How to fix racial bias in policing

We estimate the degree to which individual police officers practice racial discrimination. Traffic police regularly discount the charged speed on drivers’ tickets to avoid a discrete jump in the fine schedule. This behavior leads to an excess mass in the distribution of charged speeds just below the jump. Using a bunching estimation design and data from the Florida Highway Patrol, we show that minorities are less likely to receive this break than white drivers. We disaggregate to the individual police officer level and find significant heterogeneity across officers in their degree of discrimination, with 40% of officers explaining the entirety of the aggregate discrimination. Our measure of discrimination is easy to calculate and can be used by police departments as part of an early warning system. Using a simple personnel policy that reassigns officers across locations based on their lenience, departments can effectively reduce the aggregate disparity in treatment.

While 40% is a high number, overall I find that reassuring. That is from the job market paper of Felipe Goncalves, of Princeton University.


How about we just replace fines with community service. One speeding ticket is converted into 8 hours of community service, regardless of whether they were driving 10 or 50 miles over the speed limit.
Driving through a red light is 12 hours community service.

So speeding tickets are no longer a revenue angle, speeding trap towns no longer exist, and the sentence is so inconvenient for everyone involved (even more so for working poor whose jobs are at risk) that challenges to tickets fill the docket? Not only is this very questionable as an improvement to anything, but I strongly suspect it would lead to speeding below the reckless endangerment level being very rarely enforced except to establish probable cause for contact with the local underclass, so the racial disparity in ticketing would be expected to increase as a result.

This is something a white supremacist would say.

I point you to Abrams, D. S., Bertrand, M., and Mullainathan, S. 2012 publication "Do judges vary in their treatment of race?" which beyond a shadow of a doubt of white supremacy.

Bottom line, check your white privilege.

If you act out and run your mouth when a policeman interacts with you it is likely that the police will be less than eager to cut you a break. If you aren't taught manners at home and your upbringing consists of getting away with blaming everyone else for your problems then it is unlikely that you can stop yourself from being a "dick" when dealing with authority. But, cheer up. When you get "justice" for acting like "dick" you will be able to blame it all on discrimination or something instead of taking responsibility for your stupid actions

You talk like you haven't hear of the racism before. You do realize that slavery was replaced with Jim Crow and Jim crow was replaced by the modern day criminal justice system don't you?

White privilege - the benefit to be called a racist by someone that knows nothing about you, except that you are white.

I think we can agree. Most blacks are not murderers, rapists, and thieves. Same same most cops are not racists. .

Young cops are nervous. Older cops have seen too much.

Here's how I take advantage of my white supremacy: When interacting (recently my 70 year-old wife was pulled over for 83 in a 70 zone) with a big, ugly (in this case black) man just stepped out of a big-assed, black/white vehicle with flashing lights all over it; and who wears a badge and carries a large caliber handgun, I act real polite and keep my hands where he can see them.

If the differences in ticketing are due to driver behavior, then how would 40% of the police officers account for the entirety of the aggregate discrimination? Those police officers must be very unlucky. Those officers encounter black drivers that behave worse than white drivers, while the other 60% of officers deal with blacks that behave the same as whites?

everyone needs to deal with this video to realize what we are dealign with

watch this video and then look at all these high profile BLM protesting shootings. they usually fit the bill.

you also have to consider that we are not just police officers but marriage councilors and therapists as well. plus we have our own problems and own families. we are exhausted. all we deal with every day all day are liars and people who even when they sell drugs to an undercover cop claim "they didn't do nothing."

So quit.

That is an elegant piece of research. It may even have discovered a real effect. Although I think we would have to look closer - how much of the discounting was for neighbors? Were Black drivers the same as White drivers - I would think not. We would also want to know how much less likely Black drivers were to get this and what was the race of the policemen. After all, a whole bunch of convicted Black people have just been let out of prison because two Black police officers may have framed them. Abusive policing is more common among minority police officers.

Yeah, I agree.

I'm not sure what the public expects from us. I work in south Florida with a high number of blacks and latino's. Youth from those two groups are responsible for most of our violent crime. My job is to be alert for suspicious activity. You can pass all the legislation you want but I'm going to be more suspicious of a 18 year old black male than a 87 year old asian women. We can talk about if this is correct after the fact, but when the stakes are high and we are patrolling a high crime area, we will use the demographic information we find to our advantage.

I'm from Ecuador if it matters and my partner is African American.

None of that matters. The standard is disparate outcome, formerly known as the "effects test."

More blacks than whites paid higher fines than whites. That's all they needed.

That concept is the red-meat of contemporary racial racketeering, e.g., CRA/anti-redlining and Ivy League acceptance rates.

I'm old enough to remember when the standard was treating all fairly and equally. Now, the standard is "fair and equal" outcomes. But, what is "fair?"

"Were Black drivers the same as White drivers "

In order for this to make sense you would have to have whites being 'strategic speeders' who break the speed limit but cluster their speed just under the next jump in the fine schedule while blacks are 'careless speeders' who simply go as fast as they please but don't cluster right under the next upwards step in the fine schedule.

I'm not sure what set of incentive motivations could account for that type of difference. My thinking would be that blacks might be more reluctant to get pulled over hence would be more inclined to be a 'strategic speeder'...who breaks the law when under pressure to get somewhere but still tries to keep the potential harm at a min. This would also require whites to be very aware of the speeding schedules for where they are driving. Far more likely, IMO, is the hypothesis that cops give speeders a break by 'rounding down' their speed to be just under the step they are in.

Another possibility.....cops suspect whites will challenge their speeding tickets in court with lawyers who will start getting all technical. If someone was going 25 mph over the limit, rounding it down to 19 mph makes for an easier case to prove and makes it harder for the driver to justify trying to fight the ticket.

Or whites otherwise give off higher Socio-Economic cues, or are just better at dealing with Cops.

Speeding is an almost universal crime. So you're stopping a lot middle/high-status whites and Asians who otherwise never interact with the police. These people are going to be treated with a degree of leniency. But it's not racist; the well-dressed-and-spoken black salesman will also get a break from the traffic cop.

I agree; racism is a lazy explanation. There's lot of potential for confounding effects nested in the traffic stop demographics.

Yes, by far the most convincing (in view of the abstract only, I hadn't time to read them all) job paper presented by Tyler this season.

Why is the 40% number reassuring in this case? If that high a percentage of police are actively discriminating how do you go about getting rid of them? Does the myth of a 'training problem' still hold if 40% of the police officers are exhibiting bad behavior?

The title of the paper is, “A Few Bad Apples? Racial Bias in Policing”. It's an apt title for a very good paper. They anticipate most of the criticisms and questions folks in the comments here have made. Most importantly, they get away from the global measures and are able to give you officer-level numbers. If you aggregate across the whole state, the amount of discrimination doesn't look like much. Percentage-wise, it's not too bad. But that's because most officers don't discriminate, so the actions of the "bad apples" is nearly hidden.

In my view, even the 40% number Tyler and you chose to point to is the wrong one to emphasize. Most of the folks in that group don't exhibit much discrimination, to the point that if it were just them, I imagine the authors would have trouble pointing to a real effect. Check out Table 5. White folks stopped by a "lenient" officer get the speed listed on their ticket reduced about 30% of the time, or with a probability of 0.3. For minorities, that only drops to 23% for officers in the 75th percentile. It only gets really bad when you get the most discriminatory officers. For example, the most discriminatory 10% will only give a reduction to minorities 17% of the time or less. (Note that these values vary depending on how individually lenient the officer is; I've used 30% as a baseline, following the discussion at the end of Section 4.)

I find this result very credible. It agrees with my priors that most officers aren't discriminating, but some "bad apples" to discriminate and they do it to such a degree that minorities notice. So when cops or people who know cops hear about discrimination, they rightly think, "I'm a good person and I don't discriminate. If the study is saying that cops like me discriminate, it's wrong." And when minorities experience discrimination by an officer, they rightly think, "See? This is exactly what we're talking about. I'm being discriminated against by this cop because of my race." Both of these things can be true, and a study that's able to look at behavior at the individual officer level rather than just averaging things and throwing all the officers into one pot can address both of these viewpoints simultaneously.

+2, very useful.

So, we're left with the boring racism-real-but-negligible explanation. "1 in 10 cops are racist and make your punishment 10% worse" doesn't really have that much of ring to it. But it's the kind of finding that can infuriate all sides!

I'm a. cop and we use that ability to use our discretion to lower the fine as an incentive to get the arrested party to cooperate. In my experience, blacks are far less cooperative and far more defiant/difficult in such routine situations as a traffic stop. Hispanics are the most cooperative and respectful. Whites can be hit or miss, but that usually has to do with how wealthy they are.

Also, do we know the race of the officer?

Pigs reward sheep - gotcha.

>>> "I’m a. cop and we use that ability to use our discretion..."

'Cop Discretion' is precisely the problem with our oppressive American policing system.

Cops have way too much "discretion" in interpreting and applying law. In effect, street cops have become 'mini-judges' with wide judicial powers informally delegated to them. For example, the authority to "arrest" someone is fundamentally a judicial power exercised via a formal arrest warrant -- but American cops routinely arrest people at their own cop-whim. Cops should not have any more arrest-authority than normal citizens (via "citizens-arrest-authority" for exigent crimes). Cops should also be fully prosecuted for false arrests, as any normal citizens would be.

Your proposed policy change is being informally tested in Baltimore since about 2 1/2 years age - the results have not been pretty.

2 1/2 years ago - since the Freddy Gray riots, that is

you can't write laws to perfectly match the reality on the ground. thats why we have some discretion at our use.

"If I can't kill suspects, I won't do my job."
I understand why it is an informal policy, I wouldn't say it aloud, too.

Hua Wei, I'm not sure what your point is.

Basically, if I wanted, I could start my day at 8 AM and write 40 citations in a matter of 3 hours. I couldn't walk down 3 blocks without seeing something I could penalize someone for. But my job isn't to make the neighborhood miserable but to do worthy and good police work.

My point is, I believe in a narion of laws.

While having too many laws may be a problem, the cop still has his point: discretion is always applied.

Um, people are still getting arrested in Baltimore. The bulk of the current crime wave (that is the "excess" crime over and above the already high rate of crimes in past years) seems to be due to juveniles, not adult offenders-- a trend that was already visible back in 2013. Some of these brats (to keep it classy here) were still in grade school when the Freddie Grey thing went down and the "It's all about Freddie Grey" excuse is wearing more than a little thin. (FYI, I live in Baltimore; this is not baseless speculation from a distance).

No, no, no.

Discretion is absolutely critical to effective policing. Enforcing all the law, all the time, will alienate everybody, swamp the courts and jails, and exhaust policing resources, demoralise the police, and open a giant pit of litigation from angry arrestees.

You can't write rules that effectively cover the range of circumstances officers encounter. You just can't. Officer judgement is certainly prone to abuse, but prescriptive rules do an even worse job. You're better with a discretion + record & remedy approach.

(also, "authority to “arrest” someone is fundamentally a judicial power exercised via a formal arrest warrant" is just plain wrong. Powers of arrest are vested in all citizens in both US and UK common law.)

Studying discrimination in policing always has the problem of controlling for underlying behavioral differences in populations. So these analysts think they've come up with a clever way to do that - mph leniency on traffic tickets, which must surely be completely arbitrary, right? Of course it's not, because no single officer is lenient with every single offender, nor is this leniency 100% based on the race of the offender (the effect sizes in this study are small). Instead, the leniency is awarded based on factor x. This factor x is probably the co-operativeness of the offender, and thus the most parsimonious explanation for the disparity is behavioral differences, not race. This study does not address that issue - indeed how could it?

The other effect it is likely to be is who is doing the driving. White drivers are more likely to be suburban. Roads are more common in the suburbs. The police are more likely to pick up White suburb drivers taking their children to soccer practice than Black ones doing the same thing. On the other hand a larger percentage of Black drivers will be young, foolish and doing considerable speed. Simply because there are fewer Black suburbanites.

Isn't it reasonable that someone lets off an otherwise safe driver just going over the speed limit while giving a ticket to the young and foolish?

That doesn't really explain the 'clustering' effect. I could see the argument that a young Black driver might be reckless and just speed to a degree he wants resulting in tickets that cluster around all different MPH's over the limit. Likewise I could see the argument that White suburban driver taking kids to soccer practice might be less inclined to speed at all.

BUT this is looking at speeding tickets actually issued so for both white and black tickets, you are dealing with people who were speeders. So the question is why would the white tickets be more clustered around the lower fine level while black tickets are more spread out? A non-biased explanation, I think, would have to assert that white drivers are being very strategic speeders....deciding to speed but keeping their speeding within whatever the various 'speeding brackets' are. This seems like a very hyper-rationalist speeding model where white speeders even guess that they are going to get pulled over and somehow balance their desire to speed against their desire to keep the ticket at a min.

I think it is far more likely that it is the cop who is producing the clustering by deciding to 'round the speed down' when writing the ticket.

I wouldn't say hyper-rationalist but different rationalization process. I'm late and 15mph above the speed limit isn't really speeding vs I'm late screw the speed limit.

I think the objection is that the demographics can simply contain confounding factors. White speeders may simply interact with the cop better, but not because they are white, but because they contain a greater share of respectful / older / middle-class / well-dressed / no previous conviction types.

if you were to replace a African Americans cop with a white cop regarding how black cops talk about young black males, you would have a national riot on your hands. we have a "common sense squad" in our department who consists of African American cops who say what white cops think but can't say out loud.

That's interesting. A "common sense squad"? Can you tell use more?

Interesting. The commentary from serving officers here, especially the black and Hispanic ones, really undermines the racism explanation.

"controlling for underlying behavioral differences in populations"

First, you have to believe such a thing exists...

Are you suggesting that they don't or that we should pretend not to notice them?

I would never suggest that collective differences in behavior don't exist. They do, but a huge amount of modern economic and social science theory seems to posit that they don't.

We are talking about speeding tickets here, though. What type of 'cooperation' are you trying to get out of the standard speeding ticket?

So what is the thinking with 'rounding down' the speeding ticket? On one hand I can see a 'squeaky wheel' play where you round down the ticket on the annoying guy who makes a big fuss about it. On the other hand I could see you being more inclined to 'reward' the nice guy for making your job a bit easier.

This makes for a very complicated game if you consider it from the speeders POV. If the speeder thinks he can win over the cop by being polite, friendly and seeming like he has a lot in common, he will be inclined to go that route. If the speeder suspects the cop doesn't like him, feels contempt for him, it might make sense to go the 'squeaky wheel' route and make it clear anything you do will mean you will be dealing with him 'kicking and screaming' so he hopes you'll just cut him a break because you won't want to deal with it. What you might be seeing is the result many other interactions previous officers have had with the community. Not knowing you the speeder must make a call as to his optimal strategy...try to be Mr. Polite Guy or Mr. I Raise Hell Guy.

What I'm getting at here Blacks being less cooperative, Hispanics being very cooperative and Whites being hit or miss is almost certainly NOT some aspect of their inherent nature but a strategic response based on their experience. What strategies could be deployed to signal that the 'payoff function' has changed and a more friendly encounter is more likely to result in some type of a break than a loud argumentative one?

To Boonton,

I'm just a cop who found Tyler's MRU videos interesting so I starting following his blog. Not an economist so I don't understand all your points.

But the less cooperative white guy will talk some shit but cooperate in a reluctant way. they will be sarcastic and snide. but they cooperate with your orders.

Hispanics who don't cooperate don't speak English, lol. usually they are polite and look at the ground the entire time. like they're ashamed of something. especially if they are Mexican.

the blacks who don't cooperate repeat everything you say (we call it parroting), ask "why" every step of the detainment, try to trip you up by twisting your words, use the race card endlessly like accusing you of being racist for pulling them over in the first place, ask you how many white people you have pulled over today, refuse to roll down their windows or put out cigarettes, refuse to turn down their music, etc. etc. they either want to play Sherlock Holmes by questioning you about everything (yesterday a guy wanted me to prove to him my police vehicle was insured) or they are just say refuse. they either refuse to lawful commands verbally or at times, even refuse physically.

needless to say this isn't most blacks. but I never really see this sort of behavior from other races.

So my question/thinking is from the POV of the person stopped 'what works?' Act annoying and hope you'll go easy on them because then you can spend less time with them or act really nice and hope you go easy because you will like them and cut them a break. When I hear you encounter the no cooperation a lot that tells me part of the answer has to be 'because it works'....or at least it works better than other strategies like acting super respectful.

When I say 'works' I mean in an odds sense. Sometimes there is no approach a person can take that will talk themselves out of a ticket once they are pulled over but sometimes there is. If there is the question a speeder should ask is which tactic is more likely to get me the best outcome?

I suspect Blacks have picked up that playing the sympathy or respectful angle does not work as much as playing the very annoying angle hence that's the strategy you see more often than not. That's not a good thing because it increases the chances of a stop escalating and even if it doesn't it leaves everyone angry and upset...which carries on into the rest of life. So what strategies might be used to send the message that the odds have changed and approaching being pulled over with the nice angle is likely to work better?

On the flip side I think the best case for society is 'hit or miss'. If *everyone* is always super respectful towards getting pulled over at all times, that would imply to me cops are letting almost anyone talk themselves out of a ticket by deploying lots of flattery. That's not good for a free society either, gov't agents should be able to handle some rudeness and poor attitudes from the general public.

I think if everyone is respectful and co-operative at traffic stops in order to game the system, we might happily forgo a bit of public revenue from speeding fines.

There's a reason why pleading guilty gets lower sentences. Better social equilibrium.

"... but that usually has to do with how wealthy they are."

More or less respectful/cooperative if more wealthy?

And thanks for the real-world perspective.

Wealthy whites (especially if they are young) tend to be assholes when it comes to citations. I've been on the force for 14 years and I have only had white kids 1) belittle my profession and 2) tell me who their parents are.

Maybe it is where I police? But when I policed the white neighborhoods all the time I would get a young wealthy white kid saying to me "you have no idea what you'r doing" as I write them a ticket.

The paper seems to have taken this into account in a very creative manner:

The first task of this paper is to confirm that this disparity is evidence of officer
discrimination. Our central challenge is in ruling out that racial differences in treatment
are due to differences in criminality. Minorities may be driving faster than whites when
stopped, leading officers to treat them less leniently. To deal with this challenge, we use
the fact that one-third of officers practice no lenience. Namely, they exhibit no bunching
in their distribution of tickets.4 For these officers, we argue that their distribution of
ticketed speeds reflects the true distribution of driven speeds among stopped drivers. We
show that, conditional on location and time, driver characteristics are not predictive of
the average lenience of the officer he encounters. Non-lenient officers do not write fewer
tickets than lenient officers, and a similar share of their tickets are for speeding offenses.
These facts suggest that lenient and non-lenient officers are pulling over similar types of
drivers, and thus non-lenient officers can be used to identify the “true” distribution of

It's clear that lenient cops are preferentially lenient towards whites, yes, but the argument was that this may be due to behavioral differences in the interaction rather than due to race itself. Maybe I'm not understanding the idea, but I don't see how the non-lenient officers can be used to evaluate differences in interaction compliance, although I agree they can be used to evaluate differences in speeding rates.

The behavior of the 'zero leniency' officers may tell us what law enforcement would look like if all officers were unswayed by any bias either in terms of race or in terms of 'cooperation'.

But I think the level of cooperation is not an unbiased 'fact of life' but in fact part of the dynamic. One reason blacks may seem to be more defiant when stopped in your experience may be that blacks have figured acting meek and passive more often than not does not result in the cop giving them the benefit of a doubt and cutting them a break. Whites might be hit or miss because you are either going to get a white person who feels playing it nice will be the best way to get a break OR you will get a white person who feels highly entitled and is inclined to treat you like you are his personal servant. That still leaves open the question of why do these different populations feel inclined to play different cards ("Nice polite guy" versus "Defiant guy") as their best options in dealing with being stopped?

Optimally I think we would want a society where stops for all groups end up 'hit or miss' in their reaction to being stopped. I think we would want everyone to feel being polite is likely to work sometimes to catch a break from a cop that stops you *but* we also don't want a society where becoming adapt at excessive fawning and flattery of gov't authority is a sure ticket to getting the best break and asserting one's rights or views is punished by always getting the harshest reading of the law.

"One reason blacks may seem to be more defiant when stopped in your experience may be that blacks have figured acting meek and passive more often than not does not result in the cop giving them the benefit of a doubt and cutting them a break."

That is just laughable.

Is it? Not too long ago I was listening to a podcast that was interviewing a compulsive shoplifter. The woman openly stated that she thought she never gets caught because she's a well dressed white woman. I suspect if she did get caught she would try to play act very apologetic and passive in the hopes of being cut a break.

That story has nothing to do with your preposterous statement I quoted. The notion that black speeding offenders would find less success being respectful and cooperative than in being defiant and uncooperative is ludicrous.

I should have filled out the story more but I don't think you've made your case. Being defiant and uncooperative seems to be a foreign idea to you and something that is distasteful to you personally. I'm like that too however it is a mistake to assume that it must always be the case the world will reward you for being a nice guy and punish you for being an obnoxious guy. Reality is it often doesn't and if you can stomach being disliked by other people, you can sometimes get a better chance at getting your way by being obnoxious rather than being bend over backwards polite. This is well known by many sociopaths and salesmen.

The question then is what dynamics might cause people to try the obnoxious route rather than the sweet route? If we want to avoid an obnoxious society we should consider what signals are being sent that tell some people sweet will not be rewarded.

Based on my own experience, I have long suspected this to be true. While I am not in general a fan of speed cameras, I have always noted that using cameras instead of live officers would at least have a salutary effect on this.

Those speed cameras are set up to record black people and to gather evidence to put them in jail.

The officer's discretion is not the problem in itself. The goal is not to achieve perfect enforcement of every law down to every letter. The goal with law enforcement is good government....people want their environment to be safe and positive BUT at the same time they don't want to feel like the whole system is set up to jam them up should they make the smallest mistake. To accomplish this the officer's discretion is part of the expertise needed to do the job having a good sense of when it makes sense to escalate the situation and 'jam someone up' or 'cut them a break' the good cop does the job well.

You are hoping that technology can somehow make actual judgment unnecessary. If that was the case then we wouldn't need cops at all, we could install 'black boxes' in all cars and whenever someone commits a moving violation the fine is automatically issued to them the way your credit card automatically assesses the fees for a late payment.

.... police- officer's-discretion is a horrible fix to the mass of horrible, poorly written "laws" issued by generations of idiot politician/legislators in every state.

"Good Government" does not arise from government employees.

The laws aren't poorly written. Speeding is pretty objective. If you wanted perfect enforcement of speeding laws you could easily do it with cameras or car 'black boxes' that record and issue automatic tickets for going over the limit in any particular area.

What you are missing is discretion has always been an element of good government. There's the laws and *how* they are enforced. Both prosecutors and cops have to exercise *judgment* on the job...not just judges.

"“Good Government” does not arise from government employees."

perhaps not but it would be pretty strange to argue they aren't a key part of the picture.

Discretion is indeed an important factor in police work.

But suppose Blacks and Whites speed with equal probability, but police give Whites a break more often. While the Black drivers did indeed break the law and deserve their punishment, they are still victims of discrimination.

I can't say whether the remedy is to charge Whites more or charge Blacks less. I'm inclined to do the former. I'm worried we are becoming a lawless society because of leniency and indifference.

Let's let cameras and lasers issue speeding tickets. I'm growing opposed to police stops which all too often are pretext for unlawful searches and opportunities for intimidation. It may be possible in some cases to issue tickets for infractions without getting out of the car. Identification of the driver is the tricky part.

There's an interesting book out about the death of Eric Garner. Garner, you may recall, was a large, overweight and somewhat sloppy looking Black guy who sold 'loose' cigarettes on the street in NYC...when cops put him in an illegal chock hold to arrest him, he died causing massive protests.

The author went into some of the backstory. The neighborhood used to be filled with open drug use/dealing and was rarely well policed. Quite grimy. Recently, however, it has been policed more and more as the neighborhood has been getting upscale residents who are pushing out the more 'street' locals. At the same time 'broken window theory' policing puts extreme quotas on cops. They *must* do so many stop and frisks per week, score so many arrests each week for drugs, weapons whatnot.

The day of Garner's death, two people got into a fight in a public park. Police ignored the fight (not part of the quota), but Garner...being a local personality and not wanting more police attention to his corner...broke up the fight. Later on police came to arrest him supposedly for selling loose smokes...which he hadn't done that day but was known to do...was predictably outraged. Here he had done 'their job' earlier and now the one time he wasn't breaking a law he gets arrested because of some quota.

In an earlier age where racism was more tolerated among the NYC police, I don't think that would have happened. Even though you might have fewer minority officers and open racism tolerated by cops, the local 'cop on the beat' racist as he might have been....would have been more strategic in his use of power. He would have probably recognized the people fighting in the park was more of a threat to 'law and order' than Garner. While he might hassle Garner on a slow day, he would have probably recognized doing so on that day would have accomplished nothing but stoke neighborhood anger. He also might have recognized that a 'street character' like Garner could be used to help be his eyes and ears when he wasn't around. That's not to say Garner wouldn't have gotten arrested or hassled in an earlier age, it's still that fact that even an imperfect cop would have been expected to not just enforce the law but use his power of discretion in an intelligent manner.

It seems a lot of the trouble we have now is refusal to recognize discretion/judgment as an essential part of the cops' job and expertise. A speeding or red light camera that just issues tickets will ignore a murder or car jacking. A cop that ignores such a crime because he wants to finish writing a ticket for speeding would be fired. Trying to turn cops into cameras under the whip of simplistic quotas is to take a valid insight that broken windows had and take it in a disastrous direction.

+1 for discretion.

We have speed cameras all over in the UK. They are seen as revenue-raising devices, lack any discretion (e.g. deserted road at 3am, weather and road conditions), distract people from the driving task (looking out for cameras, maintaining legal limit +/-2mph), detract effort away from detection of more serious traffic offences, and generally undermine public confidence in the police.

America's house has become hoplessly diveded against itself and proceeds fast to demolition.

This is a clever method, but it can still be confounded. For example, the behavior of the driver when stopped may influence the leniency of the cop, with more cooperative drivers benefiting which may vary by race.

I wonder how reliable radar gun readings are? If they're frequently off by a couple of mph, then the sensitivity and specificity of the readings will differ between groups whose propensity for speeding varies, which will confound the results of a study like this.

Is there a difference in radar vs laser? If laser is dead on and radar is +-3 mph, the cop using radar might round down due to uncertainty. Before a change in fines would matter, but not if in the middle of the band. Doesn't answer the race question unless there's a difference in highway usage and race usage of the highways vs local streets.

Laser and radar are both super-accurate technologies in this context. Doppler is very good. I can't see how multi-path issues from radar could be worse than specular effects from the laser, as those introduce only ranging/bearing not speed errors.

In both cases aiming/bearing errors dominate the "technical" error sources.

The accuracy of the device is secondary to the angle at which the officer takes a reading and his ability to visually identify the target.

This assumes that the distribution of actual speeds is smoothly continuous. It's more plausible that at least some subset of drivers is aware of the steps, or at least vaguely shares the same conventions. Strategically they may keep their speeds just below the threshold. I know I'm much more likely to drive 14 over than 16.

The alternative hypothesis is that whites are more strategic thinking than blacks.

Why are Americans always fighting? Whites, Blacks, atheists, Christians, Jews, Democrats, Republicans, Trumps, never Trumps, homosexuals, heterosexuals, Northerners, Southerners, abolitionists, copperheads, evolutionists, creationists, meat--eaters vegans, etc. In Brazil, we are all brothers.

Brothers in looting the state treasury maybe.

Did you read this dear brother?

There are crimes everywhere, but, in Brazil, organized crime is fought, hence the shooting wars. There is no alliance between criminals and police, as in Japan, and crime is not triumphant, as in many parts of the United States. According to Mr. Pinker, there are parts of the United States where young boys rarely live to be eighteen years old.

The point is, in Brazil, the ideological/racial lines that divide America and set brother against brothers have faded away. Brazil is building a brotherly, democratic, fair and prosperous society.

maybe you should get out in the real world more. reading newspapers, watching t.v. and trolling interwebs is only reinforcing your biases.

Oh, those evil newspapers... Yet, someone linked one newspaper trying to prove me wrong. It is better not expect coherence from apologists of the American empire.

Unfortunately, policeman have, in the past, enforced the geographic boundaries established by politicians in the 1950's.

Highways and parks were built to separate, or protect (depending on your point of view), neighborhoods from each other. Dan Ryan Expressway, Central Park in the 1900's. So, given that geographic barrier, if you were in a neighborhood in which you did not belong, you were harassed, so you would leave.

I am currently reviewing some work on geospatial marketing--measuring or defining markets involving locational aware apps where the market is defined by movement through an area--and its amazing how you can see the effect of highways, rivers and parks on human interaction.

These decisions erecting barriers were made many years ago. The effects of these barriers persist even though public attitudes change. Even in policing, where the police may seen their role as detecting those who either don't belong there, or who look different from someone else.

I say this as lawyer who was often called at 2 am in the morning by his wife's foster brother in law who was a black Navy Seal who was often stopped in his neighborhood in San Diego, often because he looked like he didn't belong there. This was about 15 years ago, I hope it has changed since then. Just remember, though, that the guy in Minneapolis who was stopped by police and shot was stopped because he didn't look like he belonged in that area.

No, Castile was pulled over because the police thought he resembled a robbery suspect. Even if one believed that the identification could be pretext for an illegal stop, the existence of a robbery suspect at large should militate against that inference.

Yours is exactly the kind of misinformation that foments racial hatred. This is inexcusable. The common thread of every one of these shootings is deliberately false info unreasonable inferences from BLM.

Yeah, the resemblance was that he was black.

Never mind a kid was in the back seat. And his girlfriend in the front seat.

1) the paper focuses upon 15+ mph violations, yeash that’s fast.

2) the paper’s handling of the 1.5 mph difference in speed seems overly torturous and is central to the paper.

3) the paper has no access to the prior record of the driver, which is likely central to the decision.

I suspect #3 is not as big an issue as you think. Keep in mind they identified about 1/3 of officers who never exhibit 'bunching' behavior in their tickets. They assume these officers never round up or down and hence provide an accurate picture of what enforcement would look like if you had officers who simply wrote the ticket based on whatever the speed was. Since this is used as a baseline, it would capture patterns in the driver's record.

In other words, if blacks have worse driving records hence cops cut them less slack for speeding, that would be reflected in the ticket rates of the 'no leniency' cops and as a result give a higher baseline for 'unbiased' ticket writing to blacks.

Your argument makes no sense. We are talking about the tail of the distribution.

That doesn't seem to follow.

Assume the "no bunching" officers are ground truth, fine.

But the "bunching" officers may simply be taking account prior driver records or behaviours. And if the white/black demographics have different prior driver behaviours or behaviour....we explain the effect with recourse to racism.

"withOUT recourse to racism"

I'm not sure how this follows. What exactly justifies the 'bunching' officer's bunching again?

Behaviors? You mean during the stop there's some behavior that blacks systematically do that causes them to lose the 'bunching break' that non-blacks don't do. Let's imagine it's being very mouthy and argumentative since that aligns with the observation the police officer reported earlier here. If that was the case why wouldn't the word get out that you can dodge or improve a ticket by 'sweet talking' the cop? Why would only non-blacks be aware of this 'hack'? I'm sorry but 'behavior' here sounds like it's just another word for 'being black'.

Driver Records? So you assumed the 1/3 non-bunching officers are 'pure truth' when it comes to speeding. The other 2/3 opt to reward or not reward the 'bunching break' to speeders based on previous records. But then why the divergence based on race? Instead of splitting the groups into white and black speeders, you're saying if we do a 4-way split (black speeder w/good record, black speeder w/bad record, white speeder w/good record, white speeder w/bad record) we would discover the true variable is actually the record? If that's the case can you demonstrate it with the data? Should be rather easy.

BTW, I don't know about Florida but I know in NJ police do not have criminal histories available to them in stops. They can see if there are outstanding warrants but not past criminal convictions. In terms of driving I know from experience they can see if you got a *warning* from another cop in the recent past, but I don't know about actual ticket history. I'm wondering how much of a person's driving history is actually available to a cop? It actually seems rather prejudicial to me. I think the cop should decide upon a ticket based on what he saw in that moment and not by the driver's 'reputation' from past tickets.

Look there seems to be a lot of gymnastics here to find convoluted ways to explain away what is clearly racial bias. This doesn't mean anyone need be charged with being an actual racist who desires to use his power as a cop to give blacks a harsher time than whites (although there may be a few people who that best applies too). It simply means implicit bias is a thing cops have to be aware of rather than pretending it only applies to a 'racist cop' who attends KKK meetings in his off time.

If I told you more sophisticated shoplifters take advantage of various implicit biases by dressing upscale, acting like they are rich etc. in order to draw suspicion away from themselves, would you act like I was accusing you of being some type of tool of the upper class?

"Our central challenge is in ruling out that racial differences in treatment are due to differences in criminality. "

My thoughts would to be to adjust for points on the license. A cop would likely more lenient to A than B if A hadn't had a ticket in the last 5 years than if B gets one every 6 months.

Opps, as Al above alludes to in 3.

Wouldn't necessarily work if the prior tickets were also possibly the result of bias.

But it is very important, and often excluded, to control for the person's behavior. This is largely unobservable but essential to the analysis.

Here's the thing.

1. You have 1/3 of cops that exhibit no 'bunching'. They are assumed to be 'unbiased' in the sense that they don't round down speeding tickets.

2. IF your alternative theory is correct, what is going on here is 'biased' cops reward speeders with cleaner records by rounding down. It just so happens black motorists have records with more speeding tickets so they don't get the benefit.

3. There's a few easy ways to confirm/refute this alternative theory:
- First, do cops have access to driving histories? ( seems to say they just get name/address/insurance info and a flag if there are warrants outstanding in many cases and only limited driving histories are available (if at all) barring the officer making a special request. If they don't your counter theory fails.

- Second, if you think this is the case there would be a dynamite way to tell. Take the group of white and black drivers pulled over by the 'biased' cops. Split them into two groups, blacks/whites with good records and black/whites with bad records. If your theory is true the black/white difference will disappear and the correlation will be solely based on previous driving record. If the correlation doesn't disappear then you are still left with some measure of bias that appears to be driven by race. Even assuming police can and do check driving histories I would be very surprised if you could make all the variation in white/black rounding vanish by simply accounting for driving records.

Driving while black is a crime.

What's the surprise here?

And how does the researcher control for the driver acting like an ass when the officer comes to the window?

1/3 of cops have no 'bunching' in writing their speeding tickets. Presumably those cops provide us with a 'true' rate of speeding that we can compare to those who give breaks by rounding tickets down.

If the entire difference is simply due to politeness that seems pretty perplexing. You mean blacks are haven't got the word that the way to 'hack' a speeding ticket is to act really polite to the cop and there's a 2/3 chance he'll at least round the ticket down for you?

"If the entire difference is simply due to politeness that seems pretty perplexing"

If the entire difference is simply due to any one factor that seems pretty perplexing. At least to me. Looking through the paper, I can't help but notice that the author took no other factor into consideration. But given the sheer number of factors involved, there ought to be serious warning flags that the author just assumed that all other factors were evenly distributed across White and Non-White drivers. Which may be the case, but it's a real problem that the author didn't bother to even address it.

What we have here is someone who went looking for racism and then documented the steps he took.

Comments for this post are closed