Association for Public Policy Analysis & Management

Since 2009, both the rapid re-housing (RRH) and transitional housing (TH) interventions have been available to assist households experiencing homelessness in Georgia. We use statewide administrative data to compare long-term housing outcomes for three groups of households: those recently completing a RRH program, those recently completing a TH program, and those who were served only by emergency shelters (ES). Propensity score matching reduced selection bias but yielded a final sample which overrepresented households with lower barriers to housing. In separate analyses of households with and without children, we used generalized linear mixed modeling to control for several household characteristics in order to obtain intervention effect sizes. Both RRH and TH significantly reduced the likelihood of returning to shelter compared to ES; these effects were larger among households without children. However, unlike previous non-experimental studies, the difference between RRH effects and TH effects was statistically insignificant for both household types. Results suggest that both RRH and TH are effective in keeping lower-barrier households out of the shelter system over the long term, and that the cheaper RRH intervention is more cost-effective than TH for these households.