Estimations of the Economic Model of Crime Using Aggregate and Individual Level Data

Much has happened in the economics of crime since Becker's [1] seminal paper of two decades ago. The theory has been refined by, among others, Ehrlich [10; 11; 12], Block and Lind [4; 5], Block and Heinke [3], and Schmidt and Witte [30]. Attempts to test the model empirically began immediately, the most cited including Ehrlich [10; 11; 12], Sjoquist [31], Carr-Hill and Stern [7], Phillips and Votey [26], Witte [38], and Myers [22; 23]. Interest continues with recent work including Layson [19], Sandelin and Skogh [29], Ehrlich and Brower [13], Cover and Thistle [9], Viscusi [36], and Tauchen, Witte, and Griesinger [33]. There has been some controversy, however, concerning the type of data more suitable to the estimation of the economic model of crime. Most of the work has used aggregate data in which the unit of observation is a political jurisdiction. But the use of aggregate data has been sharply criticized.' The alternative is individual data. Few attempts have been made to use individual data, probably because an ideally constructed data set representative of a general population would be far too costly for most research budgets. Most of the studies that have been done have used data on releasees from prison or other criminal justice programs. They are, in effect, recidivism studies based on the economic model of crime.