Understanding the degree of geographical crime displacement is crucial for the design of crime prevention policies. This paper documents changes in automobile theft risk that were generated by the plausibly exogenous introduction of Lojack, a highly effective stolen vehicle recovery device, into a number of new Ford car models in some Mexican states, but not others. Lojack-equipped vehicles in Lojack-coverage states experienced a 48 percent reduction in theft risk due to deterrence effects. However, 18 percent of the reduction in thefts was displaced toward unprotected Lojack models in non-Lojack states, providing new evidence of geographical crime displacement in auto theft. (JEL H76, H77, K42, O17, O18, R23)