Preventing Homelessness in Alameda County, CA and New York City, NY: Investigating Effectiveness and Efficiency

Abstract

Due to lack of rigorous evaluations, there is limited evidence that homelessness prevention programs effectively reduce rates of homelessness and efficiently direct services where they can make the most difference. Effectiveness is the ability to reduce rates of homelessness among people who would otherwise experience it. Efficiency is the ability to direct services to those who would benefit most. Evidence of effectiveness requires a counterfactual – typically a comparison between a treatment group and a similar group that does not receive treatment. Evidence of efficiency necessitates development of a risk model and investigation of the levels of risk where services make the most difference. Investigations sometimes confound effectiveness and efficiency: evaluators might believe that services are effective when those services are imprecisely targeted. The current study examines effectiveness and efficiency for prevention programs in two sites. It develops risk models for homelessness using Cox proportional hazard models for 2,761 applicants for Homelessness Prevention and Rapid Re-Housing in Alameda County and for 10,220 individual applicants for HomeBase prevention services in New York City. Further, it uses a regression discontinuity design for the sample in Alameda County to examine the effectiveness of services. The findings provide limited evidence that prevention programs can reduce entries into homelessness and stronger evidence that programs can work better by focusing on individuals and families at highest risk. Triage models that exclude some applicants as too risky to serve are not supported by the data. The studies also contribute to the understanding of the causes of homelessness, via the examination of risk factors in the two sites. The results suggest that structural issues are the driving forces of homelessness in two housing markets with limited access to affordable housing. Future research is necessary to obtain more precise estimates of prevention effects and to examine similarities and differences in findings across housing markets. Prevention programs might do better not only to provide immediate prevention services for individuals and families but also to combat the structural forces that lead to high rates of homelessness.