Testing the local and spatial spillover effects of police monitored CCTV systems on crime

Objectives

Police-monitored CCTV systems are supposed to reduce on-the-street crime. The purpose of this study is to determine whether the MiCalle CCTV systems program in Mexico City had local and spatial spillover effects on delivery robberies.

Methods

When data is clustered in space and time, random effects and errors can be autocorrelated in both dimensions. To test the effectiveness of the MiCalle program on delivery robbery crime rates, a Difference-in-Differences (DID) design was specified in a maximum likelihood (ML) panel regression model with spatial lag coefficients, random effects variance estimates, and autoregressive serial correlation variance estimates.

Results

At the neighborhood level of study, we find no evidence connecting delivery robbery crime rates to the Mexico City MiCalle program.

Conclusions

The MiCalle program has been geographically unfocused and appears to have been ineffective in reducing delivery robberies. Future studies of police-monitored CCTV systems should consider the possibility of spatiotemporal interactions among variables, particularly if spatialized panel data is used.