Winner Teaching innovation Award 2017

The irrationality of ethnic profiling: abandon this police practice!

5 minutes to read
Column
Paul Mutsaers
18/04/2017

Some days ago, on a lazy Sunday afternoon,  I was sitting outside with my friends, enjoying some Belgian brews, Italian food and the last rays of sun beaming on my face. I made a remark about the culinary superiority of Italians and before I knew it, my company engaged in the all too familiar competition as to who was first with the ethnic profiling joke: “that’s ethnic profiling!” said my accountant-friend.

Occupation oriented jokes are a sure part of our repertoire. The accountant is always asked to check the bill. The general practitioner expects questions about how he would go about resuscitating this or that waitress. The civil servant is bound to explain on whose tax money he was drinking beer that Thursday 3 pm. As a police anthropologist interested in police discrimination and related stuff, I know what to expect, but normally dodge the questions.  This time, however, I had just finished a long conversation with the insurer on insurance and actuarialism and felt like explaining my opposition to ethnic profiling as an actuarial, that is, statistics-based risk-assessment tool.

These are all descent guys, so I didn’t have to substantiate the justice claims. All of them accept that good governance has to be impersonal and impartial in order to do justice to individual rights and freedom. Procedural fairness is the cornerstone of our justice apparatus and legal culture, as I wrote elsewhere on Diggit Magazine.  Without the “art of separation”—that in the context of criminal justice demands a distinction between criminal justice and popular justice, social standing and legal entanglement,  private sentiments and public tasks, and the like—justice is bound to be violated.

“But if a certain minority group is overrepresented in crime statistics in a particular area,” the accountant argued, “targeting that population is just the most efficient allocation of law enforcement resources. It’s common sense!” This, of course, is an oft-heard rational-action argument for profiling. Ethnic profiling is value for money! Ignoring the fact that this sort of rationality is inevitably at odds with our justice claims, I replied with an equally oft-heard counterargument: the self-fulfilling prophecy rationale. “In the end, crime statistics are produced by police,” I said.  But this was considered too easy an answer. Luckily for me, I happened to carry with me Bernard Harcourt’s Against Prediction, in which he uses mathematics to critique profiling. This, I thought, would convince the accountant.

Core to his argument is the so-called “ratchet effect,” which he describes as a “disproportionality that grows over time. The disproportionality in question is between the makeup of the offending population and the makeup of the carceral population—that is, the population that has criminal justice contacts such as arrest, conviction, fine, probation, imprisonment, parole or other supervision” (p.147).  While I do not have the space to explain this in detail as Harcourt has done, I will try to summarize the basics, as I did to my friends. This summary can be read as a definite statement for all the police officers, journalists, and pundits who have invited me—and will, no doubt, again invite - me to explain my opposition to ethnic profiling.

The ratchet effect occurs when these incarceration rates are subsequently used to further boost ethnic profiling

Harcourt provides the hypothetical situation of a city with 1 million residents, in which a minority group indeed has committed more offenses  than the majority group (in order to make any profiling nonspurious). The minority group (200,000) is taken to commit offenses at a higher rate of 8 percent than the majority group, which commits offences at a rate of 6 percent (see first table). As the table show, the incarceration rate is proportional to the offending rate when the police search randomly (that is, without profiling) for offenders, for instance when checking vehicles for drugs or other contraband.   

Source: Harcourt, B. (2007). Against Prediction: Profiling, Policing and Punishing in an Actuarial Age. Chicago and London: The University of Chicago Press.

Now, if the same number of searches (10,000) are nonrandomly distributed on the basis of race or ethnicity, as in the second table, it becomes immediately apparent that the incarceration rate for the minority group becomes disproportional to the offending rate of that group (47/53 versus 25/75). This, Harcourt argues, is precisely the distortion created by using actuarial methods such as ethnic or racial profiling.

Source: Harcourt, B. (2007). Against Prediction: Profiling, Policing and Punishing in an Actuarial Age. Chicago and London: The University of Chicago Press.

The ratchet effect occurs when these incarceration rates are subsequently used to further boost ethnic profiling, resulting in a so-called “multiplier effect.” Please also note that the number of successful searches has only increased with 6 percent (from 640 to 680), whereas the incarceration rate of minorities has gone up with 22 percent.

Moreover, the raison d’être of police is, contrary to what most people think, not to detect crime but to reduce it. This is an important difference and, interestingly, ethnic profiling can be understood as a factor that produces more crime, regardless of the theory we use to analyze it. If, for example, we accept a rational choice theory that states that people include risks (being caught) and opportunities (crime pays) when deciding whether or not to commit a particular crime, the deterrence of profiled groups is balanced out or even exceeded by the attractiveness of crime for nonprofiled groups. Key in this respect is what is called the “relative elasticity of offending to policing,” which assumes that different groups react similarly to the change in policing. From this perspective, it can be assumed that ethnic profiling decreases the actual offending rate of minorities, only to increase the rate among the majority, which, by definition, is much larger! The sum effect is an increase in crime.

If we take a totally opposite theory, for instance one that is centered on police legitimacy and procedural fairness (e.g. the work of Tom Tyler), we must conclude again that ethnic profiling produces crime. This theory states that people only tend to obey the law if they consider it legitimate. From this perspective, ethnic profiling lowers the elasticity of the profiled group. It turns them “resistant” in two senses of that word: no longer responding to police authority and resisting by means of crime. This squares with the massive body of literature on subversive crime, crime as protest (e.g. Quinney) and “social banditry” (e.g. Hobsbawm).

By the time I was talking about Hobsbawm and his ideas about crime as a challenge to those who lay claim to power, law and the control of resources, I had lost my friends’ attention. A new tray of Belgian brews had arrived. I do hope, however, that I have inspired others to read Harcourt’s work on the dangers of profiling.