Despite the prevalence of auto theft and associated economic loss, the offense is relatively underexplored in the literature. The current study examines auto theft in the context of social disorganization theory. Using 2000 US Census data and 2005 Houston Police Department incident reports, auto theft rates across census tracts are compared on four dimensions of neighborhood social disorganization. Multivariate analysis indicates support for social disorganization theory; concentrated disadvantage, residential stability and racial heterogeneity had a significant effect on auto theft rates. In contrast, neighborhood immigration did not affect auto theft rates. The limitations of the study and policy implications are discussed.