This article raises the question of the relationship between the probability of racist offences occurring and the sociodemographic characteristics of a particular area. It begins by presenting an overview of the American and British literature on the spatialization of racist acts, and identifying three poles (over-determination of economic variables, demographic variables and a combination of economic and demographic variables) and a series of hypotheses (threatening power, differential power, “defended neighborhood”). It then tests these hypotheses based on the French case. The analysis of geographical data from a collective study of 483 cases of racist offences appearing before two French correctional courts, and their statistical treatment (cross tabulation and logistic regression) challenge the dominant hypothesis in the existing literature, that of “defended neighborhoods”, because it only accounts for a portion of racist acts. It is preferable to identify “local racism” and “remote racism” as well as the specific configurations in which it occurs: neighborhood, “conquest”, “defense”, and territorial neutrality. By correlating spatial data with a range of indicators, this article sheds light on the influence of economic, social, and demographic factors in the occurrence of racist acts. Following on from the work done by Jean-Claude Chamboredon and Madeleine Lemaire on social conflicts in French housing estates, this article shows that spatial proximity works in connection with social and racial distance to make racist offences more likely.