This article explores key predictors of homelessness rates in different Swedish munici-palities, the accommodation provided and the coverage of accommodation to homeless-ness. In order to create realistic models, seven sets of factors represent various structural levels in stepwise regressions. These sets are urbanisation, economy, demographic com-position, housing market, aggregated individual problems, political majority and policies and organisational features of the social services. The findings show that urbanization is central to understanding variations in homelessness and has also impact on housing pro-visions. In addition, higher rent levels as well as aggregated psychiatric problems seem to increase the level of problems. Overall supply of apartments and the proportion of public housing are important factors in providing accommodation, and a demographic factor, such as the proportion of single-parent households, seems to enhance such a provision. Political majority as well as the functional organization of the social services seems to independently impact coverage rates. Despite the importance of urbanization, findings indicate that a number of factors with impact on the problem and on the possibilities to handle them are manageable by the municipalities. Key words: predicting homelessness, accommodation, housing provisions, local author-ity, social services.