Problem: The study explores predictors of homelessness in Swedish municipalities and of the amount of accommodations provided to the homeless. Theoretically, structural factors (e.g. housing market) and individual factors (e.g. social and health factors) have competed in theories explaining homelessness. Here, the focus is not on individuals, but rather the aggregated levels of such problems. Methods: Homelessness rates were collected in a national inventory (NBHW 2006.) and data on municipal support provisions were collected in a representative sample of half of the Swedish municipalities. Candidate explaining factors of homelessness studied here are variables related to urbanisation, economy of the municipality, various housing market factors, level of social and health problems and demographic factors. Level of homelessness and political majorities are also used as predictors of provisions. Results: Homelessness rate (per 10 k inhabitants) is predicted (R2= .36) by urbanisation (no of inhabitants), market (rent level) and level of psychiatric health problems in the municipality. Amount of accommodations (per 10 k inh.) is predicted (R2= .18) by homelessness rate, social welfare costs and structure of housing market, i.e. greater public owned sector in normal housing. Conclusions: Structural factors and level of individual problems both add to the explanation of homelessness. In addition, degree of urbanisation seems to be an additional factor. This however, may in turn indicate more specific factors not studied here. The amount of accommodations per inh. is, as expected, related to the level of homelessness and the level of social problems in the municipality. Additionally, the structure of the housing market seems to have important impact on the amount of provisions, independent of problem levels.