers and politicians are campaigning to reallocate police funding to other social services that promote security and safety. But short of full police abolition, we will always grant some public servants the right to use state-sanctioned violence. Statutory limits on that right are a crucial tool for placing the public interest first, and come with fewer drawbacks than taking away police rights to collective bargaining.