How many people do police in the United States kill each year, and under which circumstances? These are among the many basic questions raised by the fatal police shooting of African-American teenager Michael Brown in Ferguson, Missouri, last year. The short – and startling – answer is that nobody knows.
A new report from the US Bureau of Justice Statistics (BJS) goes a long way in explaining why. Since 2003, BJS has compiled data from the states to produce an annual national count of people killed during arrest or in police custody, under the Arrest-Related Deaths program. But according to the new report, large variations in how states collect and report their data have “resulted in a significant underestimate of the annual number of arrest-related deaths.” As a result, the US collects data on only 35-50 percent of arrest-related deaths nationally, and even this number may overestimate coverage. Moreover, the current data collection system fails to provide useful data for examining trends in death over time, geography, demographics, or any other variable.
This matters because without accurate data on the deceased, the law enforcement agency involved, and the circumstances around these incidents, law enforcement leadership and policymakers will be hard pressed to reduce the number of future deaths.
The public uproar following the Brown killing and other cases prompted the US Congress in December to pass a law clarifying the definitions of the data states are required to produce and adding a small financial penalty for non-compliance. It now explicitly requires data on those killed while “detained” by law enforcement (for example, for questioning), and not just during arrest or in police custody. (Note that Michael Brown was not in the process of being arrested or in police custody when he was killed, so his death wouldn’t have been counted under the previous parameters.)
But the updated law, while an improvement, doesn’t go far enough, since it does nothing to address the “obstacles to collecting valid and reliable information” described in the BJS report.
If those obstacles are to be surmounted, the Justice Department needs to provide clear guidance to the states, through BJS, on how to generate accurate and valid data. The financial penalty for state non-compliance should be contingent upon following such guidance. Otherwise, the law will not produce its intended results.
Documenting the number of people killed by law enforcement is not an esoteric problem. There is a methodological solution that requires only political will, in the form of resources and collaboration across agencies and jurisdictions, to implement. The last thing the US needs is more bad data on how many people are killed while in police custody.
