Can Cash Prevent Crime?

Renowned Sociologists Ran a Field Experiment to Find Out. They Concluded “No.” But They Didn’t Understand Taxes.

The trio had an incredible new plan to avoid prison. But their scheme would go terribly wrong. The year was 1976. And the plotters were not part of the Lufthansa heist crew, but three of the nation’s most prominent criminologists, sociologists who studied crime: Richard Berk, Kenneth Lenihan, and Peter Rossi. Rossi would soon serve as president of the American Sociological Association. Their paper would end up contributing to the end of the era of prevention, and laying the intellectual groundwork for the era of “tough on crime.”

The basic idea was simple, and built on a small-scale field trial one of them had run recently in Baltimore: what if, when men get out of prison, we give them money? People commit many crimes because they need money, of course, and in the researchers’ experience most recent inmates struggled to find work. Desperate, the recently-released returned to crime instead of persisting in their job search. So a cash payment should give these men time to get on their feet, and allow them to find lawful work. The Baltimore results, though small scale, were promising, with a large drop in reported property crimes.

So the researchers pitched the U.S. Department of Labor on a larger-scale study. Then they successfully persuaded officials in Texas and Georgia to implement it. Hundreds of released inmates would receive relatively large periodic payments, amounting to about sixty percent of the full-time minimum wage amount at the time.

But something unexpected happened. Taken by themselves, the payments did reduce crime, but they also reduced work. And, since work was strongly correlated with a reduced propensity to offend, the payments ended up having no net statistically measurable impact on recidivism rates. Together with other reports finding that “nothing works,” the DoL project contributed to a rising intellectual consensus that crime prevention could not be shown to be effective. This proved a convenient message for the “tough on crime” policies of the new Reagan administration.

Berk, Lenihan, and Rossi were wrong. Their project wasn’t a failure — or, at least, not in the way they thought. The headline result should have been that payments (and work) reduce crime, full stop. The impact of payments on work, and their corresponding tendency to increase crime, was not a product of the payments themselves, but of bad tax economics.

Without realizing it, the trio had imposed a one hundred percent tax on earnings. For reasons that are not clear in the paper, they decided that for every dollar a released inmate earned, they would have to give back one dollar of the cash payment. It’s not hard to see why few of the subjects took a job: for the first thirty hours or so of work every week, they would be working for zero net wages. The jobs open to recently-released offenders, most of them black, in Texas and Georgia in 1977 were … not easy work. No one was going to do them as a volunteer.

The researchers did attempt to institute a partially controlled version of this policy, but apparently there was some snafu. In one arm of the treatment, the repayment rate was only 25%: one dollar repaid for every four dollars earned. But, in a somewhat cryptic footnote, the paper reports that the participants were misinformed about this rule and believed it was the equivalent of the 100% treatment.

Do cash payments reduce crime, then? We don’t know for sure. One recent paper finds that the Alaska Permanent Dividend Fund might in fact reduce property crimes. What we really need is to repeat the Texas and Georgia experiments. Then the Berk, Lenihan, & Rossi paper would still serve as an interesting bit of evidence in public finance debates: what happens to labor supply under a 100% tax?