Re-imagining Policing Starts by Re-imagining Traffic Enforcement

Police shootings of civilians all too often begin with a traffic stop. There are more effective alternatives.

A great deal of ink is being spilled of late debating the costs and benefits of ‘defunding’ police and ‘reimagining’ policing. It’s very trendy to talk about injecting science into government by prioritizing the funding of ‘evidence-based’ policies and programs and promoting rigorous evaluations—with the current fad of legitimizing only randomized-controlled trials. But a real scientific inquiry would start more simply, by clearly defining terms.

To begin, I think the term reimagining is quite silly. I’m not sure Sir Robert Peel or the other founders of modern policing ever ‘imagined’ policing in the vision board kind of way ‘reimagining’ suggests. Still, it has some momentum in today’s popular culture, and momentum is an underrated asset in systems reform.

The first question about a reimagined police force might be ‘who are the police?’ or ‘what should the police do?’ (or, how important is the word ‘force’ in the police force?). The next set of questions might be ‘how do police perform their duties?’ and, ‘when is appropriate to use a civilian police force rather than a military force or no police at all?’. The choice between who are the police and what should they do seems perfectly endogenous—what the police do determines who they are, just as the composition of the police force determines in large part what they do. To break apart that simultaneity problem, I will simply choose to talk first about what police do.

My reimagining requires that we start by thinking about how the world is, and how it could evolve into something better, rather than just catapulting ourselves into some mythical future. Step 1 then on reimagining policing is to think about what police are good at and what someone else might be able to do better. Again, it is reasonable to start with what police currently do and in particular how they are trained and what they are trained to do.

A typical police academy spends a lot of time teaching would-be officers how to control situations, through words and action. It is fair to say that most police officers are quite good at controlling a situation, both through how they behave and through the signals (verbal and nonverbal) they send. Police are much less well trained on social work (how to identify needs and link people to appropriate services), sociology (how society and its institutions shape peoples behaviors), economics (what values people assign to their choices), psychology (why people obey or do not obey the law). Etc. And, society invests a tremendous amount in police, both in real dollars to the cost of their training, their salaries, and their equipment and in terms of their absolutely critical role as a real-time metaphor for our shared values and the terms of our social contract. Thus, their time shouldn’t be wasted on trivial things.

So, it is reasonable to start by suggesting that sworn law enforcement should get out of the business of being traffic cops and law enforcement agencies should shift traffic enforcement from police to civilians. As I write this, this experiment is beginning to go live around the country. Here's a story from the August 10th Washington Post about a proposal to do just that in one of DC's wealthy but very diverse suburbs:

Should police be in charge of traffic enforcement?

The impetus for this reform is, of course, the same as the whole reimagining police movement: a glaring racial disparity in traffic enforcement. Black drivers are more than twice as likely to be stopped by police as white drivers in the county. Once stopped, black drivers are more likely to have their car searched, to be arrested, to be held pre-trial without bail or with a high bail, to be convicted, and to receive a longer sentence than would a white driver in the exact same circumstance. Obviously, this must change. Police are trained to control situations and look for threats, even in low leverage situations such as a traffic stop. Throw in a healthy structural bias against black drivers and it’s a toxic and volatile mix.

Plus, traffic stops are trivial from a law enforcement perspective. That doesn’t mean they are unimportant, it simply means that most driving penalties are mild. So, traffic stops are a potentially toxic mismatch of police training and poor use of a sworn officer’s time.

Plus, if the idea is to efficiently improve driving behavior, police do a very poor job. Objectively, traffic cameras do a much better job. Automated traffic enforcement is vastly more effective than officers in patrol cars in responding to traffic law violations. The automated detection rate is orders of magnitude higher. Cameras focus only on increasing the certainty of detection and distribute penalties in a completely unbiased way (everyone gets a ticket). In doing so, they maximize general deterrence, which is a critical objective of law enforcement—to get as many people as possible to believe their illegal behavior will be detected and punished.

If you believe, as I do, that the primary means by which police prevent crime is through deterrence, this switch to automated and civilian traffic patrols buys a lot more deterrence. Police ticketing is haphazard and in this non-randomness actually reduces deterrence (more on that some other time). It is also racially biased, as demonstrated by data from traffic cameras which ticket more white drivers and women drivers than subjective stops by an officer. Cameras fix all of this.

Man vs. machine: An investigation of speeding ticket disparities based on gender and race This paper analyzes the extent to which police behavior in giving speeding tickets differs from the ticketing pattern of automated cameras.

Also, traffic cameras do not cause more traffic accidents and they are not a violation of civil liberties. These complaints aren’t even worth discussing.

In addition to many more cameras, red-light, speeding, etc., many traffic duties can be undertaken by civilian traffic patrols. Using national data, @Jerry_Ratcliffe has produced this amazing graphic that suggests traffic stops are 10-20% of local law enforcement calls for service (or more since 'unfounded radio calls' include traffic-related incidents). Since the article that motivated this note is about Montgomery County, MD, I should point out that Jeff Asher (@Crimealytics) and @Vox find that in MoCo specifically, traffic-related calls for service are 14% of all calls (or more, again, b/c some traffic stops are including in the catch-all 'non-criminal calls).

Take note that for all the ink I just spilled on traffic enforcement, it is a tiny bit of police work. In the traffic domain, a bigger bit is the amount of time on derelict vehicles and traffic accidents. There is simply no reason why a civilian traffic patrol could not handle these incidents. That leaves only DUI stops as the responsibility of traditional police (and even that could be reimagined somewhat).

How do we pay for all of this? Depending on how you deal with the police officers whose duties you have replaced, it is probably cost neutral in addition to being more effective. According to the same Washington Post article, it doesn't take a lot of resources. In Montgomery County, MD, there are "73 officers and 26 civilians in the police department." That's 100 or so employees in a department of 2,000 (1,300 sworn officers and 700 civilians). So, 5% of Montgomery County, MD law enforcement staff are working on traffic stops which are 15-20% of all calls for service. That’s very cost-effective.

Bottom-line. 5% of the policing budget is responsible for 15-20% of the calls for service and could be transferred to civilians. With a knock-on effect of dramatically improving deterrence. While reducing racial disparities in traffic stops and their terrible knock-on effects. That's good policy.