Correction appended

Some of the most shocking images to emerge from the demonstrations that have dominated recent headlines stem from violent interactions between law enforcement officers and peaceful protesters. They’ve also escalated calls for police reform.

But police unions tend to be resistant to such efforts, as their mandate is to protect the interests of their members — even in cases when those interests may be counter to democratic norms and values. Though an understudied topic of criminology, what research that does exist is unequivocal: “Virtually all of the published items that express an opinion on the impact of police unions regard them as having a negative effect, particularly on innovation, accountability, and police — community relations,” as a review in the journal Police Practice and Research put it.

Researchers say unionized officers draw more excessive-force complaints and are more likely to kill civilians, particularly nonwhite ones. Here are some key findings:

Unionization emboldens violent officers

A recent University of Chicago working paper found violent misconduct among sheriff’s officers increased about 40 percent after a state supreme court ruling allowed the officers to unionize. The incidents examined in this paper are among the most serious types of violent misconduct, including sexual assault and excessive force.

It’s worth noting the baseline numbers of these types of incidents are very low, such that the 40 percent increase translates into roughly one additional violent incident per sheriff’s office every five years.

Certain union-negotiated contract provisions — including time limits on misconduct investigations, expungement of misconduct records, and mechanisms allowing officers to challenge disciplinary findings — make it more difficult to detect and punish officers who abuse their position, the researchers say.

Additionally, the authors write, unionization “may increase solidarity among officers and thereby strengthen a code of silence that impedes the detection of misconduct.”

Use-of-force complaints more likely among unionized officers

A 2006 report from the federal Bureau of Justice Statistics found unionized police agencies garnered 9.9 use-of-force complaints for every 100 officers, compared with 7.3 for non-unionized agencies. During the disciplinary process, about 7 percent of those complaints were sustained, or found to have merit, in unionized agencies. In nonunion agencies, the sustain rate was more than double, at 15 percent.

In effect, officers in unionized police forces are more likely to be the subjects of an excessive-force complaint, but more likely to beat the allegations in disciplinary hearings.

Lengthy appeals processes make it more difficult to fire ‘bad apples’

Writing in the University of Pennsylvania Law Review, Stephen Rushin analyzed 656 police union contracts to examine the role of the disciplinary appeals process in misconduct cases.

“The median police department in the data set offers police officers as many as four layers of appellate review in disciplinary cases,” he found. Some provided six or seven layers of review. After those levels are exhausted, most departments then allow officers accused of misconduct to appeal to a third-party arbitrator. More than half gave the offending officers some control over the selection of the arbitrator.

The result, as detailed in a 2017 Washington Post investigation, is that a stunningly high percentage of officers fired for misconduct are eventually rehired after a lengthy appeals process. In Washington, D.C., for instance, 45 percent of the officers fired for misconduct from 2006 to 2017 were rehired on appeal. In Philadelphia, the share is 62 percent. In San Antonio, it’s 70 percent.

Other contract provisions also shield police from accountability

In a separate paper in the Duke Law Journal, Rushin analyzed 178 police union contracts and found a number of provisions that played a role in shielding police from the consequences of misconduct, including provisions that “limit officer interrogations after alleged misconduct, mandate the destruction of disciplinary records, ban civilian oversight, prevent anonymous civilian complaints, indemnify officers in the event of civil suits, and limit the length of internal investigations.”

He found that “overall, 156 of the 178 police union contracts examined in this study — around 88 percent — contained at least one provision that could thwart legitimate disciplinary actions against officers engaged in misconduct.”

Police unions advocate shielding disciplinary records from public view

Writing in the Stanford Law & Policy Review, Katherine Bies notes that “police disciplinary records are public in only 12 states,” due in no small part to lobbying efforts by police unions.

The article deals specifically with the efforts of police unions to pass laws in two states — New York’s notorious Section 50-a and a similar law in California — that shield disciplinary records from public scrutiny.

“Police unions often strategically frame any opposition to their agenda of secrecy as endangering public safety and harming the public interest,” Bies writes. “However, police unions often conflate ‘the public interest’ with the private interests of police officers.”

Unionized police may be more likely to kill civilians, particularly nonwhite ones

Economist Rob Gillezeau recently previewed his research examining the relationship between unionization and police killings of U.S. citizens. While provisional, his initial results suggest the police unionization happening in the 1950s through the 1980s led to “about 60 to 70″ additional civilians killed by police each year. The “overwhelming majority” of those civilians were nonwhite.

“With the caveat that this is very early work, it looks like collective bargaining rights are being used to protect the ability of officers to discriminate in the disproportionate use of force against the nonwhite population,” he recently said on Twitter.

Correction: Due to an editing error, a previous version of this story incorrectly characterized the rate of use-of-force complaints for unionized police agencies. The rate is 9.9 complaints for every 100 police officers, not 9.9 out of every 100 complaints.