Political geography impedes criminal prosecution of police misconduct, UB researchers say

BUFFALO, N.Y. – National data shows that police are rarely prosecuted for alleged misconduct. It is widely acknowledged that prosecutors are reluctant to charge law enforcement colleagues. After all, they rely on police for evidence, and for political support in elections. But research from the University at Buffalo School of Law and partners shows that political geography also plays an important role. 

UB law professors Guyora Binder and Anthony O'Rourke, along with University of North Carolina at Chapel Hill law professor Rick Su, observed that police staffing and arrests are concentrated in big cities. They reasoned that the voters most concerned about police misconduct would likely be found in such cities as well. Yet, they noted, prosecutors are typically elected at the county level, diluting the influence of these city-dwelling voters.  They wondered if this mismatch between the voters most affected by police and the voters who elect prosecutors could make a difference. Could election by city voters counteract the reluctance of prosecutors to prosecute their law enforcement partners? 

The research team examined the distribution of charges against police in the prosecutorial districts containing the nation's 100 largest metro areas. They report that the frequency of prosecution of police in a given prosecutorial district is predicted by the percent of district residents located in the largest city.

In the study, the authors state that “the influence of residents and police on prosecutorial decisions depends on the political geography of a prosecutor’s office.” They add: “Examining nationwide data on charging police, we find that prosecutors are indeed more likely to prosecute police when a greater proportion of their constituents are served by the same police department.”

The authors also investigated whether the racial and ethnic proportions of the district population also made a difference. They found that prevalence of Black residents increased rates of prosecution, but prevalence of Hispanic residents did not. They pointed to higher rates of voter registration, voting and mistrust of police among Black residents compared to Hispanic populations in explaining this effect. In fact, the prevalence of city residents in combination with the prevalence of Black residents better predicted prosecutions of police than either factor in isolation. Moreover, these relationships persist when controlling for measures of crime and police violence in the cities. 

In the Q&A below, Binder answers questions on their research and its implications for policy.

Can you describe a case study in which prosecutors charged police officers with crimes more so than other communities? Why did this happen?

We featured the two prosecutorial districts with the highest rates of prosecution of police in our dataset: Orleans Parish, Louisiana, and Baltimore City, Maryland. Both districts contain only their eponymous cities, with no suburban residents at all. Both have high rates of poverty and crime. Both have majority Black populations and long-established networks of Black-identified community and political organizations, including some focused on police brutality. Unsurprisingly, both have had several Black elected officials, including mayors and prosecutors. During the years covered in our statistical study, both had infamous killings of Black victims – the Danziger Bridge and Freddie Gray cases – that provoked charges against multiple officers. 

Yet New Orleans had even higher rates of police prosecution in the preceding decade. A popular Black mayor, Marc Morial, recruited a reformist Black police chief, Richard Pennington, to lead a notoriously violent, corrupt, and under-resourced force. Armed with a financial subsidy from the business community, Morial pushed a police budget increase through the city council. This allowed Pennington to offer the carrot of wage increases to his underpaid force. Morial also formed an alliance with a long-serving (white) prosecutor, Harry Connick Sr., who used the stick of prosecution to force resignation of rogue officers.  These stratagems enabled Pennington to hire new officers and promote new leaders to implement his reform agenda.

This is not a simple story of city constituents forcing a prosecutor to punish police.  Instead, city officials influenced the prosecutor to collaborate in achieving internal police reform. So favorable political geography and demography may expose prosecutors to constituent pressure to prosecute police; but some of that pressure may be channeled through other elected officials. And prosecution can affect police behavior even when it doesn't result in punishment.

Can you lead us through a case study in which prosecutors charged police officers with fewer crimes than in other communities? Why did this happen?

Here, we might usefully contrast the example of Fresno County, California. Fresno is a poor city, with high crime and a troubled police force. It is in the top third of major cities for killings by police and in the 82nd percentile for shooting at unarmed civilians. Hispanic  residents are three times as likely as other residents to be shot by police. In the period we studied, several high-profile killings by police occasioned large civil jury verdicts. Yet none of these occasioned any prosecutions. Indeed, Fresno County had the seventh-lowest rate of prosecution of police in our dataset. So, what's going on?

Well, city residents comprise only a little over half of county residents (in the bottom 40% of our dataset). While both city and county populations were majority Hispanic, both had very small Black populations. Fresno has never had a Hispanic mayor. Only 78% of eligible voters are registered, and elections are typically decided at the primary stage, with very low turnout. We found little in the way of community advocacy organizations. Indeed, the civil lawsuits against the police were organized by activists from out of town. County and city offices have long been held by a clique of leaders closely identified with law enforcement. The current prosecutor was endorsed by the city police union. The current mayor was previously police chief for 18 years and defended many of the controversial killings as justified. He was endorsed by the police union and the prosecutor. In sum, Fresno County is not an environment where we see a lot of democratic pressure for reform of police.

In some states prosecutors aren't elected at all, right?

Yes, in five states, including four on the East Coast, prosecutors are appointed by the state government. Here, rates of prosecution of police are not affected by the percent of city residents in the district. Some of these states assign prosecution of police to a prosecutor from a different geographic district to avoid conflict of interest. In New Jersey, prosecutors typically resolve these cases with resignations.

What implications does your research have for communities that want to hold police more closely to legal standards?

Our data suggests that suspected police misconduct would be prosecuted more often if more big cities had their own prosecutorial districts, separate from surrounding metropolitan regions. But such a reorganization of prosecutorial authority would typically require action at the state level – where the electorate is less likely to be concerned about police misconduct. And we should acknowledge that city-elected prosecutors have, at best, divided loyalties. They still have some incentives to favor city police. 

One way to reduce those incentives is to require that prosecution of police be conducted at the state level, by appointed prosecutors who neither work with, nor seek the electoral support of, local police. The drawback of this reform strategy is that such bureaucrats are ultimately answerable to a statewide electorate that will likely be less concerned about police misconduct than a city electorate. In sum, we probably cannot expect prosecution of police to play a leading role in police reform.