Steven Levitt is a first-rate empirical researcher whose interests span many social sciences - economics, politics, sociology, and law. His innovative empirical investigations have examined the causes of crime, and the potency of methods of deterrence. His novel identification strategies have made possible a better understanding of fundamental and longstanding puzzles.
Crime and Deterrence:Levitt's research shows that policies that increase incarceration have a greater impact on crime rates than was previously thought. His QJE (1996) paper demonstrates this enhancement by using prison-overcrowding litigation to break the endogenous correlation of crime rates and incarceration rates. This identification exploits the notion that overcrowding litigation is likely to affect crime rates only through its impact on the prison population. The JPE (1998) paper explains the recent trend toward youth rather than adult crime as a response to differential incentives. In their JEL (1999) paper, Kessler and Levitt use the introduction of sentence increases in California to produce evidence in favor of "deterrence" theories of incarceration in contrast to "incapacitation" theories. Levitt (AER, 1997) shows that increases in the police force reduce the number of violent crimes much more substantially than the number of property crimes.
This body of papers uses four distinct and novel identification strategies, each with thoughtful consideration of strengths and weaknesses, to examine the same important issue.
Levitt studies the benefits to reducing car theft from installing a hidden radio transmitter device called a Lojack in his QJE paper (1998, with Ian Ayres). He finds that the private benefit to those who install a Lojack is dwarfed by the social benefit of general deterrence.
Sources of Criminal Activity:Levitt's research on street gangs (QJE 2000, with Vekatesh) debunks the popular view that most youth crimes are the work of a few super-predators, who are largely unresponsive to incentives. It also elucidates why gang members work for low wages and what role is played by gang patriotism. Donohue and Levitt (QJE 2001) find that legalized abortion has the indirect effect of reducing crime by reducing the size of the "unwanted," and thus less cared for, children. Duggan and Levitt (AER 2002) document how corruption alters the outcome of Sumo wrestling.
Political economy: In his JPE (1994) paper, Levitt finds that a challenger's spending in an election has much less of an impact than suggested by previous research. He argues that the rewards to spending are quite similar between incumbents and challengers. Levitt (AER, 1996) used the preferences revealed from role-call voting to show that the senator's own ideology is the primary determinant of his/her voting patterns. Voter and party preferences are secondary.
Levitt has interesting work on other topics as well, including drunk driving, and testing game-theoretic models of soccer, and corruption in sumo wrestling.
SummarySteven Levitt is the most innovative empirical researcher in his cohort. He has confronted important empirical questions in the economics of crime and political economy, by finding new data and devising novel and clever identification schemes. Levitt deserves much credit for pioneering empirical research. He is fully deserving of the John Bates Clark Medal.