The process of homelessness: an event history analysis of length of stay for the Groningen shelter accommodation

This study explores the process of shelter exits for homeless individuals and for different shelter types in the public shelter system in the city of Groningen in the Netherlands. Individualised administrative data from the Groningen Homelessness Monitor (2003–2006) are used for estimating a Weibull hazard rate regression model. Results show that the different shelter types fit adequately with their policy purposes : lengths of stay in emergency shelter accommodation appear to be shorter than in supportive housing accommodation and clients tend to remain longest in guest houses. The results also indicate that the use of day shelters, as well as demographic variables such as age, gender and nationality, are associated with a differentiated likelihood of continuing homelessness. The results of this study in the Dutch context are compared with research on the process of homelessness in the United States, and finally the implications of these findings and future directions for research on the process of homelessness are discussed