Against Prediction: Profiling, Policing, and Punishing in an Actuarial Age
by Bernard E. Harcourt. Chicago: University of Chicago Press, 2007.

pdf

Reviewed by Peter Moskos in the American Journal of Sociology

Under the medieval system of tything, individuals could be held responsible for the misdeeds of others in their collective group. In the movie Minority Report, set in the near future, criminals are incarcerated before they commit their crimes. Our present system of justice, according to Bernard E. Harcourt’s Against Prediction, combines the worst of both worlds. “The quest for prediction,” Harcourt writes, “has distorted our conception of just policing by emphasizing efficiency over crime minimization. Profiling has become second nature because of our natural tendency to favor economic efficiency” (p. 188).

It wasn’t supposed to be this way. Actuarial methods “grew out of our lust to know the individual . . . to tailor punishment to the particularities and probabilities of each man. . . . Today, we generalize to particularize” (p. 193). The book is an excellent and convincing treatise against assuming that an individual’s actions are predicted by group behavior (though curiously, the ecological fallacy is never mentioned by name).

Against Prediction convincingly argues that the use of economic actuarial methods—predicting individual criminal likelihood based on the quantifiable characteristics of groups to which one belongs—is fundamentally flawed. Against Prediction is clearly trying to wrest and reclaim this field of criminal justice from economists. I like a book that batters economic theory as much as the next sociologist, and Harcourt tries to beat economists at their own game. Harcourt’s three main points are clear. Actuarial-based criminal justice, including racial profiling and extending into most aspects of sentencing, probation, and parole, (1) goes against a crime-prevention philosophy, (2) produces racial distortion in the prison population, and (3) biases our conception of fair and just punishment.

Against Prediction is weakest when it relies too much on economic assumptions of rational action vis-à-vis crime and punishment. On account of potential variability in deterrence among different groups, Harcourt proposes that profiling and other actuarial methods actually increase crime. This counterintuitive argument is intriguing, but more as theory than as practice.

The key to the book is understanding and believing that the deterrent effect of law enforcement is elastic—that is, different for different groups and higher for whites than for blacks: “If the profiled group has lower elasticity of offending to policing, profiling that group will probably increase the overall amount of crime in society” (p. 24). This is because “as that cost increases for minorities, their offending decreases, and as that cost decreases for whites, their offending increases” (p. 116). “It all,” Harcourt proclaims, “depends on elasticities” (p. 136). Yet the “clincher,” as Harcourt readily admits, is that “we have no good idea how the elasticities compare” (p. 24).

Harcourt bases his ideas on elasticity variability on a surprising faith in the classical school of criminology’s theories of rational action and deterrence. Harcourt assumes that “ordinary citizens [respond] to the changes in police practices and corresponding changes in the cost of committing crime” (p. 116). This is too Hobbesian, too rooted in assumptions of rational-action behavior. But without these assumptions, Harcourt’s discussions of elasticity become frustratingly irrelevant.

Rather than relying on the ideas of Cesare Beccaria and Jeremy Bentham, Harcourt could turn to rich qualitative works from the likes of Elijah Anderson, Howard Becker, Philippe Bourgois, Bruce Jacobs, and Sudhir Venkatesh. Rather than asserting that a system of early parole and racial profiling could increase crime by reducing the deterrent effect on whites, Harcourt could note that longer incarceration likely increases recidivism, or that many street-level criminals lack the long-term introspection (and perhaps basic math skills) to weigh variances in sentence length. And rather than focusing on the relative elasticity of white and black drug couriers, Harcourt could observe the racially segregated world of drug dealing. Most drug kingpins do not have access to a racially diverse pool of labor. It is unlikely, as Harcourt asserts, that more whites will begin transporting drugs if a slightly greater percentage of black couriers are profiled and interdicted. While an individual may be deterred from transporting drugs by the likelihood of getting caught, the wholesale drug supplier will simply send more couriers.

Harcourt concludes by advocating “randomized” law enforcement. Outside of airport security and traffic patrol, randomization is hard to visualize. It’s intriguing, but it’s not clear how this can coexist with Sir Robert Peel’s principle goal of police, “to decrease the rate of crime” (p. 124). Randomization, along with being somewhat insulting to police officers (Harcourt has never been a fan of police officer discretion), serves little purpose for officers who police areas that are exclusively minority.

Such criticisms of Against Prediction are, naturally, academic. And perhaps I do protest too much. Harcourt’s overarching goal is indeed noble and correct. By academic standards, the writing is engrossing, and a refreshing passion flows through the pages. Even with the limitations of Harcourt’s arguments, Against Prediction is very persuasive as a moral critique on questions of racial justice.

Harcourt is at his best when he worries that we, the American society, damage ourselves through actuarial methods. This analysis is clearly presented and persuasive. Harcourt expands our understanding by contrasting the goals of actuarial science with America’s shared understanding of justice: “Everyone who commits a crime should have the same likelihood of being apprehended, regardless of race, ethnicity, gender, class, or any other irrelevant group trait” (p. 38). Put simply, if you and I commit the same crime, why should you be punished any differently because of the history of other people “like you”?

Unlike the fictional police in Minority Report, we cannot predict who will commit a crime in the future. Nor, as in the past, do we assign people to be their brother’s keeper. That we continue to do so under the guise of actuarial efficiency, that we fail to see the harms of prediction, and that we proudly aspire to some quixotic goal of corrective “efficiency” is to our collective shame as much as Against Prediction is to Harcourt’s credit.