I use data on individual bank robberies to estimate the distribution of criminals’ disutility of jail. The identification rests on the money versus risk trade-off that criminals face when deciding whether to stay an additional minute while robbing the bank. The observed (optimal) duration of successful robberies identifies the individual compensating variation of jail, called disutility of jail. The distribution of the disutility which is often assumed to be degenerate, resembles instead an earnings distribution, and highlights heterogeneity in the response to deterrence. General deterrence effects are increasing in criminal’s disutility.

VIEW PAPER