How many people have been killed by the police?

Dr. Q

How many people have been killed by the police over the past few decades?

It may seem shocking, but the only honest answer to this question is that no one really knows.

The government is very good at tracking the number of police officers who are killed while at work. Every year, the FBI publishes its Law Enforcement Officers Killed and Assaulted (LEOKA) report which includes a fairly accurate estimate of the number of police officers who were killed the previous year.

Unfortunately, the government is not very good at tracking the number of people who were killed by police officers. I was able to find government research with some information about the number of people killed by the police for every year from 1976 up through 2011. Unfortunately, as we will see, all the estimates are afflicted by under-reporting.

After reviewing the research, I found that from 1976 through 2011, police killed at least 14,012 people.

When I say “at least,” I mean just that. The figure I came up is not intended to be an answer to the question posed at the beginning of this article. As I already mentioned, the government’s publicly available research on the number of people killed by police is affected by under-reporting, so the 14,012 figure should only be seen as a floor for the number of people who were killed. The actual number, which we can’t estimate with any certainty, is much higher.

The estimates I used came from a number of different sources. In some cases, I was able to find more than one source with an estimate for a particular year. In these cases, I used the highest available estimate.

The estimates I used for 1976 through 1994 (just over half the years I found data for) all came from a 2001 report by the Bureau of Justice Statistics (BJS) titled “Policing and Homicide, 1976-98.” The estimates included in this report only count homicides committed by police that were determined to be legally justified by a government agency. Furthermore, data from some states is missing for a number of years.

I believe the estimates taken from this report are the least accurate of those I was able to find. The estimates for 1995 through 1998 which were included in the report are lower than the estimates for these years taken from the FBI’s UCR data and, as we will see, the FBI’s estimates are lower than the figures from BJS’s “Arrest-Related Deaths” reports.

Year Est. Missing data
1976 415
1977 311
1978 313
1979 442
1980 457
1981 381
1982 376
1983 406
1984 332
1985 321
1986 298
1987 296
1988 339 FL, KY
1989 362 FL
1990 379 FL
1991 359 FL
1992 414
1993 453 KS
1994 459 KS
1995 382 KS, IL
1996 355 KS, DC
1997 361 FL, KS
1998 367 FL, KS, DC, WI

For 1995 through 2002 as well as 2010 and 2011, I used data from the FBI’s Uniform Crime Reports. As I observed before, the FBI’s estimates for 1995 through 1998 were higher than those in BJS’s “Policing and Homicide” report. For the other years I used the FBI’s data for, the FBI was the only source I was able to find. When I recorded the FBI’s estimates for a particular year, I always used the most recent UCR report that listed the estimate for that year because the FBI sometimes revises their estimates with more accurate figures. As the with “Policing and Homicide” report, the FBI only estimates the number of people who were “justifiably” killed by the police, so we know for a fact that their figures do not account for all the people killed by the police.

Year Est.
1995 389
1996 357
1997 366
1998 369
1999 308
2000 309
2001 378
2002 341
2003 373
2004 367
2005 347
2006 386
2007 398
2008 378
2009 414
2010 397
2011 393

The Uniform Crime Reports included data for 2003 through 2009, however, I was able to find higher estimates using the Bureau of Justice Statistics’ Deaths In Custody Reporting Program (DCRP). This probably has to do with the fact that DCRP includes homicides ruled to be unjustified as well as those ruled to be justified.

In the report “Arrest-Related Deaths in the United States, 2003-2005,” the BJS offers an interesting table (Appendix table 3 on page 12) which shows how many homicides each state reported to the BJS for the DCRP and how many each state reported to the FBI. Sometimes the FBI’s estimate for a state was higher than the DCRP’s which makes it clear that at least some states were failing to report homicides to the DCRP.

If you use take the highest figure for each particular state from the table, you get a total of at least 1,489 people killed by police from 2003 through 2005. If you relied only on the DCRP estimates, you’d get a total of 1,095. If you relied only on the FBI’s estimates, you’d get a total of 1,082. This means that if you relied on one of these programs for estimates, you’d be missing more than a quarter of the homicides committed by police (at least).

Year FBI BJS Combined est.
2003 371 366 501
2004 368 365 492
2005 343 364 496

In the report “Arrest-Related Deaths, 2003-2009,” the BJS provides estimates for 2006 through 2009, but did not include a table with the state-by-state breakdown for both the BJS and FBI estimates. Because of this, I used the BJS’s estimates which were always higher than the FBI’s estimates for those years. Despite being higher than the FBI’s estimates, these estimates are still known to be inaccurate. Several states in did not report data to the BJS for some or all of the years, the New York City Police Department did not report data for 2009, and no data was available for the number of people killed by federal law enforcement officers.

Year Est. Missing data
2006 447 GA, MD, MT, NV, NC, OK, WY
2007 455 GA, MD, MT, NV, WY
2008 404 AR, DC, GA, MD, MT, NM, WY, WI
2009 497 AR, DC, GA, MD, MT, NE, NM, WY, WI, NYC

The point of all this is to make it clear that the government doesn’t do a very good job of tracking the number of people killed by police. Thankfully, I believe there is an alternative to relying on the government. If we use news media reports (newspaper articles, TV news reports, etc.) of homicides by police, we can get much more accurate estimates.

Robert Fisher, a law professor and former FBI agent, tracked all shootings by police officers during 2011 using news media reports. He found 607 fatal shootings by police. The FBI, on the other hand, reported only 393 “justifiable homicides” by police officers for 2011. If Fisher’s number is accurate, we can say that at least 610 people were killed by police in 2011 because Fisher’s estimate only includes firearms-related homicides and the FBI reported 3 non-firearms-related homicides that year.

If at least 610 people or more were killed by the police in 2011, then we’d be short by more than a third if we relied the FBI’s estimate of 393 “justifiable homicides” as our estimate for the number of people killed by the police in 2011. We do not have an estimate for 2011 from the BJS, but if we did, I believe it would probably be lower than Fisher’s estimate too.

Using media reports probably would not give us completely accurate estimates. Homicides by police are generally seen as newsworthy events, but it is possible the media do not report on all of them. Furthermore, a researcher studying these reports could miss some of them if they are not careful. Despite the drawbacks, researching police homicides using this method would still provide us with a better idea of the number of people killed by the police each year than relying on the government to do it.

If we study media reports of homicides by police, we can also get some answers to questions that government researchers haven’t bothered asking. For example, what percentage of homicides by police are subject to internal investigations vs. independent investigations? What percentage of homicides result in criminal charges being filed against the police involved? What percentage of homicides result in lawsuits being filed against the police? We may never have the answers to these and many other questions if we wait for the government to answer them for us, but if we do the research ourselves, we may be able to start answering them soon.

The police are one of the most powerful institutions in American society. The police can detain us, search and seize our property, issue fines, arrest us, lock us in jail, and, most significantly, use deadly force against us. I think that we deserve – at the absolute, very least – to know just how many have been killed by these people whose salaries we pay and who claim to protect us.

That’s why I would like to start a community project to track all homicides by police in the United States using news media reports. I believe that by using software like WikiMedia (the software used by Wikipedia), we can create a database that anyone can contribute to. I’m still working out the details for the project, which I am tentatively titling The Police Violence Body Count Project, and I encourage anyone who is interested to share their thoughts on how this can be a success.

Leave a Reply