This paper analyses the linkage between tourism and crime with a particular focus on the distortions generated onto criminal activities by the presence of visitors. Controlling for socio-demographic and economic variables, the contribution of tourist arrivals to different types of crimes for 103 Italian provinces and for the year 2005 are empirically investigated. The possible spillover effects of crime are taken into account by testing two spatial models (one spatial lag model and one spatial error model). The hypothesis is also tested according to which different geography of tourist destinations – that is, urban, mountain, marine, etc. – alters the impact of tourism on crime. Finally, the social cost of crime associated with tourist arrivals is measured.