Social Interactions and Crime revisited: An Investigation using Individual Offender Data in Dutch Neighborhoods

摘要: Using data on the age, sex, ethnicity and criminal involvement of 14.3 million residents aged 10–89 residing in 4,007 neighborhoods Netherlands, this article tests if an individual’s is affected by proportion criminals living their neighborhood residence. We develop a binomial discrete choice model for estimate it individual data. control endogeneity that may be related to unobserved characteristics take into account possible biases result from sorting behavior. find strongly significant social interaction effects explore implications presence multiple equilibria multipliers.