In Donald Trump’s acceptance speech last week, he stated the following:
The number of police officers killed in the line of duty has risen by almost 50% compared to this point last year.
This is actually categorically false.
As you can see, the number of officers killed in the line of duty is up 8%.
But perhaps Trump meant to say that firearms-related police fatalities were up, not that overall deaths in the line of duty were up. In that case, he actually shot low—firearms-related police fatalities are up 74% compared to last year. But the main reason for that is actually that the number of police who die in firearms-related fatalities these days is so low overall. In fact, the number is so low that the eight officers killed in Dallas and Baton Rouge alone are enough to the statistics off completely. Last year at this time we had 19 firearms-related fatalities. This year we have 33. The eight Dallas and Baton Rouge deaths account for 8 of the 14 additional deaths.
Let me pause to say that I categorically condemn the murder of police officers, and that every untimely death is a tragedy. We should absolutely be concerned about officers’ safety (just as we should be concerned about citizens’ safety, both from gun violence in general and from racial profiling and bias in policing). Every life is precious. I’m also not anti-police. A well-functioning police force is critical to a stable and healthy society. But while I’m not anti-police or in favor of of violence against the police—far from it—I am a believer in context, sound statistics, and maintaining a historical lens.
When it comes to measuring police fatalities, several different metrics are used. One metric counts every death that is at all related to policing. Another metric counts only officers “feloniously” killed in the line of duty—i.e., cases where their deaths aren’t accidental. A third metric counts only firearms-related deaths. This third metric is very close to the second metric, which counts felonious deaths, but the overlap isn’t complete—some felonious deaths aren’t caused by firearms. This large number of metrics can make tracing police fatalities confusing, but we’ll make the best of it.
Let’s look, for a moment, at firearms related fatalities. This image from mid-July:
I just crunched some numbers, and if the current rate of firearms-related police fatalities continues, we will end the year with 71 firearm-related police fatalities. This number is higher than the past few years, yes, but it isn’t unprecedented—it’s in line with the numbers for 2007 and 2011, when 70 and 73 deaths were recorded, respectively. Neither of those years pretended a future increase in the police firearms-related fatality rate, as you can see in the graph above. Further, given that the number of firearms-related fatalities for 2016 so far is elevated by several mass shootings, it is unlikely that the elevated rate that has held for the first half of the year will hold for the second half as well, unless more such events occur. Without Dallas and Baton Rough, the number of firearms-related fatalities would be up 32%, which would be unsurprising given the sort of variation we see when numbers are this small.
Next, have a look at this graph:
Note that since this graph ended in 2009, the numbers have dropped still further, hitting only 33 in 2013. If we hit 71 fatalities this year, as we would if current trends were to continue (including more mass shootings of police, god forbid), that number would be typical of what we saw in the 1990s and far lower than what we saw in the 1970s. There’s also no reason to assume it would be a long-term trend—as we saw in the first graph, it’s not uncommon for numbers to spike in a single year and then decline.Given that these small numbers create a lot of variation, this image is helpful:
This graph ends in 2010, but I just ran the numbers for 2011 through 2015 through my head (the next segment given the graph’s date periodization), and came up with 49, which is actually lower than any other five-year period on this graph. In other words, we live in the safest era of American policing in over five decades. Even this year’s higher number does not change this reality unless it becomes part of a longer trend, and we’ll have to wait years to see that. Looking, again, at the first graph shown above, it’s not out of the ordinary for firearms-related fatalities to spike to the low 70s and then fall back to more typical levels.
I haven’t said anything about the elephant in the room, though. The number of police officers on our streets have not remained constant over this period. The above graphs do not take this into account, showing only the total fatalities and not the fatality rate per, say, 100,000 officers. Check out this graph to see how police numbers have increased:
The number of overall full-time sworn law enforcement officers grew by 25% between 1992 and 2008, and has reached 251 per 100,000 residents, up from 238 per 100,000 residents. This means that being a police officer has become even safer than the statistics and graphs shown earlier suggest, because the total number of officer fatalities have declined even as the total number of officers has increased. (If the rate held steady, you would expect it to increase as the number of sworn officers increased.)
Let’s run the math. During the first half of the 1990s, an average of 70 police officers were killed in firearm-related violence each year. During that same time there were roughly 610,000 sworn officers. During the second half of the 2000s, an average of 50 officers were killed in firearms-related violence each year. At the same time, there were roughly 765,ooo sworn officers. This means the firearm-related fatalities per 100,000 officers decreased from 11.5 to 6.5. In other words, firearm-related police fatalities were nearly cut in half between the early 1990s and the late 2000s.
I realize that the families of a lot of police officers are probably pretty worried about their loved ones right now. In my own city, police have been traveling in pairs, for safety. I cannot imagine how difficult it must be to be the spouse of a police officer. I am horrified by what took place in Dallas and in Baton Rouge, and I very much hope that those events do not become part of a larger pattern. Still as it stands, right now, we are in the safest era to be a police officer in many decades. Statistically, there has never been a better era to be a police officer. Even this year, with the violence in Dallas and Baton Rouge, we have not seen an increase in police fatalities high enough to put 2016 outside of ordinary year-to-year variation. So let’s not fall prey to fear-mongering.
I’m going to leave you with one last graph. As you look at it, bear in mind that the U.S. population in 1950 was less than half what it is today, and that the graph tracks total number of firearms-related police fatality, not rate. In other words, our rate of firearms-related police fatalities is actually lower today than it was during the lowest dip between prohibition and the 1970s.