Using data on recidivism after a large collective pardon in Italy, we exploit exogenous variation in sentence length generated by the pardon to estimate the degree of discounting among criminal offenders. Consistent with the idea that low future time preference is a driver of criminal behavior, our analysis implies average annual discount factors of 0.74. We also find evidence of heterogeneity in time preference across offenders based on age, educational attainment, crime type, and nationality. Our model estimates, which are in line with the reduced form evidence, imply that the marginal effect of imprisonment varies considerably with the length of the sentence.
Giovanni Mastrobuoni (Dept. of Economics)
Date & time:
25 Nov 2015 13:00 pm - 25 Nov 2015 14:00 pm
Internal seminars home