Proportional policing and criminal patrol

One of the (many) moral problems of policing in the US is that law enforcement strategies often fail a proportionality requirement. The principle of proportionality is often invoked in discussions of the morality of the use of force, whether it is force used to achieve military or police goals, or for individual self-defense. It is impermissible to use force that causes significantly more harm than the benefit caused by achieving one’s goal. Police are not permitted to shoot fleeing robbery suspects, for example. Proportionality also informs distributional concerns. Justice requires not just that we minimize errors in criminal trials, but that we achieve a particular distribution of errors. We tend to accept an increase in false findings of innocence as a cost of reducing false findings of guilt. A legal system that imposes costs on the innocent too frequently in the course of punishing the guilty will be disproportionate because the harm required to punish the guilty would be too great. Here I’ll argue that proportionality constrains not just use of force policies, but also the selection of strategies for enforcing law.

I’ve argued elsewhere that a proportionality principle constrains the tactics police officers are permitted to use when responding to protests and demonstrations. But there are some police strategies that fail a proportionality requirement even if officers rarely resort to physical force. Most strategies that rely on “filtering” are unlikely to meet the requirement.

Consider the stop and frisk strategy employed by the NYPD during the 2000s and early 2010s. This practice became intensely controversial because a majority of the New Yorkers stopped, questioned, and frisked where Black or Hispanic. At the time, Mayor Bloomberg insisted that officers stopped people based on their behavior, not their appearance. At the peak of stop and frisk, the NYPD averaged around 1200 stops per day which resulted in around 70 arrests, and only 2 on illegal gun possession charges. In other words, with a success rate of around 6 percent, the NYPD did not appear to have a great sense of which behaviors indicated criminal activity. Even putting aside the threat of force required in each stop, police stops are not trivial incursions into a person’s life. Around 94 percent of the people the NYPD stopped in a day who had to bear those costs were innocent. This means that the strategy, even granting that it reduced crime and successfully enforced legitimate laws (both of which are controversial), has a significant number of false “findings of guilt.” Given the huge number of innocent people stopped as part of the strategy, it is not proportional.

Similarly, consider the use of traffic stops for criminal law enforcement. The LAPD, like most police departments, uses traffic stops to find illegal guns and drugs. Even if this is a common way for officers to make arrests and remove illegal guns from the streets, it is unlikely that one can make accurate judgements about who is breaking the law simply by looking at the exterior of one’s vehicle. This leads to officers pulling motorists over for violating traffic laws just to examine the driver and the inside of the vehicle. There are serious concerns about racist police activity in this context too. Black drivers are less likely to be pulled over at night when the driver’s race is “masked.” Even if there weren’t concerns about racist enforcement, however, the concern about proportionality remains. The ultimate reason for the traffic stop is to investigate criminal activity, but most of the people stopped are innocent. Filtering strategies that rely on stopping large numbers of people, most of whom are innocent, to find a small number of guilty people, fail a proportionality requirement.

Evaluating these strategies in terms of proportionality is useful because it lets us avoid arguing about the presumption of innocence. You might think that these strategies undermine the presumption of innocence (I do), but that objection is less useful against strategies that rely on pretextual stops. Traffic laws, or anti-loitering laws, are tools for enabling more invasive police interactions. The goal is rarely to enforce the law that an officer witnesses one violating, but rather to look for criminal activity.

If this is right, then proportionality is also a tool for evaluating law enforcement strategies. This is useful because determining the just administration of law is separate from the question of which laws are just. Though the former has gotten less attention, it is still important. This kind of argument about proportionality plays a role in some of the work on policing I’m doing now.