Abstract
This paper uses the roll out of different modes of public transport, especially new metro and train connections, in São Paulo Metropolitan Region (SPMR), one of the most crime-prone cities of the world and examine their impact on wide-ranging crime indices. We document that initial crime in SPMR was independent of the roll out of new transport infrastructure in SPMR and exploit the exogenous variation in the timing of the introduction of new public transport (often restricted by land acquisition and other problems beyond the influence of SPMR district authorities) to obtain their causal impact. We employ a flexible event studies approach with various fixed effects to account for district, border districts as well as central districts level omitted factors over time that may also influence these crime indices. Results suggest that access to new metro and train lines that connected the central districts to more remote districts had significantly lowered car crime, both car thefts and car robbery. The crime reduction impact of car theft occurred within a year of receiving new transport while it took ten years or so for a reduction in car robbery in treated districts receiving new Metro and Train connections. No such effect is, however, observed for drug crime, homicide, rape or suicide or when we consider access to any new mode of public transport in our sample. Evidently, the reduction in car theft and car robbery were higher in treated districts with high initial car crime, greater population density, greater educated jobs and also with more than one new Metro/Train connections. We document that the observed car crime reduction effects of new metro/train connections can be attributed to substitution of private cars by new Metro/Train connections as well as income effects arising from the gradual gentrification of treated districts. Results have important implications for the design and operation of public transport system not only in other Brazilian cities, but also beyond its border