In October 2015, James B. Comey, the director of the Federal Bureau of Investigation (FBI), called the government’s effort to track deaths during police interactions “unacceptable” and “embarrassing and ridiculous.” Since 2003, the FBI has used the Arrest-Related Deaths Program to track people who died during arrest or in police custody, but a 2015 Bureau of Justice Statistics report found that the program only accounted for about half of the expected number of law enforcement homicides in the United States. The finding is consistent with the realization that the FBI figures on police shootings were significantly lower than those in independent databases assembled by the Washington Post and the Guardian newspapers.
This lack of an open, centralized, government-run database on violent police encounters limits police departments’ ability to identify challenges and work to correct them and makes it difficult for academic and nonprofit organizations to study critical questions of police-community relations. Moreover, if the public does not have the data it needs to make “fair and informed judgments” about its police, that perceived absence of transparency can undermine trust and legitimacy.
To address this challenge, the FBI announced that it will replace and improve the system for tracking fatal police shootings by 2017. The new system will cover nearly 20,000 state and local law enforcement agencies and 685 medical examiner or coroner’s offices. It will track all incidents of officer-caused serious injury or death, including shootings, the use of stun guns, and pepper spray.
For this new system to be successful, the federal government must support local law enforcement agencies to track and report this information. Although the FBI currently requires all police departments to report violent encounter data to the Bureau of Justice Statistics, only three percent of the nation’s 18,000 state and local police agencies have done so since 2011. Many police departments around the country are small and lack the resources to track and report the necessary data. As of 2008, about half (49 percent) of all police agencies in the United States employed fewer than 10 full-time officers. An open source solution can help these departments open their data without overtaxing their limited resources.
Prior to the FBI’s announcement of a system upgrade, the White House launched the Police Data Initiative (PDI), a community of practice committed to improving the relationship between citizens and police through data and increased transparency. As of April 2016, 53 state and local police jurisdictions, covering more than 41 million people, had committed to the PDI, and collectively released over 90 datasets. While many state and local police departments have used the PDI to jump-start open data and transparency efforts, others that committed to PDI have not taken any action or published any new datasets.
The White House should support and build on the FBI’s effort by announcing the Police Data Initiative 2.0 (PDI2.0). Under PDI2.0, state and local law enforcement agencies and their supporting offices would make clear commitments to fully participate in the FBI’s initiative, publicly release the data, and incorporate the data into key elements of their operations, including department metrics and individual performance reviews. In exchange, the federal government would support local law enforcement communities by providing resources and technical assistance. The FBI and White House should seek opportunities to develop open source technology and data models, perhaps in partnership with nonprofit organizations, to give local law enforcement a free, user-friendly resource to meet this reporting obligation.
Developers can look to states and municipalities already moving ahead with similar programs in California, New Orleans, and Indianapolis. For example, in September 2016 California launched URSUS, an open source, all-digital police use-of-force data collection system developed by Bayes Impact, in all 800 law enforcement agencies around the state. Other law enforcement communities could adapt the code with minimal investment to suit their purposes.
Opening data on police violent encounters in this way will enable government, nonprofits, and academics to explore critical questions about police interactions in America, set goals for improvement, and help law enforcement agencies measure themselves against those goals.