Police Killings

Richard Epstein writes:

Police officer deaths in the line of duty, year to date for 2014, were 67 of which 27 were by gunfire. For the full year of 2013, the numbers were 105 total deaths, with 30 by gunfire. It would be odd to say that police officer deaths (which are more common than deaths to citizens from police officers) should not count…

It would indeed be odd to say that police officer deaths should not count, which is perhaps why no one says this. Police officer deaths are counted but the literal truth is that we don’t count deaths to citizens. No one knows for sure exactly how many citizens are killed by police because the government doesn’t keep a count. Draw your own conclusions. What we do know, is that it is not true that police officer deaths are more common than deaths to citizens from police officers. Not even close.

105 officers were killed in the line of duty in 2013 but to be clear this includes heart attacks, falls, and automobile accidents. Deaths due to violent conflict include 30 deaths by gunfire, 5 vehicular assaults, 2 stabbings and a bomb. To be conservative, let’s say 50 deaths to police at the hands of citizens.

According to the FBI there are around 400 justifiable homicides by police every year, where justified is defined as the killing of a felon by a law enforcement officer in the line of duty. But note that if the killing of Michael Brown is found to be unjustified it won’t show up in these statistics.

The best information we have of citizens killed by the police, believe it or not, are private tabulations from newspaper accounts. On the basis of one such collection, DataLab at FiveThirtyEight estimates that police kill 1000 people a year.

Thus, killings by police seem to be on the order of 10 to 20 times higher than killings of police.


I have been taking that ~400 number approximately true. Its nice to see a basic counting methodology that sheds some light. I am not reflexively anti cop, but their statistics are inherently political so a grain of salt is always necessary. I am not ready to extrapolate one year across all years - I would surprised if there isnt a trend in one direction or another, if for no other reason population increase - but still, its good to know.

"Police officer deaths are counted but the literal truth is that we don’t count deaths to citizens."

Actually, we do. See http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2012/crime-in-the-u.s.-2012/offenses-known-to-law-enforcement/expanded-homicide/expandhomicidemain



Also see a detailed report (pdf) on the subject

"Justifiable Homicides by Law Enforcement Officers"


Wikipedia actually has detailed lists of justifiable homicide cases. See https://en.wikipedia.org/wiki/List_of_killings_by_law_enforcement_officers_in_the_United_States

Justifiable homicide as described in your link applies to the justifiable shooting of felons, NOT the accidental shooting of law-abiding citizens. Precisely Tabarrok's point.


You aren't giving the police enough credit. The police almost invariably define that deceased in a justifiable homicide case as a felon. Why wouldn't they?

Seriously, there have been numerous studies of police justifiable homicides. Try typing the search words into Google for a few. Many of the studies included detailed examinations of every case. The "felon" qualification isn't material (statistically).

See http://tacreports.org/storage/documents/2013-justifiable-homicides.pdf for a representative study (with an intended emphasis on mental illness).

Is this a bizarre suggestion that a proper police fatality rate should be in the ratio of 1:1 with police officer fatality rates. In Australia, where we have neither the gun violence or racial issues, present in American, the ratio is about the same as above.

Nobody's specifying any particular ratio. But it's not unreasonable to expect police to accept some level of personal risk rather than start blasting away. Do I exaggerate? Watch the cell-phone video of a fatal police shooting that happened in St Louis a couple of days after Ferguson. We can expect it to be ruled justified (because the mentally disturbed man did have a knife). But notice that almost no time passes between the police arriving on the scene, getting out of their car, and shooting the man dead. There are any number of things they might have done other than fire so quickly -- stay in the car, get back in the car, get the car between them and the man. Shooting should have been the very last resort. Instead it was almost the first -- they issue a command, it is not obeyed, and they open fire.


Is this *really* the level of willingness to use deadly force that you think is appropriate?

You call the cops and you create a situation where the cop who has no personal stake in your family enters the situation where his main goal is to protect himself. The emphasis on officer safety creates a don't involve the cops bias for an informed citizen who understands the incentives and trends. Maybe it should. Or maybe we need a separate line for de-escalation officers.

In fact, if I am in my home, why EXACTLY don't I have a right to wave a knife at people to tell them to back off? My family members didn't shoot me. And had they known the cops would they wouldn't have called the cops.

Notice in that video how many unarmed citizens didn't feel threatened.

Who were in fact threatened by stray bullets from the cops.

Cops roll up completely uninformed and scared for their own safety and get up in a tense situation and escalate.

Maybe there is a conceivable improvement...

Btw, let's pause for a second and remember that 15 years ago libertarians told everyone that cheap CCD cameras would bring this to a head, and it is.

If the main goal of the officer is to protect himself, he should be fired.

He/she is hired to protect the public, not him/herself.

It is his primary goal. And it isn't really an optional thing.

The objective is for policy to create the situation so that the incentives are aligned.

Here is the other thing about the ratio.

Maybe it isn't a ratio. Maybe rather than 20 to 1 it should be 20+1.

Maybe to some degree some of the wrong things that cops do that result in the 20 citizen deaths are also contributing to the cop death as well.

I don't know why the policemen behaved so unintelligently: why didn't they get out of their car on the safe side? Nor can I see why the shot the fellow.

I recall at least one video of police negotiating with a man holding a knife for quite a few minutes.

First, they were trying to figure out why he was angry. This is very smart. Usually, when people get angry, there is a reason.

When police can figure out why they are angry, they can start to talk to them. Then, they find out that they aren't dealing with a psycho, rather, that there is a human being.

Then, they can start to deal with the situation.

Eventually, a properly trained officer(s) can insert himself into the scene and use non-lethal force to take down the angry man.

Then, having ensured that public safety is looked out for, they can determine whether criminal charges may be required or whether the man was legitimately angry and the root cause of the problem needs to be resolved.

What happens is the person, psychotic, hopped up on adrenalin, testosterone or just stupidity will walk towards the cop, and since the cop is out of the car and is justified to stand his ground (you can't risk stumbling backwards) there are few options left.

You also simply don't rely on less lethal force in a lethal encounter. So expecting a cop to rely on a tazer or pepper spray or a club in a lethal encounter is unreasonable.

As you pointed out, a lot of ordinary people managed to steer clear of the crazy guy with much trouble at all, so the idea that the cops had no other feasible options but to shoot is nuts -- they should be the damn experts in how not to get themselves cornered and 'have' to shoot people.

To expand on what you said, i had a conversation with a cop friend who told me that they show videos of cops killed in the line of duty at police academy in order to scare them into not hesitating because, as he put it, the 1/10th of a second you hesitate could be the difference between life and death. Now, I am a retires Marine Corps officer, and i will tell you that we take the opposite approach when preparing Marines for a combat deployment. Before deployment, we show interviews with soldiers who participated in the My Lai Massacre during Vietnam and instruct them to be sure to make the right decisions and come home with their honor in tact. Pretty shocking to me that the police, who you'd think would take the lives of the citizenry more seriously than a military occupying a foreign country, takes the shoot first ask questions later approach whereas the military preaches restraint. .

Wow! Now if you're looking for a smoking gun, I think we're onto something.


Btw, no racial issues in Australia? Hmm, then why is your ratio so bad after correcting for racial issues?

Shockingly, the number of arrests made BY police is also significantly higher than the number of police who get arrested. We must bring these numbers in line!

And to be clear, I am broadly sympathetic with Prof. Tabarrok on this topic. But he's not going to convince anyone with these incredibly superficial arguments.


Maybe or maybe not. Ask someone who doesn't know what they think the ratio is. Then tell them the estimated ratio. What do they say?

Does what they say indicate they could never be convinced anyway because they are either pro or anti cop?

Tabarrok is correcting Epstein not taking a position on ratios.

It seems weird how wrong Epstein appears. What's up?

What's the ratio of the rate of cops killed by gunfire to the rate of regular citizens killed by gunfire? Sure cops kill more citizens than citizens kill cops, but cops are still (I would expect) more likely to get killed by gunfire than is an ordinary citizen. Isn't that the ratio that really tells us about the risk to cops?

Kevin, that was my first thought too. There are about 460,000 police officers in the US, as compared to 314 million people in the US. Of course the numbers would skew towards more civilians dying just because of the sheer numbers involved.

That's not to say I don't think there is a problem, I just don't think this data really tells us anything.

I agree. Too many cops get away with bloody murder. They are human too, and we should treat them as such.

I wonder how many fewer deaths by cop there would be if the zealous campaign against lunatic asylums had failed. And even cop deaths too.

You realize those campaigns were by mental health advocates and resulted from the institutionalization of large numbers of people, often for life, without any clear standards?

And one must never, ever reform or improve a system. One must be truly radical and abolish it? Come, come.

So little is understood about insanity that it seems to me that it would have been wiser to have taken baby steps. Put otherwise, "mental health advocates" have been responsible for an awful lot of misery and death.

I really doubt if most of these deaths are driven by lunatics. The typical schizophrenic homeless man is too inept to take down a cop. Seems much more likely to be gang bangers, who have a fair amount of training and practice at urban combat.

Germany is frequently held up as a point of comparison, because police in the entire country only shoot their guns a few dozen times in the course of a typical year. Germany has it's share of lunatics, but much fewer gang bangers.

That's the point. He could never pose a real threat to a cop, but he is plenty competent to hold a knife and be too up in his head to comply with an officer's timeline.

How many of those people suffered from disorders primarily founded in non-conformist political views in generally racist or ideologically "pure" communities, leading to long periods of confinement?

Consider The Protest Psychosis, by Jonathan Metzl, which documents how black men involved in civil rights were effectively imprisoned in mental wards after the community of quacks determined that participation in civil rights protests combined with anger at being forcibly detained was taken as evidence of schizophrenia.

This is documented as having occurred in the 1960s and 1970s. Rest assured, this type of strategy has not been completely eradicated.

But it's not the big thing.

'I wonder how many fewer deaths by cop there would be if the zealous campaign against lunatic asylums had failed.'

You aren't American, but the one thing that is notable in such statistics is not the number mentally ill (with or without substance abuse being involved) killed, it is the fact that it is the poor that get killed -

'This is the most most heinous thing I've learned in my two years compiling Fatal Encounters. You know who dies in the most population-dense areas? Black men. You know who dies in the least population dense areas? Mentally ill men. It's not to say there aren't dangerous and desperate criminals killed across the line. But African-Americans and the mentally ill people make up a huge percentage of people killed by police.

And if you want to get down to nut-cuttin' time, across the board, it's poor people who are killed by police. (And by the way, around 96 percent of people killed by police are men.)' http://gawker.com/what-ive-learned-from-two-years-collecting-data-on-poli-1625472836

However, as has been noted, the poor will always be with us, with the modern American proviso being they will also always be available for police target practice.

"...it’s poor people who are killed by police."

I think that's a mis-characterization reflecting your biases.

It would be more accurate to say people who are engaged in violent or threatening behavior are more likely to be killed by police.

Accurate, but misleading. The cops often instigate the violent and threatening behavior.

It is even beyond cops doing this as individuals. The policy is to put them in the situation where this happens.

Take, for example, the cop car chases. We can, and some places do, err on the side of noting the vehicle and not chasing them at high velocities. This is because the risk is very high to innocent bystanders.

When it comes to shootings, the risk is lower to innocent bystanders, so we are more accepting of the cop having to complete an arrest in a timely manner while minimizing their own endangerment.

Excepting those who are mentally deficient, if you make a good faith effort to obey the cops, you are extremely unlikely to end up hurt.

This is a descriptive statement, not a normative statement.

You mean dead, not hurt. You are getting hurt.

What zealous campaign against lunatic asylums.

The reason the number of lunatic asylums plunged was Reagan said it should be a state responsibility and cut the federal financing of them. Needless to say, the states did not make up the difference.

Cliff says "You realize those campaigns were by mental health advocates" - which is my understanding too, though I wouldn't have used the term he did.

Unpossible. Reagan did it, on his own, out of evil spite.


"I wonder how many fewer deaths by cop there would be if the zealous campaign against lunatic asylums had failed. And even cop deaths too"

A lot. See http://tacreports.org/storage/documents/2013-justifiable-homicides.pdf

1. Did 538 become something worth citing?

2. Is 538 all part of the liberal conspiracy now? they had kind of held themselves to a better standard than Journalist, but I see that even VOX is dissenting from the party line on some issues.

3. Are Libertarian so jealous of the liberal click bait traffic that they will go down the path of accusing the police of racism?

I agree the libertarians will throw the racism charge around when it's convenient to their arguments, even mindlessly citing the "African Americans are disproportionately..." statistics without context.

But policing policies could sure use some thoughtful analysis. There was the case last year where a young woman was grabbed in a home invasion (white girl, black perp) and used as a human shield (the police fully aware of the situation). When the perp began firing, they fired back - killing the girl. Not exactly grace under pressure.

Look, I agree and would even say that libertarians have earned some street cred on the issue, although in typical fashion they aren't interested in solutions.

I just found it interesting that both 538 and Alex are engaged in clickbaiting race-liberals.

They're not, you're just delusional

All I have is solutions. You just can't recognize them.

See Slocum's video link, and the Mike Brown shooting.

Start with cops staying in the car and using their PA system instead of jumping out and creating an altercation where there were previously a bunch of people milling about feeling very unthreatened.

Then let's move on to legalizing marijuana, focusing on violent crime, etc.

Note how all the citizens aren't in any real danger whatsoever until the cops show up. This is not an unique incident. They are all roughly the same, except some cops and some suspects bring with them a higher propensity to dominate and escalate.

It is not like cops always hit what they are aiming at either. I forget the FBI statistics but a very high percentage of shots are errant, and thus a threat to the innocent bystanders. On the far end of the spectrum was the relatively recent NYC shooting chase where not a single bullet hit anyone other than innocent bystanders. Citizens would almost never be so reckless in a justified defensive use of their firearm.

Something I learned long ago when in many parts of the world is that avoiding places of high concentrations of police is one of the best ways to stay safe.

Cant tell if you are trolling or not, but ill ask. Investigating police shootings are "liberal click bait"? And who here is accusing the police of racism?

The last Alex post had 300 or something comments and charlie apparently didn't read many of them.

Yep, that is pretty much the definition of click bait.

No, it is people like me having to repeatedly correct basic logical errors over and over and over.

I have literally never seen a libertarian accuse police of racism. I've been at this a while. That's not to say it doesn't happen, but I don't know how libertarian random anonymous commenters are. I'm talking about libertarian opinion-makers.

What Alex did like 3 posts on the topic ago was point out the racial disparity of enforcement. This is just fact. It is a wildly uncalled for leap of logic to assume this is an accusation of racism.

See pretty much all the voluntaryist Free Talk Live left-libertarian types.

So what?

We hire police to protect the general population from predatory criminals, some of whom need to be taken down with lethal force. We do not hire police to be the subject of target practice by the world's violent felons.

I don't see the point either.

Give me the ration of times a cop killed someone they shouldn't have over the times they resolved a confrontation without killing anyone.

At a minimum, it speaks to the question of how dangerous it is to be a cop. Officer safety is almost always cited as a factor in a police shooting, i think its fair to question that dialog.

Police aren't even in the top 10 of dangerous jobs.


See, it's working.

"Even well-trained officers are not consistently able to fire their weapon in time before a suspect holding a gun can raise it and fire first; this makes split-second judgments, even under “ideal” circumstances, exceptionally difficult." Link

The officer generally cannot "beat" a criminal who has decided to shoot, so must anticipate a shooting. The anticipation increases the officer's survival rate, and increases the chance of an unnecessary killing. You could require the officer to fire only when fired upon, but then your police force would consist of psychopaths,making top 1% income.

'You could require the officer to fire only when fired upon, but then your police force would consist of psychopaths,making top 1% income.'

Well, the German standard for police shooting (involving a grand total of 85 bullets fired at people in 2011) is quite close to that.

Which probably explains why the German police are psychopaths, with large incomes - in an alternative reality, that is.

As for the British police - well, they would actually need to be armed before they even need to wonder when would be the proper point to shoot someone.

The U.S. is a far more violent place than Western Europe, and has been for centuries.

Yeah, Western Europe was a regular hug-it-out zone in the first half of the 20th century and before.

UK police only fired their guns three times in 2013, killed no one.


How can they have a functioning government over there? It must be total anarchy.

"Yeah, Western Europe was a regular hug-it-out zone in the first half of the 20th century and before."

The state has monopolized violence in Europe for hundreds of years. This seems to have lead to a less violent society on the community level, but get a group of Europeans off having a war and they are quite lethal.

PD Shaw,

This is not exactly accurate. You are essentially assuming the perp has a gun, and thus the imminent threat to the officer is a given. If a suspect is holding a gun (or a knife) and makes a move with it, then few will argue with an officer defending himself EXACTLY as any citizen would be justified to do so, by the way. This is not really typical of the situation, though. However, a citizen would be expected to not go confront a neighbor who was seen with a knife just as Zimmerman was admonished for getting out of the car and presenting himself as a victim (although justified in defending himself). Nor does being "unarmed" mean a suspect is innocent and no threat, especially when they are young, huge, and mixed martial arts wannabes.

However, a police officer can't anticipate an imminent threat that is not currently imminent, because anyone can be a future imminent threat. Or, to the point, their doing that which is what people take issue with.

Yes, the study analyzed a straightforward scenario:

"In this experiment, police officers encountered a suspect armed with a gun, pointing down and not at the police officer. The police officer had his gun aimed at the suspect and ordered the suspect to drop the gun. The suspect then either surrendered or attempted to shoot the officer. The speed with which the officer fired if the suspect chose to shoot was assessed. Results suggest that the officers were generally not able to fire before the suspect."

I don't think this is necessarily a typical situation, but the addition of various confounding factors, such as visibility, uncertainty about the suspect's intentions (they may not be bimodal), and whether he has ready access to a weapon don't really improve the cop's dilemma.

We also know you can't even draw and fire at a knife-wielder who starts running at you from 20 feet away.

But that doesn't mean one is justified in drawing and firing at someone who might be a knife wielder.

So ... you advocate for cops being allowed to shoot anyone who they suspect is maaaybe thinking of shooting, whether or not a gun is yet in sight?

I do not want to live in such a country.

I only advocate shooting people who create strawmans.

I think you'd expect to see a rise in violent crime correlating with a rise in justifiable homicide by the police if that were the case, but you do no. Violent crime has been going down for decades while police homicide has been going up. I think this suggests that the police are taking more violent approaches than they had in the past leading to more deaths.

I'd sort of expect correlations with both weak significance and weak coefficients myself.

Do the innocent citizens also sometimes need to be taken down with lethal force because they talk back?

I wish people would wait until all of the facts are in. Witness testimony is unreliable--even when witnesses are honest.

I can't see at this point how anyone can tell whether Wilson was defending himself, lost control after getting punched in the face, or is a totally rogue cop who took out his frustrations on a fleeing suspect.

BTW..Shooting an unarmed suspect is not always the wrong thing to do, an unarmed suspect could maim or kill a police officer if the officer does not defend himself. Fists and arms and batons are not always reliable enough defenses.

The "unarmed" part is dumb, annoying, and neither here nor there, not to mention wrong.

Any altercation with a police officer is an armed conflict because the cop brings a gun, whether or not the suspect is actually going for the gun or not.

post hoc

"taken down with lethal force": what sort of mealy-mouthed euphemism is that? Do you mean "killed"?

Lethal force doesn't not necessitate a death, just the potential to cause it. A bit mealy-mouthed, perhaps, but there is a distinction.

This is just noise. There were 780,000 active cops in the US in 2012. 400 deaths in a year is noise.

Yeah I can't even believe that not even 1% of cops kill someone in any given year. It's shocking.

It's not "noise," which means meaningless data. It's the signal.

Actually you're using the wrong number, in fact it is over 1%, which is surprising to me.

Bleh no edit button ignore this one

The 780k may not be the correct denominator. It might be the number of arrests made. Or even better, the number of deaths caused by marijuana, which is very low if not zero.

4816 arrests related deaths were reported between 2003 and 2007 that is 1 in 20,000 arrest. http://designandgeography.com/2014/08/27/arrests-and-deaths-by-race-what-the-latest-evedence-shows/
I doubt marijuana kill 1 in 20,000 users.

Police don't kill anyone. People are killed by bullets discharged during an officer-involved shooting.

by using the Glock with it's "Safety" trigger this is almost true!

The police using Glocks LITERALLY walk around half-cocked.

You dont know what half-cocked means. You dont know how a Glock operates. Glocks cant be half cocked. They are either cocked or, like you. :)

And knives kill people, not people.

And stones kill people, not people.

And fists break noses, not people.

And feet break balls, not people.

And teeth crunch chips, not people

And stomachs digest food, not people.

And ______ ______ ______, not people.

We are not responsible for anything, ever.


(noun) (verb) (noun), not people.

This is how the world knows that the 5th to 35th percentile rightest of America is real smrt.

Being dense is how we know the 5th to 35th percentile rightest of MR comments is real smrt.

Dense is right. Obviously the 5th to 35th percentile leftest of MR comments is real smrt too.

Thus, killings by police seem to be on the order of 10 to 20 times higher than killings of police.

I would guess that the ratio is similar in Europe where both types of killings are very rare.

One department that does keep careful records of citizen shooting deaths is the New York City Police Department. The NYPD reports that in the past five years (2008-2012) they’ve shot dead 58 suspects. In that same time, 1 NYPD police officer was shot to death. See the report here (numbers are in the appendix):

These ratios are completely useless for what you are trying to use them for.

Cops are well trained shooters, most perps arent. Cops operate in teams. Cops seek perps as a job, perps dont seek cops as a job. Perps dont get radio calls to run to the scene of a cop. People dont have a number to call to have a cop killed.

The ratios favor the cops precisley because they are meant to.

The NYPD are not well train shooters. On the contrary they've proven time and time again that they need 20+ shots to hit someone 5-10 yards away.

No reason not to link to a project attempting to quantify the number of people killed by police - http://www.fatalencounters.org/

And a bit of information about this effort - 'The biggest thing I've taken away from this project is something I'll never be able to prove, but I'm convinced to my core: The lack of such a database is intentional. No government—not the federal government, and not the thousands of municipalities that give their police forces license to use deadly force—wants you to know how many people it kills and why.

It's the only conclusion that can be drawn from the evidence. What evidence? In attempting to collect this information, I was lied to and delayed by the FBI, even when I was only trying to find out the addresses of police departments to make public records requests. The government collects millions of bits of data annually about law enforcement in its Uniform Crime Report, but it doesn't collect information about the most consequential act a law enforcer can do.

I've been lied to and delayed by state, county and local law enforcement agencies—almost every time. They've blatantly broken public records laws, and then thumbed their authoritarian noses at the temerity of a citizen asking for information that might embarrass the agency. And these are the people in charge of enforcing the law.'

And when the government lies to you, what does that make of its democracy?

We can feel exceedingly comforted that highly ethical security officers can be counted upon to make good decisions about what the public needs to know.

Oversight will get in their way of doing their job and effectively terrorize the public with unnecessary information.

‘The biggest thing I’ve taken away from this project is something I’ll never be able to prove, but I’m convinced to my core: The lack of such a database is intentional. No government—not the federal government, and not the thousands of municipalities that give their police forces license to use deadly force—wants you to know how many people it kills and why.' Yup. Ditto all the obfuscation about temperature records: if the bastards won't tell you it's because they are up to mischief.

One thing I find interesting is that according to the best available statistics, 40% of the fatalities caused by police officers in the US are justifiable, using a fairly expansive definition of justifiable. I hope there is a problem with the data.

So expansive that if it had been for all the civil disturbances surrounding Michael Brown's death, the non-filed paperwork backed by public statements from the police department concerning his death would have been considered more than adequate to have the shooting of a man who had attacked a police officer to be considered a justifiable homicide.

Maybe the 12 witnesses stating Brown attacked the office and the fact the officer had a broken eye socket helped their story.

What are the names of these 12 alleged witnesses?

This is yet another example of people striving to get valid numbers, defending their counts, and then using them in a totally meaningless and useless comparison.

It really does not matter whether the police body count is higher or lower than the civilian body count. The only issue with the civilian body count is whether it can be reduced without increasing the rate of serious injury or death to both police and other civilians.

And likewise, the only issue with the police body count is how to minimize without compromising public safety.

I believe the man behind the Fatal Encounters project, which is attempting to create an accurate database of these deaths, would agree - 'But maybe most important thing I learned is that collecting this information is hard. I still firmly believe that having a large, searchable database will allow us not just better understanding of these incidents, but better training, policies and protocols for police, and consequently fewer dead people and police.' http://www.fatalencounters.org

This is not quite right because the justification for the use of deadly force is because police are in harm's way. If they aren't really in harm's way, then they are less justified. If there were zero police deaths per 40 citizen deaths they would not truly be in harm's way. 1 to 40 may be the right ratio, but it smells fishy.

A super high ratio implies that cops are really good at identifying real threats. I think we know that isn't true.

So we have 1000 killings by police, and 50 police killed, and 16,000 total homicides.
Estimate a model for the appropriate amount of police killings that minimizes total homicides.

Legalize drugs and force cops to be laser focused on violent (and potentially) crime. Then we treat them like football coaches, results or hit the bricks.

But that would solve most of the problem. Why would we want to do that?

Now another statistic, out of the total homicides committed in the U S for any year, what percent were killings by police actions (of any kind).

We have police to kill people of certain motivations so we don't have to ourselves.

Now, it is true police kill when we might not. (mental cases, etc.)
It is also true we kill when the police would not.

As has been pointed out by D. Brian Burghart at the Fatal Encounters site, 'what percent were killings by police actions (of any kind)' is absolutely impossible to answer, as no essentially no reliable data is collected by most local departments, and is definitely not collected by the federal government.

'Nowhere could I find out how many people died during interactions with police in the United States. Try as I might, I just couldn't wrap my head around that idea. How was it that, in the 21st century, this data wasn't being tracked, compiled, and made available to the public? How could journalists know if police were killing too many people in their town if they didn't have a way to compare to other cities? Hell, how could citizens or police? How could cops possibly know "best practices" for dealing with any fluid situation? They couldn't.' http://gawker.com/what-ive-learned-from-two-years-collecting-data-on-poli-1625472836

"We have police to kill people of certain motivations so we don’t have to ourselves"

Ummm. No. This would imply that cops intervene in assaults in progress. They don't. I can't even think of an appropriately narrow qualifier.
What cops do is roll up on a suspect and then create an impending assault and then justify use of deadly force because of the situation they created.

Yes, that happens far too often. An amazing thing about police killings is that they are almost always reviewed only internally, by the police department itself.

Interesting article about a Wisconsin assembly bill which would require independent, outside review of police shootings: http://www.kenoshanews.com/news/ab_409_aims_to__strike_a_balance_475445430.html

I've seen that Wisconsin story, but couldn't reconcile it with what happened recently when my high school friend was killed by local cops in what appeared to be "suicide by cops." Within a few hours, the Illinois state police had taken over the investigation because obviously the local cops couldn't investigate themselves. Is this truly not the norm?

Sorry to hear about your friend. It appears that in California the police investigate these killings themselves, with some window dressing provided by the local mayor if the press shines a light on it.

OTOH, if criminal charges are brought by the DA, then some outside investigation occurs in order to go to trial. But that is really, really rare. And it's rarer still for the DA to win a case like that, even one that looks fairly clear to outside observers. (e.g. Kelly Thomas, an Orange County man with psychological issues who was beaten to death by police) .

Better the police send a hand written invitation to the suspect to drop by the station to talk.

Works every time.

Actually, it does work, but not quite that way. And they will start doing more of it. Think the "you won free football tickets!" scam that cops use to round up warrants.

It is weird how cops who fear for their safety would prefer to instigate armed conflict over some other option. It is also a little weird to say that a cop must make an imminent arrest of someone who we are then going to assume is innocent until proven guilty in court.

If the person is an imminent threat, wonderful. But that is some rare Jack Bauer shit. A lot if no the vast majority of the time they aren't an imminent threat to anyone until the cop shows up.

"It is also true we kill when the police would not."

Aside from murders, this is also I suspect so rare as to not justify any qualifier. The cop (citizen on patrol) is usually legally insulated from major consequences (i.e., the problem) whereas the citizen who does not have the protection of government does not. Also, the cop has a responsibility to make an arrest. These things combine to militate towards a citizen deflecting a situation without having to escalate.

On top of that, the risk to yourself from the legal system is so great it is a tough proposition to even carry a gun. So, we are really left with situations where citizens are in their homes and defending their "castle" which is a way higher standard than most police shootings.

In other words, all the misguided bluster over "stand your ground" is referring to a concept all cops in even greater strength.

Here's some stats on which types of officers are more likely to use deadly force:


I want to know their high school sports history (i.e., football).

Given costs and benefits of type 1 and type 2 errors, there's an optimal ratio of cops/non-cops killed in violent interactions. There's a paper that will never be written.

Why not? That is what economists are for. We don't need pussy economists any more than we need pussy cops.

Put the equations on line and let readers choose their own weightings.

You do the stats I'll write everything else. Better for you, find an actual economist collaborator!


Interesting idea. It would be interesting to generate output using this approach, then to quiz users on ideological views, then compare this to preferred selections for parameters, etc.

Why on earth would we be counting any and all police deaths that just happened to occur while they're in uniform? A police officer crashing his motorcycle is in no way comparable to civilian killings by police.

Just doing a simple count, I see 7 heart attacks, 15 auto accidents, 1 gunfire accident, and 2 motorcycle accidents. C'mon, Epstein.

I think an underappreciated factor is "suicide by cop." Those who want to die find a way to die. Some use a gun on themselves, or kill themselves in a car "accident." Others go out with a bang, sometimes in a mass murder situation.

They grab a gun, make a crisis, maybe kill some people, and then it is "come and get me, copper!" Such people want to die, and they are implicitly asking the police to do it for them. They will kill until they are killed. (Think of the U of Texas sniper in 1966.)

One way or another, those people should be considered mentally ill. As for other sorts of mentally ill people, I wish there were a way to use tasers or tranquillizing darts to subdue them safely. Finally, those who are heavily intoxicated by alcohol or drugs may be temporarily mentally ill and a grave and violent risk to the community. Substance abuse, I have heard, plays a very large role in violent encounters between civilians and with the police. (Those who advocate drug legalization are naive in believing that it will reduce violent crime.) Michael Brown, btw, have pot in his system, and it curious that he had just stolen cigars; substances may have excited him to escalate a simple situation.

It is tragic, to be sure. But I don't blame the police for it. One way or another, those people are going to die. I am just sorry they feel entitled to take innocent people down with them.

I think the statistical problem is that the U.S. does not have an inquisitional legal system. This cop in Ferguson won't receive a determination that he was justified unless charges are brought, and the jury finds for the defendant. If charges are not brought, it doesn't mean that they might not later if additional evidence turns up. Law enforcement generally never declares someone innocent, as opposed to lacking evidence at this time . . ..

Their job is to provide evidence of guilt.

That is why presumption of innocence is sacrosanct in the legal system. It must be upheld even in many cases where it seems like a silly exercise to do so.

How many innocent men would you send to jail to nab one real badass?

It's curious that the Michael Brown episode in Ferguson, MO should spur this conversation when the details of the event are quite ambiguous. Other police killings in the recent past are much more problematical. The Miriam Carey incident near the capitol, http://nailheadtom.blogspot.com/2014/04/dc-cops-use-of-force-policy-is-secret.html and the Jose Guarena killing in South Tucson http://nailheadtom.blogspot.com/2011/05/swat-team-fires-71-shots.html are just two of the most egregious examples.

There's always been much controversy over the imposition of the death penalty. Yet individual cops are allowed to impose it in virtually any circumstances they deem fit and the departments that employ them automatically defend their actions to the bitter end. They can do this by refusing to outline their "department policies" that justify deadly force, which seems to be failure to comply with any police command and the fact that the officer involved "feels that his life was in danger".

Oddly, in a business that whole-heartedly embraces technological progress, law enforcement has made zero effort to develop or adopt non-lethal means of incapacitating dangerous individuals. It would appear that part of the attraction of being on the front lines of law enforcement is the opportunity to pump someone full of lead.

Finally, the most bizarre example of sociopathic police behavior is the Christopher Dorner affair in Southern California. http://nailheadtom.blogspot.com/2013/02/christopher-dorner-and-lapd.html

Technology: Just off the top of my head, a PA system, a camera and a pepper spray paint ball turret on the police car would eliminate any problems of anyone with a knife or less. Cop doesn't have to get out of the car to gain compliance. If perp does not comply, snap off some pictures and a few paint balls. At worst, perps watery eyes put him at a disadvantage if and when an altercation occurs.

We aren't even talking about the gooey foam or auto EMPs. I could have this up and running by the afternoon. The problem is I wouldn't be able to waste hundreds of millions of dollars in homeland security money.

Pepper spray paintballs: Amazon, $40
Paintball Gun: Wal-Mart, $100 (the government version would surely be $1000)
Camera and remote turret: $100 (again, $1000, but still)

No cop or schizophrenic deaths, priceless.


Could be operated remotely by the dispatcher.

The point is not that this is a totally viable idea, although technologically it is. The point is to back up Chuck's point that despite obtaining armored personnel carriers the biggest technological advancement police have made in citizen contacts is parking their cars cattywhompus, but I wonder if that is partly to point their dash cams away from the interaction.

Here is an example of cops using a pepper spray paintball gun effectively, if patiently.


But then we would have to train them to deal with the situations instead of letting them cover up errors all the time.

That will get expensive. We might have to reduce the number of cops and focus on quality instead.

Even economists don't seem to realize that the concept of marginal utility applies to employees as well as other things. The first cop in Dodge City was probably a pretty significant and valuable asset. The last one that they've hired probably doesn't contribute as much to the peaceful atmosphere on the banks of the Arkansas River.

Any comparison of the lawful use of force with a trial and sentencing is BS .

"Thus, killings by police seem to be on the order of 10 to 20 times higher than killings of police."

What's the police to violent criminal ratio? Or, the killing to violent confrontation ratio?

Police aren't perfect and it's certainly possible that they could be corrupt or power crazed or especially more prone to make mistakes and/or err on the side of caution during violent confrontations. But, what I don't see is evidence that is anything more than noise rather than systematic.

Chris Rock still is right. http://clashdaily.com/2014/08/dear-thugs-please-watch-get-ass-kicked-police/

In the course of playing amateur busybody detective looking into the killing of an 18-year-old violist in a local parking lot by a DEA agent, I ran into the grieving mother who was also looking for clues. I told her that from my knowledge of the area the cops' story sounded fishy and strongly advised her to consider a lawsuit. Three years later the parents won $3 million:


There are instructive similarities and differences in this example with the vastly more publicized deaths of Trayvon Martin and Michael Brown.

Three million dollars won't bring their son back and none of that money will come out of the pockets of the coercion agents involved. Concrete individuals (the victim and his parents) are in the grip of an abstraction (the government) with no personal responsibility.

In missing Steve's point you illustrate it.

What I've heard of more than once is a police officer will jump in front of or behind a moving car and then open fire because he felt threatened by the fleeing suspect.

This of course puts the lie to it because no one asked him to jump in front of the car, and if his objective was to avoid feeling threatened he wouldn't have done so in the first place.

(it should go without saying, but I'm always surprised how things must be said, this is especially problematic when the officer is plain clothesed and possibly thought to be a car jacker by the driver)

"Thus, killings by police seem to be on the order of 10 to 20 times higher than killings of police."

Other facts;
less than 1 million policeofficers
more than 300 million citizens

Who appears to be most violent?

We don't know. It would be nice to have the statistic that separates "actually justified" from "justified because the officer couldn't know he was really not justified."

This separates things like the criminal that actually had a gun from the combs and wallets the officer thought was a gun. This might instruct future "rules of engagement" policies.

It is weird that on the citizen side (castle doctrine, stand your ground) the policies are both more conservative and more well defined.

Chalk two more up...last night in Omaha police shoot robbery suspect who had an air gun and accidentally kill the COPS sound engineer covering that shift for the reality show.



Quote from the link

"The robbery suspect apparently had an air gun, a type of BB gun that looks like an actual firearm. He apparently was a prison parolee from Kansas, law enforcement sources said."

That should make an impressive edition of the show, the cops shoot and kill a crewman of the TV show that's displaying their prowess.

Well, this is interesting - 'The World-Herald has learned that at least 30 shots were fired at the Wendy’s near 43rd and Dodge Streets. Officials said it appears the only shots fired came from police.' It took the police 15 bullets to cause each fatality, apparently.

In contrast, in 2011 in Germany, 85 bullets were used by police against people - of which 49 of the shots were intentional warning shots. Of the remaining 36 bullets shot intentionally at people during that year, the toll was 15 people injured and 6 killed. Meaning that the German police, on average, require about 1.5 bullets to hit their target (assuming single shots only causing an injury or fatality - obviously, they could have missed less, but it is reasonable to believe that German police also miss their target too).

What makes this interesting as a contrast in police competence is that during the riots in Ferguson, the American police shot wooden bullets at and arrested a Bild reporter - Bild being the largest circulation paper in Europe, and one that is considered very law and order conservative in Germany. As a reaction to this style of American policing, the Bild wondered why American police were so utterly incompetent - unable to arrest rioters, while also ignoring the law and targeting reporters.

This sort of shoot first and keep shooting until you run dry policing is laughable in a civilized country.

Haven't read the above comments- BUT your point about police "killings" being 10-20 times higher is complete bovine scatology- off course police killings are higher b/c you are conflating justifiable shootings with unjustifiable. It is like saying life is the number one cause of death. Would you prefer that so many police are murdered that the overall number of deaths was equal?

"since it is very difficult to envision circumstances where killing a police officer counts as a form of justifiable homicide"
I can think of plenty. Not that the jury will believe you, or that you actually alive to present a defense for that matter.

Epstein also seems to think police officers face stiffer penalties for wrong doings, which is just bizarre, if not frightening in its obliviousness. I like me some Epstain, but the guy really needs to take some Ritalin, and cut down on the TL;DR when penning opinion pieces.

By trying to be too cute and Epsteinesque with his ponderous style, he makes an otherwise straight forward an compelling point (hey, this is a function our society needs, let's not throw the baby out with the bath water, maybe there are approaches to training that can be implemented to mitigate unnecessary harm) rather risible.

Comments for this post are closed