TY - JOUR TI - Differently rational DO - https://doi.org/doi:10.7282/T3QC028T PY - 2012 AB - This dissertation uses a range of economic tools to analyze and understand criminal behavior, particularly theft. The first chapter outlines a number of key issues and stylized facts observed in criminal behavior, and provides an outline of the chapters that follow. The second chapter proposes a simple threshold model of theft, and develops a number of structural estimators based on this model. It then tests the model against data from the National Longitudinal Study of Youth, 1997 Cohort. The evidence suggests that the key determinant of theft behavior is the costs of theft to the thief, and in particular the thief's perception of future costs. There does not seem to be significant variation in the benefits of theft; that is, there is no sign that some individuals are more capable of theft than others. The data also shows that theft behavior is usually very short-lived, with the vast majority of thieves showing activity for less than one year in adolescence. The third chapter looks at the temporal pattern of criminal behavior (frequently termed a ``criminal career'' or ``trajectory'') in individuals' lives, up to the age of 25-30. It uses three different data sets, based on several methods of observation, and finds a number of similarities. In contrast to earlier work describing criminal careers, the data suggest that the two measures of age-specific inclination and individual-specific intensity are the key to describing patterns in criminal behavior. Specifically, there is significant evidence that an individual A committing a single crime at age 14 has more in common with individual B who commits only one crime, at age 24, than C who commits three crimes at age 14. The fourth chapter looks at the decision to steal in the context of a simple model of human capital accumulation, as a way to tease out the relative role of labor substitution and impatience in individuals' decision to steal. The data support the role of impatience a significant driver. First, individuals who report stealing show a wage and labor market participation pattern strongly consistent with a low discount rate. Second, individuals who report stealing show significant underinvestment in education, with lower enrollment rates and higher grade repetition than comparable non-thieves. Finally, individuals who report stealing show a larger number of employers, suggesting underinvestment in long-term career success. KW - Economics KW - Criminal behavior KW - Criminal behavior, Prediction of KW - Differential association theory KW - Theft LA - eng ER -