Consider the duration of stay of migrants in a host country. We propose a statistical model of locally interdependent return hazards in order to examine whether interactions at the level of the neighbourhood are present and lead to social multipliers. To estimate this model we develop and study two complementary estimation strategies, demonstrate their good performance while standard non-spatial estimators are shown to be heavily biased. Using a unique large administrative panel dataset for the population of recent labour immigrants to the Netherlands, we quantify the local social multipliers in several factual and counterfactual experiments, and demonstrate that these are substantial.