Turning Prisoners Into Slaves Doesn't Deter Crime

Making life behind bars even worse does no one any good, and it's really expensive.

Each dawn I die.

Photographer: Andrew Lichtenstein/Corbis/Getty images

In recent years, people have started to notice that the economics discipline is turning away from pure theory and toward a more empirical approach. But this isn’t just an academic curiosity. It has real, positive consequences for people’s lives. Economic theory, if unchecked by data, can easily lead to very bad policy decisions.

For a good example, consider a recent paper in the National Bureau of Economic Research by Stanford University law professor A. Mitchell Polinsky. The author isn’t primarily an academic economist -- he has a courtesy appointment at Stanford’s econ department -- but the legal profession now has many people working at the intersection of the two disciplines, with Judge Richard Posner being the most famous example. Polinsky’s paper, entitled “Prison work programs in a model of deterrence,” says that making prisoners work without pay can help deter crime, by making prison more unpleasant.

Many people already object to prison labor on moral grounds. Forced labor, though allowed by the Constitution, is a form of slavery. Since smoking marijuana, for example, is a federal crime, this means that if you ever smoked, the U.S. government has the legal right to enslave you. I don’t like that one bit. Polinsky, perhaps wary of advocating this extreme policy, also constructs a model in which voluntary work programs also serve a deterrent function. But the real problem with his paper isn’t its lack of a moral grounding -- economics research often deals with difficult choices and dubious policies -- but its illusion of plausibility.

Polinsky used what he calls a standard model of how punishment deters crime. This model hearkens back to Gary Becker, the Nobel prize-winning University of Chicago theorist who gained fame for applying econ theory to many non-traditional walks of life. Becker’s model assumes that criminals are rational, forward-looking individuals, who coolly decide whether to commit a crime by carefully weighing the gains from the misdeed, the probability of getting caught, and the unpleasantness of punishment. These imaginary criminals are as rational as any business manager deciding whether to drill a new oil well.

Becker’s theory yields one key result -- the notion that it’s possible for society to substitute very harsh punishment for very strict enforcement. If criminals are rational and risk-averse, a small chance of a very severe punishment is even scarier than a large chance of a small punishment. So society can save money by employing fewer police and catching fewer criminals, but making prison sentences very long and harsh for those who get caught.

By now, anyone who has actually worked in law enforcement is probably tearing their hair out as they read this. Real-life criminals are nothing like the hyper-rational calculating machines in Becker’s model. They tend to be impulsive, short-term thinkers. They can’t really imagine what a 50-year prison sentence would really feel like. They often commit crimes in the heat of the moment and regret them bitterly years later.

Alex Tabarrok, an economist at George Mason University, writes eloquently about the failure of this sort of theory:

We have now tried [Becker’s] experiment and it didn’t work. Beginning in the 1980s we dramatically increased the punishment for crime in the United States but we did so more by increasing sentence length than by increasing the probability of being punished. In [Becker’s] theory, this should have reduced crime, reduced the costs of crime control and led to fewer people in prison. In practice…the experiment with greater punishment led to more spending on crime control and many more people in prison.

This failed experiment ruined the lives of a great many Americans. A policy that relied less on the theory of deterrence and more on hard evidence about human psychology would probably have kept many people from making bad decisions that ended up destroying their futures and costing American taxpayers billions of dollars. Theories have consequences -- that’s why they should always be backed up by lots and lots of empirical analysis.

It is in this respect that Polinsky’s theory is lacking. It has only assumptions and equations that express those assumptions. To some readers, math automatically gives the impression of science. But omit the premise of criminals as well-informed rational calculating machines and the equations become misleading fictions.

Within the economics profession itself, simple models like this, with no empirical support, wouldn’t be given much credence in this day and age. But in the shadowy borderlands where law and economics meet, they may carry much more weight. That’s very dangerous, because although academic econ is mostly confined to the ivory tower, law determines real people’s fates on a daily basis.

The public, including lawyers and judges, needs to be apprised of this threat. Economics has already moved strongly toward empiricism, evidence and proof. Disciplines that borrow from econ should do the same, with all deliberate speed.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

    To contact the author of this story:
    Noah Smith at

    To contact the editor responsible for this story:
    James Greiff at

    Before it's here, it's on the Bloomberg Terminal.