The impact of permanent supportive housing on homeless populations

Between 2007 and 2014 in the United States, national counts of the homeless population decreased as permanent supportive housing (PSH) for the homeless increased. Using a panel dataset from all communities across the country during this period, I investigate the effect of PSH on homeless population sizes. I first estimate the association between PSH beds and homeless counts, controlling for all time-invariant community-level factors, several time-varying community-level factors, and in some specifications, all time-varying state-level factors. One additional PSH bed is associated with between 0.04 and 0.12 fewer homeless people. Second, I estimate causal effects using federal funds allocated to communities for homeless assistance as instruments; funding decisions are based on information that precedes any current shocks to homeless populations. Adding one PSH bed reduces the homeless count by up to 0.10 people, and I can reject a reduction of more than 0.72 people at the 95% confidence level. Finally, I discuss several possible explanations for the relatively modest impact, including poor targeting, differential exit rates into private housing from PSH relative to homelessness, incentives for remaining homeless, errors in homeless counts, and migration in response to expanded PSH.